cannam@95: Sat Nov 24 22:37:54 EST 2012 stevenj@fftw.org cannam@95: * fixed deadlock bug caused by bogosity flag getting out of synch between processes; thanks to Michael Pippig for the bug report cannam@95: cannam@95: M ./kernel/ifftw.h +1 cannam@95: M ./kernel/planner.c -3 +6 cannam@95: M ./mpi/api.c +12 cannam@95: cannam@95: Wed Nov 21 18:34:29 EST 2012 athena@fftw.org cannam@95: * Updated NEWS cannam@95: cannam@95: M ./NEWS -2 +7 cannam@95: cannam@95: Wed Nov 21 18:33:15 EST 2012 athena@fftw.org cannam@95: * use 2x2 AVX transposition instead of individual stores. cannam@95: cannam@95: This seems to improve single-precision AVX on Sandy Bridge machines. cannam@95: cannam@95: cannam@95: M ./simd-support/simd-avx.h -2 +14 cannam@95: cannam@95: Tue Nov 20 12:18:00 EST 2012 stevenj@fftw.org cannam@95: * revert part of Taylor patch to acx_mpi.m4: do not link -lmpi if mpicc works without libraries, as -lmpi may be some completely different MPI implementation cannam@95: cannam@95: M ./m4/acx_mpi.m4 -3 +3 cannam@95: cannam@95: Tue Nov 20 11:44:57 EST 2012 stevenj@fftw.org cannam@95: * fix deadlock bug (thanks to Michael Pippig for the bug report and patch, and to Graham Dennis for the bug report) in which some processes called MPI_Alltoall and some called MPI_Alltoallv cannam@95: cannam@95: M ./mpi/transpose-alltoall.c -3 +2 cannam@95: cannam@95: Mon Oct 29 15:20:01 EDT 2012 athena@fftw.org cannam@95: * fix texinfo quirk cannam@95: cannam@95: M ./doc/tutorial.texi -2 +2 cannam@95: cannam@95: Mon Oct 29 09:16:43 EDT 2012 athena@fftw.org cannam@95: * clarify that padding only applies to in-place transforms cannam@95: cannam@95: M ./doc/tutorial.texi -5 +10 cannam@95: cannam@95: Sun Oct 28 18:42:48 EDT 2012 athena@fftw.org cannam@95: * make the index-computation logic less paranoid cannam@95: cannam@95: The problem is that for each K and for each expression of the form P[I cannam@95: + STRIDE * K] in a loop, most compilers will try to lift an induction cannam@95: variable PK := &P[I + STRIDE * K]. In large codelets we have many cannam@95: such values of K. For example, a codelet of size 32 with 4 input cannam@95: pointers will generate O(128) induction variables, which will likely cannam@95: overflow the register set, which is likely worse than doing the index cannam@95: computation in the first place. cannam@95: cannam@95: In the past we (wisely and correctly) assumed that compilers will do cannam@95: the wrong thing, and consequently we disabled the induction-variable cannam@95: "optimization" altogether by setting STRIDE ^= ZERO, where ZERO is a cannam@95: value guaranteed to be 0. Since the compiler does not know that cannam@95: ZERO=0, it cannot perform its "optimization" and it is forced to cannam@95: behave sensibly. cannam@95: cannam@95: With this patch, FFTW is a little bit less paranoid. FFTW now cannam@95: disables the induction-variable optimization" only when we estimate cannam@95: that the codelet uses more than ESTIMATED_AVAILABLE_INDEX_REGISTERS cannam@95: induction variables. cannam@95: cannam@95: Currently we set ESTIMATED_AVAILABLE_INDEX_REGISTERS=16. 16 registers ought cannam@95: to be enough for anybody (or so the amd64 and ARM ISA's seem to imply). cannam@95: cannam@95: cannam@95: M ./genfft/gen_hc2c.ml -1 +1 cannam@95: M ./genfft/gen_hc2cdft.ml -1 +1 cannam@95: M ./genfft/gen_hc2cdft_c.ml -1 +1 cannam@95: M ./genfft/gen_hc2hc.ml -1 +1 cannam@95: M ./genfft/gen_notw.ml -2 +2 cannam@95: M ./genfft/gen_notw_c.ml -2 +2 cannam@95: M ./genfft/gen_r2cb.ml -3 +3 cannam@95: M ./genfft/gen_r2cf.ml -3 +3 cannam@95: M ./genfft/gen_r2r.ml -2 +2 cannam@95: M ./genfft/gen_twiddle.ml -1 +1 cannam@95: M ./genfft/gen_twiddle_c.ml -1 +1 cannam@95: M ./genfft/gen_twidsq.ml -2 +2 cannam@95: M ./genfft/gen_twidsq_c.ml -2 +2 cannam@95: M ./genfft/genutil.ml -1 +2 cannam@95: M ./kernel/ifftw.h -3 +20 cannam@95: cannam@95: Sun Oct 28 18:33:24 EDT 2012 athena@fftw.org cannam@95: * silence warnings cannam@95: cannam@95: M ./kernel/buffered.c +1 cannam@95: M ./rdft/rank0.c +1 cannam@95: cannam@95: Sat Oct 27 09:58:49 EDT 2012 athena@fftw.org cannam@95: * bump version to 3.3.3 cannam@95: cannam@95: M ./NEWS +7 cannam@95: M ./configure.ac -1 +1 cannam@95: cannam@95: Sat Oct 27 09:55:15 EDT 2012 athena@fftw.org cannam@95: * evaluate plans for >1ms when using gettimeofday() cannam@95: cannam@95: The previous limit 10ms was too paranoid, and it made life difficult cannam@95: on machines without an "official" cycle counter, such as ARM. cannam@95: cannam@95: M ./kernel/timer.c -1 +1 cannam@95: cannam@95: Sat Oct 27 09:46:04 EDT 2012 athena@fftw.org cannam@95: * use 4-way NEON SIMD instead of 2-way cannam@95: cannam@95: Kai-Uwe Bloem tried to warn me a year ago that 128-bit NEON was better cannam@95: than 64-bit NEON even on machines with a 64-bit pipe, but I foolishly cannam@95: did not listen. Now that 128-bit NEON pipes are starting to appear on cannam@95: the market it is definitely time to switch. cannam@95: cannam@95: cannam@95: M ./simd-support/simd-neon.h -55 +100 cannam@95: cannam@95: Wed Sep 26 14:21:12 EDT 2012 athena@fftw.org cannam@95: * Note that fftw-3.3 includes MPI support cannam@95: cannam@95: M ./doc/intro.texi -5 +4 cannam@95: cannam@95: Wed Jul 18 11:25:40 EDT 2012 athena@fftw.org cannam@95: * remove obsolete unused function cannam@95: cannam@95: M ./dft/bluestein.c -14 cannam@95: cannam@95: Fri Jun 29 15:57:14 EDT 2012 stevenj@fftw.org cannam@95: * whoops, call omp_get_max_threads; thanks to Hanno Rein for the bug report cannam@95: cannam@95: M ./doc/threads.texi -1 +1 cannam@95: cannam@95: Sat Apr 28 10:55:09 EDT 2012 athena@fftw.org cannam@95: * Fix libfftw3/libfftw3_threads chicken-egg problem cannam@95: cannam@95: On most systems we want to build libfftw3 first, so that cannam@95: libfftw3_threads can depend upon libfftw3. When producing a single cannam@95: combined-thread library (e.g. on Windows) we want the opposite, cannam@95: so that libfftw3 can include libfftw3_threads. cannam@95: cannam@95: cannam@95: M ./Makefile.am -1 +18 cannam@95: M ./threads/Makefile.am +5 cannam@95: cannam@95: Sat Apr 28 10:11:28 EDT 2012 athena@fftw.org cannam@95: * updated NEWS for 3.3.2 cannam@95: cannam@95: M ./NEWS +15 cannam@95: cannam@95: Thu Apr 26 19:36:11 EDT 2012 athena@fftw.org cannam@95: * change revision to 3.3.2 cannam@95: cannam@95: M ./configure.ac -2 +2 cannam@95: cannam@95: Thu Apr 26 19:31:02 EDT 2012 athena@fftw.org cannam@95: * Remove old aligned_main() hack. cannam@95: cannam@95: On i386, in our benchmark program we used to manually aligned the cannam@95: stack to 16-byte boundary via asm trickery. This was a good idea in cannam@95: 1999 (and it was actually necessary to make things work) but the hack cannam@95: is now obsolete and it seems to break gcc-4.7. So the hack is now cannam@95: gone. cannam@95: cannam@95: cannam@95: M ./libbench2/Makefile.am -6 +6 cannam@95: R ./libbench2/aligned-main.c cannam@95: M ./libbench2/main.c -1 +1 cannam@95: cannam@95: Thu Mar 29 16:26:16 EDT 2012 athena@fftw.org cannam@95: * Bugfix: a couple of uninitialized values in the benchmark program cannam@95: cannam@95: M ./libbench2/problem.c -1 +1 cannam@95: cannam@95: Tue Mar 20 19:03:47 EDT 2012 athena@fftw.org cannam@95: * make libfftw{threads,mpi} depend upon libfftw for libtool purposes cannam@95: cannam@95: Thanks Julian Taylor for the patch cannam@95: cannam@95: M ./Makefile.am -1 +1 cannam@95: M ./m4/acx_mpi.m4 -3 +3 cannam@95: M ./mpi/Makefile.am +1 cannam@95: M ./threads/Makefile.am +1 cannam@95: cannam@95: Tue Mar 6 04:44:00 EST 2012 stevenj@fftw.org cannam@95: * formatting tweak cannam@95: cannam@95: M ./doc/reference.texi -1 +1 cannam@95: cannam@95: Mon Mar 5 21:05:27 EST 2012 athena@fftw.org cannam@95: * destroying => overwriting cannam@95: cannam@95: M ./doc/tutorial.texi -1 +1 cannam@95: cannam@95: Fri Mar 2 10:31:20 EST 2012 stevenj@fftw.org cannam@95: * note that WISDOM_ONLY is a documented flag cannam@95: cannam@95: M ./api/fftw3.h -1 +1 cannam@95: cannam@95: Fri Mar 2 10:27:08 EST 2012 stevenj@fftw.org cannam@95: * check for icc pretending to be gcc before including quad-precision decls; thanks to Michael Anselmi for the bug report cannam@95: cannam@95: M ./api/fftw3.h +1 cannam@95: cannam@95: Fri Mar 2 10:23:19 EST 2012 stevenj@fftw.org cannam@95: * foo_CFLAGS needs to manually include AM_CFLAGS; thanks to Henry Gomersall for the Windows bug report cannam@95: cannam@95: M ./threads/Makefile.am -2 +2 cannam@95: cannam@95: Sat Feb 25 15:21:39 EST 2012 athena@fftw.org cannam@95: * update for latest mingw cannam@95: cannam@95: M ./support/BUILD-MINGW32.sh -3 +3 cannam@95: M ./support/BUILD-MINGW64.sh -2 +7 cannam@95: cannam@95: Sat Feb 25 15:21:17 EST 2012 athena@fftw.org cannam@95: tagged fftw-3.3.1 cannam@95: cannam@95: cannam@95: Mon Feb 20 23:06:13 EST 2012 stevenj@fftw.org cannam@95: * added Fortran NEWS cannam@95: cannam@95: M ./NEWS +6 cannam@95: cannam@95: Mon Feb 20 23:00:13 EST 2012 stevenj@fftw.org cannam@95: * move non-portable extended/quad precision F03 interfaces into separate .f03 files (while keeping double/single in fftw3.f03 for minimal ABI breakage) cannam@95: cannam@95: M ./api/Makefile.am -3 +10 cannam@95: M ./api/f03api.sh -2 +4 cannam@95: M ./api/genf03.pl -5 +5 cannam@95: M ./doc/modern-fortran.texi -2 +39 cannam@95: M ./doc/mpi.texi -2 +2 cannam@95: M ./mpi/Makefile.am -5 +11 cannam@95: M ./mpi/f03api.sh -2 +4 cannam@95: cannam@95: Mon Feb 20 11:21:57 EST 2012 athena@fftw.org cannam@95: * rm mpi/fftw3-mpi.f03 at make clean time; thanks Tyler Luchko for the bug report. cannam@95: cannam@95: M ./mpi/Makefile.am +1 cannam@95: cannam@95: Mon Feb 20 11:18:24 EST 2012 athena@fftw.org cannam@95: * Disable a Visual Studion warning that was obnoxious enough for Sebastian Schuberth to send us a patch. cannam@95: cannam@95: M ./simd-support/simd-sse2.h +14 cannam@95: cannam@95: Mon Feb 20 11:18:06 EST 2012 athena@fftw.org cannam@95: * Change version to 3.3.1 cannam@95: cannam@95: M ./configure.ac -1 +1 cannam@95: cannam@95: Mon Feb 20 11:03:15 EST 2012 athena@fftw.org cannam@95: * Integrated Visual Studio AVX patches by Carsten Steger cannam@95: cannam@95: M ./NEWS +3 cannam@95: M ./simd-support/amd64-cpuid.h -10 +40 cannam@95: M ./simd-support/simd-avx.h +6 cannam@95: M ./simd-support/simd-sse2.h +8 cannam@95: M ./simd-support/x86-cpuid.h +7 cannam@95: cannam@95: Wed Nov 9 10:13:32 EST 2011 stevenj@fftw.org cannam@95: * typo cannam@95: cannam@95: M ./doc/modern-fortran.texi -2 +2 cannam@95: cannam@95: Tue Nov 8 22:45:09 EST 2011 stevenj@fftw.org cannam@95: * add missing F77 set_timelimit function; thanks to Martin Diehl for the bug repory cannam@95: cannam@95: M ./api/f77funcs.h +5 cannam@95: cannam@95: Sun Sep 25 10:54:56 EDT 2011 athena@fftw.org cannam@95: * note requirement of /machine:x64 in windows x64 README cannam@95: cannam@95: M ./support/BUILD-MINGW64.sh +7 cannam@95: cannam@95: Sun Sep 18 09:28:20 EDT 2011 athena@fftw.org cannam@95: * AVX detection for MSVC cannam@95: cannam@95: M ./simd-support/amd64-cpuid.h -2 +10 cannam@95: cannam@95: Tue Sep 13 14:58:29 EDT 2011 athena@fftw.org cannam@95: * compile with C89 cannam@95: cannam@95: M ./dft/bluestein.c -1 +1 cannam@95: M ./dft/generic.c -1 +1 cannam@95: M ./rdft/dht-rader.c -1 +1 cannam@95: cannam@95: Sat Sep 3 16:25:50 EDT 2011 athena@fftw.org cannam@95: * use the same search pruning heuristics for threaded plans as for nonthreaded plans cannam@95: cannam@95: M ./threads/ct.c -1 +3 cannam@95: M ./threads/hc2hc.c -1 +3 cannam@95: cannam@95: Sat Sep 3 16:12:11 EDT 2011 athena@fftw.org cannam@95: * shorten ESTIMATE planning time for certain weird sizes cannam@95: cannam@95: FFTW includes a collection of "solvers" that apply to a subset of cannam@95: "problems". Assume for simplicity that a "problem" is a single 1D cannam@95: complex transform of size N, even though real "problems" are much more cannam@95: general than that. FFTW includes three "prime" solvers called cannam@95: "generic", "bluestein", and "rader", which implement different cannam@95: algorithms for prime sizes. cannam@95: cannam@95: Now, for a "problem" of size 13 (say) FFTW also includes special code cannam@95: that handles that size at high speed. It would be a waste of time to cannam@95: measure the execution time of the prime solvers, since we know that cannam@95: the special code is way faster. However, FFTW is modular and one may cannam@95: or may not include the special code for size 13, in which case we must cannam@95: resort to one of the "prime" solvers. To address this issue, the cannam@95: "prime" solvers (and others) are proclaimed to be SLOW". When cannam@95: planning, FFTW first tries to produce a plan ignoring all the SLOW cannam@95: solvers, and if this fails FFTW tries again allowing SLOW solvers. cannam@95: cannam@95: This heuristic works ok unless the sizes are too large. For example cannam@95: for 1044000=2*2*2*2*2*3*3*5*5*5*29 FFTW explores a huge search tree of cannam@95: all zillion factorizations of 1044000/29, failing every time because cannam@95: 29 is SLOW; then it finally allows SLOW solvers and finds a solution cannam@95: immediately. cannam@95: cannam@95: This patch proclaims solvers to be SLOW only for small values of N. cannam@95: For example, the "generic" solver implements an O(n^2) DFT algorithm; cannam@95: we say that it is SLOW only for N<=16. cannam@95: cannam@95: The side effects of this choice are as follows. If one modifies FFTW to cannam@95: include a fast solver of size 17, then planning for N=17*K will be cannam@95: slower than today, because FFTW till try both the fast solver and the cannam@95: generic solver (which is SLOW today and therefore not tried, but is no cannam@95: longer SLOW after the patch). If one removes a fast solver, of size say cannam@95: 13, then he may still fall into the current exponential-search behavior cannam@95: for "problems" of size 13*HIGHLY_FACTORIZABLE_N. cannam@95: cannam@95: If somebody had compleined about transforms of size 1044000 ten years cannam@95: ago, "don't do that" would have been an acceptable answer. I guess the cannam@95: bar is higher today, so I am going to include this patch in our 3.3.1 cannam@95: release despite their side-effects for people who want to modify FFTW. cannam@95: cannam@95: cannam@95: M ./dft/bluestein.c -5 +14 cannam@95: M ./dft/generic.c -15 +6 cannam@95: M ./dft/rader.c -7 +7 cannam@95: M ./kernel/ifftw.h +8 cannam@95: M ./kernel/primes.c +6 cannam@95: M ./rdft/dht-rader.c -7 +8 cannam@95: M ./rdft/generic.c -14 +4 cannam@95: cannam@95: Sat Aug 27 13:55:24 EDT 2011 athena@fftw.org cannam@95: * Fix typo fftw_execute_dft_r2r => fftw_execute_r2r cannam@95: cannam@95: Thanks KIU Shueng Chuan for the bug report. cannam@95: cannam@95: M ./doc/reference.texi -1 +1 cannam@95: cannam@95: Fri Aug 26 06:13:55 EDT 2011 athena@fftw.org cannam@95: * In Rader's algorithm, compute the generator lazily. cannam@95: cannam@95: The planner was spending a lot of time computing generators for cannam@95: plans that were immediately discarded. Now we compute generators cannam@95: only when absolutely needed. cannam@95: cannam@95: M ./NEWS +5 cannam@95: M ./dft/rader.c -3 +4 cannam@95: M ./rdft/dht-rader.c -3 +4 cannam@95: cannam@95: Sun Aug 21 16:27:31 EDT 2011 athena@fftw.org cannam@95: tagged fftw-3.3.1-beta1 cannam@95: cannam@95: cannam@95: Sun Aug 21 16:16:38 EDT 2011 athena@fftw.org cannam@95: * Release notes for 3.3.1-beta1 cannam@95: cannam@95: M ./NEWS +4 cannam@95: M ./configure.ac -1 +1 cannam@95: M ./doc/install.texi -8 +18 cannam@95: M ./doc/other.texi -2 +3 cannam@95: cannam@95: Fri Aug 19 19:59:17 EDT 2011 stevenj@alum.mit.edu cannam@95: * make fftw_mpi_block routine 10x faster, since it is being called zillions of times (thanks to Tom Vacek for the profiling) cannam@95: cannam@95: M ./mpi/block.c -6 +2 cannam@95: cannam@95: Thu Aug 18 14:19:36 EDT 2011 athena@fftw.org cannam@95: * Implement autodetection of NEON extensions cannam@95: cannam@95: M ./simd-support/neon.c -4 +51 cannam@95: cannam@95: Sun Aug 14 14:12:29 EDT 2011 athena@fftw.org cannam@95: * Update the FSF address. cannam@95: cannam@95: The FSF moved downtown. cannam@95: cannam@95: M ./COPYRIGHT -1 +1 cannam@95: M ./api/api.h -1 +1 cannam@95: M ./api/apiplan.c -1 +1 cannam@95: M ./api/configure.c -1 +1 cannam@95: M ./api/execute-dft-c2r.c -1 +1 cannam@95: M ./api/execute-dft-r2c.c -1 +1 cannam@95: M ./api/execute-dft.c -1 +1 cannam@95: M ./api/execute-r2r.c -1 +1 cannam@95: M ./api/execute-split-dft-c2r.c -1 +1 cannam@95: M ./api/execute-split-dft-r2c.c -1 +1 cannam@95: M ./api/execute-split-dft.c -1 +1 cannam@95: M ./api/execute.c -1 +1 cannam@95: M ./api/export-wisdom-to-file.c -1 +1 cannam@95: M ./api/export-wisdom-to-string.c -1 +1 cannam@95: M ./api/export-wisdom.c -1 +1 cannam@95: M ./api/f77api.c -1 +1 cannam@95: M ./api/f77funcs.h -1 +1 cannam@95: M ./api/flops.c -1 +1 cannam@95: M ./api/forget-wisdom.c -1 +1 cannam@95: M ./api/import-system-wisdom.c -1 +1 cannam@95: M ./api/import-wisdom-from-file.c -1 +1 cannam@95: M ./api/import-wisdom-from-string.c -1 +1 cannam@95: M ./api/import-wisdom.c -1 +1 cannam@95: M ./api/malloc.c -1 +1 cannam@95: M ./api/map-r2r-kind.c -1 +1 cannam@95: M ./api/mapflags.c -1 +1 cannam@95: M ./api/mkprinter-file.c -1 +1 cannam@95: M ./api/mktensor-iodims.h -1 +1 cannam@95: M ./api/mktensor-rowmajor.c -1 +1 cannam@95: M ./api/plan-dft-1d.c -1 +1 cannam@95: M ./api/plan-dft-2d.c -1 +1 cannam@95: M ./api/plan-dft-3d.c -1 +1 cannam@95: M ./api/plan-dft-c2r-1d.c -1 +1 cannam@95: M ./api/plan-dft-c2r-2d.c -1 +1 cannam@95: M ./api/plan-dft-c2r-3d.c -1 +1 cannam@95: M ./api/plan-dft-c2r.c -1 +1 cannam@95: M ./api/plan-dft-r2c-1d.c -1 +1 cannam@95: M ./api/plan-dft-r2c-2d.c -1 +1 cannam@95: M ./api/plan-dft-r2c-3d.c -1 +1 cannam@95: M ./api/plan-dft-r2c.c -1 +1 cannam@95: M ./api/plan-dft.c -1 +1 cannam@95: M ./api/plan-guru-dft-c2r.h -1 +1 cannam@95: M ./api/plan-guru-dft-r2c.h -1 +1 cannam@95: M ./api/plan-guru-dft.h -1 +1 cannam@95: M ./api/plan-guru-r2r.h -1 +1 cannam@95: M ./api/plan-guru-split-dft-c2r.h -1 +1 cannam@95: M ./api/plan-guru-split-dft-r2c.h -1 +1 cannam@95: M ./api/plan-guru-split-dft.h -1 +1 cannam@95: M ./api/plan-many-dft-c2r.c -1 +1 cannam@95: M ./api/plan-many-dft-r2c.c -1 +1 cannam@95: M ./api/plan-many-dft.c -1 +1 cannam@95: M ./api/plan-many-r2r.c -1 +1 cannam@95: M ./api/plan-r2r-1d.c -1 +1 cannam@95: M ./api/plan-r2r-2d.c -1 +1 cannam@95: M ./api/plan-r2r-3d.c -1 +1 cannam@95: M ./api/plan-r2r.c -1 +1 cannam@95: M ./api/print-plan.c -1 +1 cannam@95: M ./api/rdft2-pad.c -1 +1 cannam@95: M ./api/the-planner.c -1 +1 cannam@95: M ./api/version.c -1 +1 cannam@95: M ./api/x77.h -1 +1 cannam@95: M ./commercialize.sh -2 +2 cannam@95: M ./dft/bluestein.c -1 +1 cannam@95: M ./dft/buffered.c -1 +1 cannam@95: M ./dft/codelet-dft.h -1 +1 cannam@95: M ./dft/conf.c -1 +1 cannam@95: M ./dft/ct.c -1 +1 cannam@95: M ./dft/ct.h -1 +1 cannam@95: M ./dft/dft.h -1 +1 cannam@95: M ./dft/dftw-direct.c -1 +1 cannam@95: M ./dft/dftw-directsq.c -1 +1 cannam@95: M ./dft/dftw-generic.c -1 +1 cannam@95: M ./dft/dftw-genericbuf.c -1 +1 cannam@95: M ./dft/direct.c -1 +1 cannam@95: M ./dft/generic.c -1 +1 cannam@95: M ./dft/indirect-transpose.c -1 +1 cannam@95: M ./dft/indirect.c -1 +1 cannam@95: M ./dft/kdft-dif.c -1 +1 cannam@95: M ./dft/kdft-difsq.c -1 +1 cannam@95: M ./dft/kdft-dit.c -1 +1 cannam@95: M ./dft/kdft.c -1 +1 cannam@95: M ./dft/nop.c -1 +1 cannam@95: M ./dft/plan.c -1 +1 cannam@95: M ./dft/problem.c -1 +1 cannam@95: M ./dft/rader.c -1 +1 cannam@95: M ./dft/rank-geq2.c -1 +1 cannam@95: M ./dft/scalar/n.c -1 +1 cannam@95: M ./dft/scalar/n.h -1 +1 cannam@95: M ./dft/scalar/t.c -1 +1 cannam@95: M ./dft/scalar/t.h -1 +1 cannam@95: M ./dft/simd/common/genus.c -1 +1 cannam@95: M ./dft/simd/n1b.h -1 +1 cannam@95: M ./dft/simd/n1f.h -1 +1 cannam@95: M ./dft/simd/n2b.h -1 +1 cannam@95: M ./dft/simd/n2f.h -1 +1 cannam@95: M ./dft/simd/n2s.h -1 +1 cannam@95: M ./dft/simd/q1b.h -1 +1 cannam@95: M ./dft/simd/q1f.h -1 +1 cannam@95: M ./dft/simd/t1b.h -1 +1 cannam@95: M ./dft/simd/t1bu.h -1 +1 cannam@95: M ./dft/simd/t1f.h -1 +1 cannam@95: M ./dft/simd/t1fu.h -1 +1 cannam@95: M ./dft/simd/t2b.h -1 +1 cannam@95: M ./dft/simd/t2f.h -1 +1 cannam@95: M ./dft/simd/t3b.h -1 +1 cannam@95: M ./dft/simd/t3f.h -1 +1 cannam@95: M ./dft/simd/ts.h -1 +1 cannam@95: M ./dft/solve.c -1 +1 cannam@95: M ./dft/vrank-geq1.c -1 +1 cannam@95: M ./dft/zero.c -1 +1 cannam@95: M ./doc/f77_wisdom.f -1 +1 cannam@95: M ./doc/license.texi -3 +3 cannam@95: M ./genfft/algsimp.ml -1 +1 cannam@95: M ./genfft/algsimp.mli -1 +1 cannam@95: M ./genfft/annotate.ml -1 +1 cannam@95: M ./genfft/annotate.mli -1 +1 cannam@95: M ./genfft/assoctable.ml -1 +1 cannam@95: M ./genfft/assoctable.mli -1 +1 cannam@95: M ./genfft/c.ml -1 +1 cannam@95: M ./genfft/c.mli -1 +1 cannam@95: M ./genfft/complex.ml -1 +1 cannam@95: M ./genfft/complex.mli -1 +1 cannam@95: M ./genfft/conv.ml -1 +1 cannam@95: M ./genfft/conv.mli -1 +1 cannam@95: M ./genfft/dag.ml -1 +1 cannam@95: M ./genfft/dag.mli -1 +1 cannam@95: M ./genfft/expr.ml -1 +1 cannam@95: M ./genfft/expr.mli -1 +1 cannam@95: M ./genfft/fft.ml -1 +1 cannam@95: M ./genfft/fft.mli -1 +1 cannam@95: M ./genfft/gen_hc2c.ml -1 +1 cannam@95: M ./genfft/gen_hc2cdft.ml -1 +1 cannam@95: M ./genfft/gen_hc2cdft_c.ml -1 +1 cannam@95: M ./genfft/gen_hc2hc.ml -1 +1 cannam@95: M ./genfft/gen_mdct.ml -1 +1 cannam@95: M ./genfft/gen_notw.ml -1 +1 cannam@95: M ./genfft/gen_notw_c.ml -1 +1 cannam@95: M ./genfft/gen_r2cb.ml -1 +1 cannam@95: M ./genfft/gen_r2cf.ml -1 +1 cannam@95: M ./genfft/gen_r2r.ml -1 +1 cannam@95: M ./genfft/gen_twiddle.ml -1 +1 cannam@95: M ./genfft/gen_twiddle_c.ml -1 +1 cannam@95: M ./genfft/gen_twidsq.ml -1 +1 cannam@95: M ./genfft/gen_twidsq_c.ml -1 +1 cannam@95: M ./genfft/genutil.ml -1 +1 cannam@95: M ./genfft/littlesimp.ml -1 +1 cannam@95: M ./genfft/littlesimp.mli -1 +1 cannam@95: M ./genfft/magic.ml -1 +1 cannam@95: M ./genfft/monads.ml -1 +1 cannam@95: M ./genfft/number.ml -1 +1 cannam@95: M ./genfft/number.mli -1 +1 cannam@95: M ./genfft/oracle.ml -1 +1 cannam@95: M ./genfft/oracle.mli -1 +1 cannam@95: M ./genfft/schedule.ml -1 +1 cannam@95: M ./genfft/schedule.mli -1 +1 cannam@95: M ./genfft/simd.ml -1 +1 cannam@95: M ./genfft/simd.mli -1 +1 cannam@95: M ./genfft/simdmagic.ml -1 +1 cannam@95: M ./genfft/to_alist.ml -1 +1 cannam@95: M ./genfft/to_alist.mli -1 +1 cannam@95: M ./genfft/trig.ml -1 +1 cannam@95: M ./genfft/trig.mli -1 +1 cannam@95: M ./genfft/twiddle.ml -1 +1 cannam@95: M ./genfft/twiddle.mli -1 +1 cannam@95: M ./genfft/unique.ml -1 +1 cannam@95: M ./genfft/unique.mli -1 +1 cannam@95: M ./genfft/util.ml -1 +1 cannam@95: M ./genfft/util.mli -1 +1 cannam@95: M ./genfft/variable.ml -1 +1 cannam@95: M ./genfft/variable.mli -1 +1 cannam@95: M ./kernel/align.c -1 +1 cannam@95: M ./kernel/alloc.c -1 +1 cannam@95: M ./kernel/assert.c -1 +1 cannam@95: M ./kernel/awake.c -1 +1 cannam@95: M ./kernel/buffered.c -1 +1 cannam@95: M ./kernel/cpy1d.c -1 +1 cannam@95: M ./kernel/cpy2d-pair.c -1 +1 cannam@95: M ./kernel/cpy2d.c -1 +1 cannam@95: M ./kernel/ct.c -1 +1 cannam@95: M ./kernel/debug.c -1 +1 cannam@95: M ./kernel/extract-reim.c -1 +1 cannam@95: M ./kernel/hash.c -1 +1 cannam@95: M ./kernel/iabs.c -1 +1 cannam@95: M ./kernel/ifftw.h -1 +1 cannam@95: M ./kernel/kalloc.c -1 +1 cannam@95: M ./kernel/md5-1.c -1 +1 cannam@95: M ./kernel/md5.c -1 +1 cannam@95: M ./kernel/minmax.c -1 +1 cannam@95: M ./kernel/ops.c -1 +1 cannam@95: M ./kernel/pickdim.c -1 +1 cannam@95: M ./kernel/plan.c -1 +1 cannam@95: M ./kernel/planner.c -1 +1 cannam@95: M ./kernel/primes.c -1 +1 cannam@95: M ./kernel/print.c -1 +1 cannam@95: M ./kernel/problem.c -1 +1 cannam@95: M ./kernel/rader.c -1 +1 cannam@95: M ./kernel/scan.c -1 +1 cannam@95: M ./kernel/solver.c -1 +1 cannam@95: M ./kernel/solvtab.c -1 +1 cannam@95: M ./kernel/stride.c -1 +1 cannam@95: M ./kernel/tensor.c -1 +1 cannam@95: M ./kernel/tensor1.c -1 +1 cannam@95: M ./kernel/tensor2.c -1 +1 cannam@95: M ./kernel/tensor3.c -1 +1 cannam@95: M ./kernel/tensor4.c -1 +1 cannam@95: M ./kernel/tensor5.c -1 +1 cannam@95: M ./kernel/tensor7.c -1 +1 cannam@95: M ./kernel/tensor8.c -1 +1 cannam@95: M ./kernel/tensor9.c -1 +1 cannam@95: M ./kernel/tile2d.c -1 +1 cannam@95: M ./kernel/timer.c -1 +1 cannam@95: M ./kernel/transpose.c -1 +1 cannam@95: M ./kernel/trig.c -1 +1 cannam@95: M ./kernel/twiddle.c -1 +1 cannam@95: M ./libbench2/aligned-main.c -1 +1 cannam@95: M ./libbench2/bench-main.c -1 +1 cannam@95: M ./libbench2/bench-user.h -1 +1 cannam@95: M ./libbench2/bench.h -1 +1 cannam@95: M ./libbench2/can-do.c -1 +1 cannam@95: M ./libbench2/dotens2.c -1 +1 cannam@95: M ./libbench2/info.c -1 +1 cannam@95: M ./libbench2/main.c -1 +1 cannam@95: M ./libbench2/my-getopt.c -1 +1 cannam@95: M ./libbench2/my-getopt.h -1 +1 cannam@95: M ./libbench2/problem.c -1 +1 cannam@95: M ./libbench2/report.c -1 +1 cannam@95: M ./libbench2/speed.c -1 +1 cannam@95: M ./libbench2/tensor.c -1 +1 cannam@95: M ./libbench2/timer.c -1 +1 cannam@95: M ./libbench2/useropt.c -1 +1 cannam@95: M ./libbench2/util.c -1 +1 cannam@95: M ./libbench2/verify-dft.c -1 +1 cannam@95: M ./libbench2/verify-lib.c -1 +1 cannam@95: M ./libbench2/verify-r2r.c -1 +1 cannam@95: M ./libbench2/verify-rdft2.c -1 +1 cannam@95: M ./libbench2/verify.c -1 +1 cannam@95: M ./libbench2/verify.h -1 +1 cannam@95: M ./libbench2/zero.c -1 +1 cannam@95: M ./mpi/any-true.c -1 +1 cannam@95: M ./mpi/api.c -1 +1 cannam@95: M ./mpi/block.c -1 +1 cannam@95: M ./mpi/choose-radix.c -1 +1 cannam@95: M ./mpi/conf.c -1 +1 cannam@95: M ./mpi/dft-problem.c -1 +1 cannam@95: M ./mpi/dft-rank-geq2-transposed.c -1 +1 cannam@95: M ./mpi/dft-rank-geq2.c -1 +1 cannam@95: M ./mpi/dft-rank1-bigvec.c -1 +1 cannam@95: M ./mpi/dft-rank1.c -1 +1 cannam@95: M ./mpi/dft-serial.c -1 +1 cannam@95: M ./mpi/dft-solve.c -1 +1 cannam@95: M ./mpi/dtensor.c -1 +1 cannam@95: M ./mpi/ifftw-mpi.h -1 +1 cannam@95: M ./mpi/mpi-dft.h -1 +1 cannam@95: M ./mpi/mpi-rdft.h -1 +1 cannam@95: M ./mpi/mpi-rdft2.h -1 +1 cannam@95: M ./mpi/mpi-transpose.h -1 +1 cannam@95: M ./mpi/rdft-problem.c -1 +1 cannam@95: M ./mpi/rdft-rank-geq2-transposed.c -1 +1 cannam@95: M ./mpi/rdft-rank-geq2.c -1 +1 cannam@95: M ./mpi/rdft-rank1-bigvec.c -1 +1 cannam@95: M ./mpi/rdft-serial.c -1 +1 cannam@95: M ./mpi/rdft-solve.c -1 +1 cannam@95: M ./mpi/rdft2-problem.c -1 +1 cannam@95: M ./mpi/rdft2-rank-geq2-transposed.c -1 +1 cannam@95: M ./mpi/rdft2-rank-geq2.c -1 +1 cannam@95: M ./mpi/rdft2-serial.c -1 +1 cannam@95: M ./mpi/rdft2-solve.c -1 +1 cannam@95: M ./mpi/rearrange.c -1 +1 cannam@95: M ./mpi/testsched.c -1 +1 cannam@95: M ./mpi/transpose-alltoall.c -1 +1 cannam@95: M ./mpi/transpose-pairwise.c -1 +1 cannam@95: M ./mpi/transpose-problem.c -1 +1 cannam@95: M ./mpi/transpose-recurse.c -1 +1 cannam@95: M ./mpi/transpose-solve.c -1 +1 cannam@95: M ./mpi/wisdom-api.c -1 +1 cannam@95: M ./rdft/buffered.c -1 +1 cannam@95: M ./rdft/buffered2.c -1 +1 cannam@95: M ./rdft/codelet-rdft.h -1 +1 cannam@95: M ./rdft/conf.c -1 +1 cannam@95: M ./rdft/ct-hc2c-direct.c -1 +1 cannam@95: M ./rdft/ct-hc2c.c -1 +1 cannam@95: M ./rdft/ct-hc2c.h -1 +1 cannam@95: M ./rdft/dft-r2hc.c -1 +1 cannam@95: M ./rdft/dht-r2hc.c -1 +1 cannam@95: M ./rdft/dht-rader.c -1 +1 cannam@95: M ./rdft/direct-r2c.c -1 +1 cannam@95: M ./rdft/direct-r2r.c -1 +1 cannam@95: M ./rdft/direct2.c -1 +1 cannam@95: M ./rdft/generic.c -1 +1 cannam@95: M ./rdft/hc2hc-direct.c -1 +1 cannam@95: M ./rdft/hc2hc-generic.c -1 +1 cannam@95: M ./rdft/hc2hc.c -1 +1 cannam@95: M ./rdft/hc2hc.h -1 +1 cannam@95: M ./rdft/indirect.c -1 +1 cannam@95: M ./rdft/khc2c.c -1 +1 cannam@95: M ./rdft/khc2hc.c -1 +1 cannam@95: M ./rdft/kr2c.c -1 +1 cannam@95: M ./rdft/kr2r.c -1 +1 cannam@95: M ./rdft/nop.c -1 +1 cannam@95: M ./rdft/nop2.c -1 +1 cannam@95: M ./rdft/plan.c -1 +1 cannam@95: M ./rdft/plan2.c -1 +1 cannam@95: M ./rdft/problem.c -1 +1 cannam@95: M ./rdft/problem2.c -1 +1 cannam@95: M ./rdft/rank-geq2-rdft2.c -1 +1 cannam@95: M ./rdft/rank-geq2.c -1 +1 cannam@95: M ./rdft/rank0-rdft2.c -1 +1 cannam@95: M ./rdft/rank0.c -1 +1 cannam@95: M ./rdft/rdft-dht.c -1 +1 cannam@95: M ./rdft/rdft.h -1 +1 cannam@95: M ./rdft/rdft2-inplace-strides.c -1 +1 cannam@95: M ./rdft/rdft2-rdft.c -1 +1 cannam@95: M ./rdft/rdft2-strides.c -1 +1 cannam@95: M ./rdft/rdft2-tensor-max-index.c -1 +1 cannam@95: M ./rdft/scalar/hb.h -1 +1 cannam@95: M ./rdft/scalar/hc2c.c -1 +1 cannam@95: M ./rdft/scalar/hc2cb.h -1 +1 cannam@95: M ./rdft/scalar/hc2cf.h -1 +1 cannam@95: M ./rdft/scalar/hf.h -1 +1 cannam@95: M ./rdft/scalar/hfb.c -1 +1 cannam@95: M ./rdft/scalar/r2c.c -1 +1 cannam@95: M ./rdft/scalar/r2cb.h -1 +1 cannam@95: M ./rdft/scalar/r2cbIII.h -1 +1 cannam@95: M ./rdft/scalar/r2cf.h -1 +1 cannam@95: M ./rdft/scalar/r2cfII.h -1 +1 cannam@95: M ./rdft/scalar/r2r.c -1 +1 cannam@95: M ./rdft/scalar/r2r.h -1 +1 cannam@95: M ./rdft/simd/common/genus.c -1 +1 cannam@95: M ./rdft/simd/hc2cbv.h -1 +1 cannam@95: M ./rdft/simd/hc2cfv.h -1 +1 cannam@95: M ./rdft/solve.c -1 +1 cannam@95: M ./rdft/solve2.c -1 +1 cannam@95: M ./rdft/vrank-geq1-rdft2.c -1 +1 cannam@95: M ./rdft/vrank-geq1.c -1 +1 cannam@95: M ./rdft/vrank3-transpose.c -1 +1 cannam@95: M ./reodft/conf.c -1 +1 cannam@95: M ./reodft/redft00e-r2hc-pad.c -1 +1 cannam@95: M ./reodft/redft00e-r2hc.c -1 +1 cannam@95: M ./reodft/reodft.h -1 +1 cannam@95: M ./reodft/reodft00e-splitradix.c -1 +1 cannam@95: M ./reodft/reodft010e-r2hc.c -1 +1 cannam@95: M ./reodft/reodft11e-r2hc-odd.c -1 +1 cannam@95: M ./reodft/reodft11e-r2hc.c -1 +1 cannam@95: M ./reodft/reodft11e-radix2.c -1 +1 cannam@95: M ./reodft/rodft00e-r2hc-pad.c -1 +1 cannam@95: M ./reodft/rodft00e-r2hc.c -1 +1 cannam@95: M ./simd-support/altivec.c -1 +1 cannam@95: M ./simd-support/amd64-cpuid.h -1 +1 cannam@95: M ./simd-support/avx.c -1 +1 cannam@95: M ./simd-support/neon.c -1 +1 cannam@95: M ./simd-support/simd-altivec.h -1 +1 cannam@95: M ./simd-support/simd-avx.h -1 +1 cannam@95: M ./simd-support/simd-common.h -1 +1 cannam@95: M ./simd-support/simd-neon.h -1 +1 cannam@95: M ./simd-support/simd-sse2.h -1 +1 cannam@95: M ./simd-support/sse2-nonportable.c -1 +1 cannam@95: M ./simd-support/sse2.c -1 +1 cannam@95: M ./simd-support/taint.c -1 +1 cannam@95: M ./simd-support/x86-cpuid.h -1 +1 cannam@95: M ./simd/altivec.c -1 +1 cannam@95: M ./simd/mips_ps.c -1 +1 cannam@95: M ./simd/mips_ps.h -1 +1 cannam@95: M ./simd/nonportable/sse.c -1 +1 cannam@95: M ./simd/nonportable/sse2.c -1 +1 cannam@95: M ./simd/simd-altivec.h -1 +1 cannam@95: M ./simd/simd-mips_ps.h -1 +1 cannam@95: M ./simd/simd-sse.h -1 +1 cannam@95: M ./simd/simd-sse2.h -1 +1 cannam@95: M ./simd/simd.h -1 +1 cannam@95: M ./simd/sse.c -1 +1 cannam@95: M ./simd/sse2.c -1 +1 cannam@95: M ./simd/taint.c -1 +1 cannam@95: M ./simd/x86-cpuid.h -1 +1 cannam@95: M ./threads/api.c -1 +1 cannam@95: M ./threads/conf.c -1 +1 cannam@95: M ./threads/ct.c -1 +1 cannam@95: M ./threads/dft-vrank-geq1.c -1 +1 cannam@95: M ./threads/f77api.c -1 +1 cannam@95: M ./threads/f77funcs.h -1 +1 cannam@95: M ./threads/hc2hc.c -1 +1 cannam@95: M ./threads/openmp.c -1 +1 cannam@95: M ./threads/rdft-vrank-geq1.c -1 +1 cannam@95: M ./threads/threads.c -1 +1 cannam@95: M ./threads/threads.h -1 +1 cannam@95: M ./threads/vrank-geq1-rdft2.c -1 +1 cannam@95: M ./tools/fftw-wisdom-to-conf.1 -1 +1 cannam@95: M ./tools/fftw-wisdom-to-conf.in -1 +1 cannam@95: M ./tools/fftw-wisdom.c -1 +1 cannam@95: M ./tools/fftw_wisdom.1.in -1 +1 cannam@95: cannam@95: Thu Aug 11 14:54:38 EDT 2011 stevenj@fftw.org cannam@95: * allow specifying TRANSPOSE_{IN/OUT} transpose plans, since libbench does not canonicalize rnk=1 n=1 plans as rnk=0 cannam@95: cannam@95: M ./mpi/mpi-bench.c -2 +2 cannam@95: cannam@95: Thu Aug 11 14:17:24 EDT 2011 stevenj@fftw.org cannam@95: * check.pl should occasionally check DESTROY_INPUT problems too (especially since those enable slightly different algorithms in MPI) cannam@95: cannam@95: M ./tests/check.pl +1 cannam@95: cannam@95: Thu Aug 11 12:37:51 EDT 2011 stevenj@fftw.org cannam@95: * unify post-MPI transpose handling in pairwise and alltoall solvers; should make the former faster in the destroy-input out-of-place case, and the latter more widely applicable cannam@95: cannam@95: M ./mpi/mpi-transpose.h +5 cannam@95: M ./mpi/transpose-alltoall.c -69 +24 cannam@95: M ./mpi/transpose-pairwise.c -66 +97 cannam@95: cannam@95: Mon Aug 8 10:06:14 EDT 2011 athena@fftw.org cannam@95: * Add support for ARM NEON cannam@95: cannam@95: M ./Makefile.am -1 +6 cannam@95: M ./api/version.c +4 cannam@95: M ./configure.ac -5 +23 cannam@95: M ./dft/codelet-dft.h +1 cannam@95: M ./dft/conf.c +4 cannam@95: M ./dft/simd/Makefile.am -1 +1 cannam@95: A ./dft/simd/neon/ cannam@95: A ./dft/simd/neon/Makefile.am cannam@95: M ./kernel/ifftw.h +2 cannam@95: M ./m4/ax_cc_maxopt.m4 +10 cannam@95: M ./rdft/codelet-rdft.h +1 cannam@95: M ./rdft/conf.c +4 cannam@95: M ./rdft/simd/Makefile.am -1 +1 cannam@95: A ./rdft/simd/neon/ cannam@95: A ./rdft/simd/neon/Makefile.am cannam@95: M ./simd-support/Makefile.am -1 +2 cannam@95: A ./simd-support/neon.c cannam@95: M ./simd-support/simd-common.h +3 cannam@95: A ./simd-support/simd-neon.h cannam@95: cannam@95: Fri Aug 5 17:25:32 EDT 2011 stevenj@fftw.org cannam@95: * more C++ paranoia cannam@95: cannam@95: M ./tests/fftw-bench.h +10 cannam@95: cannam@95: Fri Aug 5 17:02:00 EDT 2011 stevenj@fftw.org cannam@95: * tentative version bump for 3.3.1 cannam@95: cannam@95: M ./NEWS +5 cannam@95: M ./configure.ac -2 +2 cannam@95: cannam@95: Fri Aug 5 16:52:28 EDT 2011 stevenj@fftw.org cannam@95: * fixes so that MPI code compiles when MPICC is a C++ compiler, even if the serial code is compiled with a C compiler; thanks to Kyle Spyksma for the bug report cannam@95: cannam@95: M ./api/api.h +12 cannam@95: M ./api/plan-guru-r2r.h -2 cannam@95: M ./api/plan-many-r2r.c -2 cannam@95: M ./dft/dft.h +9 cannam@95: M ./kernel/ifftw.h +9 cannam@95: M ./mpi/api.c -3 +1 cannam@95: M ./mpi/dtensor.c -1 +1 cannam@95: M ./mpi/ifftw-mpi.h -1 +4 cannam@95: M ./mpi/mpi-bench.c -1 +1 cannam@95: M ./rdft/rdft.h +9 cannam@95: cannam@95: Fri Aug 5 16:04:06 EDT 2011 stevenj@fftw.org cannam@95: * use correct precision in f03-wrap.c, avoiding a (harmless) implicit pointer cast that prevented compilation under C++; thanks to Kyle Spyksma for the bug report cannam@95: cannam@95: M ./mpi/genf03-wrap.pl +2 cannam@95: cannam@95: Fri Aug 5 14:04:32 EDT 2011 stevenj@fftw.org cannam@95: * manual typo cannam@95: cannam@95: M ./doc/modern-fortran.texi -1 +1 cannam@95: cannam@95: Wed Jul 27 12:50:25 EDT 2011 athena@fftw.org cannam@95: tagged fftw-3.3 cannam@95: cannam@95: cannam@95: Tue Jul 26 20:55:45 EDT 2011 athena@fftw.org cannam@95: * Honor WITH_OUR_MALLOC in libbench2 cannam@95: cannam@95: M ./libbench2/util.c -6 +5 cannam@95: cannam@95: Tue Jul 26 20:27:28 EDT 2011 athena@fftw.org cannam@95: * fixed typo: incorrect name of combined threads library on Windows cannam@95: cannam@95: M ./Makefile.am -1 +1 cannam@95: cannam@95: Mon Jul 25 14:38:20 EDT 2011 stevenj@fftw.org cannam@95: * 3.3 version bump & NEWS cannam@95: cannam@95: M ./NEWS -5 +16 cannam@95: M ./configure.ac -1 +1 cannam@95: cannam@95: Mon Jul 25 14:37:48 EDT 2011 stevenj@fftw.org cannam@95: * use int(..., C_SIZE_T) rather than declaring another variable in the Fortran examples cannam@95: cannam@95: M ./doc/modern-fortran.texi -12 +9 cannam@95: cannam@95: Wed Jul 13 05:02:32 EDT 2011 stevenj@fftw.org cannam@95: * typo, thanks to Rhys Ulerich for the comment cannam@95: cannam@95: M ./doc/mpi.texi -1 +1 cannam@95: cannam@95: Mon Jul 11 14:39:52 EDT 2011 athena@fftw.org cannam@95: * Fix bug in bubblesort cannam@95: cannam@95: Bubblesort was not sorting. This was a bug in the benchmark library cannam@95: (not in FFTW per se), and it impacted the benchmark program cannam@95: with --report-time and --report-mflops causing it to output cannam@95: an incorrect value for the median. (The minimum, maximum, and cannam@95: average value were correct.) Thanks Dima Baksheev of Intel for cannam@95: reporting this bug. cannam@95: cannam@95: cannam@95: M ./libbench2/report.c -2 +2 cannam@95: cannam@95: Fri Jul 8 13:35:59 EDT 2011 stevenj@fftw.org cannam@95: * small manual typos cannam@95: cannam@95: M ./doc/modern-fortran.texi -2 +2 cannam@95: cannam@95: Wed Jul 6 10:49:40 EDT 2011 athena@fftw.org cannam@95: * Detection of altivec.h requires $ALTIVEC_CFLAGS cannam@95: cannam@95: M ./configure.ac -5 +4 cannam@95: cannam@95: Tue Jul 5 19:58:47 EDT 2011 athena@fftw.org cannam@95: * Introduce fake dependency so that my-getopt.c is recompiled cannam@95: cannam@95: my-getopt.c does not depend on anything, and so it is not rebuilt when cannam@95: reconfiguring for a different ISA (e.g., CC="gcc -m32" vs CC="gcc cannam@95: -m64"). Add a fake dependency on so that the file is cannam@95: recompiled. cannam@95: cannam@95: cannam@95: cannam@95: M ./libbench2/my-getopt.c +1 cannam@95: cannam@95: Tue Jul 5 18:53:36 EDT 2011 stevenj@fftw.org cannam@95: * support compiling/installing --enable-threads --enable-openmp at the same time, although in this case the test program only uses the threads variety. Update documentation accordingly, and in general expand the documentation of the OpenMP support cannam@95: cannam@95: M ./Makefile.am -1 +1 cannam@95: M ./configure.ac -21 +13 cannam@95: M ./doc/install.texi -14 +13 cannam@95: M ./doc/threads.texi -19 +43 cannam@95: M ./mpi/Makefile.am -2 +6 cannam@95: M ./tests/Makefile.am -2 +6 cannam@95: M ./tests/fftw-bench.c -1 +3 cannam@95: M ./threads/Makefile.am -6 +20 cannam@95: M ./threads/conf.c -4 cannam@95: M ./threads/openmp.c -4 cannam@95: M ./threads/threads.c -3 cannam@95: M ./tools/Makefile.am -4 +8 cannam@95: cannam@95: Tue Jul 5 16:04:03 EDT 2011 stevenj@fftw.org cannam@95: * call omp_set_num_threads in fftw-bench so that the number of OpenMP threads corresponds with the number of FFTW threads cannam@95: cannam@95: M ./tests/fftw-bench.c +5 cannam@95: cannam@95: Tue Jul 5 16:03:06 EDT 2011 stevenj@fftw.org cannam@95: * when --enable-openmp, install as fftw3_omp rather than fftw3_threads, so that both the POSIX threads and OpenMP variants of FFTW can be installed at once cannam@95: cannam@95: M ./Makefile.am -1 +1 cannam@95: M ./configure.ac +4 cannam@95: M ./threads/Makefile.am -4 +4 cannam@95: cannam@95: Sat Jul 2 02:21:22 EDT 2011 stevenj@fftw.org cannam@95: * don't even declare an fftw_execute interface in Fortran, since it is unsafe and we recommend against it anyway; thanks to Arjen Markus for the suggestion cannam@95: cannam@95: M ./api/f03api.sh -1 +1 cannam@95: M ./doc/modern-fortran.texi -2 +4 cannam@95: cannam@95: Fri Jul 1 14:35:44 EDT 2011 athena@fftw.org cannam@95: * consistently use the order single, double, long double cannam@95: cannam@95: M ./support/BUILD-MINGW32.sh -1 +1 cannam@95: M ./support/BUILD-MINGW64.sh -1 +1 cannam@95: cannam@95: Wed Jun 29 17:27:06 EDT 2011 athena@fftw.org cannam@95: * MSVC AVX 64-bit detection does not work, punt for now. cannam@95: cannam@95: M ./simd-support/amd64-cpuid.h -16 +2 cannam@95: cannam@95: Wed Jun 29 15:52:27 EDT 2011 stevenj@fftw.org cannam@95: * fixed typo, added note on transposed flags for r2c/c2r; thanks to Rhys Ulerich for the suggestions cannam@95: cannam@95: M ./doc/mpi.texi -1 +14 cannam@95: cannam@95: Wed Jun 29 09:41:39 EDT 2011 athena@fftw.org cannam@95: * fixes for compiling with MSVC (untested) cannam@95: cannam@95: M ./simd-support/amd64-cpuid.h -2 +16 cannam@95: M ./simd-support/simd-avx.h -11 +16 cannam@95: M ./simd-support/x86-cpuid.h -1 +7 cannam@95: cannam@95: Tue Jun 28 16:48:36 EDT 2011 athena@fftw.org cannam@95: * comment cannam@95: cannam@95: M ./simd-support/sse2-nonportable.c +3 cannam@95: cannam@95: Mon Jun 27 21:01:56 EDT 2011 stevenj@fftw.org cannam@95: * rm extraneous line break in HTML output ... I hate texinfo cannam@95: cannam@95: M ./doc/reference.texi -6 +3 cannam@95: cannam@95: Mon Jun 27 06:43:57 EDT 2011 athena@fftw.org cannam@95: tagged fftw-3.3-beta1 cannam@95: cannam@95: cannam@95: Mon Jun 27 00:47:33 EDT 2011 stevenj@fftw.org cannam@95: * maintainer-clean should delete html directory (otherwwise we keep obsolete HTML files in the dist tarball, sigh) cannam@95: cannam@95: M ./doc/Makefile.am -2 +2 cannam@95: cannam@95: Mon Jun 27 00:45:28 EDT 2011 stevenj@fftw.org cannam@95: tagged fftw-3.3-beta1 cannam@95: cannam@95: cannam@95: Sun Jun 26 23:36:32 EDT 2011 stevenj@fftw.org cannam@95: * update copyright year in manual cannam@95: cannam@95: M ./doc/license.texi -2 +2 cannam@95: cannam@95: Sun Jun 26 22:52:54 EDT 2011 stevenj@fftw.org cannam@95: * whoops, don't dist .f03 headers, since those are built by the user's Makefile cannam@95: cannam@95: M ./api/Makefile.am -1 +2 cannam@95: M ./mpi/Makefile.am -1 +2 cannam@95: cannam@95: Sun Jun 26 22:43:49 EDT 2011 stevenj@fftw.org cannam@95: * fix embarrassing deadlock/crashing bug in my previous nowisdom_hook fix -- I forgot to handle the case where one process has wisdom and another one doesn't, requiring a nowisdom_hook in the latter case; this should only affect MPI transforms since otherwise these hook functions are NULL cannam@95: cannam@95: M ./kernel/ifftw.h +1 cannam@95: M ./kernel/planner.c -44 +48 cannam@95: M ./mpi/api.c +14 cannam@95: cannam@95: Sun Jun 26 21:02:15 EDT 2011 stevenj@fftw.org cannam@95: * subsubheadings, MPI transpose reference cannam@95: cannam@95: M ./doc/mpi.texi -7 +53 cannam@95: cannam@95: Sun Jun 26 20:48:53 EDT 2011 stevenj@fftw.org cannam@95: * add MPI plan reference cannam@95: cannam@95: M ./doc/mpi.texi -5 +197 cannam@95: cannam@95: Sun Jun 26 17:07:21 EDT 2011 stevenj@fftw.org cannam@95: * portions of MPI reference docs; tweaks to NEWS cannam@95: cannam@95: M ./NEWS -2 +3 cannam@95: M ./doc/mpi.texi -12 +264 cannam@95: cannam@95: Sun Jun 26 12:40:43 EDT 2011 stevenj@fftw.org cannam@95: * use $(CHECK_PL_OPTS) more consistently cannam@95: cannam@95: M ./tests/Makefile.am -1 +1 cannam@95: cannam@95: Sun Jun 26 10:04:54 EDT 2011 athena@fftw.org cannam@95: * accept \r\n as well as \n. Grrr... cannam@95: cannam@95: M ./tests/check.pl -1 +2 cannam@95: cannam@95: Sun Jun 26 09:52:11 EDT 2011 athena@fftw.org cannam@95: * new configure option --with-incoming-stack-boundary=N cannam@95: cannam@95: This option selects CFLAGS to align the stack at all externally-callable cannam@95: functions. This currently comprises api/* and threads/* cannam@95: cannam@95: cannam@95: M ./api/Makefile.am +1 cannam@95: M ./configure.ac +14 cannam@95: M ./support/BUILD-MINGW32.sh -1 +1 cannam@95: M ./threads/Makefile.am +1 cannam@95: cannam@95: Sun Jun 26 09:51:37 EDT 2011 athena@fftw.org cannam@95: * add -fomit-frame-pointer back cannam@95: cannam@95: Somehow -O3 does not imply -fomit-frame-pointer on ia32 cannam@95: cannam@95: cannam@95: M ./m4/ax_cc_maxopt.m4 -2 +3 cannam@95: cannam@95: Sun Jun 26 07:20:27 EDT 2011 athena@fftw.org cannam@95: * Note that removal of mips-ps is temporary. cannam@95: cannam@95: M ./NEWS -3 +5 cannam@95: cannam@95: Sat Jun 25 23:15:03 EDT 2011 stevenj@fftw.org cannam@95: * update copyright year cannam@95: cannam@95: M ./COPYRIGHT -2 +2 cannam@95: M ./api/api.h -2 +2 cannam@95: M ./api/apiplan.c -2 +2 cannam@95: M ./api/configure.c -2 +2 cannam@95: M ./api/execute-dft-c2r.c -2 +2 cannam@95: M ./api/execute-dft-r2c.c -2 +2 cannam@95: M ./api/execute-dft.c -2 +2 cannam@95: M ./api/execute-r2r.c -2 +2 cannam@95: M ./api/execute-split-dft-c2r.c -2 +2 cannam@95: M ./api/execute-split-dft-r2c.c -2 +2 cannam@95: M ./api/execute-split-dft.c -2 +2 cannam@95: M ./api/execute.c -2 +2 cannam@95: M ./api/export-wisdom-to-file.c -2 +2 cannam@95: M ./api/export-wisdom-to-string.c -2 +2 cannam@95: M ./api/export-wisdom.c -2 +2 cannam@95: M ./api/f77api.c -2 +2 cannam@95: M ./api/f77funcs.h -2 +2 cannam@95: M ./api/fftw3.h -2 +2 cannam@95: M ./api/flops.c -2 +2 cannam@95: M ./api/forget-wisdom.c -2 +2 cannam@95: M ./api/import-system-wisdom.c -2 +2 cannam@95: M ./api/import-wisdom-from-file.c -2 +2 cannam@95: M ./api/import-wisdom-from-string.c -2 +2 cannam@95: M ./api/import-wisdom.c -2 +2 cannam@95: M ./api/malloc.c -2 +2 cannam@95: M ./api/map-r2r-kind.c -2 +2 cannam@95: M ./api/mapflags.c -2 +2 cannam@95: M ./api/mkprinter-file.c -2 +2 cannam@95: M ./api/mktensor-iodims.h -2 +2 cannam@95: M ./api/mktensor-rowmajor.c -2 +2 cannam@95: M ./api/plan-dft-1d.c -2 +2 cannam@95: M ./api/plan-dft-2d.c -2 +2 cannam@95: M ./api/plan-dft-3d.c -2 +2 cannam@95: M ./api/plan-dft-c2r-1d.c -2 +2 cannam@95: M ./api/plan-dft-c2r-2d.c -2 +2 cannam@95: M ./api/plan-dft-c2r-3d.c -2 +2 cannam@95: M ./api/plan-dft-c2r.c -2 +2 cannam@95: M ./api/plan-dft-r2c-1d.c -2 +2 cannam@95: M ./api/plan-dft-r2c-2d.c -2 +2 cannam@95: M ./api/plan-dft-r2c-3d.c -2 +2 cannam@95: M ./api/plan-dft-r2c.c -2 +2 cannam@95: M ./api/plan-dft.c -2 +2 cannam@95: M ./api/plan-guru-dft-c2r.h -2 +2 cannam@95: M ./api/plan-guru-dft-r2c.h -2 +2 cannam@95: M ./api/plan-guru-dft.h -2 +2 cannam@95: M ./api/plan-guru-r2r.h -2 +2 cannam@95: M ./api/plan-guru-split-dft-c2r.h -2 +2 cannam@95: M ./api/plan-guru-split-dft-r2c.h -2 +2 cannam@95: M ./api/plan-guru-split-dft.h -2 +2 cannam@95: M ./api/plan-many-dft-c2r.c -2 +2 cannam@95: M ./api/plan-many-dft-r2c.c -2 +2 cannam@95: M ./api/plan-many-dft.c -2 +2 cannam@95: M ./api/plan-many-r2r.c -2 +2 cannam@95: M ./api/plan-r2r-1d.c -2 +2 cannam@95: M ./api/plan-r2r-2d.c -2 +2 cannam@95: M ./api/plan-r2r-3d.c -2 +2 cannam@95: M ./api/plan-r2r.c -2 +2 cannam@95: M ./api/print-plan.c -2 +2 cannam@95: M ./api/rdft2-pad.c -2 +2 cannam@95: M ./api/the-planner.c -2 +2 cannam@95: M ./api/version.c -2 +2 cannam@95: M ./api/x77.h -2 +2 cannam@95: M ./dft/bluestein.c -2 +2 cannam@95: M ./dft/buffered.c -2 +2 cannam@95: M ./dft/codelet-dft.h -2 +2 cannam@95: M ./dft/conf.c -2 +2 cannam@95: M ./dft/ct.c -2 +2 cannam@95: M ./dft/ct.h -2 +2 cannam@95: M ./dft/dft.h -2 +2 cannam@95: M ./dft/dftw-direct.c -2 +2 cannam@95: M ./dft/dftw-directsq.c -2 +2 cannam@95: M ./dft/dftw-generic.c -2 +2 cannam@95: M ./dft/dftw-genericbuf.c -2 +2 cannam@95: M ./dft/direct.c -2 +2 cannam@95: M ./dft/generic.c -2 +2 cannam@95: M ./dft/indirect-transpose.c -2 +2 cannam@95: M ./dft/indirect.c -2 +2 cannam@95: M ./dft/kdft-dif.c -2 +2 cannam@95: M ./dft/kdft-difsq.c -2 +2 cannam@95: M ./dft/kdft-dit.c -2 +2 cannam@95: M ./dft/kdft.c -2 +2 cannam@95: M ./dft/nop.c -2 +2 cannam@95: M ./dft/plan.c -2 +2 cannam@95: M ./dft/problem.c -2 +2 cannam@95: M ./dft/rader.c -2 +2 cannam@95: M ./dft/rank-geq2.c -2 +2 cannam@95: M ./dft/scalar/n.c -2 +2 cannam@95: M ./dft/scalar/n.h -2 +2 cannam@95: M ./dft/scalar/t.c -2 +2 cannam@95: M ./dft/scalar/t.h -2 +2 cannam@95: M ./dft/simd/common/genus.c -2 +2 cannam@95: M ./dft/simd/n1b.h -2 +2 cannam@95: M ./dft/simd/n1f.h -2 +2 cannam@95: M ./dft/simd/n2b.h -2 +2 cannam@95: M ./dft/simd/n2f.h -2 +2 cannam@95: M ./dft/simd/n2s.h -2 +2 cannam@95: M ./dft/simd/q1b.h -2 +2 cannam@95: M ./dft/simd/q1f.h -2 +2 cannam@95: M ./dft/simd/t1b.h -2 +2 cannam@95: M ./dft/simd/t1bu.h -2 +2 cannam@95: M ./dft/simd/t1f.h -2 +2 cannam@95: M ./dft/simd/t1fu.h -2 +2 cannam@95: M ./dft/simd/t2b.h -2 +2 cannam@95: M ./dft/simd/t2f.h -2 +2 cannam@95: M ./dft/simd/t3b.h -2 +2 cannam@95: M ./dft/simd/t3f.h -2 +2 cannam@95: M ./dft/simd/ts.h -2 +2 cannam@95: M ./dft/solve.c -2 +2 cannam@95: M ./dft/vrank-geq1.c -2 +2 cannam@95: M ./dft/zero.c -2 +2 cannam@95: M ./doc/f77_wisdom.f -2 +2 cannam@95: M ./genfft/algsimp.ml -2 +2 cannam@95: M ./genfft/algsimp.mli -2 +2 cannam@95: M ./genfft/annotate.ml -2 +2 cannam@95: M ./genfft/annotate.mli -2 +2 cannam@95: M ./genfft/assoctable.ml -2 +2 cannam@95: M ./genfft/assoctable.mli -2 +2 cannam@95: M ./genfft/c.ml -2 +2 cannam@95: M ./genfft/c.mli -2 +2 cannam@95: M ./genfft/complex.ml -2 +2 cannam@95: M ./genfft/complex.mli -2 +2 cannam@95: M ./genfft/conv.ml -2 +2 cannam@95: M ./genfft/conv.mli -2 +2 cannam@95: M ./genfft/dag.ml -2 +2 cannam@95: M ./genfft/dag.mli -2 +2 cannam@95: M ./genfft/expr.ml -2 +2 cannam@95: M ./genfft/expr.mli -2 +2 cannam@95: M ./genfft/fft.ml -2 +2 cannam@95: M ./genfft/fft.mli -2 +2 cannam@95: M ./genfft/gen_hc2c.ml -2 +2 cannam@95: M ./genfft/gen_hc2cdft.ml -2 +2 cannam@95: M ./genfft/gen_hc2cdft_c.ml -2 +2 cannam@95: M ./genfft/gen_hc2hc.ml -2 +2 cannam@95: M ./genfft/gen_mdct.ml -2 +2 cannam@95: M ./genfft/gen_notw.ml -2 +2 cannam@95: M ./genfft/gen_notw_c.ml -2 +2 cannam@95: M ./genfft/gen_r2cb.ml -2 +2 cannam@95: M ./genfft/gen_r2cf.ml -2 +2 cannam@95: M ./genfft/gen_r2r.ml -2 +2 cannam@95: M ./genfft/gen_twiddle.ml -2 +2 cannam@95: M ./genfft/gen_twiddle_c.ml -2 +2 cannam@95: M ./genfft/gen_twidsq.ml -2 +2 cannam@95: M ./genfft/gen_twidsq_c.ml -2 +2 cannam@95: M ./genfft/genutil.ml -2 +2 cannam@95: M ./genfft/littlesimp.ml -2 +2 cannam@95: M ./genfft/littlesimp.mli -2 +2 cannam@95: M ./genfft/magic.ml -2 +2 cannam@95: M ./genfft/monads.ml -2 +2 cannam@95: M ./genfft/number.ml -2 +2 cannam@95: M ./genfft/number.mli -2 +2 cannam@95: M ./genfft/oracle.ml -2 +2 cannam@95: M ./genfft/oracle.mli -2 +2 cannam@95: M ./genfft/schedule.ml -2 +2 cannam@95: M ./genfft/schedule.mli -2 +2 cannam@95: M ./genfft/simd.ml -2 +2 cannam@95: M ./genfft/simd.mli -2 +2 cannam@95: M ./genfft/simdmagic.ml -2 +2 cannam@95: M ./genfft/to_alist.ml -2 +2 cannam@95: M ./genfft/to_alist.mli -2 +2 cannam@95: M ./genfft/trig.ml -2 +2 cannam@95: M ./genfft/trig.mli -2 +2 cannam@95: M ./genfft/twiddle.ml -2 +2 cannam@95: M ./genfft/twiddle.mli -2 +2 cannam@95: M ./genfft/unique.ml -2 +2 cannam@95: M ./genfft/unique.mli -2 +2 cannam@95: M ./genfft/util.ml -2 +2 cannam@95: M ./genfft/util.mli -2 +2 cannam@95: M ./genfft/variable.ml -2 +2 cannam@95: M ./genfft/variable.mli -2 +2 cannam@95: M ./kernel/align.c -2 +2 cannam@95: M ./kernel/alloc.c -2 +2 cannam@95: M ./kernel/assert.c -2 +2 cannam@95: M ./kernel/awake.c -2 +2 cannam@95: M ./kernel/buffered.c -2 +2 cannam@95: M ./kernel/cpy1d.c -2 +2 cannam@95: M ./kernel/cpy2d-pair.c -2 +2 cannam@95: M ./kernel/cpy2d.c -2 +2 cannam@95: M ./kernel/ct.c -2 +2 cannam@95: M ./kernel/cycle.h -2 +2 cannam@95: M ./kernel/debug.c -2 +2 cannam@95: M ./kernel/extract-reim.c -2 +2 cannam@95: M ./kernel/hash.c -2 +2 cannam@95: M ./kernel/iabs.c -2 +2 cannam@95: M ./kernel/ifftw.h -2 +2 cannam@95: M ./kernel/kalloc.c -2 +2 cannam@95: M ./kernel/md5-1.c -2 +2 cannam@95: M ./kernel/md5.c -2 +2 cannam@95: M ./kernel/minmax.c -2 +2 cannam@95: M ./kernel/ops.c -2 +2 cannam@95: M ./kernel/pickdim.c -2 +2 cannam@95: M ./kernel/plan.c -2 +2 cannam@95: M ./kernel/primes.c -2 +2 cannam@95: M ./kernel/print.c -2 +2 cannam@95: M ./kernel/problem.c -2 +2 cannam@95: M ./kernel/rader.c -2 +2 cannam@95: M ./kernel/scan.c -2 +2 cannam@95: M ./kernel/solver.c -2 +2 cannam@95: M ./kernel/solvtab.c -2 +2 cannam@95: M ./kernel/stride.c -2 +2 cannam@95: M ./kernel/tensor.c -2 +2 cannam@95: M ./kernel/tensor1.c -2 +2 cannam@95: M ./kernel/tensor2.c -2 +2 cannam@95: M ./kernel/tensor3.c -2 +2 cannam@95: M ./kernel/tensor4.c -2 +2 cannam@95: M ./kernel/tensor5.c -2 +2 cannam@95: M ./kernel/tensor7.c -2 +2 cannam@95: M ./kernel/tensor8.c -2 +2 cannam@95: M ./kernel/tensor9.c -2 +2 cannam@95: M ./kernel/tile2d.c -2 +2 cannam@95: M ./kernel/timer.c -2 +2 cannam@95: M ./kernel/transpose.c -2 +2 cannam@95: M ./kernel/trig.c -2 +2 cannam@95: M ./kernel/twiddle.c -2 +2 cannam@95: M ./libbench2/dotens2.c -2 +2 cannam@95: M ./libbench2/my-getopt.c -2 +2 cannam@95: M ./libbench2/my-getopt.h -2 +2 cannam@95: M ./libbench2/verify-dft.c -2 +2 cannam@95: M ./libbench2/verify-lib.c -2 +2 cannam@95: M ./libbench2/verify-r2r.c -2 +2 cannam@95: M ./libbench2/verify-rdft2.c -2 +2 cannam@95: M ./libbench2/verify.h -2 +2 cannam@95: M ./mpi/any-true.c -2 +2 cannam@95: M ./mpi/api.c -2 +2 cannam@95: M ./mpi/block.c -2 +2 cannam@95: M ./mpi/choose-radix.c -2 +2 cannam@95: M ./mpi/conf.c -2 +2 cannam@95: M ./mpi/dft-problem.c -2 +2 cannam@95: M ./mpi/dft-rank-geq2-transposed.c -2 +2 cannam@95: M ./mpi/dft-rank-geq2.c -2 +2 cannam@95: M ./mpi/dft-rank1-bigvec.c -2 +2 cannam@95: M ./mpi/dft-rank1.c -2 +2 cannam@95: M ./mpi/dft-serial.c -2 +2 cannam@95: M ./mpi/dft-solve.c -2 +2 cannam@95: M ./mpi/dtensor.c -2 +2 cannam@95: M ./mpi/fftw3-mpi.h -2 +2 cannam@95: M ./mpi/ifftw-mpi.h -2 +2 cannam@95: M ./mpi/mpi-dft.h -2 +2 cannam@95: M ./mpi/mpi-rdft.h -2 +2 cannam@95: M ./mpi/mpi-rdft2.h -2 +2 cannam@95: M ./mpi/mpi-transpose.h -2 +2 cannam@95: M ./mpi/rdft-problem.c -2 +2 cannam@95: M ./mpi/rdft-rank-geq2-transposed.c -2 +2 cannam@95: M ./mpi/rdft-rank-geq2.c -2 +2 cannam@95: M ./mpi/rdft-rank1-bigvec.c -2 +2 cannam@95: M ./mpi/rdft-serial.c -2 +2 cannam@95: M ./mpi/rdft-solve.c -2 +2 cannam@95: M ./mpi/rdft2-problem.c -2 +2 cannam@95: M ./mpi/rdft2-rank-geq2-transposed.c -2 +2 cannam@95: M ./mpi/rdft2-rank-geq2.c -2 +2 cannam@95: M ./mpi/rdft2-serial.c -2 +2 cannam@95: M ./mpi/rdft2-solve.c -2 +2 cannam@95: M ./mpi/rearrange.c -2 +2 cannam@95: M ./mpi/testsched.c -1 +1 cannam@95: M ./mpi/transpose-alltoall.c -2 +2 cannam@95: M ./mpi/transpose-pairwise.c -2 +2 cannam@95: M ./mpi/transpose-problem.c -2 +2 cannam@95: M ./mpi/transpose-recurse.c -2 +2 cannam@95: M ./mpi/transpose-solve.c -2 +2 cannam@95: M ./mpi/wisdom-api.c -2 +2 cannam@95: M ./rdft/buffered.c -2 +2 cannam@95: M ./rdft/buffered2.c -2 +2 cannam@95: M ./rdft/codelet-rdft.h -2 +2 cannam@95: M ./rdft/conf.c -2 +2 cannam@95: M ./rdft/ct-hc2c-direct.c -2 +2 cannam@95: M ./rdft/ct-hc2c.c -2 +2 cannam@95: M ./rdft/ct-hc2c.h -2 +2 cannam@95: M ./rdft/dft-r2hc.c -2 +2 cannam@95: M ./rdft/dht-r2hc.c -2 +2 cannam@95: M ./rdft/dht-rader.c -2 +2 cannam@95: M ./rdft/direct-r2c.c -2 +2 cannam@95: M ./rdft/direct-r2r.c -2 +2 cannam@95: M ./rdft/direct2.c -2 +2 cannam@95: M ./rdft/generic.c -2 +2 cannam@95: M ./rdft/hc2hc-direct.c -2 +2 cannam@95: M ./rdft/hc2hc-generic.c -2 +2 cannam@95: M ./rdft/hc2hc.c -2 +2 cannam@95: M ./rdft/hc2hc.h -2 +2 cannam@95: M ./rdft/indirect.c -2 +2 cannam@95: M ./rdft/khc2c.c -2 +2 cannam@95: M ./rdft/khc2hc.c -2 +2 cannam@95: M ./rdft/kr2c.c -2 +2 cannam@95: M ./rdft/kr2r.c -2 +2 cannam@95: M ./rdft/nop.c -2 +2 cannam@95: M ./rdft/nop2.c -2 +2 cannam@95: M ./rdft/plan.c -2 +2 cannam@95: M ./rdft/plan2.c -2 +2 cannam@95: M ./rdft/problem.c -2 +2 cannam@95: M ./rdft/problem2.c -2 +2 cannam@95: M ./rdft/rank-geq2-rdft2.c -2 +2 cannam@95: M ./rdft/rank-geq2.c -2 +2 cannam@95: M ./rdft/rank0-rdft2.c -2 +2 cannam@95: M ./rdft/rank0.c -2 +2 cannam@95: M ./rdft/rdft-dht.c -2 +2 cannam@95: M ./rdft/rdft.h -2 +2 cannam@95: M ./rdft/rdft2-inplace-strides.c -2 +2 cannam@95: M ./rdft/rdft2-rdft.c -2 +2 cannam@95: M ./rdft/rdft2-strides.c -2 +2 cannam@95: M ./rdft/rdft2-tensor-max-index.c -2 +2 cannam@95: M ./rdft/scalar/hb.h -2 +2 cannam@95: M ./rdft/scalar/hc2c.c -2 +2 cannam@95: M ./rdft/scalar/hc2cb.h -2 +2 cannam@95: M ./rdft/scalar/hc2cf.h -2 +2 cannam@95: M ./rdft/scalar/hf.h -2 +2 cannam@95: M ./rdft/scalar/hfb.c -2 +2 cannam@95: M ./rdft/scalar/r2c.c -2 +2 cannam@95: M ./rdft/scalar/r2cb.h -2 +2 cannam@95: M ./rdft/scalar/r2cbIII.h -2 +2 cannam@95: M ./rdft/scalar/r2cf.h -2 +2 cannam@95: M ./rdft/scalar/r2cfII.h -2 +2 cannam@95: M ./rdft/scalar/r2r.c -2 +2 cannam@95: M ./rdft/scalar/r2r.h -2 +2 cannam@95: M ./rdft/simd/common/genus.c -2 +2 cannam@95: M ./rdft/simd/hc2cbv.h -2 +2 cannam@95: M ./rdft/simd/hc2cfv.h -2 +2 cannam@95: M ./rdft/solve.c -2 +2 cannam@95: M ./rdft/solve2.c -2 +2 cannam@95: M ./rdft/vrank-geq1-rdft2.c -2 +2 cannam@95: M ./rdft/vrank-geq1.c -2 +2 cannam@95: M ./rdft/vrank3-transpose.c -2 +2 cannam@95: M ./reodft/conf.c -2 +2 cannam@95: M ./reodft/redft00e-r2hc-pad.c -2 +2 cannam@95: M ./reodft/redft00e-r2hc.c -2 +2 cannam@95: M ./reodft/reodft.h -2 +2 cannam@95: M ./reodft/reodft010e-r2hc.c -2 +2 cannam@95: M ./reodft/reodft11e-r2hc-odd.c -2 +2 cannam@95: M ./reodft/reodft11e-r2hc.c -2 +2 cannam@95: M ./reodft/reodft11e-radix2.c -2 +2 cannam@95: M ./reodft/rodft00e-r2hc-pad.c -2 +2 cannam@95: M ./reodft/rodft00e-r2hc.c -2 +2 cannam@95: M ./simd/altivec.c -2 +2 cannam@95: M ./simd/nonportable/sse.c -2 +2 cannam@95: M ./simd/nonportable/sse2.c -2 +2 cannam@95: M ./simd/simd-altivec.h -2 +2 cannam@95: M ./simd/simd-sse.h -2 +2 cannam@95: M ./simd/simd-sse2.h -2 +2 cannam@95: M ./simd/simd.h -2 +2 cannam@95: M ./simd/sse.c -2 +2 cannam@95: M ./simd/sse2.c -2 +2 cannam@95: M ./simd/taint.c -2 +2 cannam@95: M ./simd/x86-cpuid.h -2 +2 cannam@95: M ./simd-support/altivec.c -2 +2 cannam@95: M ./simd-support/amd64-cpuid.h -2 +2 cannam@95: M ./simd-support/avx.c -2 +2 cannam@95: M ./simd-support/simd-altivec.h -2 +2 cannam@95: M ./simd-support/simd-avx.h -2 +2 cannam@95: M ./simd-support/simd-common.h -2 +2 cannam@95: M ./simd-support/simd-sse2.h -2 +2 cannam@95: M ./simd-support/sse2-nonportable.c -2 +2 cannam@95: M ./simd-support/sse2.c -2 +2 cannam@95: M ./simd-support/taint.c -2 +2 cannam@95: M ./simd-support/x86-cpuid.h -2 +2 cannam@95: M ./threads/api.c -2 +2 cannam@95: M ./threads/conf.c -2 +2 cannam@95: M ./threads/ct.c -2 +2 cannam@95: M ./threads/dft-vrank-geq1.c -2 +2 cannam@95: M ./threads/f77api.c -2 +2 cannam@95: M ./threads/f77funcs.h -2 +2 cannam@95: M ./threads/hc2hc.c -2 +2 cannam@95: M ./threads/openmp.c -2 +2 cannam@95: M ./threads/rdft-vrank-geq1.c -2 +2 cannam@95: M ./threads/threads.c -2 +2 cannam@95: M ./threads/threads.h -2 +2 cannam@95: M ./threads/vrank-geq1-rdft2.c -2 +2 cannam@95: M ./tools/fftw-wisdom-to-conf.1 -4 +4 cannam@95: M ./tools/fftw-wisdom-to-conf.in -2 +2 cannam@95: M ./tools/fftw-wisdom.c -2 +2 cannam@95: M ./tools/fftw_wisdom.1.in -4 +4 cannam@95: cannam@95: Sat Jun 25 21:33:13 EDT 2011 stevenj@fftw.org cannam@95: * updated NEWS cannam@95: cannam@95: M ./NEWS -4 +53 cannam@95: cannam@95: Sat Jun 25 20:29:55 EDT 2011 stevenj@fftw.org cannam@95: * fixes to Fortran interface and docs cannam@95: cannam@95: M ./api/genf03.pl -12 +22 cannam@95: M ./doc/modern-fortran.texi -5 +14 cannam@95: M ./doc/mpi.texi -13 +16 cannam@95: cannam@95: Sat Jun 25 17:43:31 EDT 2011 stevenj@fftw.org cannam@95: * initial stab at MPI Fortran docs cannam@95: cannam@95: M ./doc/mpi.texi +138 cannam@95: cannam@95: Sat Jun 25 16:43:31 EDT 2011 stevenj@fftw.org cannam@95: * correct description of what MPI standard says about I/O (I can't believe this crap) cannam@95: cannam@95: M ./doc/mpi.texi -11 +21 cannam@95: cannam@95: Sat Jun 25 15:14:07 EDT 2011 stevenj@fftw.org cannam@95: * more MPI documentation; mention `fftw_alloc' functions earlier in the manual cannam@95: cannam@95: M ./doc/fftw3.texi +12 cannam@95: M ./doc/mpi.texi -30 +158 cannam@95: M ./doc/other.texi +9 cannam@95: M ./doc/tutorial.texi -11 +14 cannam@95: cannam@95: Sat Jun 25 13:40:19 EDT 2011 stevenj@fftw.org cannam@95: * clarification about --enable-sse2 cannam@95: cannam@95: M ./doc/install.texi -6 +8 cannam@95: cannam@95: Sat Jun 25 13:31:25 EDT 2011 athena@fftw.org cannam@95: * Update mingw build scripts for fftw-3.3 cannam@95: cannam@95: M ./kernel/kalloc.c -2 +2 cannam@95: M ./support/BUILD-MINGW32.sh -3 +3 cannam@95: M ./support/BUILD-MINGW64.sh -4 +4 cannam@95: cannam@95: Sat Jun 25 08:52:13 EDT 2011 athena@fftw.org cannam@95: * Fix typo: EXTRADIST => EXTRA_DIST cannam@95: cannam@95: M ./mpi/Makefile.am -1 +1 cannam@95: cannam@95: Fri Jun 24 23:52:19 EDT 2011 stevenj@fftw.org cannam@95: * finished draft "modern fortran" chapter cannam@95: cannam@95: M ./doc/modern-fortran.texi -9 +91 cannam@95: cannam@95: Fri Jun 24 20:47:49 EDT 2011 stevenj@fftw.org cannam@95: * include FFTW_EXTERN prototypes for wrappers, so that they are properly exported to DLLs on Windows (sigh) cannam@95: cannam@95: M ./mpi/f03-wrap.sh +4 cannam@95: cannam@95: Fri Jun 24 16:52:30 EDT 2011 athena@fftw.org cannam@95: * use malloc() instead of alloca() for large buffers cannam@95: cannam@95: The proximate cause for this patch is that OpenBSD/i386 reserves 256KB cannam@95: stack size per thread. We were allocating a buffer of size cannam@95: 128*130*sizeof(fftw_complex) that exceeds the stack. cannam@95: cannam@95: While 128*130*sizeof(fftw_complex) = 260KiB is the worst case for cannam@95: normal configurations, it is a good idea to limit stack allocation cannam@95: just in case. Also, the generic solver might in principle generate cannam@95: unbounded buffers, even though it is normally disabled for n > 137. cannam@95: cannam@95: So, as an added precaution, we now never stack-allocate buffers larger cannam@95: than 64KiB, which ought to be enough for anybody. cannam@95: cannam@95: cannam@95: M ./dft/dftw-direct.c -2 +3 cannam@95: M ./dft/direct.c -2 +3 cannam@95: M ./dft/generic.c -2 +3 cannam@95: M ./kernel/ifftw.h -7 +31 cannam@95: M ./rdft/ct-hc2c-direct.c -2 +3 cannam@95: M ./rdft/direct-r2c.c -2 +3 cannam@95: M ./rdft/generic.c -4 +6 cannam@95: M ./rdft/hc2hc-direct.c -2 +3 cannam@95: cannam@95: Fri Jun 24 16:32:30 EDT 2011 stevenj@fftw.org cannam@95: * don't imply that AVX is available on Pentium III; note that MIPS Paired Single is currently only in FFTW 3.2.x cannam@95: cannam@95: M ./doc/install.texi -3 +2 cannam@95: M ./doc/other.texi -3 +4 cannam@95: cannam@95: Fri Jun 24 16:05:27 EDT 2011 stevenj@fftw.org cannam@95: * silence annoying gfortran warnings cannam@95: cannam@95: M ./api/genf03.pl -1 +6 cannam@95: M ./configure.ac +14 cannam@95: M ./mpi/Makefile.am -3 +6 cannam@95: cannam@95: Fri Jun 24 14:59:30 EDT 2011 stevenj@fftw.org cannam@95: * a couple MPI Fortran 2003 fixes; changed MPI flags to not use 1<<31 since Fortran (not having unsigned integers) does not allow us to declare that constant in a portable way cannam@95: cannam@95: M ./mpi/api.c -1 +1 cannam@95: M ./mpi/f03api.sh -2 +2 cannam@95: M ./mpi/fftw3-mpi.h -4 +4 cannam@95: cannam@95: Fri Jun 24 15:05:05 EDT 2011 athena@fftw.org cannam@95: * Fix libtool shared version info. cannam@95: cannam@95: FFTW-3.3.x should be a direct drop-in replacement for all FFTW-3.x.y cannam@95: versions. cannam@95: cannam@95: cannam@95: M ./configure.ac -1 +4 cannam@95: cannam@95: Fri Jun 24 14:38:47 EDT 2011 stevenj@fftw.org cannam@95: * add MPI Fortran API and wrappers cannam@95: cannam@95: M ./api/Makefile.am -2 +3 cannam@95: M ./api/f03api.sh -3 +5 cannam@95: M ./api/genf03.pl -2 +14 cannam@95: M ./mpi/Makefile.am -3 +14 cannam@95: R ./mpi/README cannam@95: A ./mpi/f03-wrap.sh cannam@95: A ./mpi/f03api.sh cannam@95: M ./mpi/fftw3-mpi.h -4 +4 cannam@95: A ./mpi/genf03-wrap.pl cannam@95: cannam@95: Fri Jun 24 14:51:12 EDT 2011 athena@fftw.org cannam@95: * Do not require fig2dev on the user's machine cannam@95: cannam@95: Distribute the manual's figures in PDF/PS/PNG form instead. cannam@95: cannam@95: M ./doc/Makefile.am -10 +16 cannam@95: cannam@95: Fri Jun 24 11:52:44 EDT 2011 athena@fftw.org cannam@95: * Remove --enable-portable-binary, --with-gcc-arch from documentation. cannam@95: cannam@95: M ./NEWS +3 cannam@95: M ./doc/install.texi -23 +2 cannam@95: cannam@95: Fri Jun 24 11:48:48 EDT 2011 athena@fftw.org cannam@95: * Forget about specifying nonportable CFLAGS. Let the user do it if he wants. cannam@95: cannam@95: M ./m4/Makefile.am -4 +3 cannam@95: M ./m4/ax_cc_maxopt.m4 -24 +10 cannam@95: R ./m4/ax_gcc_archflag.m4 cannam@95: R ./m4/ax_gcc_x86_cpuid.m4 cannam@95: cannam@95: Fri Jun 24 11:48:25 EDT 2011 athena@fftw.org cannam@95: * Add "-avx" to version string when appropriate. cannam@95: cannam@95: M ./api/version.c +4 cannam@95: cannam@95: Fri Jun 24 10:26:38 EDT 2011 athena@fftw.org cannam@95: * change 3.3-alpha => 3.3-beta1 cannam@95: cannam@95: M ./configure.ac -1 +1 cannam@95: cannam@95: Fri Jun 24 09:25:49 EDT 2011 athena@fftw.org cannam@95: * Extend OUR_MALLOC16 to larger alignments cannam@95: cannam@95: Make it work for 32-byte alignment and beyond, as needed by AVX. cannam@95: Rename --with-our-malloc16 to --with-our-malloc. Keep old --with-our-malloc16 cannam@95: flag for compatibility. cannam@95: cannam@95: cannam@95: M ./configure.ac -3 +4 cannam@95: M ./kernel/kalloc.c -9 +10 cannam@95: cannam@95: Fri Jun 24 09:19:38 EDT 2011 athena@fftw.org cannam@95: * Fix typo cannam@95: cannam@95: M ./doc/install.texi -1 +1 cannam@95: cannam@95: Fri Jun 24 09:10:26 EDT 2011 athena@fftw.org cannam@95: * One pass over the manual. cannam@95: cannam@95: M ./NEWS -1 +6 cannam@95: M ./doc/Makefile.am -5 +2 cannam@95: M ./doc/install.texi -16 +10 cannam@95: M ./doc/intro.texi -1 +1 cannam@95: M ./doc/other.texi -2 +2 cannam@95: M ./doc/reference.texi -5 +5 cannam@95: M ./doc/tutorial.texi -32 +32 cannam@95: cannam@95: Fri Jun 24 08:19:03 EDT 2011 athena@fftw.org cannam@95: * eliminate the WITH_ALIGNED_STACK hack cannam@95: cannam@95: This is 2011 and I have no system with incorrect stack alignment. cannam@95: cannam@95: cannam@95: M ./TODO -2 cannam@95: M ./api/apiplan.c -9 +4 cannam@95: M ./api/execute-dft-c2r.c -2 +2 cannam@95: M ./api/execute-dft-r2c.c -2 +2 cannam@95: M ./api/execute-dft.c -2 +2 cannam@95: M ./api/execute-r2r.c -2 +2 cannam@95: M ./api/execute-split-dft-c2r.c -2 +2 cannam@95: M ./api/execute-split-dft-r2c.c -2 +2 cannam@95: M ./api/execute-split-dft.c -2 +2 cannam@95: M ./api/execute.c -2 +2 cannam@95: M ./api/f77funcs.h -16 +16 cannam@95: M ./doc/Makefile.am -2 +2 cannam@95: M ./doc/fftw3.texi -1 +1 cannam@95: M ./doc/intro.texi -3 +4 cannam@95: M ./doc/legacy-fortran.texi -1 +1 cannam@95: M ./doc/other.texi -51 +4 cannam@95: M ./doc/reference.texi -1 +1 cannam@95: M ./doc/tutorial.texi -9 +5 cannam@95: M ./kernel/ifftw.h -51 cannam@95: M ./threads/ct.c -2 +2 cannam@95: M ./threads/dft-vrank-geq1.c -2 +2 cannam@95: M ./threads/hc2hc.c -2 +2 cannam@95: M ./threads/rdft-vrank-geq1.c -2 +2 cannam@95: M ./threads/vrank-geq1-rdft2.c -2 +2 cannam@95: cannam@95: Fri Jun 24 07:49:47 EDT 2011 athena@fftw.org cannam@95: * enable both threaded and unthreaded wisdom in tools/fftw-wisdom cannam@95: cannam@95: M ./TODO -7 +2 cannam@95: M ./tools/fftw-wisdom.c -18 +14 cannam@95: cannam@95: Fri Jun 24 02:40:04 EDT 2011 stevenj@fftw.org cannam@95: * clarification cannam@95: cannam@95: M ./doc/modern-fortran.texi -2 +2 cannam@95: cannam@95: Fri Jun 24 02:24:01 EDT 2011 stevenj@fftw.org cannam@95: * check for error code in example cannam@95: cannam@95: M ./doc/modern-fortran.texi +2 cannam@95: cannam@95: Fri Jun 24 02:22:18 EDT 2011 stevenj@fftw.org cannam@95: * cleanup - since NATIVE_MALLOC is always malloc, delete this #define cannam@95: cannam@95: M ./api/export-wisdom-to-string.c -1 +1 cannam@95: M ./kernel/ifftw.h -2 cannam@95: cannam@95: Fri Jun 24 02:19:44 EDT 2011 stevenj@fftw.org cannam@95: * document wisdom string import/export in Fortran cannam@95: cannam@95: M ./doc/modern-fortran.texi +49 cannam@95: cannam@95: Fri Jun 24 02:11:40 EDT 2011 stevenj@fftw.org cannam@95: * bug fix - NATIVE_MALLOC should always be plain malloc, even in debug_malloc mode, because it is used in the API to return things that should be deallocated with free(); correspondingly, be sure to use free() ansd not X(free) with this cannam@95: cannam@95: M ./kernel/ifftw.h -1 +1 cannam@95: M ./mpi/wisdom-api.c -1 +1 cannam@95: cannam@95: Fri Jun 24 01:35:27 EDT 2011 stevenj@fftw.org cannam@95: * declaration style cannam@95: cannam@95: M ./doc/modern-fortran.texi -2 +4 cannam@95: cannam@95: Fri Jun 24 01:25:36 EDT 2011 stevenj@fftw.org cannam@95: * document wisdom file export/import from Fortran; add export/import_to/from_filename functions for convenience cannam@95: cannam@95: M ./api/export-wisdom-to-file.c +11 cannam@95: M ./api/fftw3.h +2 cannam@95: M ./api/import-wisdom-from-file.c +10 cannam@95: M ./doc/modern-fortran.texi -9 +79 cannam@95: M ./doc/other.texi -5 +6 cannam@95: M ./doc/reference.texi -11 +22 cannam@95: cannam@95: Thu Jun 23 19:19:43 EDT 2011 stevenj@fftw.org cannam@95: * more fortran docs cannam@95: cannam@95: M ./doc/legacy-fortran.texi -5 +6 cannam@95: M ./doc/modern-fortran.texi -10 +167 cannam@95: M ./doc/reference.texi +2 cannam@95: cannam@95: Thu Jun 23 17:50:30 EDT 2011 stevenj@fftw.org cannam@95: * enforce 132-character line-length limit that is the default in Fortran cannam@95: cannam@95: M ./api/genf03.pl -3 +26 cannam@95: cannam@95: Wed Jun 22 23:27:31 EDT 2011 stevenj@fftw.org cannam@95: * the F03 standard is ambiguous about whether types can be assigned to wider types as formal parameters with VALUE attributes, and e.g. gfortran interprets it to disallow this code cannam@95: cannam@95: M ./doc/modern-fortran.texi -2 +4 cannam@95: cannam@95: Thu Jun 23 18:12:10 EDT 2011 athena@fftw.org cannam@95: * Add md5 hash of fftw's configuration to wisdom file cannam@95: cannam@95: People were already confused by threaded vs unthreaded wisdom, and now cannam@95: things will be even worse because we enable/disable AVX codelets at cannam@95: runtime. Accept incoming wisdom only if it was produced by the same cannam@95: configuration (modulo MD5). cannam@95: cannam@95: M ./kernel/planner.c -2 +33 cannam@95: cannam@95: Thu Jun 23 09:01:27 EDT 2011 athena@fftw.org cannam@95: * distribute fftw3.f03.in cannam@95: cannam@95: M ./api/Makefile.am -1 +1 cannam@95: cannam@95: Wed Jun 22 22:02:18 EDT 2011 stevenj@fftw.org cannam@95: * more Fortran documentation cannam@95: cannam@95: M ./doc/modern-fortran.texi -1 +95 cannam@95: cannam@95: Wed Jun 22 20:10:39 EDT 2011 stevenj@fftw.org cannam@95: * correct comment cannam@95: cannam@95: M ./api/genf03.pl -1 +2 cannam@95: cannam@95: Wed Jun 22 20:26:18 EDT 2011 athena@fftw.org cannam@95: * Use "sh FOO.sh" instead of "./FOO.sh" to avoid chmod +x. cannam@95: cannam@95: M ./api/Makefile.am -1 +1 cannam@95: cannam@95: Wed Jun 22 19:19:05 EDT 2011 stevenj@fftw.org cannam@95: * document fftw_alloc_real/complex ... should we switch to using these in the tutorial examples? cannam@95: cannam@95: M ./doc/reference.texi +19 cannam@95: cannam@95: Wed Jun 22 19:07:49 EDT 2011 stevenj@fftw.org cannam@95: * whoops, added missing file cannam@95: cannam@95: A ./api/genf03.pl cannam@95: cannam@95: Wed Jun 22 18:46:01 EDT 2011 athena@fftw.org cannam@95: * Note addition of AVX. cannam@95: cannam@95: M ./NEWS +2 cannam@95: cannam@95: Wed Jun 22 18:43:43 EDT 2011 athena@fftw.org cannam@95: * In SSE2, AVX: use FMA macros when applicable. cannam@95: cannam@95: Makes it easier to play with fma4 and fma3 when it comes out. cannam@95: cannam@95: M ./simd-support/simd-avx.h -21 +21 cannam@95: M ./simd-support/simd-sse2.h -19 +19 cannam@95: cannam@95: Wed Jun 22 18:16:45 EDT 2011 stevenj@fftw.org cannam@95: * all modern Fortran compilers can call FFTW's C interfface directly -- support this, and in particular generate a Fortran 2003 interface file from fftw3.h so that Fortran code calling FFTW can be typechecked ((addressing the source of a lot of Fortran-user problems) cannam@95: cannam@95: ./doc/fortran.texi -> ./doc/legacy-fortran.texi cannam@95: M ./api/Makefile.am -3 +12 cannam@95: A ./api/f03api.sh cannam@95: M ./api/fftw3.h -3 +11 cannam@95: M ./api/malloc.c +18 cannam@95: M ./configure.ac +12 cannam@95: M ./doc/Makefile.am -1 +1 cannam@95: M ./doc/fftw3.texi -2 +4 cannam@95: M ./doc/install.texi -2 +2 cannam@95: M ./doc/intro.texi -5 +6 cannam@95: M ./doc/legacy-fortran.texi -23 +38 cannam@95: A ./doc/modern-fortran.texi cannam@95: M ./doc/mpi.texi -1 +1 cannam@95: M ./doc/other.texi -2 +4 cannam@95: M ./doc/upgrading.texi -1 +1 cannam@95: cannam@95: Wed Jun 22 13:10:02 EDT 2011 athena@fftw.org cannam@95: * some cleanup of SSE2 macros cannam@95: cannam@95: M ./simd-support/simd-sse2.h -27 +16 cannam@95: cannam@95: Wed Jun 22 07:38:18 EDT 2011 athena@fftw.org cannam@95: * don't use -xHost on ICC cannam@95: cannam@95: -xHost with ICC is problematic. On icc-12.0.0, "-mavx -xHost" cannam@95: overrides -mavx with -xHost, generating SSE2 code instead of AVX code. cannam@95: ICC does not seem to support -mtune=host or equivalent non-ABI cannam@95: changing flag. cannam@95: cannam@95: cannam@95: M ./m4/ax_cc_maxopt.m4 -31 +5 cannam@95: cannam@95: Tue Jun 21 20:35:36 EDT 2011 athena@fftw.org cannam@95: * Complete AVX implementation for split codelets cannam@95: cannam@95: M ./configure.ac -36 +38 cannam@95: M ./simd-support/simd-avx.h -41 +26 cannam@95: cannam@95: Tue Jun 21 19:37:14 EDT 2011 stevenj@fftw.org cannam@95: * whoops, missing altivec conf patches cannam@95: cannam@95: M ./dft/codelet-dft.h +1 cannam@95: M ./dft/conf.c +4 cannam@95: M ./rdft/codelet-rdft.h +1 cannam@95: M ./rdft/conf.c +4 cannam@95: cannam@95: Tue Jun 21 19:12:45 EDT 2011 stevenj@fftw.org cannam@95: * some BSD ar versions (e.g. on MacOS X) give an error if there are no object files, so we cannot build empty libraries cannam@95: cannam@95: M ./dft/simd/altivec/Makefile.am -4 +3 cannam@95: M ./dft/simd/avx/Makefile.am -3 +4 cannam@95: M ./dft/simd/sse2/Makefile.am -4 +3 cannam@95: M ./rdft/simd/altivec/Makefile.am -4 +3 cannam@95: M ./rdft/simd/avx/Makefile.am -3 +4 cannam@95: M ./rdft/simd/sse2/Makefile.am -3 +4 cannam@95: cannam@95: Tue Jun 21 19:12:12 EDT 2011 stevenj@fftw.org cannam@95: * re-insertion of Altivec code cannam@95: cannam@95: M ./Makefile.am -1 +6 cannam@95: M ./configure.ac -8 +13 cannam@95: M ./dft/simd/Makefile.am -1 +1 cannam@95: A ./dft/simd/altivec/ cannam@95: A ./dft/simd/altivec/Makefile.am cannam@95: M ./kernel/ifftw.h -2 +9 cannam@95: M ./rdft/simd/Makefile.am -1 +1 cannam@95: A ./rdft/simd/altivec/ cannam@95: A ./rdft/simd/altivec/Makefile.am cannam@95: M ./simd-support/Makefile.am -1 +1 cannam@95: A ./simd-support/altivec.c cannam@95: A ./simd-support/simd-altivec.h cannam@95: M ./simd-support/simd-common.h +3 cannam@95: cannam@95: Tue Jun 21 16:26:09 EDT 2011 athena@fftw.org cannam@95: * Implement faster AVX loads/stores. cannam@95: cannam@95: M ./simd-support/simd-avx.h -21 +28 cannam@95: cannam@95: Tue Jun 21 16:03:24 EDT 2011 athena@fftw.org cannam@95: * Initial AVX256/single implementation cannam@95: cannam@95: This should be correct but slow. I need to figure out how to implement cannam@95: noncontiguous loads/stores efficiently. cannam@95: cannam@95: M ./simd-support/simd-avx.h -53 +133 cannam@95: cannam@95: Tue Jun 21 14:13:57 EDT 2011 athena@fftw.org cannam@95: * fix AVX alignment cannam@95: cannam@95: M ./simd-support/simd-common.h -6 +8 cannam@95: cannam@95: Tue Jun 21 14:07:28 EDT 2011 athena@fftw.org cannam@95: * rename avx256d -> avx cannam@95: cannam@95: AVX will work in both double and single precision, like SSE2. cannam@95: cannam@95: ./dft/simd/avx256d -> ./dft/simd/avx cannam@95: ./rdft/simd/avx256d -> ./rdft/simd/avx cannam@95: ./simd-support/avx256d.c -> ./simd-support/avx.c cannam@95: ./simd-support/simd-avx256d.h -> ./simd-support/simd-avx.h cannam@95: M ./Makefile.am -4 +4 cannam@95: M ./configure.ac -11 +6 cannam@95: M ./dft/codelet-dft.h -1 +1 cannam@95: M ./dft/conf.c -3 +3 cannam@95: M ./dft/simd/Makefile.am -1 +1 cannam@95: M ./dft/simd/avx/Makefile.am -4 +4 cannam@95: M ./kernel/ifftw.h -2 +2 cannam@95: M ./rdft/codelet-rdft.h -1 +1 cannam@95: M ./rdft/conf.c -3 +3 cannam@95: M ./rdft/simd/Makefile.am -1 +1 cannam@95: M ./rdft/simd/avx/Makefile.am -4 +4 cannam@95: M ./simd-support/Makefile.am -1 +1 cannam@95: M ./simd-support/avx.c -3 +3 cannam@95: M ./simd-support/simd-avx.h -2 +2 cannam@95: cannam@95: Tue Jun 21 13:52:20 EDT 2011 athena@fftw.org cannam@95: * remove CODELET_OPTIM cannam@95: cannam@95: In the old 32-bit gcc-3.x days we used to play games with gcc to force cannam@95: it to produce decent code. Now gcc has gotten smarter and it produces cannam@95: indecent code no matter what we do, so it is safe to remove these hacks. cannam@95: cannam@95: cannam@95: M ./api/version.c -5 +5 cannam@95: M ./configure.ac -40 cannam@95: M ./support/Makefile.codelets -6 cannam@95: cannam@95: Tue Jun 21 09:57:31 EDT 2011 athena@fftw.org cannam@95: * work around gcc/icc quirks cannam@95: cannam@95: M ./simd-support/simd-avx256d.h -14 +45 cannam@95: cannam@95: Tue Jun 21 09:56:07 EDT 2011 athena@fftw.org cannam@95: * Add remarks in places where we work around gcc quirks cannam@95: cannam@95: M ./simd-support/simd-sse2.h -7 +16 cannam@95: cannam@95: Mon Jun 20 21:17:59 EDT 2011 stevenj@fftw.org cannam@95: * remove the libbench directory (which we have kept lingering in the repository for years due to CVS's inability to remove directories) cannam@95: cannam@95: R ./libbench/Makefile.am cannam@95: R ./libbench/accopy-from.c cannam@95: R ./libbench/accopy-to.c cannam@95: R ./libbench/acopy.c cannam@95: R ./libbench/allocate.c cannam@95: R ./libbench/ascale.c cannam@95: R ./libbench/aset.c cannam@95: R ./libbench/bench-main.c cannam@95: R ./libbench/bench-user.h cannam@95: R ./libbench/bench.h cannam@95: R ./libbench/caadd.c cannam@95: R ./libbench/cacopy.c cannam@95: R ./libbench/can-do.c cannam@95: R ./libbench/cascale.c cannam@95: R ./libbench/caset.c cannam@95: R ./libbench/casub.c cannam@95: R ./libbench/ccopy-from.c cannam@95: R ./libbench/ccopy-to.c cannam@95: R ./libbench/copy-c2c-from.c cannam@95: R ./libbench/copy-c2c-to.c cannam@95: R ./libbench/copy-c2h-1d-fftpack.c cannam@95: R ./libbench/copy-c2h-1d-halfcomplex.c cannam@95: R ./libbench/copy-c2h-1d-packed.c cannam@95: R ./libbench/copy-c2h-1d-unpacked-ri.c cannam@95: R ./libbench/copy-c2h-unpacked.c cannam@95: R ./libbench/copy-c2h.c cannam@95: R ./libbench/copy-c2r-packed.c cannam@95: R ./libbench/copy-c2r-unpacked.c cannam@95: R ./libbench/copy-c2r.c cannam@95: R ./libbench/copy-c2ri.c cannam@95: R ./libbench/copy-h2c-1d-fftpack.c cannam@95: R ./libbench/copy-h2c-1d-halfcomplex.c cannam@95: R ./libbench/copy-h2c-1d-packed.c cannam@95: R ./libbench/copy-h2c-1d-unpacked-ri.c cannam@95: R ./libbench/copy-h2c-unpacked.c cannam@95: R ./libbench/copy-h2c.c cannam@95: R ./libbench/copy-r2c-packed.c cannam@95: R ./libbench/copy-r2c-unpacked.c cannam@95: R ./libbench/copy-r2c.c cannam@95: R ./libbench/copy-ri2c.c cannam@95: R ./libbench/deallocate.c cannam@95: R ./libbench/getopt-utils.c cannam@95: R ./libbench/getopt.c cannam@95: R ./libbench/getopt.h cannam@95: R ./libbench/getopt1.c cannam@95: R ./libbench/info.c cannam@95: R ./libbench/log2.c cannam@95: R ./libbench/main.c cannam@95: R ./libbench/mflops.c cannam@95: R ./libbench/mp.c cannam@95: R ./libbench/ovtpvt.c cannam@95: R ./libbench/pow2.c cannam@95: R ./libbench/prime.c cannam@95: R ./libbench/problem.c cannam@95: R ./libbench/report.c cannam@95: R ./libbench/speed.c cannam@95: R ./libbench/timer.c cannam@95: R ./libbench/unnormalize.c cannam@95: R ./libbench/util.c cannam@95: R ./libbench/verify.c cannam@95: R ./libbench/zero.c cannam@95: R ./libbench/ cannam@95: cannam@95: Mon Jun 20 21:17:14 EDT 2011 stevenj@fftw.org cannam@95: * update URLs cannam@95: cannam@95: M ./doc/install.texi -2 +2 cannam@95: M ./doc/intro.texi -1 +1 cannam@95: M ./doc/license.texi -1 +1 cannam@95: M ./doc/mpi.texi -1 +1 cannam@95: M ./doc/other.texi -1 +1 cannam@95: cannam@95: Mon Jun 20 20:53:31 EDT 2011 stevenj@fftw.org cannam@95: * whoops, forgot to check in alignment change cannam@95: cannam@95: M ./simd-support/simd-common.h -3 +6 cannam@95: cannam@95: Mon Jun 20 20:22:23 EDT 2011 athena@fftw.org cannam@95: * "test X = Y" requires spaces around "=" cannam@95: cannam@95: M ./configure.ac -1 +1 cannam@95: cannam@95: Mon Jun 20 19:18:52 EDT 2011 stevenj@fftw.org cannam@95: * indenting cannam@95: cannam@95: M ./simd-support/simd-sse2.h -3 +3 cannam@95: cannam@95: Mon Jun 20 18:57:10 EDT 2011 stevenj@fftw.org cannam@95: * merge back in SSE support, now combined with SSE2; --enable-sse2 now works in both single and double precision, and simd-sse2.h contains both the double- and single-precision code (which overlap a lot); in single precision it is still compiled for SSE-only (SSE2 is only required for double) cannam@95: cannam@95: M ./configure.ac -16 +9 cannam@95: M ./simd-support/simd-sse2.h -54 +181 cannam@95: M ./simd-support/sse2-nonportable.c -2 +6 cannam@95: M ./simd-support/sse2.c -4 +11 cannam@95: cannam@95: Mon Jun 20 16:02:07 EDT 2011 athena@fftw.org cannam@95: * Implement AVX autodetection (gcc-only so far) cannam@95: cannam@95: M ./simd-support/Makefile.am -1 +1 cannam@95: A ./simd-support/amd64-cpuid.h cannam@95: M ./simd-support/avx256d.c -2 +32 cannam@95: M ./simd-support/x86-cpuid.h +33 cannam@95: cannam@95: Mon Jun 20 14:25:54 EDT 2011 athena@fftw.org cannam@95: * Add VZEROUPPER at the end of AVX codelets cannam@95: cannam@95: If the Intel Optimization Manual is to be believed, we need to wave a cannam@95: dead chicken before transitioning from AVX code to SSE code. I am cannam@95: supposed to believe that there is a transition penalty for doing so, cannam@95: unless one uses a magic VZEROUPPER instruction that apparently has cannam@95: zero cost. Whatever. cannam@95: cannam@95: cannam@95: M ./genfft/c.ml +2 cannam@95: M ./genfft/c.mli +1 cannam@95: M ./genfft/gen_hc2c.ml -1 +1 cannam@95: M ./genfft/gen_hc2cdft.ml -1 +1 cannam@95: M ./genfft/gen_hc2cdft_c.ml -1 +1 cannam@95: M ./genfft/gen_hc2hc.ml -1 +1 cannam@95: M ./genfft/gen_mdct.ml -1 +1 cannam@95: M ./genfft/gen_notw.ml -1 +1 cannam@95: M ./genfft/gen_notw_c.ml -1 +1 cannam@95: M ./genfft/gen_r2cb.ml -1 +1 cannam@95: M ./genfft/gen_r2cf.ml -1 +1 cannam@95: M ./genfft/gen_r2r.ml -1 +1 cannam@95: M ./genfft/gen_twiddle.ml -1 +1 cannam@95: M ./genfft/gen_twiddle_c.ml -1 +1 cannam@95: M ./genfft/gen_twidsq.ml -1 +1 cannam@95: M ./genfft/gen_twidsq_c.ml -1 +1 cannam@95: M ./genfft/genutil.ml -2 +2 cannam@95: M ./genfft/simd.ml +1 cannam@95: M ./simd-support/simd-avx256d.h +5 cannam@95: M ./simd-support/simd-sse2.h +2 cannam@95: cannam@95: Mon Jun 20 10:21:25 EDT 2011 athena@fftw.org cannam@95: * Move RDFT to new simd scheme cannam@95: cannam@95: A ./rdft/simd/avx256d/ cannam@95: R ./rdft/simd/codelets/Makefile.am cannam@95: R ./rdft/simd/codelets/ cannam@95: A ./rdft/simd/common/ cannam@95: A ./rdft/simd/sse2/ cannam@95: M ./Makefile.am -2 +4 cannam@95: M ./configure.ac +4 cannam@95: M ./dft/simd/Makefile.am -1 +1 cannam@95: M ./genfft/gen_hc2cdft_c.ml -2 +2 cannam@95: M ./kernel/ifftw.h +1 cannam@95: M ./rdft/Makefile.am -1 +1 cannam@95: M ./rdft/codelet-rdft.h -1 +2 cannam@95: M ./rdft/conf.c -2 +7 cannam@95: M ./rdft/simd/Makefile.am -6 +2 cannam@95: A ./rdft/simd/avx256d/Makefile.am cannam@95: A ./rdft/simd/codlist.mk cannam@95: A ./rdft/simd/common/Makefile.am cannam@95: A ./rdft/simd/common/genus.c cannam@95: R ./rdft/simd/hc2cbv.c cannam@95: M ./rdft/simd/hc2cbv.h -2 +2 cannam@95: R ./rdft/simd/hc2cfv.c cannam@95: M ./rdft/simd/hc2cfv.h -2 +2 cannam@95: A ./rdft/simd/simd.mk cannam@95: A ./rdft/simd/sse2/Makefile.am cannam@95: cannam@95: Mon Jun 20 09:23:38 EDT 2011 athena@fftw.org cannam@95: * New SIMD build system cannam@95: cannam@95: We now support multiple SIMD extensions in the same binary, e.g. cannam@95: --enable-sse2 --enable-avx. This patch adds the necessary cannam@95: infrastructure for SSE2/AVX and complex DFT. Later patches will add cannam@95: RDFT and SSE/ALTIVEC/etc. cannam@95: cannam@95: cannam@95: A ./dft/simd/avx256d/ cannam@95: A ./simd-support/ cannam@95: R ./dft/simd/codelets/Makefile.am cannam@95: R ./dft/simd/codelets/ cannam@95: A ./dft/simd/common/ cannam@95: A ./dft/simd/sse2/ cannam@95: M ./Makefile.am -11 +13 cannam@95: M ./configure.ac -47 +74 cannam@95: M ./dft/codelet-dft.h -4 +2 cannam@95: M ./dft/conf.c -2 +7 cannam@95: M ./dft/scalar/codelets/Makefile.am -1 +1 cannam@95: M ./dft/simd/Makefile.am -6 +3 cannam@95: A ./dft/simd/avx256d/Makefile.am cannam@95: A ./dft/simd/codlist.mk cannam@95: A ./dft/simd/common/Makefile.am cannam@95: A ./dft/simd/common/genus.c cannam@95: R ./dft/simd/n1b.c cannam@95: M ./dft/simd/n1b.h -2 +2 cannam@95: R ./dft/simd/n1f.c cannam@95: M ./dft/simd/n1f.h -2 +2 cannam@95: R ./dft/simd/n2b.c cannam@95: M ./dft/simd/n2b.h -2 +2 cannam@95: R ./dft/simd/n2f.c cannam@95: M ./dft/simd/n2f.h -2 +2 cannam@95: R ./dft/simd/n2s.c cannam@95: M ./dft/simd/n2s.h -2 +2 cannam@95: R ./dft/simd/q1b.c cannam@95: M ./dft/simd/q1b.h -2 +2 cannam@95: R ./dft/simd/q1f.c cannam@95: M ./dft/simd/q1f.h -2 +2 cannam@95: A ./dft/simd/simd.mk cannam@95: A ./dft/simd/sse2/Makefile.am cannam@95: R ./dft/simd/t.c cannam@95: M ./dft/simd/t1b.h -2 +2 cannam@95: M ./dft/simd/t1bu.h -2 +2 cannam@95: M ./dft/simd/t1f.h -2 +2 cannam@95: M ./dft/simd/t1fu.h -2 +2 cannam@95: M ./dft/simd/t2b.h -2 +2 cannam@95: M ./dft/simd/t2f.h -2 +2 cannam@95: M ./dft/simd/t3b.h -2 +2 cannam@95: M ./dft/simd/t3f.h -2 +2 cannam@95: R ./dft/simd/ts.c cannam@95: M ./dft/simd/ts.h -2 +2 cannam@95: M ./genfft/gen_twiddle.ml -2 +2 cannam@95: M ./genfft/genutil.ml -2 +5 cannam@95: M ./kernel/ifftw.h -2 +6 cannam@95: M ./rdft/Makefile.am -1 +1 cannam@95: M ./rdft/conf.c -1 +1 cannam@95: M ./rdft/scalar/r2cb/Makefile.am +1 cannam@95: M ./rdft/scalar/r2cf/Makefile.am +1 cannam@95: M ./rdft/scalar/r2r/Makefile.am +1 cannam@95: A ./simd-support/Makefile.am cannam@95: A ./simd-support/avx256d.c cannam@95: A ./simd-support/simd-avx256d.h cannam@95: A ./simd-support/simd-common.h cannam@95: A ./simd-support/simd-sse2.h cannam@95: A ./simd-support/sse2-nonportable.c cannam@95: A ./simd-support/sse2.c cannam@95: A ./simd-support/taint.c cannam@95: A ./simd-support/x86-cpuid.h cannam@95: M ./support/Makefile.codelets -21 +22 cannam@95: cannam@95: Sun Jun 19 12:29:27 EDT 2011 stevenj@fftw.org cannam@95: * italicize Latin quote cannam@95: cannam@95: M ./doc/reference.texi -2 +2 cannam@95: cannam@95: Sun Jun 19 12:26:34 EDT 2011 stevenj@fftw.org cannam@95: * work around incredibly annoying makeinfo bug -- for HTML output, in any paragraph ending with an @index command, two blank lines are needed to create a paragraph break ... our HTML output has apparently been screwed up for years cannam@95: cannam@95: M ./doc/acknowledgements.texi +2 cannam@95: M ./doc/fortran.texi +1 cannam@95: M ./doc/install.texi +2 cannam@95: M ./doc/intro.texi +1 cannam@95: M ./doc/mpi.texi +20 cannam@95: M ./doc/other.texi +2 cannam@95: M ./doc/reference.texi +21 cannam@95: M ./doc/threads.texi +2 cannam@95: M ./doc/tutorial.texi +13 cannam@95: cannam@95: Sun Jun 19 12:01:39 EDT 2011 stevenj@fftw.org cannam@95: * conjugate-pair algorithm turns out not to be due to djb, but it was pointed out to us by djb cannam@95: cannam@95: M ./doc/intro.texi -3 +4 cannam@95: cannam@95: Sun Jun 19 11:59:21 EDT 2011 stevenj@fftw.org cannam@95: * new-array execute functions are *not* the same for MPI, since the problems are different cannam@95: cannam@95: M ./doc/mpi.texi -2 +3 cannam@95: cannam@95: Sun Jun 19 11:57:52 EDT 2011 stevenj@fftw.org cannam@95: * tweaks to MPI manual cannam@95: cannam@95: M ./doc/mpi.texi -19 +44 cannam@95: M ./doc/reference.texi -1 +1 cannam@95: cannam@95: Sun Jun 19 00:47:23 EDT 2011 stevenj@fftw.org cannam@95: * only call MPI_Init_thread for MPI version >= 2 cannam@95: cannam@95: M ./mpi/mpi-bench.c +5 cannam@95: cannam@95: Sun Jun 19 00:35:44 EDT 2011 stevenj@fftw.org cannam@95: * document quad precision in Fortran cannam@95: cannam@95: M ./doc/fortran.texi -9 +12 cannam@95: cannam@95: Sun Jun 19 00:16:10 EDT 2011 stevenj@fftw.org cannam@95: * use -lquadmath for quad-precision library in pkgconfig file cannam@95: cannam@95: M ./configure.ac +2 cannam@95: M ./fftw.pc.in -1 +1 cannam@95: cannam@95: Sun Jun 19 00:10:33 EDT 2011 stevenj@fftw.org cannam@95: * document quad precision cannam@95: cannam@95: M ./doc/install.texi -1 +11 cannam@95: M ./doc/reference.texi +8 cannam@95: cannam@95: Sat Jun 18 21:26:24 EDT 2011 stevenj@fftw.org cannam@95: * there is currently no standard quad-precision type in MPI, so don't allow MPI support to be compiled with --enable-quad-precision cannam@95: cannam@95: M ./configure.ac +3 cannam@95: M ./mpi/ifftw-mpi.h +2 cannam@95: M ./mpi/mpi-bench.c +2 cannam@95: cannam@95: Sat Jun 18 21:25:51 EDT 2011 stevenj@fftw.org cannam@95: * make quad-precision library libfftwq cannam@95: cannam@95: M ./configure.ac +1 cannam@95: cannam@95: Sat Jun 18 21:19:50 EDT 2011 stevenj@fftw.org cannam@95: * use --estimate in check script with --enable-random-estimator cannam@95: cannam@95: M ./configure.ac +2 cannam@95: M ./mpi/Makefile.am -1 +1 cannam@95: M ./tests/Makefile.am -5 +5 cannam@95: cannam@95: Sat Jun 18 20:36:51 EDT 2011 stevenj@fftw.org cannam@95: * bug fix - correct crashing interaction between threads and debug-malloc cannam@95: cannam@95: M ./threads/threads.c +4 cannam@95: cannam@95: Sat Jun 18 18:41:48 EDT 2011 stevenj@fftw.org cannam@95: * whoops, typo cannam@95: cannam@95: M ./doc/reference.texi -1 +1 cannam@95: cannam@95: Sat Jun 18 18:35:14 EDT 2011 stevenj@fftw.org cannam@95: * clarification of when fftw_cost may return 0 cannam@95: cannam@95: M ./doc/reference.texi -2 +5 cannam@95: cannam@95: Sat Jun 18 18:31:26 EDT 2011 stevenj@fftw.org cannam@95: * corrected manual and test program for proper interaction of MPI and threads cannam@95: cannam@95: M ./doc/mpi.texi -20 +59 cannam@95: M ./mpi/mpi-bench.c +13 cannam@95: M ./tests/fftw-bench.c -4 +10 cannam@95: M ./tests/fftw-bench.h +5 cannam@95: cannam@95: Sat Jun 18 17:13:52 EDT 2011 stevenj@fftw.org cannam@95: * need --mpi restriction in mpi+threads check cannam@95: cannam@95: M ./mpi/Makefile.am -1 +1 cannam@95: cannam@95: Sat Jun 18 12:09:04 EDT 2011 stevenj@fftw.org cannam@95: * split fftw3.texi into multiple files for ease of editing cannam@95: cannam@95: M ./doc/Makefile.am -2 +1 cannam@95: A ./doc/acknowledgements.texi cannam@95: A ./doc/cindex.texi cannam@95: M ./doc/fftw3.texi -6158 +14 cannam@95: A ./doc/findex.texi cannam@95: A ./doc/fortran.texi cannam@95: A ./doc/install.texi cannam@95: A ./doc/intro.texi cannam@95: A ./doc/license.texi cannam@95: A ./doc/mpi.texi cannam@95: A ./doc/other.texi cannam@95: A ./doc/reference.texi cannam@95: A ./doc/threads.texi cannam@95: A ./doc/tutorial.texi cannam@95: A ./doc/upgrading.texi cannam@95: cannam@95: Fri Jun 17 17:26:50 EDT 2011 stevenj@fftw.org cannam@95: * merge recent Cell deletion with MPI branch cannam@95: cannam@95: M ./Makefile.am -3 +2 cannam@95: M ./NEWS +40 cannam@95: M ./configure.ac -1 +1 cannam@95: M ./doc/fftw3.texi -95 +909 cannam@95: cannam@95: Fri Jun 17 01:52:59 EDT 2011 stevenj@fftw.org cannam@95: * whoops, incorrect assertion cannam@95: cannam@95: M ./mpi/transpose-problem.c -2 +2 cannam@95: cannam@95: Fri Jun 17 01:52:51 EDT 2011 stevenj@fftw.org cannam@95: * comment fix cannam@95: cannam@95: M ./mpi/block.c -2 +2 cannam@95: cannam@95: Thu Jun 16 23:30:27 EDT 2011 stevenj@fftw.org cannam@95: * check if pln creation failed (e.g. for split input) bbefore calling setup_gather_scatter, to prevent crashes cannam@95: cannam@95: M ./mpi/mpi-bench.c -1 +1 cannam@95: cannam@95: Thu Jun 16 23:26:48 EDT 2011 stevenj@fftw.org cannam@95: * bug fix -- transpose-recurse is only applicable if subtransposes fit in the same space (unless I change the allocation routine, but this would seem to require looking at all possible recursive invocations of transpose-recurse) cannam@95: cannam@95: M ./mpi/transpose-recurse.c +28 cannam@95: cannam@95: Mon Apr 11 17:58:03 EDT 2011 stevenj@fftw.org cannam@95: * yikes, any_true check on subplan creation should be in comm, not comm2, so that all processes know if failure occurred cannam@95: cannam@95: M ./mpi/transpose-recurse.c -6 +4 cannam@95: cannam@95: Mon Apr 11 17:00:46 EDT 2011 stevenj@fftw.org cannam@95: * add wisdom_ok_hook to enforce wisdom synchronization on MPI problems, apparently fixing a longstanding deadlock/crash bug cannam@95: cannam@95: M ./kernel/ifftw.h +1 cannam@95: M ./kernel/planner.c +7 cannam@95: M ./mpi/api.c -13 +61 cannam@95: cannam@95: Mon Apr 11 15:10:22 EDT 2011 stevenj@fftw.org cannam@95: * add a check (in DEBUG mode only) that all processes produce the same hash of MPI problems; don't include alignment in MPI problem hash because it may differ between processes for unaligned malloc cannam@95: cannam@95: M ./mpi/any-true.c +21 cannam@95: M ./mpi/dft-problem.c -2 +6 cannam@95: M ./mpi/ifftw-mpi.h +1 cannam@95: M ./mpi/rdft-problem.c -2 +6 cannam@95: M ./mpi/rdft2-problem.c -2 +6 cannam@95: M ./mpi/transpose-problem.c -2 +6 cannam@95: cannam@95: Fri Apr 8 18:46:54 EDT 2011 stevenj@fftw.org cannam@95: * use cost_hook in random_estimate cannam@95: cannam@95: M ./kernel/planner.c -1 +1 cannam@95: cannam@95: Sun Mar 6 23:33:53 EST 2011 stevenj@fftw.org cannam@95: * added mpi new-array execute functions; thanks to Guo Luo for the bug report cannam@95: cannam@95: M ./mpi/api.c +23 cannam@95: M ./mpi/fftw3-mpi.h -1 +6 cannam@95: cannam@95: Wed Feb 9 21:29:17 EST 2011 stevenj@fftw.org cannam@95: * MPI may not support tags > 2^15-1 (e.g. Cray MPI requires tags < 2^24); thanks to Jonathan Bentz for the bug report. cannam@95: cannam@95: M ./mpi/transpose-pairwise.c -4 +4 cannam@95: cannam@95: Wed Feb 2 12:21:30 EST 2011 stevenj@fftw.org cannam@95: * fix merge conflicts cannam@95: cannam@95: M! ./configure.ac -1 +1 cannam@95: cannam@95: Sat Nov 15 22:33:20 EST 2008 stevenj@fftw.org cannam@95: tagged fftw-3.3alpha1 cannam@95: cannam@95: cannam@95: Sat Nov 15 20:33:33 EST 2008 stevenj@fftw.org cannam@95: * version bump for 3.3alpha1 cannam@95: cannam@95: M! ./NEWS -40 cannam@95: M! ./configure.ac -2 +2 cannam@95: cannam@95: Sun Oct 26 22:47:07 EDT 2008 stevenj@fftw.org cannam@95: * re-added mpi/Makefile cannam@95: cannam@95: M ./configure.ac +2 cannam@95: cannam@95: Sat Oct 25 17:14:42 EDT 2008 stevenj@fftw.org cannam@95: * re-add MPI to dist cannam@95: cannam@95: M! ./Makefile.am -2 +3 cannam@95: M! ./doc/fftw3.texi -9 +134 cannam@95: cannam@95: Sat Jun 18 08:50:13 EDT 2011 athena@fftw.org cannam@95: * remove obsolete Cell code cannam@95: cannam@95: M ./kernel/align.c -1 +1 cannam@95: M ./kernel/ifftw.h -9 +2 cannam@95: M ./rdft/rank0.c -83 cannam@95: M ./simd/simd.h -6 cannam@95: M ./simd/taint.c -1 +1 cannam@95: M ./tests/fftw-bench.c -7 cannam@95: cannam@95: Fri Jun 17 23:31:33 EDT 2011 stevenj@fftw.org cannam@95: * bug fix in accuracy test, which prevented us from consiistently determining accuracy in > double precision cannam@95: cannam@95: M ./libbench2/mp.c -2 +2 cannam@95: cannam@95: Fri Jun 17 20:05:13 EDT 2011 athena@fftw.org cannam@95: * do not check for gcc version before checking for gcc cannam@95: cannam@95: M ./configure.ac -1 +1 cannam@95: cannam@95: Fri Jun 17 18:56:37 EDT 2011 stevenj@fftw.org cannam@95: * require gcc 4.6.0 or later for --enable-quad-precision, to match fftw3.h header file; no need to mark this as EXPERIMENTAL (make check passes, and support in gcc 4.6 seems reasonably complete) cannam@95: cannam@95: M ./configure.ac -1 +2 cannam@95: cannam@95: Fri Jun 17 18:51:41 EDT 2011 stevenj@fftw.org cannam@95: * need ugly __attribute__ to use __float128 with _Complex, ugh cannam@95: cannam@95: M ./api/fftw3.h +9 cannam@95: cannam@95: Fri Jun 17 18:23:05 EDT 2011 stevenj@fftw.org cannam@95: * --verify tolerance in quad precision changed to 1e-29 cannam@95: cannam@95: M ./libbench2/bench-main.c -1 +1 cannam@95: cannam@95: Fri Jun 17 18:22:38 EDT 2011 stevenj@fftw.org cannam@95: * quad-precision F77 api should use "qfftw" prefix cannam@95: cannam@95: M ./api/x77.h +3 cannam@95: cannam@95: Fri Jun 17 18:22:27 EDT 2011 stevenj@fftw.org cannam@95: * rm extraneous space from fftw3.h cannam@95: cannam@95: M ./api/fftw3.h -1 +1 cannam@95: cannam@95: Fri Jun 17 18:05:10 EDT 2011 stevenj@fftw.org cannam@95: * use cosq etcetera with libquadmath in libbench2, so that --verify correctly gives ~33 decimal places in shift test cannam@95: cannam@95: M ./libbench2/verify.h -1 +4 cannam@95: cannam@95: Fri Jun 17 17:52:51 EDT 2011 stevenj@fftw.org cannam@95: * libquadmath ships with gcc 4.6.0, so we should require this library for sinq/cosq with --enable-quad-precision; also, include the __float128 FFTW functions in the header file for gcc >= 4.6 on i86/x86??_64/ia64 cannam@95: cannam@95: M ./Makefile.am +5 cannam@95: M ./api/fftw3.h -1 +6 cannam@95: M ./configure.ac -1 +5 cannam@95: M ./kernel/trig.c -13 +4 cannam@95: cannam@95: Fri Jun 17 16:54:01 EDT 2011 stevenj@fftw.org cannam@95: * typo in manual for fftw_cost cannam@95: cannam@95: M ./doc/fftw3.texi -1 +1 cannam@95: cannam@95: Fri Jun 17 16:48:24 EDT 2011 stevenj@fftw.org cannam@95: * fix fftw_cost function: pcost needs to be saved in mkapiplan, since the plan is re-created from wisdom cannam@95: cannam@95: M ./api/apiplan.c +5 cannam@95: M ./tests/fftw-bench.c -2 +3 cannam@95: cannam@95: Fri Jun 17 16:42:25 EDT 2011 athena@fftw.org cannam@95: * removed support for the Cell Broadband Engine cannam@95: cannam@95: R ./cell/spu/Makefile.am cannam@95: R ./cell/spu/alloc.spuc cannam@95: R ./cell/spu/copy.spuc cannam@95: R ./cell/spu/dft.spuc cannam@95: R ./cell/spu/dma.spuc cannam@95: R ./cell/spu/execute.spuc cannam@95: R ./cell/spu/fftw-spu.h cannam@95: R ./cell/spu/main.spuc cannam@95: R ./cell/spu/planner.spuc cannam@95: R ./cell/spu/spu-double.h cannam@95: R ./cell/spu/spu-single.h cannam@95: R ./cell/spu/transpose.spuc cannam@95: R ./cell/spu/ cannam@95: R ./cell/Makefile.am cannam@95: R ./cell/cell.c cannam@95: R ./cell/conf.c cannam@95: R ./cell/copy.c cannam@95: R ./cell/dft-direct-cell.c cannam@95: R ./cell/fftw-cell.h cannam@95: R ./cell/plans-double.c cannam@95: R ./cell/plans-single.c cannam@95: R ./cell/spufftw-embed.S cannam@95: R ./cell/transpose.c cannam@95: R ./cell/ cannam@95: M ./AUTHORS -2 +3 cannam@95: M ./Makefile.am -11 +5 cannam@95: M ./NEWS +2 cannam@95: R ./README.Cell cannam@95: M ./api/configure.c -3 cannam@95: M ./configure.ac -18 +1 cannam@95: M ./dft/dft.h -4 cannam@95: M ./dft/rank-geq2.c -6 cannam@95: M ./doc/fftw3.texi -128 +4 cannam@95: cannam@95: Tue May 24 06:51:07 EDT 2011 athena@fftw.org cannam@95: * Undo previous change; the typo was not a typo after all. cannam@95: cannam@95: M ./doc/fftw3.texi -1 +1 cannam@95: cannam@95: Mon May 23 05:08:05 EDT 2011 athena@fftw.org cannam@95: * Fix typo in manual cannam@95: cannam@95: M ./doc/fftw3.texi -1 +1 cannam@95: cannam@95: Sat May 21 17:37:50 EDT 2011 athena@fftw.org cannam@95: * clarify intent about canonicalization of tensor in tensor_compress_contiguous() cannam@95: cannam@95: M ./kernel/tensor7.c -1 +9 cannam@95: cannam@95: Sat May 21 17:30:31 EDT 2011 athena@fftw.org cannam@95: * avoid useless canonicalization in tensor_compress_contiguous() cannam@95: cannam@95: M ./kernel/tensor7.c -1 +1 cannam@95: cannam@95: Sat May 21 17:24:57 EDT 2011 athena@fftw.org cannam@95: * Fix tensor_compress_contiguous cannam@95: cannam@95: tensor_compress_contiguous() was supposed to sort dimensions by cannam@95: descending istride, and then compress adjacent dimensions. This cannam@95: property was lost once we changed the canonical order of strides to be cannam@95: sorted by descending min{istride,ostride}. cannam@95: cannam@95: Change tensor_compress_contiguous() to sort by descending istride cannam@95: again, which is necessary for its correctness, and then canonicalize cannam@95: at the end. cannam@95: cannam@95: cannam@95: M ./kernel/tensor7.c -20 +41 cannam@95: cannam@95: Sun May 8 18:47:26 EDT 2011 athena@fftw.org cannam@95: * Don't distribute obsolete .depend cannam@95: cannam@95: M ./genfft/Makefile.am -2 +2 cannam@95: cannam@95: Sun May 8 18:05:36 EDT 2011 athena@fftw.org cannam@95: * Use ocamlbuild for building genfft cannam@95: cannam@95: Remove the old Makefile cruft to support ocaml, and use ocamlbuild cannam@95: instead. cannam@95: cannam@95: M ./bootstrap.sh -2 cannam@95: M ./configure.ac -10 +1 cannam@95: M ./genfft/Makefile.am -151 +8 cannam@95: M ./m4/Makefile.am -1 +1 cannam@95: R ./m4/ocaml.m4 cannam@95: M ./support/Makefile.codelets -13 +13 cannam@95: cannam@95: Sun May 8 18:03:07 EDT 2011 athena@fftw.org cannam@95: * Do not use __float128 unless BENCHFFT_QUAD is defined cannam@95: cannam@95: Otherwise, compilation fails on compilers that do not support cannam@95: __float128. cannam@95: cannam@95: cannam@95: M ./libbench2/bench-user.h +5 cannam@95: cannam@95: Fri Apr 8 13:15:54 EDT 2011 stevenj@fftw.org cannam@95: * fix configure --help string for --disable-alloca (since default is enabled) cannam@95: cannam@95: M ./configure.ac -1 +1 cannam@95: cannam@95: Fri Apr 8 13:09:56 EDT 2011 stevenj@fftw.org cannam@95: * add "random estimator" for debugging purposes; note that this is best used with ESTIMATE_PATIENT mode cannam@95: cannam@95: M ./api/apiplan.c +4 cannam@95: M ./configure.ac +5 cannam@95: M ./kernel/ifftw.h +4 cannam@95: M ./kernel/planner.c +27 cannam@95: M ./tests/fftw-bench.c +7 cannam@95: cannam@95: Tue Apr 5 14:47:56 EDT 2011 stevenj@fftw.org cannam@95: * add AC_CHECK_DECLS for srand48; thanks to Ralf Wildenhues for the bug report cannam@95: cannam@95: M ./configure.ac -1 +1 cannam@95: M ./libbench2/util.c +3 cannam@95: cannam@95: Sat Feb 5 17:00:40 EST 2011 stevenj@fftw.org cannam@95: * experimental support for gcc's __float128 quad-precision type cannam@95: cannam@95: M ./api/fftw3.h +3 cannam@95: M ./configure.ac -2 +14 cannam@95: M ./kernel/ifftw.h -2 +10 cannam@95: M ./kernel/trig.c -1 +16 cannam@95: M ./libbench2/bench-main.c +2 cannam@95: M ./libbench2/bench-user.h +4 cannam@95: M ./libbench2/info.c -1 +2 cannam@95: M ./libbench2/verify.h +6 cannam@95: M ./simd/simd-sse2.h -1 +1 cannam@95: M ./tests/fftw-bench.h +2 cannam@95: M ./tools/fftw-wisdom.c +2 cannam@95: cannam@95: Sun Oct 24 14:33:59 EDT 2010 athena@fftw.org cannam@95: * guarantee that "timelimit < 0" means "no timeout" cannam@95: "timelimit < 0" was always meant to be equivalent to cannam@95: "timelimit = HUGENUM", but this was not true in all cases, cannam@95: causing some obscure wisdom behavior. cannam@95: cannam@95: Thanks William Andrew Burnson for the bug report. cannam@95: cannam@95: cannam@95: M ./api/mapflags.c -1 +1 cannam@95: cannam@95: Sun Oct 24 14:32:20 EDT 2010 athena@fftw.org cannam@95: * compile with --enable-fma and SSE, SSE2 cannam@95: Allow compilation with --enable-fma and --enable-sse, --enable-sse2. cannam@95: This is a bad idea performance-wise, but people will try anyway. cannam@95: cannam@95: M ./simd/simd-sse.h +5 cannam@95: M ./simd/simd-sse2.h +5 cannam@95: cannam@95: Sun Jul 11 13:34:06 EDT 2010 athena@fftw.org cannam@95: * Make threads.c compiler with c++ cannam@95: cannam@95: M ./threads/threads.c -1 +1 cannam@95: cannam@95: Sun Jul 11 10:05:05 EDT 2010 athena@fftw.org cannam@95: * Attempt at clarifying the advanced interface doc. cannam@95: cannam@95: M ./doc/fftw3.texi -72 +151 cannam@95: cannam@95: Sun Jul 11 07:37:27 EDT 2010 athena@fftw.org cannam@95: * rename rfftwnd html picture cannam@95: cannam@95: It turns out that texinfo with pdf output reads .png cannam@95: files in preference to .pdf files (when did this change?). cannam@95: I renamed the .png figure to avoid producing an ugly pdf file. cannam@95: cannam@95: M ./doc/Makefile.am -4 +4 cannam@95: M ./doc/fftw3.texi +5 cannam@95: cannam@95: Tue Mar 30 19:43:22 EDT 2010 stevenj@fftw.org cannam@95: * added fftw_cost function; this is the second time people have asked for this, and there is a reasonable use for it in comparing e.g. oout-of-place vs. in-place plans cannam@95: cannam@95: M ./NEWS +4 cannam@95: M ./api/f77funcs.h +10 cannam@95: M ./api/fftw3.h +1 cannam@95: M ./api/flops.c +5 cannam@95: M ./doc/fftw3.texi +17 cannam@95: cannam@95: Tue Mar 2 18:55:49 EST 2010 stevenj@fftw.org cannam@95: * documented that --enable-debug-malloc causes fftw_execute to be thread-unsafe (thanks to Alexis Rohou for the problem report) cannam@95: cannam@95: M ./doc/fftw3.texi +8 cannam@95: cannam@95: Fri Jan 22 19:42:08 EST 2010 athena@fftw.org cannam@95: * Added FAQ about how to transpose matrices using FFTW. cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn +33 cannam@95: cannam@95: Thu Jan 7 20:16:57 EST 2010 stevenj@fftw.org cannam@95: * catch FMS (instead of generating FMA(_,_,NEG(_)) with h -generic-arith option cannam@95: cannam@95: M ./genfft/c.ml +2 cannam@95: cannam@95: Fri Dec 11 07:01:26 EST 2009 athena@fftw.org cannam@95: * note future wisdom enhancements. cannam@95: cannam@95: M ./TODO +5 cannam@95: cannam@95: Mon Oct 19 20:21:05 EDT 2009 athena@fftw.org cannam@95: * Use SIMD flags when checking for xmmintrin.h cannam@95: cannam@95: This prevents an obnoxious warning from configure. cannam@95: cannam@95: M ./configure.ac -3 +3 cannam@95: cannam@95: Sat Aug 29 20:47:56 EDT 2009 athena@fftw.org cannam@95: * new bug cannam@95: cannam@95: M ./TODO +4 cannam@95: cannam@95: Sun Jul 26 00:40:11 EDT 2009 stevenj@fftw.org cannam@95: * typo (s/man1/many) cannam@95: cannam@95: M ./doc/fftw3.texi -1 +1 cannam@95: cannam@95: Tue Jul 14 14:19:08 EDT 2009 stevenj@fftw.org cannam@95: * BUILD-MINGW32 script, updated Windows README cannam@95: cannam@95: A ./support/BUILD-MINGW32.sh cannam@95: M ./support/BUILD-MINGW64.sh -4 +4 cannam@95: cannam@95: Mon Jul 13 09:40:38 EDT 2009 fftw@fftw.org cannam@95: * cleanup BUILD-MINGW64.sh cannam@95: cannam@95: M ./support/BUILD-MINGW64.sh -19 +27 cannam@95: cannam@95: Sun Jul 12 06:34:46 EDT 2009 athena@fftw.org cannam@95: * Update NEWS, version number for 3.2.2 release. cannam@95: cannam@95: M ./NEWS +3 cannam@95: M ./configure.ac -2 +2 cannam@95: cannam@95: Sat Jul 11 22:28:38 EDT 2009 athena@fftw.org cannam@95: * Reintroduce the pruning heuristic in ESTIMATE mode for r2r problems. cannam@95: cannam@95: Somehow, we lost this feature between fftw-3.1.3 and fftw-3.2. cannam@95: cannam@95: M ./rdft/hc2hc-direct.c +3 cannam@95: cannam@95: Thu Jun 25 07:39:04 EDT 2009 athena@fftw.org cannam@95: * don't use pshared=1 in sem_init cannam@95: cannam@95: pshared is really not necessary, and it is not supported on cannam@95: GNU/kFreeBSD. Thanks Petr Salinger for the bug report. cannam@95: cannam@95: M ./threads/threads.c -2 +2 cannam@95: cannam@95: Thu Jun 11 19:35:40 EDT 2009 fftw@fftw.org cannam@95: * Add mingw64 build file so that we can track it. cannam@95: cannam@95: A ./support/BUILD-MINGW64.sh cannam@95: cannam@95: Wed Jun 10 12:10:58 EDT 2009 fftw@fftw.org cannam@95: * note 3.2.2 NEWS cannam@95: cannam@95: M ./NEWS +7 cannam@95: cannam@95: Wed Jun 10 12:04:54 EDT 2009 fftw@fftw.org cannam@95: * add --disable-alloca to configure cannam@95: cannam@95: It looks like alloca() is broken on mingw64, and thus cannam@95: we need to disable it explicitly. cannam@95: cannam@95: M ./configure.ac -1 +6 cannam@95: M ./kernel/ifftw.h -1 +1 cannam@95: cannam@95: Sun Apr 26 16:33:10 EDT 2009 athena@fftw.org cannam@95: * Note in FAQ that --enable-k7 has been discontinued. cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -4 +7 cannam@95: cannam@95: Tue Mar 24 09:16:18 EDT 2009 athena@fftw.org cannam@95: * clarified small confusion in fftw_cleanup documentation cannam@95: cannam@95: M ./doc/fftw3.texi -2 +3 cannam@95: cannam@95: Thu Mar 19 13:18:06 EDT 2009 stevenj@fftw.org cannam@95: * fix documentation of dfftw_init_threads to indicate thaat it takes an argument (since the C version returns a value); thanks t Hans Johnnston for the bug report cannam@95: cannam@95: M ./doc/fftw3.texi -1 +5 cannam@95: cannam@95: Thu Mar 12 13:12:13 EDT 2009 fftw@fftw.org cannam@95: * if possible, use a 128-bit type for copy cannam@95: cannam@95: M ./configure.ac -1 +1 cannam@95: M ./kernel/cpy2d.c -9 +38 cannam@95: cannam@95: Tue Mar 10 12:49:51 EDT 2009 fftw@fftw.org cannam@95: * add size-128 simd codelets cannam@95: cannam@95: It's about time cannam@95: cannam@95: M ./dft/simd/codelets/Makefile.am -2 +2 cannam@95: cannam@95: Mon Mar 9 20:29:16 EDT 2009 athena@fftw.org cannam@95: * copy two floats as a double when possible cannam@95: cannam@95: Resurrect the old hack of copying two floats as a double, cannam@95: which makes some difference in these days of 64 bit boxes. cannam@95: cannam@95: cannam@95: M ./kernel/cpy2d.c -16 +34 cannam@95: cannam@95: Sun Mar 8 18:08:04 EDT 2009 athena@fftw.org cannam@95: * fixed (harmless) confusion of strides cannam@95: cannam@95: RS and VS were swapped in dftw-direct.c. This is a bug, but cannam@95: it is harmless unless one uses fixed-stride codelets, which we cannam@95: do not. cannam@95: cannam@95: M ./dft/dftw-direct.c -7 +7 cannam@95: cannam@95: Sun Mar 8 10:29:49 EDT 2009 athena@fftw.org cannam@95: * oops, I checked in debug code accidentally. cannam@95: cannam@95: M ./kernel/ct.c -1 cannam@95: cannam@95: Sat Feb 14 19:01:00 EST 2009 athena@fftw.org cannam@95: * Change TLO email address since Magdalen is no longer there. cannam@95: cannam@95: M ./commercialize.sh -1 +1 cannam@95: cannam@95: Sat Feb 14 18:18:45 EST 2009 stevenj@fftw.org cannam@95: * quote arguments to bench in test script on the off-chance that '*' would be expanded by the shell into a valid filename, and also to avoid shell confusion on Cygwin that "//" begins the name of a Windows network mountpoint cannam@95: cannam@95: M ./tests/check.pl -1 +1 cannam@95: cannam@95: Sat Feb 14 18:17:23 EST 2009 stevenj@fftw.org cannam@95: tagged fftw-3.2.1 cannam@95: cannam@95: cannam@95: Sun Mar 8 10:02:59 EDT 2009 athena@fftw.org cannam@95: * stricter conditions for Cooley-Tukey being ugly cannam@95: cannam@95: It turns out that m=2 in the leaf of Cooley-Tukey may be cannam@95: advantageous in certain cases, eg. i512v512 on AMD Shanghai: cannam@95: cannam@95: (dft-buffered-512-x128/512-6 cannam@95: (dft-ct-dit/4 cannam@95: (dftw-direct-4/24-x128 "t2fv_4") cannam@95: (dft-vrank>=1-x4/1 cannam@95: (dft-ct-dit/64 cannam@95: (dftw-direct-64/504-x128 "t2fv_64") cannam@95: (dft-vrank>=1-x64/1 cannam@95: (dft-direct-2-x128 "n2fv_2"))))) cannam@95: (dft-r2hc-1 cannam@95: (rdft-rank0-tiled/2-x128-x512)) cannam@95: (dft-nop)) cannam@95: cannam@95: Presumably this works around the 2 way associativity of the L1 cache. cannam@95: cannam@95: M ./dft/dftw-direct.c -1 +1 cannam@95: M ./kernel/ct.c -2 +3 cannam@95: M ./kernel/ifftw.h -1 +1 cannam@95: M ./rdft/ct-hc2c-direct.c -1 +1 cannam@95: M ./rdft/hc2hc-direct.c -3 +3 cannam@95: cannam@95: Mon Feb 9 19:46:00 EST 2009 stevenj@fftw.org cannam@95: * disable Windows QueryPerformanceCounter code, since it requires us to pull in windows.h in ifftw.h and causes namespace conflicts; gettimeofday seems to work well enough and has had few complaints cannam@95: cannam@95: M ./kernel/ifftw.h -4 +1 cannam@95: M ./kernel/timer.c -1 +8 cannam@95: cannam@95: Wed Feb 4 22:55:54 EST 2009 stevenj@fftw.org cannam@95: * version bump for 3.2.1, updated NEWS cannam@95: cannam@95: M ./NEWS +10 cannam@95: M ./configure.ac -1 +1 cannam@95: cannam@95: Wed Feb 4 22:27:28 EST 2009 stevenj@fftw.org cannam@95: * recommend that users avoid fftw_execute in Fortran, instead using dfftw_execute_dft and friends so that the compiler knows that the input/output arrays are used cannam@95: cannam@95: M ./doc/fftw3.texi -10 +80 cannam@95: cannam@95: Wed Jan 21 16:02:08 EST 2009 stevenj@fftw.org cannam@95: * prefer windows queryperformancecounter to gettimeofday on Windows, thanks to David Price for the suggestion cannam@95: cannam@95: M ./kernel/ifftw.h -1 +4 cannam@95: M ./kernel/timer.c -1 +20 cannam@95: cannam@95: Sun Feb 1 14:34:49 EST 2009 athena@fftw.org cannam@95: * compilation fixes in case snprintf() is defined as a macro. cannam@95: cannam@95: M ./libbench2/report.c -8 +7 cannam@95: cannam@95: Wed Jan 28 20:19:04 EST 2009 athena@fftw.org cannam@95: * Automake does not like continuation lines beginning with a comment. cannam@95: cannam@95: M ./rdft/scalar/r2cb/Makefile.am -1 +1 cannam@95: cannam@95: Wed Jan 28 18:24:39 EST 2009 athena@fftw.org cannam@95: * Add r2cb_2.c cannam@95: cannam@95: r2cb_2.c is needed for problem rb2, which is not equivalent to cannam@95: rf2 (unlike kb2, which is equivalent to kf2). cannam@95: cannam@95: This change would not matter much except that rb2 is generated cannam@95: when reducing backward rdft2 to dft, and the absence of the codelet was cannam@95: preventing radix 2 from being employed at all in this case. cannam@95: cannam@95: M ./rdft/scalar/r2cb/Makefile.am -6 +4 cannam@95: cannam@95: Sat Jan 10 06:47:22 EST 2009 athena@fftw.org cannam@95: * handle the case vecsz->rnk == 0 correctly. cannam@95: cannam@95: M ./dft/buffered.c -1 +3 cannam@95: M ./rdft/buffered.c -1 +3 cannam@95: M ./rdft/buffered2.c -1 +3 cannam@95: cannam@95: Fri Dec 19 15:20:36 EST 2008 stevenj@fftw.org cannam@95: * Macs are no longer ppc-based; thanks to Charles Collicutt for the FAQ update cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -3 +3 cannam@95: cannam@95: Mon Dec 8 18:08:33 EST 2008 stevenj@fftw.org cannam@95: * use new multiple-nbuf code in rdft/buffered, like for dft/buffered cannam@95: cannam@95: M ./rdft/buffered.c -11 +25 cannam@95: M ./rdft/buffered2.c -12 +26 cannam@95: cannam@95: Sat Dec 6 16:34:36 EST 2008 stevenj@fftw.org cannam@95: * make x86_cpuid macro work on x86_64 cannam@95: cannam@95: M ./m4/ax_gcc_x86_cpuid.m4 -1 +15 cannam@95: cannam@95: Sat Dec 6 09:20:37 EST 2008 athena@fftw.org cannam@95: * Allow automatic choice of buffer size in dft/buffered.c cannam@95: cannam@95: Try a couple of different buffer sizes in buffered transforms, cannam@95: since this seems to make a difference on some Core2 models. cannam@95: cannam@95: cannam@95: M ./api/mapflags.c -2 cannam@95: M ./dft/buffered.c -10 +25 cannam@95: M ./kernel/buffered.c -3 +19 cannam@95: M ./kernel/ifftw.h -1 +5 cannam@95: M ./rdft/buffered.c -2 +3 cannam@95: M ./rdft/buffered2.c -2 +3 cannam@95: M ./rdft/rdft2-rdft.c -1 +1 cannam@95: cannam@95: Tue Dec 2 19:18:30 EST 2008 athena@fftw.org cannam@95: * libbench2: do not assume that split-complex arrays are stride-1 cannam@95: cannam@95: M ./libbench2/verify-dft.c -4 +3 cannam@95: M ./libbench2/verify.c -1 +3 cannam@95: M ./tests/bench.c -3 +2 cannam@95: cannam@95: Tue Dec 2 18:39:43 EST 2008 stevenj@fftw.org cannam@95: * updated NEWS cannam@95: cannam@95: M ./NEWS +14 cannam@95: cannam@95: Tue Dec 2 18:30:00 EST 2008 stevenj@fftw.org cannam@95: * date fix cannam@95: cannam@95: M ./m4/ax_gcc_archflag.m4 -1 +1 cannam@95: cannam@95: Tue Dec 2 18:29:06 EST 2008 stevenj@fftw.org cannam@95: * updated icc flags -- now prefer -xHost (-xN etc. seem t be obsolete), check for new spelling -ansi-alias, and use -malign-double like we do for gcc cannam@95: cannam@95: M ./m4/ax_cc_maxopt.m4 -2 +14 cannam@95: cannam@95: Tue Dec 2 18:28:03 EST 2008 stevenj@fftw.org cannam@95: * use $ax_cv_c_compiler_vendor rather than $GCC, as the former is more reliable (icc incorrectly self-identifies as gcc on MacOS where we don't use -no-gcc) cannam@95: cannam@95: M ./configure.ac -1 +1 cannam@95: cannam@95: Tue Dec 2 17:55:36 EST 2008 stevenj@fftw.org cannam@95: * don't use -no-gcc for icc on MacOS cannam@95: cannam@95: M ./configure.ac -2 +7 cannam@95: cannam@95: Tue Dec 2 17:34:04 EST 2008 stevenj@fftw.org cannam@95: * document some more bench options cannam@95: cannam@95: M ./tests/README -1 +17 cannam@95: cannam@95: Wed Nov 19 16:55:13 EST 2008 stevenj@fftw.org cannam@95: * make it clearer that --enable-openmp and --enable-threads are mutually exclusive; thanks to Long To for his comments cannam@95: cannam@95: M ./doc/fftw3.texi -6 +10 cannam@95: cannam@95: Mon Nov 17 20:16:28 EST 2008 stevenj@fftw.org cannam@95: * version bump to 3.2.1, use explicit Makefile.am for m4 subdirectory so that tarball does not include random files in there when you do 'make dist' cannam@95: cannam@95: M ./Makefile.am -3 +2 cannam@95: M ./configure.ac -2 +4 cannam@95: A ./m4/Makefile.am cannam@95: cannam@95: Sat Nov 15 21:12:58 EST 2008 stevenj@fftw.org cannam@95: * document behavior of FFTW guru arrays, and in particular the odd behavior of the plan_guru_r2r routine in Fortran (thanks to Alexander Pozdneev for the bug report) cannam@95: cannam@95: M ./doc/fftw3.texi +10 cannam@95: cannam@95: Sat Nov 15 21:03:14 EST 2008 stevenj@fftw.org cannam@95: tagged fftw-3.2 cannam@95: cannam@95: cannam@95: Mon Nov 10 20:21:32 EST 2008 stevenj@fftw.org cannam@95: * version bump to 3.2, updated copyright year cannam@95: cannam@95: M ./COPYRIGHT -2 +2 cannam@95: M ./NEWS -1 +6 cannam@95: M ./api/api.h -2 +2 cannam@95: M ./api/apiplan.c -2 +2 cannam@95: M ./api/configure.c -2 +2 cannam@95: M ./api/execute-dft-c2r.c -2 +2 cannam@95: M ./api/execute-dft-r2c.c -2 +2 cannam@95: M ./api/execute-dft.c -2 +2 cannam@95: M ./api/execute-r2r.c -2 +2 cannam@95: M ./api/execute-split-dft-c2r.c -2 +2 cannam@95: M ./api/execute-split-dft-r2c.c -2 +2 cannam@95: M ./api/execute-split-dft.c -2 +2 cannam@95: M ./api/execute.c -2 +2 cannam@95: M ./api/export-wisdom-to-file.c -2 +2 cannam@95: M ./api/export-wisdom-to-string.c -2 +2 cannam@95: M ./api/export-wisdom.c -2 +2 cannam@95: M ./api/f77api.c -2 +2 cannam@95: M ./api/f77funcs.h -2 +2 cannam@95: M ./api/fftw3.h -2 +2 cannam@95: M ./api/flops.c -2 +2 cannam@95: M ./api/forget-wisdom.c -2 +2 cannam@95: M ./api/import-system-wisdom.c -2 +2 cannam@95: M ./api/import-wisdom-from-file.c -2 +2 cannam@95: M ./api/import-wisdom-from-string.c -2 +2 cannam@95: M ./api/import-wisdom.c -2 +2 cannam@95: M ./api/malloc.c -2 +2 cannam@95: M ./api/map-r2r-kind.c -2 +2 cannam@95: M ./api/mapflags.c -2 +2 cannam@95: M ./api/mkprinter-file.c -2 +2 cannam@95: M ./api/mktensor-iodims.h -2 +2 cannam@95: M ./api/mktensor-rowmajor.c -2 +2 cannam@95: M ./api/plan-dft-1d.c -2 +2 cannam@95: M ./api/plan-dft-2d.c -2 +2 cannam@95: M ./api/plan-dft-3d.c -2 +2 cannam@95: M ./api/plan-dft-c2r-1d.c -2 +2 cannam@95: M ./api/plan-dft-c2r-2d.c -2 +2 cannam@95: M ./api/plan-dft-c2r-3d.c -2 +2 cannam@95: M ./api/plan-dft-c2r.c -2 +2 cannam@95: M ./api/plan-dft-r2c-1d.c -2 +2 cannam@95: M ./api/plan-dft-r2c-2d.c -2 +2 cannam@95: M ./api/plan-dft-r2c-3d.c -2 +2 cannam@95: M ./api/plan-dft-r2c.c -2 +2 cannam@95: M ./api/plan-dft.c -2 +2 cannam@95: M ./api/plan-guru-dft-c2r.h -2 +2 cannam@95: M ./api/plan-guru-dft-r2c.h -2 +2 cannam@95: M ./api/plan-guru-dft.h -2 +2 cannam@95: M ./api/plan-guru-r2r.h -2 +2 cannam@95: M ./api/plan-guru-split-dft-c2r.h -2 +2 cannam@95: M ./api/plan-guru-split-dft-r2c.h -2 +2 cannam@95: M ./api/plan-guru-split-dft.h -2 +2 cannam@95: M ./api/plan-many-dft-c2r.c -2 +2 cannam@95: M ./api/plan-many-dft-r2c.c -2 +2 cannam@95: M ./api/plan-many-dft.c -2 +2 cannam@95: M ./api/plan-many-r2r.c -2 +2 cannam@95: M ./api/plan-r2r-1d.c -2 +2 cannam@95: M ./api/plan-r2r-2d.c -2 +2 cannam@95: M ./api/plan-r2r-3d.c -2 +2 cannam@95: M ./api/plan-r2r.c -2 +2 cannam@95: M ./api/print-plan.c -2 +2 cannam@95: M ./api/rdft2-pad.c -2 +2 cannam@95: M ./api/the-planner.c -2 +2 cannam@95: M ./api/version.c -2 +2 cannam@95: M ./api/x77.h -2 +2 cannam@95: M ./configure.ac -2 +2 cannam@95: M ./dft/bluestein.c -2 +2 cannam@95: M ./dft/buffered.c -2 +2 cannam@95: M ./dft/codelet-dft.h -2 +2 cannam@95: M ./dft/conf.c -2 +2 cannam@95: M ./dft/ct.c -2 +2 cannam@95: M ./dft/ct.h -2 +2 cannam@95: M ./dft/dft.h -2 +2 cannam@95: M ./dft/dftw-direct.c -2 +2 cannam@95: M ./dft/dftw-directsq.c -2 +2 cannam@95: M ./dft/dftw-generic.c -2 +2 cannam@95: M ./dft/dftw-genericbuf.c -2 +2 cannam@95: M ./dft/direct.c -2 +2 cannam@95: M ./dft/generic.c -2 +2 cannam@95: M ./dft/indirect-transpose.c -2 +2 cannam@95: M ./dft/indirect.c -2 +2 cannam@95: M ./dft/kdft-dif.c -2 +2 cannam@95: M ./dft/kdft-difsq.c -2 +2 cannam@95: M ./dft/kdft-dit.c -2 +2 cannam@95: M ./dft/kdft.c -2 +2 cannam@95: M ./dft/nop.c -2 +2 cannam@95: M ./dft/plan.c -2 +2 cannam@95: M ./dft/problem.c -2 +2 cannam@95: M ./dft/rader.c -2 +2 cannam@95: M ./dft/rank-geq2.c -2 +2 cannam@95: M ./dft/scalar/n.c -2 +2 cannam@95: M ./dft/scalar/n.h -2 +2 cannam@95: M ./dft/scalar/t.c -2 +2 cannam@95: M ./dft/scalar/t.h -2 +2 cannam@95: M ./dft/simd/n1b.c -2 +2 cannam@95: M ./dft/simd/n1b.h -2 +2 cannam@95: M ./dft/simd/n1f.c -2 +2 cannam@95: M ./dft/simd/n1f.h -2 +2 cannam@95: M ./dft/simd/n2b.c -2 +2 cannam@95: M ./dft/simd/n2b.h -2 +2 cannam@95: M ./dft/simd/n2f.c -2 +2 cannam@95: M ./dft/simd/n2f.h -2 +2 cannam@95: M ./dft/simd/n2s.c -2 +2 cannam@95: M ./dft/simd/n2s.h -2 +2 cannam@95: M ./dft/simd/q1b.c -2 +2 cannam@95: M ./dft/simd/q1b.h -2 +2 cannam@95: M ./dft/simd/q1f.c -2 +2 cannam@95: M ./dft/simd/q1f.h -2 +2 cannam@95: M ./dft/simd/t.c -2 +2 cannam@95: M ./dft/simd/t1b.h -2 +2 cannam@95: M ./dft/simd/t1bu.h -2 +2 cannam@95: M ./dft/simd/t1f.h -2 +2 cannam@95: M ./dft/simd/t1fu.h -2 +2 cannam@95: M ./dft/simd/t2b.h -2 +2 cannam@95: M ./dft/simd/t2f.h -2 +2 cannam@95: M ./dft/simd/t3b.h -2 +2 cannam@95: M ./dft/simd/t3f.h -2 +2 cannam@95: M ./dft/simd/ts.c -2 +2 cannam@95: M ./dft/simd/ts.h -2 +2 cannam@95: M ./dft/solve.c -2 +2 cannam@95: M ./dft/vrank-geq1.c -2 +2 cannam@95: M ./dft/zero.c -2 +2 cannam@95: M ./doc/f77_wisdom.f -2 +2 cannam@95: M ./genfft/algsimp.ml -2 +2 cannam@95: M ./genfft/algsimp.mli -2 +2 cannam@95: M ./genfft/annotate.ml -2 +2 cannam@95: M ./genfft/annotate.mli -2 +2 cannam@95: M ./genfft/assoctable.ml -2 +2 cannam@95: M ./genfft/assoctable.mli -2 +2 cannam@95: M ./genfft/c.ml -2 +2 cannam@95: M ./genfft/c.mli -2 +2 cannam@95: M ./genfft/complex.ml -2 +2 cannam@95: M ./genfft/complex.mli -2 +2 cannam@95: M ./genfft/conv.ml -2 +2 cannam@95: M ./genfft/conv.mli -2 +2 cannam@95: M ./genfft/dag.ml -2 +2 cannam@95: M ./genfft/dag.mli -2 +2 cannam@95: M ./genfft/expr.ml -2 +2 cannam@95: M ./genfft/expr.mli -2 +2 cannam@95: M ./genfft/fft.ml -2 +2 cannam@95: M ./genfft/fft.mli -2 +2 cannam@95: M ./genfft/gen_hc2c.ml -2 +2 cannam@95: M ./genfft/gen_hc2cdft.ml -2 +2 cannam@95: M ./genfft/gen_hc2cdft_c.ml -2 +2 cannam@95: M ./genfft/gen_hc2hc.ml -2 +2 cannam@95: M ./genfft/gen_mdct.ml -2 +2 cannam@95: M ./genfft/gen_notw.ml -2 +2 cannam@95: M ./genfft/gen_notw_c.ml -2 +2 cannam@95: M ./genfft/gen_r2cb.ml -2 +2 cannam@95: M ./genfft/gen_r2cf.ml -2 +2 cannam@95: M ./genfft/gen_r2r.ml -2 +2 cannam@95: M ./genfft/gen_twiddle.ml -2 +2 cannam@95: M ./genfft/gen_twiddle_c.ml -2 +2 cannam@95: M ./genfft/gen_twidsq.ml -2 +2 cannam@95: M ./genfft/gen_twidsq_c.ml -2 +2 cannam@95: M ./genfft/genutil.ml -2 +2 cannam@95: M ./genfft/littlesimp.ml -2 +2 cannam@95: M ./genfft/littlesimp.mli -2 +2 cannam@95: M ./genfft/magic.ml -2 +2 cannam@95: M ./genfft/monads.ml -2 +2 cannam@95: M ./genfft/number.ml -2 +2 cannam@95: M ./genfft/number.mli -2 +2 cannam@95: M ./genfft/oracle.ml -2 +2 cannam@95: M ./genfft/oracle.mli -2 +2 cannam@95: M ./genfft/schedule.ml -2 +2 cannam@95: M ./genfft/schedule.mli -2 +2 cannam@95: M ./genfft/simd.ml -2 +2 cannam@95: M ./genfft/simd.mli -2 +2 cannam@95: M ./genfft/simdmagic.ml -2 +2 cannam@95: M ./genfft/to_alist.ml -2 +2 cannam@95: M ./genfft/to_alist.mli -2 +2 cannam@95: M ./genfft/trig.ml -2 +2 cannam@95: M ./genfft/trig.mli -2 +2 cannam@95: M ./genfft/twiddle.ml -2 +2 cannam@95: M ./genfft/twiddle.mli -2 +2 cannam@95: M ./genfft/unique.ml -2 +2 cannam@95: M ./genfft/unique.mli -2 +2 cannam@95: M ./genfft/util.ml -2 +2 cannam@95: M ./genfft/util.mli -2 +2 cannam@95: M ./genfft/variable.ml -2 +2 cannam@95: M ./genfft/variable.mli -2 +2 cannam@95: M ./kernel/align.c -2 +2 cannam@95: M ./kernel/alloc.c -2 +2 cannam@95: M ./kernel/assert.c -2 +2 cannam@95: M ./kernel/awake.c -2 +2 cannam@95: M ./kernel/buffered.c -2 +2 cannam@95: M ./kernel/cpy1d.c -2 +2 cannam@95: M ./kernel/cpy2d-pair.c -2 +2 cannam@95: M ./kernel/cpy2d.c -2 +2 cannam@95: M ./kernel/ct.c -2 +2 cannam@95: M ./kernel/cycle.h -2 +2 cannam@95: M ./kernel/debug.c -2 +2 cannam@95: M ./kernel/extract-reim.c -2 +2 cannam@95: M ./kernel/hash.c -2 +2 cannam@95: M ./kernel/iabs.c -2 +2 cannam@95: M ./kernel/ifftw.h -2 +2 cannam@95: M ./kernel/kalloc.c -2 +2 cannam@95: M ./kernel/md5-1.c -2 +2 cannam@95: M ./kernel/md5.c -2 +2 cannam@95: M ./kernel/minmax.c -2 +2 cannam@95: M ./kernel/ops.c -2 +2 cannam@95: M ./kernel/pickdim.c -2 +2 cannam@95: M ./kernel/plan.c -2 +2 cannam@95: M ./kernel/primes.c -2 +2 cannam@95: M ./kernel/print.c -2 +2 cannam@95: M ./kernel/problem.c -2 +2 cannam@95: M ./kernel/rader.c -2 +2 cannam@95: M ./kernel/scan.c -2 +2 cannam@95: M ./kernel/solver.c -2 +2 cannam@95: M ./kernel/solvtab.c -2 +2 cannam@95: M ./kernel/stride.c -2 +2 cannam@95: M ./kernel/tensor.c -2 +2 cannam@95: M ./kernel/tensor1.c -2 +2 cannam@95: M ./kernel/tensor2.c -2 +2 cannam@95: M ./kernel/tensor3.c -2 +2 cannam@95: M ./kernel/tensor4.c -2 +2 cannam@95: M ./kernel/tensor5.c -2 +2 cannam@95: M ./kernel/tensor7.c -2 +2 cannam@95: M ./kernel/tensor8.c -2 +2 cannam@95: M ./kernel/tensor9.c -2 +2 cannam@95: M ./kernel/tile2d.c -2 +2 cannam@95: M ./kernel/timer.c -2 +2 cannam@95: M ./kernel/transpose.c -2 +2 cannam@95: M ./kernel/trig.c -2 +2 cannam@95: M ./kernel/twiddle.c -2 +2 cannam@95: M ./libbench2/dotens2.c -2 +2 cannam@95: M ./libbench2/my-getopt.c -2 +2 cannam@95: M ./libbench2/my-getopt.h -2 +2 cannam@95: M ./libbench2/verify-dft.c -2 +2 cannam@95: M ./libbench2/verify-lib.c -2 +2 cannam@95: M ./libbench2/verify-r2r.c -2 +2 cannam@95: M ./libbench2/verify-rdft2.c -2 +2 cannam@95: M ./libbench2/verify.h -2 +2 cannam@95: M ./mpi/any-true.c -2 +2 cannam@95: M ./mpi/api.c -2 +2 cannam@95: M ./mpi/block.c -2 +2 cannam@95: M ./mpi/choose-radix.c -2 +2 cannam@95: M ./mpi/conf.c -2 +2 cannam@95: M ./mpi/dft-problem.c -2 +2 cannam@95: M ./mpi/dft-rank-geq2-transposed.c -2 +2 cannam@95: M ./mpi/dft-rank-geq2.c -2 +2 cannam@95: M ./mpi/dft-rank1-bigvec.c -2 +2 cannam@95: M ./mpi/dft-rank1.c -2 +2 cannam@95: M ./mpi/dft-serial.c -2 +2 cannam@95: M ./mpi/dft-solve.c -2 +2 cannam@95: M ./mpi/dtensor.c -2 +2 cannam@95: M ./mpi/fftw3-mpi.h -2 +2 cannam@95: M ./mpi/ifftw-mpi.h -2 +2 cannam@95: M ./mpi/mpi-dft.h -2 +2 cannam@95: M ./mpi/mpi-rdft.h -2 +2 cannam@95: M ./mpi/mpi-rdft2.h -2 +2 cannam@95: M ./mpi/mpi-transpose.h -2 +2 cannam@95: M ./mpi/rdft-problem.c -2 +2 cannam@95: M ./mpi/rdft-rank-geq2-transposed.c -2 +2 cannam@95: M ./mpi/rdft-rank-geq2.c -2 +2 cannam@95: M ./mpi/rdft-rank1-bigvec.c -2 +2 cannam@95: M ./mpi/rdft-serial.c -2 +2 cannam@95: M ./mpi/rdft-solve.c -2 +2 cannam@95: M ./mpi/rdft2-problem.c -2 +2 cannam@95: M ./mpi/rdft2-rank-geq2-transposed.c -2 +2 cannam@95: M ./mpi/rdft2-rank-geq2.c -2 +2 cannam@95: M ./mpi/rdft2-serial.c -2 +2 cannam@95: M ./mpi/rdft2-solve.c -2 +2 cannam@95: M ./mpi/rearrange.c -2 +2 cannam@95: M ./mpi/testsched.c -2 +2 cannam@95: M ./mpi/transpose-alltoall.c -2 +2 cannam@95: M ./mpi/transpose-pairwise.c -2 +2 cannam@95: M ./mpi/transpose-problem.c -2 +2 cannam@95: M ./mpi/transpose-recurse.c -2 +2 cannam@95: M ./mpi/transpose-solve.c -2 +2 cannam@95: M ./mpi/wisdom-api.c -2 +2 cannam@95: M ./rdft/buffered.c -2 +2 cannam@95: M ./rdft/buffered2.c -2 +2 cannam@95: M ./rdft/codelet-rdft.h -2 +2 cannam@95: M ./rdft/conf.c -2 +2 cannam@95: M ./rdft/ct-hc2c-direct.c -2 +2 cannam@95: M ./rdft/ct-hc2c.c -2 +2 cannam@95: M ./rdft/ct-hc2c.h -2 +2 cannam@95: M ./rdft/dft-r2hc.c -2 +2 cannam@95: M ./rdft/dht-r2hc.c -2 +2 cannam@95: M ./rdft/dht-rader.c -2 +2 cannam@95: M ./rdft/direct-r2c.c -2 +2 cannam@95: M ./rdft/direct-r2r.c -2 +2 cannam@95: M ./rdft/direct2.c -2 +2 cannam@95: M ./rdft/generic.c -2 +2 cannam@95: M ./rdft/hc2hc-direct.c -2 +2 cannam@95: M ./rdft/hc2hc-generic.c -2 +2 cannam@95: M ./rdft/hc2hc.c -2 +2 cannam@95: M ./rdft/hc2hc.h -2 +2 cannam@95: M ./rdft/indirect.c -2 +2 cannam@95: M ./rdft/khc2c.c -2 +2 cannam@95: M ./rdft/khc2hc.c -2 +2 cannam@95: M ./rdft/kr2c.c -2 +2 cannam@95: M ./rdft/kr2r.c -2 +2 cannam@95: M ./rdft/nop.c -2 +2 cannam@95: M ./rdft/nop2.c -2 +2 cannam@95: M ./rdft/plan.c -2 +2 cannam@95: M ./rdft/plan2.c -2 +2 cannam@95: M ./rdft/problem.c -2 +2 cannam@95: M ./rdft/problem2.c -2 +2 cannam@95: M ./rdft/rank-geq2-rdft2.c -2 +2 cannam@95: M ./rdft/rank-geq2.c -2 +2 cannam@95: M ./rdft/rank0-rdft2.c -2 +2 cannam@95: M ./rdft/rank0.c -2 +2 cannam@95: M ./rdft/rdft-dht.c -2 +2 cannam@95: M ./rdft/rdft.h -2 +2 cannam@95: M ./rdft/rdft2-inplace-strides.c -2 +2 cannam@95: M ./rdft/rdft2-rdft.c -2 +2 cannam@95: M ./rdft/rdft2-strides.c -2 +2 cannam@95: M ./rdft/rdft2-tensor-max-index.c -2 +2 cannam@95: M ./rdft/scalar/hb.h -2 +2 cannam@95: M ./rdft/scalar/hc2c.c -2 +2 cannam@95: M ./rdft/scalar/hc2cb.h -2 +2 cannam@95: M ./rdft/scalar/hc2cf.h -2 +2 cannam@95: M ./rdft/scalar/hf.h -2 +2 cannam@95: M ./rdft/scalar/hfb.c -2 +2 cannam@95: M ./rdft/scalar/r2c.c -2 +2 cannam@95: M ./rdft/scalar/r2cb.h -2 +2 cannam@95: M ./rdft/scalar/r2cbIII.h -2 +2 cannam@95: M ./rdft/scalar/r2cf.h -2 +2 cannam@95: M ./rdft/scalar/r2cfII.h -2 +2 cannam@95: M ./rdft/scalar/r2r.c -2 +2 cannam@95: M ./rdft/scalar/r2r.h -2 +2 cannam@95: M ./rdft/simd/hc2cbv.c -2 +2 cannam@95: M ./rdft/simd/hc2cbv.h -2 +2 cannam@95: M ./rdft/simd/hc2cfv.c -2 +2 cannam@95: M ./rdft/simd/hc2cfv.h -2 +2 cannam@95: M ./rdft/solve.c -2 +2 cannam@95: M ./rdft/solve2.c -2 +2 cannam@95: M ./rdft/vrank-geq1-rdft2.c -2 +2 cannam@95: M ./rdft/vrank-geq1.c -2 +2 cannam@95: M ./rdft/vrank3-transpose.c -2 +2 cannam@95: M ./reodft/conf.c -2 +2 cannam@95: M ./reodft/redft00e-r2hc-pad.c -2 +2 cannam@95: M ./reodft/redft00e-r2hc.c -2 +2 cannam@95: M ./reodft/reodft.h -2 +2 cannam@95: M ./reodft/reodft010e-r2hc.c -2 +2 cannam@95: M ./reodft/reodft11e-r2hc-odd.c -2 +2 cannam@95: M ./reodft/reodft11e-r2hc.c -2 +2 cannam@95: M ./reodft/reodft11e-radix2.c -2 +2 cannam@95: M ./reodft/rodft00e-r2hc-pad.c -2 +2 cannam@95: M ./reodft/rodft00e-r2hc.c -2 +2 cannam@95: M ./simd/altivec.c -2 +2 cannam@95: M ./simd/nonportable/sse.c -2 +2 cannam@95: M ./simd/nonportable/sse2.c -2 +2 cannam@95: M ./simd/simd-altivec.h -2 +2 cannam@95: M ./simd/simd-sse.h -2 +2 cannam@95: M ./simd/simd-sse2.h -2 +2 cannam@95: M ./simd/simd.h -2 +2 cannam@95: M ./simd/sse.c -2 +2 cannam@95: M ./simd/sse2.c -2 +2 cannam@95: M ./simd/taint.c -2 +2 cannam@95: M ./simd/x86-cpuid.h -2 +2 cannam@95: M ./threads/api.c -2 +2 cannam@95: M ./threads/conf.c -2 +2 cannam@95: M ./threads/ct.c -2 +2 cannam@95: M ./threads/dft-vrank-geq1.c -2 +2 cannam@95: M ./threads/f77api.c -2 +2 cannam@95: M ./threads/f77funcs.h -2 +2 cannam@95: M ./threads/hc2hc.c -2 +2 cannam@95: M ./threads/openmp.c -2 +2 cannam@95: M ./threads/rdft-vrank-geq1.c -2 +2 cannam@95: M ./threads/threads.c -2 +2 cannam@95: M ./threads/threads.h -2 +2 cannam@95: M ./threads/vrank-geq1-rdft2.c -2 +2 cannam@95: M ./tools/fftw-wisdom-to-conf.1 -4 +4 cannam@95: M ./tools/fftw-wisdom-to-conf.in -2 +2 cannam@95: M ./tools/fftw-wisdom.c -2 +2 cannam@95: M ./tools/fftw_wisdom.1.in -4 +4 cannam@95: cannam@95: Wed Nov 5 16:40:31 EST 2008 athena@fftw.org cannam@95: * Store GPLv2 in darcs because automake installs GPLv3 these days. cannam@95: cannam@95: A ./COPYING cannam@95: cannam@95: Thu Oct 30 15:03:41 EDT 2008 athena@fftw.org cannam@95: * stylistic changes, comments cannam@95: cannam@95: M ./threads/threads.c -13 +17 cannam@95: cannam@95: Thu Oct 30 14:40:14 EDT 2008 athena@fftw.org cannam@95: * oops cannam@95: cannam@95: M ./threads/threads.c -27 +28 cannam@95: cannam@95: Thu Oct 30 14:30:08 EDT 2008 athena@fftw.org cannam@95: * simplification of the threading machinery cannam@95: cannam@95: M ./threads/threads.c -12 +7 cannam@95: cannam@95: Thu Oct 30 14:22:40 EDT 2008 athena@fftw.org cannam@95: * typo cannam@95: cannam@95: M ./threads/threads.c -1 +1 cannam@95: cannam@95: Thu Oct 30 13:42:07 EDT 2008 athena@fftw.org cannam@95: * [SECOND ATTEMPT] do not assume that a semaphore can be freed just because nobody is using it cannam@95: cannam@95: Let S be a semaphore, initially 0. Let thread A execute UP(S); cannam@95: let thread B execute DOWN(S); free(&S); It is unclear whether this cannam@95: code is correct with posix semaphores. The problem is whether UP() cannam@95: uses S after allowing DOWN() to continue; this seems to be the cannam@95: case in the glibc-2.7 implementation, and thus the pattern above cannam@95: seems to be incorrect. Avoid using such a pattern, and introduce cannam@95: a global semaphore for the unavoidable case when nothing else cannam@95: can be depended upon. cannam@95: cannam@95: cannam@95: M ./threads/threads.c -61 +80 cannam@95: cannam@95: Wed Oct 29 20:09:39 EDT 2008 stevenj@fftw.org cannam@95: * updated cpu codes from x86-1.21 cannam@95: cannam@95: M ./m4/ax_gcc_archflag.m4 -5 +7 cannam@95: cannam@95: Wed Oct 29 16:24:16 EDT 2008 athena@fftw.org cannam@95: * Previous change was bogus, need to find another way. cannam@95: cannam@95: M ./threads/threads.c -56 +42 cannam@95: cannam@95: Wed Oct 29 12:22:20 EDT 2008 athena@fftw.org cannam@95: * do not assume that a semaphore can be freed just because nobody is using it cannam@95: cannam@95: Let S be a semaphore, initially 0. Let thread A execute UP(S); cannam@95: let thread B execute DOWN(S); free(&S); It is unclear whether this cannam@95: code is correct with posix semaphores. The problem is whether UP() cannam@95: uses S after allowing DOWN() to continue; this seems to be the cannam@95: case in the glibc-2.7 implementation, and thus the pattern above cannam@95: seems to be incorrect. Avoid using such a pattern, and introduce cannam@95: a global semaphore for the unavoidable case when nothing else cannam@95: can be depended upon. cannam@95: cannam@95: cannam@95: M ./threads/threads.c -42 +56 cannam@95: cannam@95: Mon Oct 27 23:38:02 EDT 2008 stevenj@fftw.org cannam@95: * don't need PROG_AS any more cannam@95: cannam@95: M ./configure.ac -1 cannam@95: R ./m4/amx_prog_as.m4 cannam@95: cannam@95: Sun Oct 26 23:41:11 EDT 2008 stevenj@fftw.org cannam@95: * use AC_CONFIG_MACRO_DIR macro cannam@95: cannam@95: M ./Makefile.am -1 +1 cannam@95: M ./configure.ac +1 cannam@95: cannam@95: Sun Oct 26 10:08:44 EDT 2008 athena@fftw.org cannam@95: * Remove mpi/Makefile from configure.ac cannam@95: Otherwise, the tarball breaks because mpi/ is not in cannam@95: the distribution. cannam@95: cannam@95: M ./configure.ac -2 cannam@95: cannam@95: Sat Oct 25 17:13:50 EDT 2008 stevenj@fftw.org cannam@95: * remove MPI from dist until FFTW 3.3 cannam@95: cannam@95: M ./Makefile.am -1 +1 cannam@95: M ./NEWS +3 cannam@95: M ./doc/fftw3.texi -948 +9 cannam@95: cannam@95: Sat Oct 25 17:12:35 EDT 2008 stevenj@fftw.org cannam@95: * use MPIRUN even for -np 1 cannam@95: cannam@95: M ./mpi/Makefile.am -1 +1 cannam@95: cannam@95: Fri Jul 18 17:17:08 EDT 2008 stevenj@fftw.org cannam@95: * use new gcc arch=native flag as fallback cannam@95: cannam@95: M ./m4/ax_gcc_archflag.m4 -5 +5 cannam@95: cannam@95: Sat Oct 25 13:36:40 EDT 2008 athena@fftw.org cannam@95: * Use sem_t to implement mutexes cannam@95: Use sem_t instead of pthread_mutex_t to implement mutexes. cannam@95: It seems like pthread mutexes hang on linux-2.6.22 after several cannam@95: days of tests; the hang does not occur on linux >= 2.6.24 cannam@95: or when we use sem_t instead of pthread_mutex_t. The cannam@95: situation is still quite mysterious but this code seems to cannam@95: work. cannam@95: cannam@95: M ./threads/threads.c -22 +34 cannam@95: cannam@95: Thu Oct 23 13:32:45 EDT 2008 athena@fftw.org cannam@95: * print informative message when pstring is NULL. cannam@95: cannam@95: M ./libbench2/verify.c -1 +2 cannam@95: cannam@95: Sun Oct 19 16:00:07 EDT 2008 athena@fftw.org cannam@95: * Fix incorrect alignment in dftw-generic. cannam@95: cannam@95: Multithreaded dftw-generic is supposed to process only a slice cannam@95: of the array, but we were planning with the alignment of the cannam@95: original array rather than the slice. This led to unaligned cannam@95: accesses in certain obscure situations. cannam@95: cannam@95: cannam@95: M ./NEWS +3 cannam@95: M ./configure.ac -1 +1 cannam@95: M ./dft/dftw-generic.c -1 +2 cannam@95: cannam@95: Mon Aug 18 17:27:26 EDT 2008 Matteo Frigo cannam@95: * Paranoia: do not create OS threads while holding locks. cannam@95: cannam@95: Glibc at least plays silly games such as keeping a global variable cannam@95: that records whether there is more than one thread in the process, and cannam@95: it does not perform atomic operations if the variable says that there cannam@95: is only one thread. Who knows how this interacts with creating cannam@95: threads while holding a lock. Some day some genius will come up with cannam@95: some ``optimization'' that breaks everything. cannam@95: cannam@95: M ./threads/threads.c -7 +13 cannam@95: cannam@95: Wed Aug 6 07:41:46 EDT 2008 athena@fftw.org cannam@95: * Welcome to the quadcore era cannam@95: cannam@95: M ./mkdist.sh -1 +1 cannam@95: cannam@95: Mon Jun 16 16:46:39 EDT 2008 stevenj@fftw.org cannam@95: * backslash is technically not allowed in "echo" arguments; thanks to Debian Bug#486046 for pointing out problem and solution (and Raphael Geissert and Vincent Zweije, in particular) cannam@95: cannam@95: M ./tools/fftw-wisdom-to-conf.in -1 +1 cannam@95: cannam@95: Sun May 4 12:15:24 EDT 2008 stevenj@fftw.org cannam@95: * note problem with test program in gcc 4.1.2-4.2; thanks to Raymond Rogers for reporting it cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn +6 cannam@95: cannam@95: Fri May 2 19:21:30 EDT 2008 stevenj@fftw.org cannam@95: * output count of constants along with other statistics cannam@95: cannam@95: M ./genfft/c.ml -3 +8 cannam@95: cannam@95: Sat Apr 19 14:15:03 EDT 2008 athena@fftw.org cannam@95: * Lower priority of unaligned SIMD codelets. cannam@95: List t1[fb]uv_* codelets before the corresponding cannam@95: aligned codelets, since the estimator picks the cannam@95: latter ones in case of a tie and aligned codelets cannam@95: are preferable. cannam@95: cannam@95: In other words, this is a hack. cannam@95: cannam@95: M ./dft/simd/codelets/Makefile.am -3 +2 cannam@95: cannam@95: Sat Apr 19 08:55:46 EDT 2008 Matteo Frigo cannam@95: * There is no point in using higher radices for unaligned codelets. cannam@95: cannam@95: cannam@95: M ./dft/simd/codelets/Makefile.am -4 +2 cannam@95: cannam@95: Fri Apr 18 19:01:27 EDT 2008 stevenj@fftw.org cannam@95: * support generating loopless, strideless r2r codelets cannam@95: cannam@95: M ./genfft/gen_r2r.ml -7 +17 cannam@95: cannam@95: Fri Apr 18 19:00:25 EDT 2008 stevenj@fftw.org cannam@95: * added Magic.threemult to use 3+3 complex-multiply variant when possible cannam@95: cannam@95: M ./genfft/complex.ml -2 +13 cannam@95: M ./genfft/magic.ml +4 cannam@95: cannam@95: Thu Apr 10 19:53:31 EDT 2008 stevenj@fftw.org cannam@95: * fix documentation bug - export_wisdom_to_string returns a string that should be deallocated with free, not fftw_free (thanks to Stein Vidar Hagfors Haugan for the bug report) cannam@95: cannam@95: M ./doc/fftw3.texi -1 +1 cannam@95: cannam@95: Mon Jan 21 01:11:44 EST 2008 stevenj@fftw.org cannam@95: * bsd calls x86_64 "amd64"; thanks to Fernando Herrero Carron for the bug report cannam@95: cannam@95: M ./m4/ax_gcc_archflag.m4 -3 +3 cannam@95: cannam@95: Tue Jan 1 12:29:56 EST 2008 stevenj@fftw.org cannam@95: * fix typo in manual, thanks to Yinon Ehrlich cannam@95: cannam@95: M ./doc/fftw3.texi -1 +1 cannam@95: cannam@95: Mon Dec 3 13:57:13 EST 2007 stevenj@fftw.org cannam@95: * note problem with gcc 3.4.4 on x86_64, thanks to Uwe Hollerbach for the report cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -2 +4 cannam@95: cannam@95: Tue Nov 13 16:19:22 EST 2007 stevenj@fftw.org cannam@95: * bump shared-lib revision cannam@95: cannam@95: M ./configure.ac -1 +1 cannam@95: cannam@95: Tue Nov 13 16:16:49 EST 2007 stevenj@fftw.org cannam@95: * update NEWS for alpha3 cannam@95: cannam@95: M ./NEWS +23 cannam@95: cannam@95: Tue Oct 2 13:53:04 EDT 2007 stevenj@fftw.org cannam@95: * fixed URL cannam@95: cannam@95: M ./m4/ax_openmp.m4 -1 +1 cannam@95: cannam@95: Mon Sep 17 19:38:29 EDT 2007 stevenj@fftw.org cannam@95: * added missing prototype cannam@95: cannam@95: M ./tests/fftw-bench.c +4 cannam@95: cannam@95: Tue Aug 14 22:35:06 EDT 2007 stevenj@fftw.org cannam@95: * terminology tweak cannam@95: cannam@95: M ./doc/fftw3.texi -1 +1 cannam@95: cannam@95: Wed Aug 1 18:44:21 EDT 2007 stevenj@fftw.org cannam@95: * check for pathscale compilers (thanks to Julian Cummings) cannam@95: cannam@95: M ./m4/ax_compiler_vendor.m4 -2 +2 cannam@95: cannam@95: Sat Sep 15 18:02:32 EDT 2007 athena@fftw.org cannam@95: * Avoid possible conflict with Windows include files. cannam@95: cannam@95: M ./threads/threads.c +5 cannam@95: cannam@95: Tue Aug 7 21:26:05 EDT 2007 athena@fftw.org cannam@95: * Distribute codlist.c for SIMD codelets in the commercial tarball. cannam@95: cannam@95: M ./commercialize.sh -1 +1 cannam@95: cannam@95: Wed Aug 1 10:33:41 EDT 2007 stevenj@fftw.org cannam@95: * some documentation clarifications, and documented FFTW_WISDOM_ONLY, at the suggestion of Mario Emmenlauer and Phil Dumont cannam@95: cannam@95: M ./api/apiplan.c -3 +3 cannam@95: M ./doc/fftw3.texi -3 +24 cannam@95: cannam@95: Tue Jul 31 16:52:56 EDT 2007 stevenj@fftw.org cannam@95: * bug fix in test program for vrank-3 transpose plans with vl=1 cannam@95: cannam@95: M ./mpi/mpi-bench.c -1 +1 cannam@95: cannam@95: Sun Jul 29 17:02:46 EDT 2007 stevenj@fftw.org cannam@95: * only run mpi checks for --enable-mpi cannam@95: cannam@95: M ./mpi/Makefile.am +4 cannam@95: cannam@95: Sun Jul 29 16:45:30 EDT 2007 stevenj@fftw.org cannam@95: * check for NULL return from spe_context_create in case SPE_MAP_PS not supported cannam@95: cannam@95: M ./cell/cell.c -4 +9 cannam@95: cannam@95: Sun Jul 29 15:56:57 EDT 2007 stevenj@fftw.org cannam@95: * use problem-state pointer to write SPE mailbox with lower latency (makes a significant performance difference for N < 32k), thanks to Jan Wagner for suggestion cannam@95: cannam@95: M ./cell/cell.c +16 cannam@95: cannam@95: Sun Jul 29 14:22:08 EDT 2007 stevenj@fftw.org cannam@95: * port cell code to SDK2.1 (libspe2), since libspe1 API is deprecated and can't be used in code that also uses libspe2 API cannam@95: cannam@95: M ./cell/cell.c -3 +49 cannam@95: M ./cell/copy.c -1 +1 cannam@95: M ./cell/dft-direct-cell.c -1 +1 cannam@95: M ./cell/fftw-cell.h -4 +5 cannam@95: M ./cell/spu/main.spuc -4 +4 cannam@95: M ./cell/transpose.c -1 +1 cannam@95: M ./configure.ac -2 +2 cannam@95: cannam@95: Sun Jul 29 11:46:24 EDT 2007 stevenj@fftw.org cannam@95: * bug fix: ego->W allocated with cell_aligned_malloc, so deallocate with free, not X(ifree0) cannam@95: cannam@95: M ./cell/dft-direct-cell.c -1 +1 cannam@95: cannam@95: Mon Jul 2 15:57:12 EDT 2007 stevenj@fftw.org cannam@95: * removed obsolete reference to CVS id cannam@95: cannam@95: M ./api/fftw3.h -4 cannam@95: cannam@95: Mon May 21 14:25:39 EDT 2007 athena@fftw.org cannam@95: * cycle counter for sun compiler cannam@95: cannam@95: M ./kernel/cycle.h -1 +1 cannam@95: cannam@95: Wed May 9 19:49:11 EDT 2007 stevenj@fftw.org cannam@95: * use __inline instead of inline for AIX routines (__inline is supported by gcc and xlc, whereas apparently "inline" is only supported by xlc if you specify -qlanglvl=stdc99 or similar); thanks to Jeff Haferman for the bug report cannam@95: cannam@95: M ./kernel/cycle.h -2 +2 cannam@95: cannam@95: Mon Apr 30 15:37:56 EDT 2007 stevenj@fftw.org cannam@95: * fixed incorrect type prefix (fftw_ vs. X(...)) in mpi/wisdom-api.c; thanks to Eric A. Borisch for the bug report cannam@95: cannam@95: M ./mpi/wisdom-api.c -6 +6 cannam@95: cannam@95: Wed Apr 25 21:21:39 EDT 2007 stevenj@fftw.org cannam@95: * some cleanups in MPI make check cannam@95: cannam@95: M ./mpi/Makefile.am -7 +11 cannam@95: cannam@95: Wed Apr 25 21:19:27 EDT 2007 stevenj@fftw.org cannam@95: * re-enable heuristic in the common case where we are not compiling for Cell cannam@95: cannam@95: M ./dft/rank-geq2.c -1 +3 cannam@95: cannam@95: Tue Apr 24 17:42:43 EDT 2007 athena@fftw.org cannam@95: * Removed duplicate codelet names, was breaking linker. cannam@95: cannam@95: M ./dft/simd/codelets/Makefile.am -4 +4 cannam@95: cannam@95: Tue Apr 24 11:38:16 EDT 2007 stevenj@fftw.org cannam@95: * added more codelets of sizes 5/10/20/25 to improve speed for round decimal sizes (speed improvements of 10-20%, at cost of 10-30% in library size) cannam@95: cannam@95: M ./dft/scalar/codelets/Makefile.am -3 +6 cannam@95: M ./dft/simd/codelets/Makefile.am -14 +26 cannam@95: M ./rdft/scalar/r2cb/Makefile.am -8 +16 cannam@95: M ./rdft/scalar/r2cf/Makefile.am -8 +16 cannam@95: M ./rdft/simd/codelets/Makefile.am -2 +4 cannam@95: cannam@95: Sat Mar 24 18:40:47 EDT 2007 stevenj@fftw.org cannam@95: * for 1d prime sizes, punt and return serial plan cannam@95: cannam@95: M ./mpi/api.c +4 cannam@95: cannam@95: Sat Mar 24 18:24:55 EDT 2007 stevenj@fftw.org cannam@95: * output reminders of the problem during bench --verify cannam@95: cannam@95: M ./libbench2/verify.c -2 +5 cannam@95: cannam@95: Sat Mar 24 18:10:24 EDT 2007 stevenj@fftw.org cannam@95: * bug fix - missing solver->destroy initializer in rdft2-rdft cannam@95: cannam@95: M ./rdft/rdft2-rdft.c -1 +1 cannam@95: cannam@95: Fri Mar 23 11:12:19 EDT 2007 stevenj@fftw.org cannam@95: * -static, in --enable-debug, doesn't work on MacOS X (according to Daniel Oberhoff) cannam@95: cannam@95: M ./configure.ac -1 cannam@95: cannam@95: Wed Mar 21 22:23:06 EDT 2007 stevenj@fftw.org cannam@95: * fix MPI r2c/c2r to work with howmany > 1 cannam@95: cannam@95: M ./mpi/rdft2-rank-geq2-transposed.c -1 +2 cannam@95: M ./mpi/rdft2-rank-geq2.c -1 +2 cannam@95: M ./mpi/rdft2-serial.c -2 +3 cannam@95: M ./tests/check.pl -1 cannam@95: cannam@95: Wed Mar 21 18:44:41 EDT 2007 stevenj@fftw.org cannam@95: * rm MPI version from TODO cannam@95: cannam@95: M ./TODO -2 cannam@95: cannam@95: Wed Mar 21 18:34:40 EDT 2007 stevenj@fftw.org cannam@95: * added 'make bigcheck' for MPI (no paranoid-check, unfortunately), and properly get MPIRUN from configure cannam@95: cannam@95: M ./configure.ac +2 cannam@95: M ./mpi/Makefile.am -7 +10 cannam@95: cannam@95: Wed Mar 21 18:23:18 EDT 2007 stevenj@fftw.org cannam@95: * bug fix - incorrect local_size returned for 1d bigvec case cannam@95: cannam@95: M ./mpi/api.c -1 +1 cannam@95: cannam@95: Wed Mar 21 03:13:54 EDT 2007 stevenj@fftw.org cannam@95: * hack to specify MPI_TRANSPOSED_IN/OUT via "[" and "]" in libbench2 problem cannam@95: cannam@95: M ./libbench2/bench-user.h +1 cannam@95: M ./libbench2/problem.c +6 cannam@95: M ./mpi/mpi-bench.c +12 cannam@95: M ./tests/check.pl -2 +7 cannam@95: cannam@95: Wed Mar 21 02:58:11 EDT 2007 stevenj@fftw.org cannam@95: * added MPI 'make check', still needs a bit of work cannam@95: cannam@95: M ./mpi/Makefile.am +28 cannam@95: M ./tests/check.pl +31 cannam@95: cannam@95: Wed Mar 21 02:47:10 EDT 2007 stevenj@fftw.org cannam@95: * bug fix in r2r transposed-input case cannam@95: cannam@95: M ./mpi/rdft-problem.c +3 cannam@95: cannam@95: Wed Mar 21 02:46:25 EDT 2007 stevenj@fftw.org cannam@95: * don't output more than 300 erroneous outputs (unless verbose > 2) cannam@95: cannam@95: M ./libbench2/verify-lib.c -2 +3 cannam@95: M ./libbench2/verify-r2r.c -2 +3 cannam@95: cannam@95: Wed Mar 21 01:48:54 EDT 2007 stevenj@fftw.org cannam@95: * fixed bug in transposed-in c2r MPI transforms ... seems to be working, finally cannam@95: cannam@95: M ./mpi/rdft2-problem.c -5 +6 cannam@95: cannam@95: Wed Mar 21 00:41:32 EDT 2007 stevenj@fftw.org cannam@95: * some fixes to MPI r2c/c2r transforms with transposed output/input cannam@95: cannam@95: M ./mpi/rdft2-rank-geq2-transposed.c -11 +17 cannam@95: cannam@95: Wed Mar 21 00:40:25 EDT 2007 stevenj@fftw.org cannam@95: * typos cannam@95: cannam@95: M ./NEWS -4 +4 cannam@95: cannam@95: Tue Mar 20 19:53:02 EDT 2007 stevenj@fftw.org cannam@95: * bug fix for mpi-bench with r2c/c2r: allocate a little bit extra to make sure that padding is allocated cannam@95: cannam@95: M ./libbench2/allocate.c -2 +4 cannam@95: cannam@95: Tue Mar 20 19:19:13 EDT 2007 stevenj@fftw.org cannam@95: * fix typo, thanks to Ernest Turro for the bug report cannam@95: cannam@95: M ./m4/ax_gcc_archflag.m4 -2 +2 cannam@95: cannam@95: Tue Mar 20 01:39:06 EDT 2007 stevenj@fftw.org cannam@95: * spacing tweaks cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Tue Mar 20 00:53:11 EDT 2007 stevenj@fftw.org cannam@95: * Ralf Wildenhues is the one who pointed out that the self-communication could fill in the stalls in the pairwise schedule cannam@95: cannam@95: M ./mpi/testsched.c -3 +2 cannam@95: cannam@95: Tue Mar 20 00:22:25 EDT 2007 stevenj@fftw.org cannam@95: * add TRANSPOSED_OUT/IN support for r2c/c2r, respectively cannam@95: cannam@95: M ./doc/fftw3.texi -4 +1 cannam@95: M ./mpi/Makefile.am -1 +1 cannam@95: M ./mpi/README -3 cannam@95: M ./mpi/api.c -6 +15 cannam@95: M ./mpi/conf.c -1 +1 cannam@95: M ./mpi/rdft2-problem.c -18 +9 cannam@95: A ./mpi/rdft2-rank-geq2-transposed.c cannam@95: cannam@95: Mon Mar 19 21:45:34 EDT 2007 stevenj@fftw.org cannam@95: * yikes! fixed likely deadlock bug in MPI cannam@95: cannam@95: M ./mpi/transpose-recurse.c -2 +2 cannam@95: cannam@95: Mon Mar 19 21:38:52 EDT 2007 stevenj@fftw.org cannam@95: * comment cannam@95: cannam@95: M ./mpi/transpose-pairwise.c -2 +8 cannam@95: cannam@95: Mon Mar 19 21:30:44 EDT 2007 stevenj@fftw.org cannam@95: * s/alpha1/alpha2/ cannam@95: cannam@95: M ./NEWS -1 +1 cannam@95: cannam@95: Mon Mar 19 00:39:47 EDT 2007 stevenj@fftw.org cannam@95: * include README in dist tarball cannam@95: cannam@95: M ./mpi/Makefile.am -1 +1 cannam@95: cannam@95: Mon Mar 19 00:35:43 EDT 2007 stevenj@fftw.org cannam@95: * added MPI r2c/c2r transforms, some more documentation cannam@95: cannam@95: M ./NEWS -4 +9 cannam@95: M ./doc/fftw3.texi -6 +73 cannam@95: M ./libbench2/bench-user.h +2 cannam@95: M ./libbench2/problem.c -2 +4 cannam@95: M ./mpi/Makefile.am -1 +2 cannam@95: A ./mpi/README cannam@95: M ./mpi/api.c +140 cannam@95: M ./mpi/conf.c +4 cannam@95: M ./mpi/fftw3-mpi.h -88 +116 cannam@95: M ./mpi/mpi-bench.c -2 +102 cannam@95: A ./mpi/mpi-rdft2.h cannam@95: A ./mpi/rdft2-problem.c cannam@95: A ./mpi/rdft2-rank-geq2.c cannam@95: A ./mpi/rdft2-serial.c cannam@95: A ./mpi/rdft2-solve.c cannam@95: cannam@95: Sun Mar 18 23:14:29 EDT 2007 stevenj@fftw.org cannam@95: * set version to 3.2alpha2 cannam@95: cannam@95: M ./configure.ac -1 +1 cannam@95: cannam@95: Sun Mar 18 19:12:18 EDT 2007 stevenj@fftw.org cannam@95: * changed --enable-mips_ps to --enable-mips-ps; added Cell section to manual (from README.Cell); many minor updates to manual cannam@95: cannam@95: M ./configure.ac -1 +1 cannam@95: M ./doc/fftw3.texi -15 +243 cannam@95: cannam@95: Sun Mar 18 15:27:06 EDT 2007 stevenj@fftw.org cannam@95: * whoops, need to sync costs in problem_mpi_rdft cannam@95: cannam@95: M ./mpi/api.c +3 cannam@95: cannam@95: Sun Mar 18 12:44:49 EDT 2007 stevenj@fftw.org cannam@95: * documented guru64 interface cannam@95: cannam@95: M ./NEWS -1 +1 cannam@95: M ./doc/fftw3.texi -11 +81 cannam@95: cannam@95: Sun Mar 18 02:57:46 EDT 2007 stevenj@fftw.org cannam@95: * typo cannam@95: cannam@95: M ./doc/fftw3.texi -1 +1 cannam@95: cannam@95: Sun Mar 18 02:45:09 EDT 2007 stevenj@fftw.org cannam@95: * bumped copyright year to 2007 cannam@95: cannam@95: M ./COPYRIGHT -2 +2 cannam@95: M ./api/api.h -2 +2 cannam@95: M ./api/apiplan.c -2 +2 cannam@95: M ./api/configure.c -2 +2 cannam@95: M ./api/execute-dft-c2r.c -2 +2 cannam@95: M ./api/execute-dft-r2c.c -2 +2 cannam@95: M ./api/execute-dft.c -2 +2 cannam@95: M ./api/execute-r2r.c -2 +2 cannam@95: M ./api/execute-split-dft-c2r.c -2 +2 cannam@95: M ./api/execute-split-dft-r2c.c -2 +2 cannam@95: M ./api/execute-split-dft.c -2 +2 cannam@95: M ./api/execute.c -2 +2 cannam@95: M ./api/export-wisdom-to-file.c -2 +2 cannam@95: M ./api/export-wisdom-to-string.c -2 +2 cannam@95: M ./api/export-wisdom.c -2 +2 cannam@95: M ./api/f77api.c -2 +2 cannam@95: M ./api/f77funcs.h -2 +2 cannam@95: M ./api/fftw3.h -2 +2 cannam@95: M ./api/flops.c -2 +2 cannam@95: M ./api/forget-wisdom.c -2 +2 cannam@95: M ./api/import-system-wisdom.c -2 +2 cannam@95: M ./api/import-wisdom-from-file.c -2 +2 cannam@95: M ./api/import-wisdom-from-string.c -2 +2 cannam@95: M ./api/import-wisdom.c -2 +2 cannam@95: M ./api/malloc.c -2 +2 cannam@95: M ./api/map-r2r-kind.c -2 +2 cannam@95: M ./api/mapflags.c -2 +2 cannam@95: M ./api/mkprinter-file.c -2 +2 cannam@95: M ./api/mktensor-iodims.h -2 +2 cannam@95: M ./api/mktensor-rowmajor.c -2 +2 cannam@95: M ./api/plan-dft-1d.c -2 +2 cannam@95: M ./api/plan-dft-2d.c -2 +2 cannam@95: M ./api/plan-dft-3d.c -2 +2 cannam@95: M ./api/plan-dft-c2r-1d.c -2 +2 cannam@95: M ./api/plan-dft-c2r-2d.c -2 +2 cannam@95: M ./api/plan-dft-c2r-3d.c -2 +2 cannam@95: M ./api/plan-dft-c2r.c -2 +2 cannam@95: M ./api/plan-dft-r2c-1d.c -2 +2 cannam@95: M ./api/plan-dft-r2c-2d.c -2 +2 cannam@95: M ./api/plan-dft-r2c-3d.c -2 +2 cannam@95: M ./api/plan-dft-r2c.c -2 +2 cannam@95: M ./api/plan-dft.c -2 +2 cannam@95: M ./api/plan-guru-dft-c2r.h -2 +2 cannam@95: M ./api/plan-guru-dft-r2c.h -2 +2 cannam@95: M ./api/plan-guru-dft.h -2 +2 cannam@95: M ./api/plan-guru-r2r.h -2 +2 cannam@95: M ./api/plan-guru-split-dft-c2r.h -2 +2 cannam@95: M ./api/plan-guru-split-dft-r2c.h -2 +2 cannam@95: M ./api/plan-guru-split-dft.h -2 +2 cannam@95: M ./api/plan-many-dft-c2r.c -2 +2 cannam@95: M ./api/plan-many-dft-r2c.c -2 +2 cannam@95: M ./api/plan-many-dft.c -2 +2 cannam@95: M ./api/plan-many-r2r.c -2 +2 cannam@95: M ./api/plan-r2r-1d.c -2 +2 cannam@95: M ./api/plan-r2r-2d.c -2 +2 cannam@95: M ./api/plan-r2r-3d.c -2 +2 cannam@95: M ./api/plan-r2r.c -2 +2 cannam@95: M ./api/print-plan.c -2 +2 cannam@95: M ./api/rdft2-pad.c -2 +2 cannam@95: M ./api/the-planner.c -2 +2 cannam@95: M ./api/version.c -2 +2 cannam@95: M ./api/x77.h -2 +2 cannam@95: M ./dft/bluestein.c -2 +2 cannam@95: M ./dft/buffered.c -2 +2 cannam@95: M ./dft/codelet-dft.h -2 +2 cannam@95: M ./dft/conf.c -2 +2 cannam@95: M ./dft/ct.c -2 +2 cannam@95: M ./dft/ct.h -2 +2 cannam@95: M ./dft/dft.h -2 +2 cannam@95: M ./dft/dftw-direct.c -2 +2 cannam@95: M ./dft/dftw-directsq.c -2 +2 cannam@95: M ./dft/dftw-generic.c -2 +2 cannam@95: M ./dft/dftw-genericbuf.c -2 +2 cannam@95: M ./dft/direct.c -2 +2 cannam@95: M ./dft/generic.c -2 +2 cannam@95: M ./dft/indirect-transpose.c -2 +2 cannam@95: M ./dft/indirect.c -2 +2 cannam@95: M ./dft/kdft-dif.c -2 +2 cannam@95: M ./dft/kdft-difsq.c -2 +2 cannam@95: M ./dft/kdft-dit.c -2 +2 cannam@95: M ./dft/kdft.c -2 +2 cannam@95: M ./dft/nop.c -2 +2 cannam@95: M ./dft/plan.c -2 +2 cannam@95: M ./dft/problem.c -2 +2 cannam@95: M ./dft/rader.c -2 +2 cannam@95: M ./dft/rank-geq2.c -2 +2 cannam@95: M ./dft/scalar/n.c -2 +2 cannam@95: M ./dft/scalar/n.h -2 +2 cannam@95: M ./dft/scalar/t.c -2 +2 cannam@95: M ./dft/scalar/t.h -2 +2 cannam@95: M ./dft/simd/n1b.c -2 +2 cannam@95: M ./dft/simd/n1b.h -2 +2 cannam@95: M ./dft/simd/n1f.c -2 +2 cannam@95: M ./dft/simd/n1f.h -2 +2 cannam@95: M ./dft/simd/n2b.c -2 +2 cannam@95: M ./dft/simd/n2b.h -2 +2 cannam@95: M ./dft/simd/n2f.c -2 +2 cannam@95: M ./dft/simd/n2f.h -2 +2 cannam@95: M ./dft/simd/n2s.c -2 +2 cannam@95: M ./dft/simd/n2s.h -2 +2 cannam@95: M ./dft/simd/q1b.c -2 +2 cannam@95: M ./dft/simd/q1b.h -2 +2 cannam@95: M ./dft/simd/q1f.c -2 +2 cannam@95: M ./dft/simd/q1f.h -2 +2 cannam@95: M ./dft/simd/t.c -2 +2 cannam@95: M ./dft/simd/t1b.h -2 +2 cannam@95: M ./dft/simd/t1bu.h -2 +2 cannam@95: M ./dft/simd/t1f.h -2 +2 cannam@95: M ./dft/simd/t1fu.h -2 +2 cannam@95: M ./dft/simd/t2b.h -2 +2 cannam@95: M ./dft/simd/t2f.h -2 +2 cannam@95: M ./dft/simd/t3b.h -2 +2 cannam@95: M ./dft/simd/t3f.h -2 +2 cannam@95: M ./dft/simd/ts.c -2 +2 cannam@95: M ./dft/simd/ts.h -2 +2 cannam@95: M ./dft/solve.c -2 +2 cannam@95: M ./dft/vrank-geq1.c -2 +2 cannam@95: M ./dft/zero.c -2 +2 cannam@95: M ./doc/f77_wisdom.f -2 +2 cannam@95: M ./genfft/algsimp.ml -2 +2 cannam@95: M ./genfft/algsimp.mli -2 +2 cannam@95: M ./genfft/annotate.ml -2 +2 cannam@95: M ./genfft/annotate.mli -2 +2 cannam@95: M ./genfft/assoctable.ml -2 +2 cannam@95: M ./genfft/assoctable.mli -2 +2 cannam@95: M ./genfft/c.ml -2 +2 cannam@95: M ./genfft/c.mli -2 +2 cannam@95: M ./genfft/complex.ml -2 +2 cannam@95: M ./genfft/complex.mli -2 +2 cannam@95: M ./genfft/conv.ml -2 +2 cannam@95: M ./genfft/conv.mli -2 +2 cannam@95: M ./genfft/dag.ml -2 +2 cannam@95: M ./genfft/dag.mli -2 +2 cannam@95: M ./genfft/expr.ml -2 +2 cannam@95: M ./genfft/expr.mli -2 +2 cannam@95: M ./genfft/fft.ml -2 +2 cannam@95: M ./genfft/fft.mli -2 +2 cannam@95: M ./genfft/gen_hc2c.ml -2 +2 cannam@95: M ./genfft/gen_hc2cdft.ml -2 +2 cannam@95: M ./genfft/gen_hc2cdft_c.ml -2 +2 cannam@95: M ./genfft/gen_hc2hc.ml -2 +2 cannam@95: M ./genfft/gen_mdct.ml -2 +2 cannam@95: M ./genfft/gen_notw.ml -2 +2 cannam@95: M ./genfft/gen_notw_c.ml -2 +2 cannam@95: M ./genfft/gen_r2cb.ml -2 +2 cannam@95: M ./genfft/gen_r2cf.ml -2 +2 cannam@95: M ./genfft/gen_r2r.ml -2 +2 cannam@95: M ./genfft/gen_twiddle.ml -2 +2 cannam@95: M ./genfft/gen_twiddle_c.ml -2 +2 cannam@95: M ./genfft/gen_twidsq.ml -2 +2 cannam@95: M ./genfft/gen_twidsq_c.ml -2 +2 cannam@95: M ./genfft/genutil.ml -2 +2 cannam@95: M ./genfft/littlesimp.ml -2 +2 cannam@95: M ./genfft/littlesimp.mli -2 +2 cannam@95: M ./genfft/magic.ml -2 +2 cannam@95: M ./genfft/monads.ml -2 +2 cannam@95: M ./genfft/number.ml -2 +2 cannam@95: M ./genfft/number.mli -2 +2 cannam@95: M ./genfft/oracle.ml -2 +2 cannam@95: M ./genfft/oracle.mli -2 +2 cannam@95: M ./genfft/schedule.ml -2 +2 cannam@95: M ./genfft/schedule.mli -2 +2 cannam@95: M ./genfft/simd.ml -2 +2 cannam@95: M ./genfft/simd.mli -2 +2 cannam@95: M ./genfft/simdmagic.ml -2 +2 cannam@95: M ./genfft/to_alist.ml -2 +2 cannam@95: M ./genfft/to_alist.mli -2 +2 cannam@95: M ./genfft/trig.ml -2 +2 cannam@95: M ./genfft/trig.mli -2 +2 cannam@95: M ./genfft/twiddle.ml -2 +2 cannam@95: M ./genfft/twiddle.mli -2 +2 cannam@95: M ./genfft/unique.ml -2 +2 cannam@95: M ./genfft/unique.mli -2 +2 cannam@95: M ./genfft/util.ml -2 +2 cannam@95: M ./genfft/util.mli -2 +2 cannam@95: M ./genfft/variable.ml -2 +2 cannam@95: M ./genfft/variable.mli -2 +2 cannam@95: M ./kernel/align.c -2 +2 cannam@95: M ./kernel/alloc.c -2 +2 cannam@95: M ./kernel/assert.c -2 +2 cannam@95: M ./kernel/awake.c -2 +2 cannam@95: M ./kernel/buffered.c -2 +2 cannam@95: M ./kernel/cpy1d.c -2 +2 cannam@95: M ./kernel/cpy2d-pair.c -2 +2 cannam@95: M ./kernel/cpy2d.c -2 +2 cannam@95: M ./kernel/ct.c -2 +2 cannam@95: M ./kernel/cycle.h -2 +2 cannam@95: M ./kernel/debug.c -2 +2 cannam@95: M ./kernel/extract-reim.c -2 +2 cannam@95: M ./kernel/hash.c -2 +2 cannam@95: M ./kernel/iabs.c -2 +2 cannam@95: M ./kernel/ifftw.h -2 +2 cannam@95: M ./kernel/kalloc.c -2 +2 cannam@95: M ./kernel/md5-1.c -2 +2 cannam@95: M ./kernel/md5.c -2 +2 cannam@95: M ./kernel/minmax.c -2 +2 cannam@95: M ./kernel/ops.c -2 +2 cannam@95: M ./kernel/pickdim.c -2 +2 cannam@95: M ./kernel/plan.c -2 +2 cannam@95: M ./kernel/primes.c -2 +2 cannam@95: M ./kernel/print.c -2 +2 cannam@95: M ./kernel/problem.c -2 +2 cannam@95: M ./kernel/rader.c -2 +2 cannam@95: M ./kernel/scan.c -2 +2 cannam@95: M ./kernel/solver.c -2 +2 cannam@95: M ./kernel/solvtab.c -2 +2 cannam@95: M ./kernel/stride.c -2 +2 cannam@95: M ./kernel/tensor.c -2 +2 cannam@95: M ./kernel/tensor1.c -2 +2 cannam@95: M ./kernel/tensor2.c -2 +2 cannam@95: M ./kernel/tensor3.c -2 +2 cannam@95: M ./kernel/tensor4.c -2 +2 cannam@95: M ./kernel/tensor5.c -2 +2 cannam@95: M ./kernel/tensor7.c -2 +2 cannam@95: M ./kernel/tensor8.c -2 +2 cannam@95: M ./kernel/tensor9.c -2 +2 cannam@95: M ./kernel/tile2d.c -2 +2 cannam@95: M ./kernel/timer.c -2 +2 cannam@95: M ./kernel/transpose.c -2 +2 cannam@95: M ./kernel/trig.c -2 +2 cannam@95: M ./kernel/twiddle.c -2 +2 cannam@95: M ./libbench2/dotens2.c -2 +2 cannam@95: M ./libbench2/my-getopt.c -2 +2 cannam@95: M ./libbench2/my-getopt.h -2 +2 cannam@95: M ./libbench2/verify-dft.c -2 +2 cannam@95: M ./libbench2/verify-lib.c -2 +2 cannam@95: M ./libbench2/verify-r2r.c -2 +2 cannam@95: M ./libbench2/verify-rdft2.c -2 +2 cannam@95: M ./libbench2/verify.h -2 +2 cannam@95: M ./mpi/any-true.c -2 +2 cannam@95: M ./mpi/api.c -2 +2 cannam@95: M ./mpi/block.c -2 +2 cannam@95: M ./mpi/choose-radix.c -2 +2 cannam@95: M ./mpi/conf.c -2 +2 cannam@95: M ./mpi/dft-problem.c -2 +2 cannam@95: M ./mpi/dft-rank-geq2-transposed.c -2 +2 cannam@95: M ./mpi/dft-rank-geq2.c -2 +2 cannam@95: M ./mpi/dft-rank1-bigvec.c -2 +2 cannam@95: M ./mpi/dft-rank1.c -2 +2 cannam@95: M ./mpi/dft-serial.c -2 +2 cannam@95: M ./mpi/dft-solve.c -2 +2 cannam@95: M ./mpi/dtensor.c -2 +2 cannam@95: M ./mpi/fftw3-mpi.h -2 +2 cannam@95: M ./mpi/ifftw-mpi.h -2 +2 cannam@95: M ./mpi/mpi-dft.h -2 +2 cannam@95: M ./mpi/mpi-rdft.h -2 +2 cannam@95: M ./mpi/mpi-transpose.h -2 +2 cannam@95: M ./mpi/rdft-problem.c -2 +2 cannam@95: M ./mpi/rdft-rank-geq2-transposed.c -2 +2 cannam@95: M ./mpi/rdft-rank-geq2.c -2 +2 cannam@95: M ./mpi/rdft-rank1-bigvec.c -2 +2 cannam@95: M ./mpi/rdft-serial.c -2 +2 cannam@95: M ./mpi/rdft-solve.c -2 +2 cannam@95: M ./mpi/rearrange.c -2 +2 cannam@95: M ./mpi/testsched.c -2 +2 cannam@95: M ./mpi/transpose-alltoall.c -2 +2 cannam@95: M ./mpi/transpose-pairwise.c -2 +2 cannam@95: M ./mpi/transpose-problem.c -2 +2 cannam@95: M ./mpi/transpose-recurse.c -2 +2 cannam@95: M ./mpi/transpose-solve.c -2 +2 cannam@95: M ./mpi/wisdom-api.c -2 +2 cannam@95: M ./rdft/buffered.c -2 +2 cannam@95: M ./rdft/buffered2.c -2 +2 cannam@95: M ./rdft/codelet-rdft.h -2 +2 cannam@95: M ./rdft/conf.c -2 +2 cannam@95: M ./rdft/ct-hc2c-direct.c -2 +2 cannam@95: M ./rdft/ct-hc2c.c -2 +2 cannam@95: M ./rdft/ct-hc2c.h -2 +2 cannam@95: M ./rdft/dft-r2hc.c -2 +2 cannam@95: M ./rdft/dht-r2hc.c -2 +2 cannam@95: M ./rdft/dht-rader.c -2 +2 cannam@95: M ./rdft/direct-r2c.c -2 +2 cannam@95: M ./rdft/direct-r2r.c -2 +2 cannam@95: M ./rdft/direct2.c -2 +2 cannam@95: M ./rdft/generic.c -2 +2 cannam@95: M ./rdft/hc2hc-direct.c -2 +2 cannam@95: M ./rdft/hc2hc-generic.c -2 +2 cannam@95: M ./rdft/hc2hc.c -2 +2 cannam@95: M ./rdft/hc2hc.h -2 +2 cannam@95: M ./rdft/indirect.c -2 +2 cannam@95: M ./rdft/khc2c.c -2 +2 cannam@95: M ./rdft/khc2hc.c -2 +2 cannam@95: M ./rdft/kr2c.c -2 +2 cannam@95: M ./rdft/kr2r.c -2 +2 cannam@95: M ./rdft/nop.c -2 +2 cannam@95: M ./rdft/nop2.c -2 +2 cannam@95: M ./rdft/plan.c -2 +2 cannam@95: M ./rdft/plan2.c -2 +2 cannam@95: M ./rdft/problem.c -2 +2 cannam@95: M ./rdft/problem2.c -2 +2 cannam@95: M ./rdft/rank-geq2-rdft2.c -2 +2 cannam@95: M ./rdft/rank-geq2.c -2 +2 cannam@95: M ./rdft/rank0-rdft2.c -2 +2 cannam@95: M ./rdft/rank0.c -2 +2 cannam@95: M ./rdft/rdft-dht.c -2 +2 cannam@95: M ./rdft/rdft.h -2 +2 cannam@95: M ./rdft/rdft2-inplace-strides.c -2 +2 cannam@95: M ./rdft/rdft2-rdft.c -2 +2 cannam@95: M ./rdft/rdft2-strides.c -2 +2 cannam@95: M ./rdft/rdft2-tensor-max-index.c -2 +2 cannam@95: M ./rdft/scalar/hb.h -2 +2 cannam@95: M ./rdft/scalar/hc2c.c -2 +2 cannam@95: M ./rdft/scalar/hc2cb.h -2 +2 cannam@95: M ./rdft/scalar/hc2cf.h -2 +2 cannam@95: M ./rdft/scalar/hf.h -2 +2 cannam@95: M ./rdft/scalar/hfb.c -2 +2 cannam@95: M ./rdft/scalar/r2c.c -2 +2 cannam@95: M ./rdft/scalar/r2cb.h -2 +2 cannam@95: M ./rdft/scalar/r2cbIII.h -2 +2 cannam@95: M ./rdft/scalar/r2cf.h -2 +2 cannam@95: M ./rdft/scalar/r2cfII.h -2 +2 cannam@95: M ./rdft/scalar/r2r.c -2 +2 cannam@95: M ./rdft/scalar/r2r.h -2 +2 cannam@95: M ./rdft/simd/hc2cbv.c -2 +2 cannam@95: M ./rdft/simd/hc2cbv.h -2 +2 cannam@95: M ./rdft/simd/hc2cfv.c -2 +2 cannam@95: M ./rdft/simd/hc2cfv.h -2 +2 cannam@95: M ./rdft/solve.c -2 +2 cannam@95: M ./rdft/solve2.c -2 +2 cannam@95: M ./rdft/vrank-geq1-rdft2.c -2 +2 cannam@95: M ./rdft/vrank-geq1.c -2 +2 cannam@95: M ./rdft/vrank3-transpose.c -2 +2 cannam@95: M ./reodft/conf.c -2 +2 cannam@95: M ./reodft/redft00e-r2hc-pad.c -2 +2 cannam@95: M ./reodft/redft00e-r2hc.c -2 +2 cannam@95: M ./reodft/reodft.h -2 +2 cannam@95: M ./reodft/reodft010e-r2hc.c -2 +2 cannam@95: M ./reodft/reodft11e-r2hc-odd.c -2 +2 cannam@95: M ./reodft/reodft11e-r2hc.c -2 +2 cannam@95: M ./reodft/reodft11e-radix2.c -2 +2 cannam@95: M ./reodft/rodft00e-r2hc-pad.c -2 +2 cannam@95: M ./reodft/rodft00e-r2hc.c -2 +2 cannam@95: M ./simd/altivec.c -2 +2 cannam@95: M ./simd/nonportable/sse.c -2 +2 cannam@95: M ./simd/nonportable/sse2.c -2 +2 cannam@95: M ./simd/simd-altivec.h -2 +2 cannam@95: M ./simd/simd-sse.h -2 +2 cannam@95: M ./simd/simd-sse2.h -2 +2 cannam@95: M ./simd/simd.h -2 +2 cannam@95: M ./simd/sse.c -2 +2 cannam@95: M ./simd/sse2.c -2 +2 cannam@95: M ./simd/taint.c -2 +2 cannam@95: M ./simd/x86-cpuid.h -2 +2 cannam@95: M ./threads/api.c -2 +2 cannam@95: M ./threads/conf.c -2 +2 cannam@95: M ./threads/ct.c -2 +2 cannam@95: M ./threads/dft-vrank-geq1.c -2 +2 cannam@95: M ./threads/f77api.c -2 +2 cannam@95: M ./threads/f77funcs.h -2 +2 cannam@95: M ./threads/hc2hc.c -2 +2 cannam@95: M ./threads/openmp.c -2 +2 cannam@95: M ./threads/rdft-vrank-geq1.c -2 +2 cannam@95: M ./threads/threads.c -2 +2 cannam@95: M ./threads/threads.h -2 +2 cannam@95: M ./threads/vrank-geq1-rdft2.c -2 +2 cannam@95: M ./tools/fftw-wisdom-to-conf.1 -4 +4 cannam@95: M ./tools/fftw-wisdom-to-conf.in -2 +2 cannam@95: M ./tools/fftw-wisdom.c -2 +2 cannam@95: M ./tools/fftw_wisdom.1.in -4 +4 cannam@95: cannam@95: Sun Mar 18 01:41:40 EDT 2007 stevenj@fftw.org cannam@95: * noted CodeSourcery in AUTHORS cannam@95: cannam@95: M ./AUTHORS -1 +2 cannam@95: cannam@95: Sun Mar 18 01:25:00 EDT 2007 stevenj@fftw.org cannam@95: * more MPI documentation cannam@95: cannam@95: M ./doc/fftw3.texi -2 +61 cannam@95: cannam@95: Sat Mar 17 23:15:04 EDT 2007 stevenj@fftw.org cannam@95: * added MPI multi-dimensional rdft solvers & tests cannam@95: cannam@95: M ./mpi/Makefile.am -2 +4 cannam@95: M ./mpi/api.c +104 cannam@95: M ./mpi/conf.c +5 cannam@95: M ./mpi/fftw3-mpi.h -73 +90 cannam@95: M ./mpi/mpi-bench.c -1 +94 cannam@95: A ./mpi/mpi-rdft.h cannam@95: A ./mpi/rdft-problem.c cannam@95: A ./mpi/rdft-rank-geq2-transposed.c cannam@95: A ./mpi/rdft-rank-geq2.c cannam@95: A ./mpi/rdft-rank1-bigvec.c cannam@95: A ./mpi/rdft-serial.c cannam@95: A ./mpi/rdft-solve.c cannam@95: cannam@95: Sat Mar 17 22:52:00 EDT 2007 stevenj@fftw.org cannam@95: * whoops cannam@95: cannam@95: M ./mpi/dft-rank1-bigvec.c -1 +1 cannam@95: cannam@95: Sat Mar 17 22:43:54 EDT 2007 stevenj@fftw.org cannam@95: * clarification - fftw_mpi_init should be called before importing wisdom cannam@95: cannam@95: M ./doc/fftw3.texi -2 +4 cannam@95: cannam@95: Sat Mar 17 19:49:37 EDT 2007 stevenj@fftw.org cannam@95: * kindx/y/z -> kind0/1/2 for consistency cannam@95: cannam@95: M ./api/fftw3.h -3 +3 cannam@95: M ./doc/fftw3.texi -9 +9 cannam@95: cannam@95: Sat Mar 17 19:34:02 EDT 2007 stevenj@fftw.org cannam@95: * typo cannam@95: cannam@95: M ./mpi/dft-rank-geq2-transposed.c -1 +1 cannam@95: cannam@95: Sat Mar 17 19:14:16 EDT 2007 stevenj@fftw.org cannam@95: * some refactoring in preparation for mpi-rdft cannam@95: cannam@95: M ./mpi/Makefile.am -1 +1 cannam@95: M ./mpi/api.c -1 +1 cannam@95: M ./mpi/dft-problem.c -2 +2 cannam@95: M ./mpi/dft-rank1-bigvec.c -53 +5 cannam@95: M ./mpi/dtensor.c -3 +3 cannam@95: M ./mpi/ifftw-mpi.h -1 +29 cannam@95: A ./mpi/rearrange.c cannam@95: cannam@95: Sat Mar 17 18:12:45 EDT 2007 stevenj@fftw.org cannam@95: * documented more stuff for MPI cannam@95: cannam@95: M ./doc/fftw3.texi -2 +112 cannam@95: cannam@95: Sat Mar 17 15:41:23 EDT 2007 stevenj@fftw.org cannam@95: * added NEWS for 3.2alpha cannam@95: cannam@95: M ./NEWS -1 +37 cannam@95: cannam@95: Sat Mar 17 14:50:22 EDT 2007 stevenj@fftw.org cannam@95: * documented MPI transpose routines cannam@95: cannam@95: M ./doc/fftw3.texi -1 +130 cannam@95: cannam@95: Sat Mar 17 08:57:30 EDT 2007 athena@fftw.org cannam@95: * Removed unused variables cannam@95: cannam@95: M ./rdft/direct-r2c.c -3 cannam@95: cannam@95: Fri Mar 16 14:47:10 EDT 2007 athena@fftw.org cannam@95: * Preparing for interim release of Cell code. cannam@95: cannam@95: M ./AUTHORS +2 cannam@95: M ./configure.ac -1 +1 cannam@95: M ./doc/fftw3.texi +3 cannam@95: cannam@95: Thu Feb 8 12:23:43 EST 2007 athena@fftw.org cannam@95: * Added README.Cell cannam@95: cannam@95: A ./README.Cell cannam@95: M ./Makefile.am -1 +2 cannam@95: M ./README.Cell +103 cannam@95: cannam@95: Sat Mar 10 19:17:40 EST 2007 athena@fftw.org cannam@95: * Synchronized with main branch cannam@95: cannam@95: M ./simd/simd.h +10 cannam@95: cannam@95: Mon Jan 22 17:43:56 EST 2007 athena@fftw.org cannam@95: * Adapted vrecur heuristic to Cell. cannam@95: cannam@95: M ./cell/dft-direct-cell.c -1 +28 cannam@95: M ./dft/dft.h +4 cannam@95: cannam@95: Sun Jan 21 19:09:33 EST 2007 athena@fftw.org cannam@95: * synchronize with main branch. cannam@95: cannam@95: cannam@95: Sun Jan 21 14:42:00 EST 2007 athena@fftw.org cannam@95: * synchronized with main branch cannam@95: cannam@95: cannam@95: Thu Jan 18 20:29:22 EST 2007 athena@fftw.org cannam@95: * Increased MAX_N to 32K/sizeof(R). cannam@95: cannam@95: M ./cell/dft-direct-cell.c -23 +35 cannam@95: M ./cell/fftw-cell.h -3 +3 cannam@95: M ./cell/plans-double.c -9 +1033 cannam@95: M ./cell/plans-single.c -13 +525 cannam@95: M ./cell/spu/dft.spuc -2 +8 cannam@95: cannam@95: Thu Jan 18 13:43:51 EST 2007 Matteo Frigo cannam@95: * Added pointer to solver->destroy which is used in the Cell branch. cannam@95: cannam@95: M ./mpi/dft-rank-geq2-transposed.c -1 +1 cannam@95: M ./mpi/dft-rank1-bigvec.c -1 +1 cannam@95: M ./mpi/dft-rank1.c -1 +1 cannam@95: M ./mpi/transpose-recurse.c -1 +1 cannam@95: cannam@95: Thu Jan 18 12:09:26 EST 2007 athena@fftw.org cannam@95: * Updated copyright notices cannam@95: cannam@95: R ./mpi/transpose-radix2.c cannam@95: M ./cell/cell.c -2 +1 cannam@95: M ./cell/conf.c -2 +1 cannam@95: M ./cell/copy.c -2 +1 cannam@95: M ./cell/dft-direct-cell.c -2 +1 cannam@95: M ./cell/fftw-cell.h +19 cannam@95: M ./cell/spu/alloc.spuc +19 cannam@95: M ./cell/spu/copy.spuc +19 cannam@95: M ./cell/spu/dft.spuc +18 cannam@95: M ./cell/spu/dma.spuc +19 cannam@95: M ./cell/spu/execute.spuc +19 cannam@95: M ./cell/spu/fftw-spu.h +19 cannam@95: M ./cell/spu/main.spuc +19 cannam@95: M ./cell/spu/planner.spuc +19 cannam@95: M ./cell/spu/spu-double.h +19 cannam@95: M ./cell/spu/spu-single.h +19 cannam@95: M ./cell/spu/transpose.spuc +19 cannam@95: M ./cell/spufftw-embed.S +19 cannam@95: M ./cell/transpose.c -2 +1 cannam@95: cannam@95: Fri Jan 12 12:54:43 EST 2007 athena@fftw.org cannam@95: * Use mfc_read_tag_status_all() instead of spu_mfcstat(2), since the former seems to be standardized. cannam@95: cannam@95: M ./cell/spu/dma.spuc -1 +1 cannam@95: cannam@95: Thu Jan 11 14:55:08 EST 2007 athena@fftw.org cannam@95: * Silence some int/INT warnings. cannam@95: cannam@95: M ./cell/dft-direct-cell.c -21 +24 cannam@95: cannam@95: Wed Jan 10 18:19:53 EST 2007 athena@fftw.org cannam@95: * Note incompatibility of --enable-cell with --enable-threads cannam@95: cannam@95: M ./configure.ac +3 cannam@95: cannam@95: Wed Jan 10 17:57:10 EST 2007 athena@fftw.org cannam@95: * forgot to add file cannam@95: cannam@95: A ./cell/spufftw-embed.S cannam@95: cannam@95: Wed Jan 10 17:45:16 EST 2007 athena@fftw.org cannam@95: * 64-bit cleanup cannam@95: cannam@95: M ./cell/Makefile.am -4 +3 cannam@95: M ./cell/cell.c -7 +21 cannam@95: M ./cell/copy.c -2 +2 cannam@95: M ./cell/dft-direct-cell.c -11 +28 cannam@95: M ./cell/fftw-cell.h -8 +10 cannam@95: M ./cell/spu/dft.spuc -3 +3 cannam@95: M ./cell/spu/dma.spuc -7 +9 cannam@95: M ./cell/spu/fftw-spu.h -2 +2 cannam@95: M ./cell/transpose.c -1 +1 cannam@95: M ./configure.ac +3 cannam@95: M ./m4/ax_gcc_archflag.m4 -1 +1 cannam@95: cannam@95: Wed Jan 10 13:47:20 EST 2007 athena@fftw.org cannam@95: * Use -mcpu=cell where appropriate. cannam@95: cannam@95: M ./m4/ax_gcc_archflag.m4 +1 cannam@95: cannam@95: Tue Dec 26 21:35:59 EST 2006 athena@fftw.org cannam@95: * synchronized with main cannam@95: cannam@95: M ./dft/buffered.c -14 +1 cannam@95: cannam@95: Sun Dec 24 20:58:25 EST 2006 athena@fftw.org cannam@95: * synchronized with main branch cannam@95: cannam@95: M ./Makefile.am -12 +14 cannam@95: cannam@95: Sun Dec 24 13:47:37 EST 2006 athena@fftw.org cannam@95: * synchronized with main branch, updated to new sdk. cannam@95: cannam@95: M! ./cell/spu/Makefile.am -4 +3 cannam@95: M! ./cell/spu/dma.spuc -1 cannam@95: M! ./cell/spu/execute.spuc -2 +2 cannam@95: M! ./cell/spu/fftw-spu.h -29 +29 cannam@95: cannam@95: Thu Dec 21 17:17:41 EST 2006 athena@fftw.org cannam@95: * removed obsolete file cannam@95: cannam@95: R ./rdft/rdft2-radix2.c cannam@95: cannam@95: Tue Dec 19 15:17:20 EST 2006 athena@fftw.org cannam@95: * synchronized with main branch cannam@95: cannam@95: R ./rdft/direct.c cannam@95: M ./rdft/direct-r2c.c -1 +1 cannam@95: M ./rdft/direct-r2r.c -1 +1 cannam@95: cannam@95: Tue Dec 19 11:27:38 EST 2006 athena@fftw.org cannam@95: * Synchronized with main branch cannam@95: cannam@95: M! ./rdft/ct-hc2c.c -1 +1 cannam@95: M! ./rdft/direct2.c -13 +1 cannam@95: cannam@95: Fri Dec 15 16:04:31 EST 2006 athena@fftw.org cannam@95: * resolved conflict with main branch cannam@95: cannam@95: M ./tests/fftw-bench.c -3 +2 cannam@95: cannam@95: Fri Dec 8 14:43:50 EST 2006 athena@fftw.org cannam@95: * Fixes for compilation in subdirectories cannam@95: cannam@95: M ./cell/spu/Makefile.am -5 +3 cannam@95: cannam@95: Fri Dec 8 12:46:00 EST 2006 athena@fftw.org cannam@95: * Silence warning cannam@95: cannam@95: M ./rdft/rank0.c -2 cannam@95: cannam@95: Fri Dec 8 12:24:19 EST 2006 athena@fftw.org cannam@95: * silence warning cannam@95: cannam@95: M ./rdft/rank0.c -1 +1 cannam@95: cannam@95: Thu Dec 7 15:18:17 EST 2006 athena@fftw.org cannam@95: * Commented a particularly obscure piece of code. cannam@95: cannam@95: M ./cell/spu/dma.spuc -1 +14 cannam@95: cannam@95: Thu Dec 7 11:53:29 EST 2006 athena@fftw.org cannam@95: * Reorganized, clarified conditions for applicability of the DFT solver. cannam@95: cannam@95: M ./cell/dft-direct-cell.c -193 +178 cannam@95: cannam@95: Mon Dec 4 21:33:49 EST 2006 athena@fftw.org cannam@95: * Minor changes cannam@95: cannam@95: M ./cell/dft-direct-cell.c -1 +6 cannam@95: M ./cell/spu/dft.spuc -1 +7 cannam@95: cannam@95: Mon Dec 4 17:43:28 EST 2006 athena@fftw.org cannam@95: * Clarified comment cannam@95: cannam@95: M ./cell/dft-direct-cell.c -2 +2 cannam@95: cannam@95: Mon Dec 4 16:49:06 EST 2006 athena@fftw.org cannam@95: * Less incorrect conditions for fitting into local store. cannam@95: cannam@95: M ./cell/dft-direct-cell.c -5 +12 cannam@95: cannam@95: Mon Dec 4 16:08:24 EST 2006 athena@fftw.org cannam@95: * Implemented DECDIF+TRANSPOSE on Cell cannam@95: cannam@95: R ./dft/ctsq.c cannam@95: M ./cell/dft-direct-cell.c -93 +142 cannam@95: M ./cell/fftw-cell.h -8 +3 cannam@95: M ./cell/spu/dft.spuc -12 +21 cannam@95: M ./dft/indirect-transpose.c -1 +2 cannam@95: cannam@95: Fri Dec 1 17:42:55 EST 2006 athena@fftw.org cannam@95: * relaxed conditions of applicability of SPE cannam@95: cannam@95: M ./cell/dft-direct-cell.c -31 +42 cannam@95: cannam@95: Fri Dec 1 16:28:10 EST 2006 athena@fftw.org cannam@95: * tweaks cannam@95: cannam@95: M ./cell/dft-direct-cell.c -3 +2 cannam@95: cannam@95: Fri Dec 1 14:35:17 EST 2006 athena@fftw.org cannam@95: * Implemented Cell opcounts cannam@95: cannam@95: M ./cell/dft-direct-cell.c -4 +55 cannam@95: M ./cell/spu/execute.spuc +1 cannam@95: cannam@95: Fri Dec 1 13:38:44 EST 2006 athena@fftw.org cannam@95: * minor cleanup cannam@95: cannam@95: M ./cell/dft-direct-cell.c -9 +4 cannam@95: cannam@95: Fri Dec 1 11:16:52 EST 2006 athena@fftw.org cannam@95: * use [c0 s0 c1 s1] format for Cell twiddle factors, rather than [c0 c1 s0 s1]. This makes life easier and there is no speed penalty on Cell (unlike Altivec). cannam@95: cannam@95: M ./cell/copy.c -6 +6 cannam@95: M ./cell/dft-direct-cell.c -46 +15 cannam@95: M ./cell/spu/planner.spuc -2 +3 cannam@95: M ./cell/spu/spu-single.h -4 +4 cannam@95: cannam@95: Wed Nov 29 18:02:54 EST 2006 athena@fftw.org cannam@95: * Implemented SPE-accelerated copies cannam@95: cannam@95: A ./cell/copy.c cannam@95: M ./cell/Makefile.am -2 +2 cannam@95: M ./cell/copy.c +81 cannam@95: M ./cell/fftw-cell.h +14 cannam@95: M ./cell/spu/copy.spuc -69 +31 cannam@95: M ./cell/spu/dma.spuc -2 +78 cannam@95: M ./cell/spu/fftw-spu.h -4 +6 cannam@95: M ./cell/spu/main.spuc +4 cannam@95: M ./cell/spu/transpose.spuc -14 +6 cannam@95: M ./cell/transpose.c -4 +4 cannam@95: M ./kernel/ifftw.h -1 +3 cannam@95: M ./rdft/rank0.c +52 cannam@95: cannam@95: Wed Nov 29 12:11:08 EST 2006 athena@fftw.org cannam@95: * allow SPEs to compute vrank-0 problems. cannam@95: cannam@95: M ./cell/dft-direct-cell.c -10 +8 cannam@95: M ./cell/spu/dma.spuc -4 +4 cannam@95: cannam@95: Tue Nov 28 18:03:07 EST 2006 athena@fftw.org cannam@95: * eliminated DMA lists cannam@95: cannam@95: M ./cell/spu/dft.spuc -4 +4 cannam@95: M ./cell/spu/dma.spuc -119 +62 cannam@95: M ./cell/spu/fftw-spu.h -3 cannam@95: M ./cell/spu/main.spuc -2 +2 cannam@95: M ./cell/spu/transpose.spuc -6 +6 cannam@95: cannam@95: Tue Nov 28 14:22:05 EST 2006 athena@fftw.org cannam@95: * Conservatively force all dimensions to be 0 (mod VL) in cell, since otherwise it is too hard to get all cases right. cannam@95: cannam@95: M ./cell/dft-direct-cell.c -11 +10 cannam@95: M ./cell/spu/dft.spuc -1 +2 cannam@95: cannam@95: Tue Nov 28 12:39:01 EST 2006 athena@fftw.org cannam@95: * Check alignment of strides when transposing on Cell. cannam@95: cannam@95: M ./rdft/rank0.c -1 +10 cannam@95: cannam@95: Tue Nov 28 12:19:09 EST 2006 athena@fftw.org cannam@95: * consistent usage of FFT_SIGN cannam@95: cannam@95: M ./cell/dft-direct-cell.c -3 +3 cannam@95: M ./cell/spu/dft.spuc -2 +2 cannam@95: cannam@95: Tue Nov 28 11:35:38 EST 2006 athena@fftw.org cannam@95: * clever transposition algorithm without buffering cannam@95: cannam@95: M ./cell/spu/copy.spuc -178 +59 cannam@95: M ./cell/spu/dft.spuc -52 +8 cannam@95: M ./cell/spu/dma.spuc -54 +97 cannam@95: M ./cell/spu/fftw-spu.h -13 +3 cannam@95: M ./cell/spu/transpose.spuc -13 +11 cannam@95: cannam@95: Mon Nov 27 14:08:28 EST 2006 athena@fftw.org cannam@95: * Fixed tracking of dependencies cannam@95: cannam@95: M ./cell/spu/Makefile.am -1 +3 cannam@95: cannam@95: Mon Nov 27 14:03:53 EST 2006 athena@fftw.org cannam@95: * implemented 1D transforms, various tweaks cannam@95: cannam@95: M ./cell/cell.c -4 +14 cannam@95: M ./cell/conf.c +1 cannam@95: M ./cell/dft-direct-cell.c -15 +196 cannam@95: M ./cell/fftw-cell.h -2 +13 cannam@95: M ./cell/spu/Makefile.am +1 cannam@95: M ./cell/spu/alloc.spuc +1 cannam@95: M ./cell/spu/copy.spuc -34 +151 cannam@95: M ./cell/spu/dft.spuc -2 +37 cannam@95: M ./cell/spu/dma.spuc -13 +9 cannam@95: M ./cell/spu/fftw-spu.h -5 +9 cannam@95: M ./cell/spu/main.spuc -6 +3 cannam@95: cannam@95: Wed Nov 22 15:43:36 EST 2006 athena@fftw.org cannam@95: * no need to poll mailbox on spu side cannam@95: cannam@95: M ./cell/spu/main.spuc -1 cannam@95: cannam@95: Wed Nov 22 14:08:24 EST 2006 athena@fftw.org cannam@95: * increased maximum size handled by spe cannam@95: cannam@95: M ./cell/fftw-cell.h -2 +2 cannam@95: M ./cell/plans-double.c -2 +1027 cannam@95: M ./cell/plans-single.c -2 +514 cannam@95: M ./cell/spu/alloc.spuc -1 +1 cannam@95: M ./cell/spu/planner.spuc -1 +1 cannam@95: cannam@95: Tue Nov 21 16:23:17 EST 2006 athena@fftw.org cannam@95: * allow vrank<=2 problems in SPEs to avoid the vecloop overhead (grrr...) cannam@95: cannam@95: M ./cell/dft-direct-cell.c -40 +114 cannam@95: M ./cell/fftw-cell.h -1 +15 cannam@95: M ./cell/spu/dft.spuc -27 +37 cannam@95: M ./dft/rank-geq2.c +4 cannam@95: cannam@95: Mon Nov 20 14:41:45 EST 2006 athena@fftw.org cannam@95: * added emacs mode cannam@95: cannam@95: M ./cell/spu/copy.spuc +1 cannam@95: cannam@95: Mon Nov 20 09:34:12 EST 2006 athena@fftw.org cannam@95: * revised transpose, cleanup cannam@95: cannam@95: A ./cell/spu/transpose.spuc cannam@95: A ./cell/spu/dft.spuc cannam@95: M ./cell/spu/Makefile.am -1 +1 cannam@95: M ./cell/spu/dft.spuc +165 cannam@95: M ./cell/spu/fftw-spu.h +5 cannam@95: M ./cell/spu/main.spuc -224 +2 cannam@95: M ./cell/spu/transpose.spuc +70 cannam@95: cannam@95: Sun Nov 19 20:20:23 EST 2006 athena@fftw.org cannam@95: * added file cannam@95: cannam@95: A ./cell/spu/Makefile.am cannam@95: cannam@95: Sun Nov 19 20:18:35 EST 2006 athena@fftw.org cannam@95: * removed file cannam@95: cannam@95: R ./cell/spu/generate.sh cannam@95: cannam@95: Sun Nov 19 20:15:38 EST 2006 athena@fftw.org cannam@95: * better automake integration cannam@95: cannam@95: M! ./cell/Makefile.am -9 +4 cannam@95: R! ./cell/spu/alloc.c cannam@95: A! ./cell/spu/alloc.spuc cannam@95: R! ./cell/spu/copy.c cannam@95: A! ./cell/spu/copy.spuc cannam@95: R! ./cell/spu/dma.c cannam@95: A! ./cell/spu/dma.spuc cannam@95: R! ./cell/spu/execute.c cannam@95: A! ./cell/spu/execute.spuc cannam@95: R! ./cell/spu/main.c cannam@95: A! ./cell/spu/main.spuc cannam@95: R! ./cell/spu/planner.c cannam@95: A! ./cell/spu/planner.spuc cannam@95: R! ./cell/spu/spu_n2fv_10.c cannam@95: R! ./cell/spu/spu_n2fv_11.c cannam@95: R! ./cell/spu/spu_n2fv_12.c cannam@95: R! ./cell/spu/spu_n2fv_13.c cannam@95: R! ./cell/spu/spu_n2fv_14.c cannam@95: R! ./cell/spu/spu_n2fv_15.c cannam@95: R! ./cell/spu/spu_n2fv_16.c cannam@95: R! ./cell/spu/spu_n2fv_2.c cannam@95: R! ./cell/spu/spu_n2fv_3.c cannam@95: R! ./cell/spu/spu_n2fv_32.c cannam@95: R! ./cell/spu/spu_n2fv_4.c cannam@95: R! ./cell/spu/spu_n2fv_5.c cannam@95: R! ./cell/spu/spu_n2fv_6.c cannam@95: R! ./cell/spu/spu_n2fv_7.c cannam@95: R! ./cell/spu/spu_n2fv_8.c cannam@95: R! ./cell/spu/spu_n2fv_9.c cannam@95: R! ./cell/spu/spu_t1fv_10.c cannam@95: R! ./cell/spu/spu_t1fv_12.c cannam@95: R! ./cell/spu/spu_t1fv_15.c cannam@95: R! ./cell/spu/spu_t1fv_16.c cannam@95: R! ./cell/spu/spu_t1fv_2.c cannam@95: R! ./cell/spu/spu_t1fv_3.c cannam@95: R! ./cell/spu/spu_t1fv_32.c cannam@95: R! ./cell/spu/spu_t1fv_4.c cannam@95: R! ./cell/spu/spu_t1fv_5.c cannam@95: R! ./cell/spu/spu_t1fv_6.c cannam@95: R! ./cell/spu/spu_t1fv_7.c cannam@95: R! ./cell/spu/spu_t1fv_8.c cannam@95: R! ./cell/spu/spu_t1fv_9.c cannam@95: M! ./dft/indirect-transpose.c -1 +1 cannam@95: M! ./rdft/direct2.c -1 +13 cannam@95: cannam@95: Sat Nov 18 20:14:29 EST 2006 athena@fftw.org cannam@95: * changed algorithm for computing chunk size cannam@95: cannam@95: M ./cell/spu/main.c -32 +36 cannam@95: cannam@95: Sat Nov 18 19:18:11 EST 2006 athena@fftw.org cannam@95: * implemented transpose, various fixes. cannam@95: cannam@95: M! ./cell/Makefile.am -2 +2 cannam@95: M! ./cell/cell.c -2 +10 cannam@95: M! ./cell/conf.c -1 +1 cannam@95: M! ./cell/dft-direct-cell.c -13 +29 cannam@95: M! ./cell/fftw-cell.h -3 +13 cannam@95: A! ./cell/spu/copy.c cannam@95: A! ./cell/spu/dma.c cannam@95: M! ./cell/spu/fftw-spu.h +27 cannam@95: M! ./cell/spu/main.c -169 +83 cannam@95: A! ./cell/transpose.c cannam@95: M! ./configure.ac -2 +5 cannam@95: M! ./kernel/ifftw.h +5 cannam@95: M! ./rdft/rank0.c +24 cannam@95: M! ./tests/fftw-bench.c +4 cannam@95: cannam@95: Thu Nov 16 16:33:50 EST 2006 athena@fftw.org cannam@95: * Added explicit destructor to all solvers to help with the cell port. cannam@95: cannam@95: M! ./Makefile.am -2 +6 cannam@95: M! ./cell/cell.c -4 +22 cannam@95: M! ./cell/dft-direct-cell.c -1 +8 cannam@95: M! ./cell/fftw-cell.h -1 +3 cannam@95: M! ./configure.ac +1 cannam@95: M! ./dft/bluestein.c -1 +1 cannam@95: M! ./dft/buffered.c -1 +1 cannam@95: M! ./dft/ct.c -1 +1 cannam@95: A! ./dft/ctsq.c cannam@95: M! ./dft/direct.c -1 +1 cannam@95: M! ./dft/generic.c -1 +1 cannam@95: M! ./dft/indirect-transpose.c -1 +1 cannam@95: M! ./dft/indirect.c -1 +1 cannam@95: M! ./dft/nop.c -1 +1 cannam@95: M! ./dft/rader.c -1 +1 cannam@95: M! ./dft/rank-geq2.c -1 +1 cannam@95: M! ./dft/vrank-geq1.c -1 +1 cannam@95: M! ./kernel/ifftw.h +1 cannam@95: M! ./kernel/solver.c -1 +4 cannam@95: M! ./mpi/dft-rank-geq2.c -1 +1 cannam@95: M! ./mpi/dft-serial.c -1 +1 cannam@95: M! ./mpi/transpose-alltoall.c -1 +1 cannam@95: M! ./mpi/transpose-pairwise.c -1 +1 cannam@95: A! ./mpi/transpose-radix2.c cannam@95: M! ./rdft/buffered.c -1 +1 cannam@95: M! ./rdft/buffered2.c -1 +1 cannam@95: M! ./rdft/dft-r2hc.c -1 +1 cannam@95: M! ./rdft/dht-r2hc.c -1 +1 cannam@95: M! ./rdft/dht-rader.c -1 +1 cannam@95: A! ./rdft/direct.c cannam@95: M! ./rdft/direct2.c -1 +1 cannam@95: M! ./rdft/generic.c -1 +1 cannam@95: M! ./rdft/hc2hc.c -1 +1 cannam@95: M! ./rdft/indirect.c -1 +1 cannam@95: M! ./rdft/nop.c -1 +1 cannam@95: M! ./rdft/nop2.c -1 +1 cannam@95: M! ./rdft/rank-geq2-rdft2.c -1 +1 cannam@95: M! ./rdft/rank-geq2.c -1 +1 cannam@95: M! ./rdft/rank0-rdft2.c -1 +1 cannam@95: M! ./rdft/rank0.c -1 +1 cannam@95: M! ./rdft/rdft-dht.c -1 +1 cannam@95: A! ./rdft/rdft2-radix2.c cannam@95: M! ./rdft/vrank-geq1-rdft2.c -1 +1 cannam@95: M! ./rdft/vrank-geq1.c -1 +1 cannam@95: M! ./rdft/vrank3-transpose.c -1 +1 cannam@95: M! ./reodft/redft00e-r2hc-pad.c -1 +1 cannam@95: M! ./reodft/redft00e-r2hc.c -1 +1 cannam@95: M! ./reodft/reodft00e-splitradix.c -1 +1 cannam@95: M! ./reodft/reodft010e-r2hc.c -1 +1 cannam@95: M! ./reodft/reodft11e-r2hc-odd.c -1 +1 cannam@95: M! ./reodft/reodft11e-r2hc.c -1 +1 cannam@95: M! ./reodft/reodft11e-radix2.c -1 +1 cannam@95: M! ./reodft/rodft00e-r2hc-pad.c -1 +1 cannam@95: M! ./reodft/rodft00e-r2hc.c -1 +1 cannam@95: M! ./tests/fftw-bench.c -1 cannam@95: M! ./threads/ct.c -1 +1 cannam@95: M! ./threads/dft-vrank-geq1.c -1 +1 cannam@95: M! ./threads/hc2hc.c -1 +1 cannam@95: M! ./threads/rdft-vrank-geq1.c -1 +1 cannam@95: M! ./threads/vrank-geq1-rdft2.c -1 +1 cannam@95: cannam@95: Thu Nov 16 15:22:15 EST 2006 athena@fftw.org cannam@95: * consistent use of #if vs. #ifdef cannam@95: cannam@95: M ./api/configure.c +2 cannam@95: M ./cell/cell.c -1 +1 cannam@95: M ./cell/conf.c -1 +1 cannam@95: M ./cell/dft-direct-cell.c -1 +1 cannam@95: M ./cell/plans-double.c -1 +1 cannam@95: M ./cell/plans-single.c -1 +1 cannam@95: cannam@95: Thu Nov 16 15:15:34 EST 2006 athena@fftw.org cannam@95: * Additional Cell double codelets, better automake integration cannam@95: cannam@95: A! ./cell/spu/spu_n2fv_11.c cannam@95: A! ./cell/spu/spu_n2fv_13.c cannam@95: A! ./cell/spu/spu_n2fv_15.c cannam@95: A! ./cell/spu/spu_n2fv_3.c cannam@95: A! ./cell/spu/spu_n2fv_5.c cannam@95: A! ./cell/spu/spu_n2fv_7.c cannam@95: A! ./cell/spu/spu_n2fv_9.c cannam@95: M! ./Makefile.am -7 +5 cannam@95: M! ./cell/Makefile.am -2 +7 cannam@95: M! ./cell/cell.c -11 +7 cannam@95: M! ./cell/conf.c +26 cannam@95: M! ./cell/dft-direct-cell.c -9 +14 cannam@95: M! ./cell/fftw-cell.h -2 cannam@95: M! ./cell/plans-double.c -196 +201 cannam@95: M! ./cell/plans-single.c -3 +9 cannam@95: M! ./cell/spu/execute.c +10 cannam@95: M! ./cell/spu/fftw-spu.h -8 +22 cannam@95: M! ./cell/spu/generate.sh -1 +6 cannam@95: M! ./cell/spu/main.c -2 +1 cannam@95: M! ./cell/spu/planner.c +3 cannam@95: M! ./cell/spu/spu_n2fv_11.c +118 cannam@95: M! ./cell/spu/spu_n2fv_13.c +155 cannam@95: M! ./cell/spu/spu_n2fv_15.c +132 cannam@95: M! ./cell/spu/spu_n2fv_3.c +37 cannam@95: M! ./cell/spu/spu_n2fv_5.c +52 cannam@95: M! ./cell/spu/spu_n2fv_7.c +68 cannam@95: M! ./cell/spu/spu_n2fv_9.c +103 cannam@95: M! ./configure.ac +6 cannam@95: M! ./dft/buffered.c -1 +14 cannam@95: M! ./kernel/align.c -1 +1 cannam@95: M! ./kernel/ifftw.h -2 +2 cannam@95: M! ./simd/simd.h -4 cannam@95: M! ./simd/taint.c -1 +1 cannam@95: cannam@95: Thu Nov 16 12:43:34 EST 2006 athena@fftw.org cannam@95: * Use dma lists. cannam@95: cannam@95: M ./cell/spu/main.c -21 +86 cannam@95: cannam@95: Thu Nov 16 11:03:46 EST 2006 athena@fftw.org cannam@95: * converted to automake cannam@95: cannam@95: R ./cell/spu/Makefile.in cannam@95: M ./cell/Makefile.am +1 cannam@95: cannam@95: Wed Nov 15 18:00:12 EST 2006 athena@fftw.org cannam@95: * Initial port to Cell Broadband Engine. cannam@95: cannam@95: A! ./cell/ cannam@95: A! ./cell/spu/ cannam@95: A! ./cell/Makefile.am cannam@95: A! ./cell/fftw-cell.h cannam@95: A! ./cell/plans-single.c cannam@95: A! ./cell/spu/Makefile.in cannam@95: A! ./cell/spu/execute.c cannam@95: A! ./cell/spu/fftw-spu.h cannam@95: A! ./cell/spu/generate.sh cannam@95: A! ./cell/spu/main.c cannam@95: A! ./cell/spu/planner.c cannam@95: A! ./cell/spu/spu-single.h cannam@95: A! ./cell/spu/spu_n2fv_10.c cannam@95: A! ./cell/spu/spu_n2fv_12.c cannam@95: A! ./cell/spu/spu_n2fv_14.c cannam@95: A! ./cell/spu/spu_n2fv_16.c cannam@95: A! ./cell/spu/spu_n2fv_2.c cannam@95: A! ./cell/spu/spu_n2fv_32.c cannam@95: A! ./cell/spu/spu_n2fv_4.c cannam@95: A! ./cell/spu/spu_n2fv_6.c cannam@95: A! ./cell/spu/spu_n2fv_8.c cannam@95: A! ./cell/spu/spu_t1fv_10.c cannam@95: A! ./cell/spu/spu_t1fv_12.c cannam@95: A! ./cell/spu/spu_t1fv_15.c cannam@95: A! ./cell/spu/spu_t1fv_16.c cannam@95: A! ./cell/spu/spu_t1fv_2.c cannam@95: A! ./cell/spu/spu_t1fv_3.c cannam@95: A! ./cell/spu/spu_t1fv_32.c cannam@95: A! ./cell/spu/spu_t1fv_4.c cannam@95: A! ./cell/spu/spu_t1fv_5.c cannam@95: A! ./cell/spu/spu_t1fv_6.c cannam@95: A! ./cell/spu/spu_t1fv_7.c cannam@95: A! ./cell/spu/spu_t1fv_8.c cannam@95: A! ./cell/spu/spu_t1fv_9.c cannam@95: A! ./cell/cell.c cannam@95: A! ./cell/conf.c cannam@95: A! ./cell/dft-direct-cell.c cannam@95: A! ./cell/spu/spu-double.h cannam@95: A! ./cell/plans-double.c cannam@95: A! ./cell/spu/alloc.c cannam@95: M! ./Makefile.am -14 +14 cannam@95: M! ./api/configure.c +1 cannam@95: M! ./cell/Makefile.am +22 cannam@95: M! ./cell/cell.c +90 cannam@95: M! ./cell/conf.c +12 cannam@95: M! ./cell/dft-direct-cell.c +302 cannam@95: M! ./cell/fftw-cell.h +64 cannam@95: M! ./cell/plans-double.c +2056 cannam@95: M! ./cell/plans-single.c +1032 cannam@95: M! ./cell/spu/Makefile.in +35 cannam@95: M! ./cell/spu/alloc.c +24 cannam@95: M! ./cell/spu/execute.c +68 cannam@95: M! ./cell/spu/fftw-spu.h +86 cannam@95: M! ./cell/spu/generate.sh +16 cannam@95: M! ./cell/spu/main.c +281 cannam@95: M! ./cell/spu/planner.c +187 cannam@95: M! ./cell/spu/spu-double.h +105 cannam@95: M! ./cell/spu/spu-single.h +120 cannam@95: M! ./cell/spu/spu_n2fv_10.c +100 cannam@95: M! ./cell/spu/spu_n2fv_12.c +109 cannam@95: M! ./cell/spu/spu_n2fv_14.c +142 cannam@95: M! ./cell/spu/spu_n2fv_16.c +146 cannam@95: M! ./cell/spu/spu_n2fv_2.c +33 cannam@95: M! ./cell/spu/spu_n2fv_32.c +313 cannam@95: M! ./cell/spu/spu_n2fv_4.c +44 cannam@95: M! ./cell/spu/spu_n2fv_6.c +62 cannam@95: M! ./cell/spu/spu_n2fv_8.c +75 cannam@95: M! ./cell/spu/spu_t1fv_10.c +104 cannam@95: M! ./cell/spu/spu_t1fv_12.c +114 cannam@95: M! ./cell/spu/spu_t1fv_15.c +162 cannam@95: M! ./cell/spu/spu_t1fv_16.c +153 cannam@95: M! ./cell/spu/spu_t1fv_2.c +32 cannam@95: M! ./cell/spu/spu_t1fv_3.c +41 cannam@95: M! ./cell/spu/spu_t1fv_32.c +329 cannam@95: M! ./cell/spu/spu_t1fv_4.c +44 cannam@95: M! ./cell/spu/spu_t1fv_5.c +60 cannam@95: M! ./cell/spu/spu_t1fv_6.c +63 cannam@95: M! ./cell/spu/spu_t1fv_7.c +81 cannam@95: M! ./cell/spu/spu_t1fv_8.c +77 cannam@95: M! ./cell/spu/spu_t1fv_9.c +120 cannam@95: M! ./configure.ac +2 cannam@95: M! ./kernel/ifftw.h -17 +21 cannam@95: M! ./libbench2/util.c -1 +1 cannam@95: M! ./tests/fftw-bench.c +1 cannam@95: cannam@95: Wed Mar 14 10:19:53 EDT 2007 athena@fftw.org cannam@95: * Remove Codesourcery contributions from commercial tarball. cannam@95: cannam@95: M ./commercialize.sh +7 cannam@95: cannam@95: Wed Mar 14 08:59:18 EDT 2007 athena@fftw.org cannam@95: * Added FFTW_WISDOM_ONLY, at the request of Phil Dumont. cannam@95: cannam@95: M ./api/apiplan.c -19 +29 cannam@95: M ./api/fftw3.h +1 cannam@95: M ./tests/fftw-bench.c +1 cannam@95: cannam@95: Tue Mar 13 00:32:05 EDT 2007 stevenj@fftw.org cannam@95: * fixed potential MPI deadlock if timer misbehaves cannam@95: cannam@95: M ./kernel/timer.c -3 +3 cannam@95: cannam@95: Mon Mar 12 23:31:52 EDT 2007 stevenj@fftw.org cannam@95: * more work on MPI documentation cannam@95: cannam@95: M ./doc/fftw3.texi -31 +309 cannam@95: M ./mpi/api.c -3 +4 cannam@95: M ./mpi/fftw3-mpi.h -19 +19 cannam@95: cannam@95: Tue Feb 27 13:48:43 EST 2007 stevenj@fftw.org cannam@95: * index cannam@95: cannam@95: M ./doc/fftw3.texi +1 cannam@95: cannam@95: Tue Feb 27 13:46:45 EST 2007 stevenj@fftw.org cannam@95: * rename "new-data execute" to "new-array execute", since of course you do not need a new array to have new data cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -1 +2 cannam@95: M ./doc/fftw3.texi -12 +12 cannam@95: cannam@95: Tue Feb 27 13:43:55 EST 2007 stevenj@fftw.org cannam@95: * consistency with manual (guru execute -> new-data execute) cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -1 +1 cannam@95: cannam@95: Tue Feb 27 13:42:24 EST 2007 stevenj@fftw.org cannam@95: * texinfo fixes; renamed "guru execute" section to "new-data execute", since previously it seemed to lead to endless confusion with the guru planner API cannam@95: cannam@95: M ./doc/fftw3.texi -60 +83 cannam@95: cannam@95: Mon Feb 26 18:57:11 EST 2007 stevenj@fftw.org cannam@95: * consistently use n0/n1/.. everywhere instead of nx/ny/... (for consistency with d-dimensional case n[0], n[1], ...) ... first start at MPI documentation cannam@95: cannam@95: M ./api/fftw3.h -10 +10 cannam@95: M ./doc/fftw3.texi -138 +374 cannam@95: M ./mpi/fftw3-mpi.h -27 +27 cannam@95: cannam@95: Sat Mar 10 18:48:05 EST 2007 athena@fftw.org cannam@95: * Changed C++-style comment into K&R cannam@95: cannam@95: M ./kernel/cycle.h -1 +1 cannam@95: cannam@95: Sat Mar 10 18:47:12 EST 2007 athena@fftw.org cannam@95: * Forgot to add file cannam@95: cannam@95: A ./simd/simd-mips_ps.h cannam@95: cannam@95: Sat Mar 10 18:44:39 EST 2007 athena@fftw.org cannam@95: * Note removal of K7 support. cannam@95: cannam@95: M ./AUTHORS -1 +4 cannam@95: cannam@95: Sat Mar 10 18:41:52 EST 2007 athena@fftw.org cannam@95: * Updated manual for MIPS PS cannam@95: cannam@95: M ./doc/fftw3.texi -9 +12 cannam@95: cannam@95: Sat Mar 10 18:37:07 EST 2007 athena@fftw.org cannam@95: * Adopted MIPS_PS patches from Codesourcery. cannam@95: cannam@95: A ./simd/mips_ps.c cannam@95: A ./simd/mips_ps.h cannam@95: M ./configure.ac +23 cannam@95: M ./kernel/cycle.h +43 cannam@95: M ./kernel/ifftw.h -1 +2 cannam@95: M ./kernel/timer.c -1 +3 cannam@95: M ./simd/Makefile.am -1 +1 cannam@95: M ./simd/mips_ps.c +45 cannam@95: M ./simd/mips_ps.h +273 cannam@95: M ./simd/simd.h +4 cannam@95: cannam@95: Sun Feb 25 11:34:51 EST 2007 athena@fftw.org cannam@95: * Incorrect initialization of win32 semaphores cannam@95: cannam@95: M ./threads/threads.c -1 +1 cannam@95: cannam@95: Tue Jan 30 11:43:09 EST 2007 stevenj@fftw.org cannam@95: * win32 fixes (I think, still untested) cannam@95: cannam@95: M ./threads/threads.c -9 +10 cannam@95: cannam@95: Fri Jan 19 17:31:47 EST 2007 stevenj@fftw.org cannam@95: * message-size heuristic in tranpose-recurse cannam@95: cannam@95: M ./mpi/transpose-recurse.c -3 +15 cannam@95: cannam@95: Tue Jan 30 08:53:55 EST 2007 athena@fftw.org cannam@95: * Threading layer for Win32, completely untested. cannam@95: cannam@95: M ./threads/threads.c +69 cannam@95: cannam@95: Mon Jan 29 14:26:30 EST 2007 athena@fftw.org cannam@95: * Check for EINTR after sem_wait(), as suggested by Chip Salzenberg. cannam@95: cannam@95: M ./threads/threads.c -1 +11 cannam@95: cannam@95: Mon Jan 22 13:58:23 EST 2007 athena@fftw.org cannam@95: * Force vector recursion by means of a separate function pointer. I need this for Cell. cannam@95: cannam@95: M ./dft/ct.c -3 +7 cannam@95: M ./dft/ct.h -2 +8 cannam@95: M ./dft/dft.h -15 cannam@95: M ./dft/dftw-direct.c -2 +2 cannam@95: M ./dft/dftw-directsq.c -2 +3 cannam@95: M ./dft/dftw-generic.c -2 +2 cannam@95: M ./dft/dftw-genericbuf.c -2 +2 cannam@95: M ./dft/vrank-geq1.c -4 cannam@95: M ./threads/ct.c -1 +3 cannam@95: M ./threads/threads.h -1 +3 cannam@95: cannam@95: Mon Jan 22 09:28:35 EST 2007 athena@fftw.org cannam@95: * Merge multiplications by twiddle with multiplications by i for faster r2c transforms. cannam@95: cannam@95: M ./genfft/algsimp.ml +4 cannam@95: M ./genfft/simd.ml +4 cannam@95: M ./simd/simd-altivec.h +24 cannam@95: M ./simd/simd-sse.h +18 cannam@95: M ./simd/simd-sse2.h +18 cannam@95: cannam@95: Sun Jan 21 19:02:44 EST 2007 athena@fftw.org cannam@95: * Disabled vector recursion, too messy. cannam@95: cannam@95: M ./dft/dft.h -2 +13 cannam@95: cannam@95: Sun Jan 21 14:23:35 EST 2007 athena@fftw.org cannam@95: * Changed heuristics for vector recursion. cannam@95: cannam@95: As in fftw-3.1, NO_VRECURSE disables vector recursion. As an cannam@95: exception, however, vector recursion is allowed when the predicate cannam@95: VRECURSE_ANYWAYP is true. We need some form of vector recursion to cannam@95: obtain decent plans on Cell, and this solution captures the common cannam@95: cases without increasing planning time too much. cannam@95: cannam@95: cannam@95: M ./api/fftw3.h -1 +1 cannam@95: M ./api/mapflags.c -2 +2 cannam@95: M ./dft/ct.c -10 +6 cannam@95: M ./dft/dft.h +4 cannam@95: M ./dft/vrank-geq1.c -6 +3 cannam@95: M ./kernel/ifftw.h -2 +2 cannam@95: M ./kernel/planner.c -1 +1 cannam@95: M ./rdft/ct-hc2c.c -9 +4 cannam@95: M ./rdft/hc2hc.c -9 +4 cannam@95: M ./rdft/vrank-geq1.c -7 cannam@95: cannam@95: Sun Dec 17 22:31:17 EST 2006 athena@fftw.org cannam@95: * fixed hc2c for vector-recursion branch cannam@95: cannam@95: M ./rdft/ct-hc2c.c -2 +7 cannam@95: cannam@95: Tue Dec 5 12:52:36 EST 2006 athena@fftw.org cannam@95: * switch to default vector recursion cannam@95: cannam@95: M ./api/fftw3.h -1 +1 cannam@95: M ./api/mapflags.c -2 +2 cannam@95: M ./dft/ct.c -4 +6 cannam@95: M ./dft/vrank-geq1.c +7 cannam@95: M ./kernel/ifftw.h -2 +2 cannam@95: M ./kernel/planner.c -1 +1 cannam@95: M ./rdft/hc2hc.c -2 +7 cannam@95: M ./rdft/vrank-geq1.c -2 +9 cannam@95: cannam@95: Sat Jan 20 23:37:33 EST 2007 athena@fftw.org cannam@95: * Smarter algorithm for selection of nbuf. cannam@95: cannam@95: M ./kernel/buffered.c -8 +6 cannam@95: cannam@95: Sat Jan 20 22:15:33 EST 2007 athena@fftw.org cannam@95: * Increased buffer sizes according to Moore's law. cannam@95: cannam@95: M ./kernel/buffered.c -3 +5 cannam@95: cannam@95: Fri Jan 19 16:02:00 EST 2007 stevenj@fftw.org cannam@95: * fix another MPI synchronization bug -- several more places where cost_hook must be called to synchronized process timings (sigh) cannam@95: cannam@95: M ./api/flops.c -3 +3 cannam@95: M ./kernel/ifftw.h -4 +5 cannam@95: M ./kernel/planner.c -10 +6 cannam@95: M ./kernel/timer.c -4 +15 cannam@95: M ./mpi/api.c -3 +1 cannam@95: cannam@95: Fri Jan 19 12:08:07 EST 2007 athena@fftw.org cannam@95: * Set havewisdom=0 when calling forget_wisdom() in the test program. cannam@95: cannam@95: M ./tests/fftw-bench.c -1 +3 cannam@95: cannam@95: Fri Jan 19 10:29:56 EST 2007 stevenj@fftw.org cannam@95: * remove redundant check cannam@95: cannam@95: M ./kernel/planner.c -4 +1 cannam@95: cannam@95: Thu Jan 18 22:37:59 EST 2007 stevenj@fftw.org cannam@95: * fixed potential (unlikely) bug in wisdom import (triggered when importing impatient wisdom after creating more patient plans, but apparently only for nonstandard configure.c configurations) cannam@95: cannam@95: M ./kernel/planner.c -1 +5 cannam@95: cannam@95: Thu Jan 18 21:50:14 EST 2007 stevenj@fftw.org cannam@95: * added functions to gather/broadcast wisdom for MPI cannam@95: cannam@95: M ./configure.ac -3 +8 cannam@95: M ./mpi/Makefile.am -1 +1 cannam@95: M ./mpi/api.c -3 +3 cannam@95: M ./mpi/fftw3-mpi.h -1 +5 cannam@95: M ./mpi/mpi-bench.c -19 +40 cannam@95: A ./mpi/wisdom-api.c cannam@95: M ./tests/bench.c +10 cannam@95: M ./tests/fftw-bench.c -3 +3 cannam@95: M ./tests/fftw-bench.h +2 cannam@95: cannam@95: Thu Jan 11 18:33:17 EST 2007 stevenj@fftw.org cannam@95: * whoops, another int/INT bug cannam@95: cannam@95: M ./mpi/transpose-pairwise.c -1 +1 cannam@95: cannam@95: Thu Jan 11 17:42:24 EST 2007 stevenj@fftw.org cannam@95: * whoops, fixed bug in transpose-recurse for r != m cannam@95: cannam@95: M ./mpi/transpose-recurse.c -10 +10 cannam@95: cannam@95: Thu Jan 11 17:25:36 EST 2007 stevenj@fftw.org cannam@95: * canonicalize mpi-transposed flags by setting TRANSPOSED_IN/OUT where possible cannam@95: cannam@95: M ./mpi/transpose-problem.c -1 +8 cannam@95: cannam@95: Thu Jan 11 17:16:24 EST 2007 stevenj@fftw.org cannam@95: * replace transpose-radix2 with much more general transpose-recurse solver cannam@95: cannam@95: M ./kernel/Makefile.am -3 +3 cannam@95: M ./kernel/ifftw.h +5 cannam@95: M ./kernel/tensor2.c -21 cannam@95: A ./kernel/tensor3.c cannam@95: M ./mpi/Makefile.am -1 +1 cannam@95: M ./mpi/conf.c -1 +1 cannam@95: M ./mpi/dft-rank-geq2-transposed.c -1 cannam@95: M ./mpi/mpi-bench.c +1 cannam@95: M ./mpi/mpi-transpose.h -1 +1 cannam@95: R ./mpi/transpose-radix2.c cannam@95: A ./mpi/transpose-recurse.c cannam@95: cannam@95: Wed Jan 10 20:23:48 EST 2007 stevenj@fftw.org cannam@95: * rename transpose-inplace to transpose-pairwise, as the algorithm is not restricted to inplace operation cannam@95: cannam@95: ./mpi/transpose-inplace.c -> ./mpi/transpose-pairwise.c cannam@95: M ./mpi/Makefile.am -1 +1 cannam@95: M ./mpi/conf.c -1 +1 cannam@95: M ./mpi/mpi-transpose.h -1 +1 cannam@95: M ./mpi/testsched.c -1 +1 cannam@95: M ./mpi/transpose-pairwise.c -9 +9 cannam@95: cannam@95: Wed Jan 10 14:39:08 EST 2007 stevenj@fftw.org cannam@95: * whoops, some int/INT bugs cannam@95: cannam@95: M ./mpi/api.c -2 +2 cannam@95: M ./mpi/block.c -3 +3 cannam@95: M ./mpi/ifftw-mpi.h -5 +5 cannam@95: cannam@95: Tue Jan 9 18:50:07 EST 2007 stevenj@fftw.org cannam@95: * fix FAQ Makefile for vpath builds cannam@95: cannam@95: M ./doc/FAQ/Makefile.am -4 +7 cannam@95: M ./doc/FAQ/fftw-faq.bfnn -1 +1 cannam@95: cannam@95: Tue Jan 9 20:22:11 EST 2007 athena@fftw.org cannam@95: * Missing ``static'' keyword. cannam@95: cannam@95: M ./rdft/direct-r2c.c -2 +2 cannam@95: cannam@95: Tue Jan 9 20:13:18 EST 2007 athena@fftw.org cannam@95: * Minor cleanup. cannam@95: cannam@95: M ./dft/direct.c -27 +16 cannam@95: cannam@95: Tue Jan 9 00:04:03 EST 2007 stevenj@fftw.org cannam@95: * interleave twiddle mults with DFTs (should we use dftw?) cannam@95: cannam@95: M ./mpi/dft-rank1.c -41 +44 cannam@95: cannam@95: Mon Jan 8 18:35:41 EST 2007 stevenj@fftw.org cannam@95: * simplified (and somewhat sped up) dft-rank1 by exploiting dft-rank1-bigvec cannam@95: cannam@95: M ./mpi/dft-rank-geq2-transposed.c -1 +1 cannam@95: M ./mpi/dft-rank-geq2.c -1 +1 cannam@95: M ./mpi/dft-rank1-bigvec.c -2 +4 cannam@95: M ./mpi/dft-rank1.c -175 +181 cannam@95: M ./mpi/transpose-alltoall.c -6 +3 cannam@95: M ./mpi/transpose-inplace.c -6 +3 cannam@95: M ./mpi/transpose-radix2.c -7 +2 cannam@95: cannam@95: Sun Jan 7 00:31:31 EST 2007 stevenj@fftw.org cannam@95: * rearranged TRANSPOSED format, numerous speedups cannam@95: cannam@95: Split the TRANSPOSED and non-TRANSPOSED rank-geq2 solvers, and changed cannam@95: the DFT TRANSPOSED format to be more like fftw2 (both globally and cannam@95: locally transposed). In general, more emphasis on arranging the data cannam@95: contiguously for the DFTs, and more flexibility in intermediate cannam@95: transposed formats. Also disable NO_SLOW when planning transposes, cannam@95: since otherwise non-square in-place transposes gratuitously put the cannam@95: planner in SLOW mode. cannam@95: cannam@95: Currently, dft-rank1-bigvec has 5 variants (or 10, if DESTROY_INPUT). cannam@95: It looks like only 2 of these are commonly used, so I should probably cannam@95: add some UGLY tags once I do more benchmarking. cannam@95: cannam@95: M ./mpi/Makefile.am -1 +1 cannam@95: M ./mpi/api.c -14 +34 cannam@95: M ./mpi/conf.c +1 cannam@95: M ./mpi/dft-problem.c -2 +12 cannam@95: A ./mpi/dft-rank-geq2-transposed.c cannam@95: M ./mpi/dft-rank-geq2.c -131 +60 cannam@95: M ./mpi/dft-rank1-bigvec.c -77 +123 cannam@95: M ./mpi/dft-rank1.c -2 +2 cannam@95: M ./mpi/ifftw-mpi.h -2 +3 cannam@95: M ./mpi/mpi-bench.c +13 cannam@95: M ./mpi/mpi-dft.h -1 +2 cannam@95: M ./mpi/transpose-alltoall.c -32 +39 cannam@95: M ./mpi/transpose-inplace.c -53 +69 cannam@95: M ./mpi/transpose-radix2.c -37 +43 cannam@95: cannam@95: Thu Jan 4 19:13:17 EST 2007 stevenj@fftw.org cannam@95: * add bench_cost_postprocess to prevent deadlocks in mpi-bench cannam@95: cannam@95: M ./libbench2/Makefile.am -5 +6 cannam@95: A ./libbench2/bench-cost-postprocess.c cannam@95: M ./libbench2/bench-user.h +1 cannam@95: M ./libbench2/speed.c -2 +2 cannam@95: M ./mpi/mpi-bench.c +7 cannam@95: cannam@95: Thu Jan 4 16:46:29 EST 2007 stevenj@fftw.org cannam@95: * whoops cannam@95: cannam@95: M ./m4/acx_pthread.m4 -1 +1 cannam@95: cannam@95: Wed Jan 3 14:23:42 EST 2007 stevenj@fftw.org cannam@95: * pass proper pointer types as arguments, so that ACX_PTHREAD still works with C++ and -Werror (thanks to Ewald Arnold for the suggestion) cannam@95: cannam@95: M ./m4/acx_pthread.m4 -3 +6 cannam@95: cannam@95: Mon Jan 1 19:30:43 EST 2007 athena@fftw.org cannam@95: * Renamed [io]vs => [io]vs_by_nbuf, which is more appropriate and would have saved me 30mins debugging. cannam@95: cannam@95: M ./dft/buffered.c -6 +6 cannam@95: cannam@95: Mon Jan 1 18:52:38 EST 2007 stevenj@fftw.org cannam@95: * add --with-g77-wrappers option & always include g77 wrappers on GNU systems and/or with gfortran cannam@95: cannam@95: Upcoming GNU/Linux distros will most likely switch to configuring FFTW cannam@95: with gfortran by default, since g77 isn't even included with recent gcc cannam@95: versions. However, we still want to include g77-compatible wrappers in cannam@95: this case (two underscores) in addition to gfortran wrappers (one cannam@95: underscore) lest we silently break binary compatibility and provoke cannam@95: lots of annoying emails. cannam@95: cannam@95: M ./api/x77.h +6 cannam@95: M ./configure.ac +11 cannam@95: cannam@95: Mon Jan 1 16:48:36 EST 2007 stevenj@fftw.org cannam@95: * use AC_HELP_STRING for --disable-fortran cannam@95: cannam@95: M ./configure.ac -1 +1 cannam@95: cannam@95: Mon Jan 1 15:56:12 EST 2007 stevenj@fftw.org cannam@95: * terminology cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -1 +1 cannam@95: cannam@95: Sat Dec 30 16:18:35 EST 2006 athena@fftw.org cannam@95: * Free buffers before calling cldrest. cannam@95: cannam@95: M ./dft/buffered.c -2 +2 cannam@95: M ./rdft/buffered.c -4 +4 cannam@95: cannam@95: Fri Dec 29 10:52:15 EST 2006 athena@fftw.org cannam@95: * Removed obsolete code. cannam@95: cannam@95: M ./simd/simd-sse.h -22 +9 cannam@95: cannam@95: Thu Dec 28 21:37:48 EST 2006 athena@fftw.org cannam@95: * Attempt to work-around old gcc bugs in a more efficient fashion that does not lose performance on newer gcc's. cannam@95: cannam@95: M ./simd/simd-sse.h -13 +24 cannam@95: cannam@95: Thu Dec 28 16:10:33 EST 2006 athena@fftw.org cannam@95: * Make sure that the speed() input is zero even in paranoid mode. cannam@95: cannam@95: M ./libbench2/speed.c +6 cannam@95: cannam@95: Thu Dec 28 11:41:46 EST 2006 athena@fftw.org cannam@95: * cld0 and cldm problems must be tainted because they are used in a v-loop. cannam@95: cannam@95: M ./rdft/ct-hc2c-direct.c -2 +5 cannam@95: M ./rdft/hc2hc-direct.c -2 +3 cannam@95: cannam@95: Wed Dec 27 17:17:45 EST 2006 athena@fftw.org cannam@95: * Run paranoid-check in patient mode. cannam@95: cannam@95: M ./tests/Makefile.am -5 +5 cannam@95: cannam@95: Wed Dec 27 10:51:42 EST 2006 athena@fftw.org cannam@95: * Fixed incorrect initialization to zero. cannam@95: cannam@95: M ./rdft/problem2.c -1 +1 cannam@95: cannam@95: Wed Dec 27 09:33:02 EST 2006 athena@fftw.org cannam@95: * Fixed wrong TAINT() cannam@95: cannam@95: M ./rdft/buffered2.c -1 +1 cannam@95: cannam@95: Tue Dec 26 22:50:28 EST 2006 athena@fftw.org cannam@95: * Grrrrr... cannam@95: cannam@95: M ./libbench2/verify-rdft2.c -1 +1 cannam@95: cannam@95: Tue Dec 26 22:48:44 EST 2006 athena@fftw.org cannam@95: * Give up trying to verify rdft2 when vrank=-infinity. cannam@95: cannam@95: M ./libbench2/verify-rdft2.c +3 cannam@95: cannam@95: Tue Dec 26 22:31:38 EST 2006 athena@fftw.org cannam@95: * typo cannam@95: cannam@95: M ./dft/direct.c -1 +2 cannam@95: cannam@95: Tue Dec 26 21:54:53 EST 2006 athena@fftw.org cannam@95: * Correctly verify rdft2 when vrank = -infinity. cannam@95: cannam@95: M ./libbench2/verify-rdft2.c -1 +1 cannam@95: cannam@95: Tue Dec 26 21:25:02 EST 2006 athena@fftw.org cannam@95: * rdft/buffered2.c now generates rdft2 subproblems, not rdft. cannam@95: cannam@95: The old rdft2->rdft reduction is now in rdft/rdft2-rdft.c cannam@95: and still does way too much. cannam@95: cannam@95: M ./TODO -3 cannam@95: M ./rdft/Makefile.am -1 +1 cannam@95: M ./rdft/buffered.c -1 +1 cannam@95: M ./rdft/buffered2.c -190 +200 cannam@95: M ./rdft/conf.c +1 cannam@95: M ./rdft/problem2.c -5 +6 cannam@95: M ./rdft/rdft.h -1 +2 cannam@95: M ./rdft/rdft2-inplace-strides.c -5 +5 cannam@95: A ./rdft/rdft2-rdft.c cannam@95: M ./rdft/rdft2-strides.c -5 +5 cannam@95: M ./rdft/vrank-geq1-rdft2.c -11 +10 cannam@95: cannam@95: Tue Dec 26 14:03:27 EST 2006 athena@fftw.org cannam@95: * Buffer the input in hc2r problems, as opposed to the output. cannam@95: cannam@95: M ./rdft/buffered.c -34 +104 cannam@95: M ./rdft/buffered2.c -1 cannam@95: cannam@95: Tue Dec 26 10:02:59 EST 2006 athena@fftw.org cannam@95: * streamlined buffered solvers cannam@95: cannam@95: M ./dft/buffered.c -66 +14 cannam@95: M ./kernel/buffered.c -4 +25 cannam@95: M ./kernel/ifftw.h -1 +4 cannam@95: M ./kernel/primes.c +11 cannam@95: M ./rdft/buffered.c -63 +13 cannam@95: M ./rdft/buffered2.c -54 +7 cannam@95: cannam@95: Mon Dec 25 16:08:22 EST 2006 athena@fftw.org cannam@95: * c++ compatibility cannam@95: cannam@95: M ./libbench2/verify-r2r.c -2 +2 cannam@95: cannam@95: Sun Dec 24 20:27:23 EST 2006 athena@fftw.org cannam@95: * Gratuitous renaming of directories and files since the old naming was becoming too inconsistent for my taste. cannam@95: cannam@95: ./dft/codelets -> ./dft/scalar cannam@95: ./dft/scalar/standard -> ./dft/scalar/codelets cannam@95: ./rdft/codelets -> ./rdft/scalar cannam@95: M ./Makefile.am -14 +15 cannam@95: M ./configure.ac -6 +6 cannam@95: M ./dft/Makefile.am -1 +1 cannam@95: M ./dft/scalar/Makefile.am -3 +3 cannam@95: M ./dft/scalar/codelets/Makefile.am -3 +3 cannam@95: M ./dft/simd/codelets/Makefile.am -2 +2 cannam@95: M ./doc/fftw3.texi -4 +4 cannam@95: M ./rdft/Makefile.am -1 +1 cannam@95: M ./rdft/scalar/Makefile.am -2 +2 cannam@95: M ./rdft/scalar/r2cb/Makefile.am -3 +3 cannam@95: M ./rdft/scalar/r2cf/Makefile.am -3 +3 cannam@95: M ./rdft/scalar/r2r/Makefile.am -3 +3 cannam@95: cannam@95: Sun Dec 24 20:11:50 EST 2006 athena@fftw.org cannam@95: * Fixed another dftw bug (sigh) cannam@95: cannam@95: M ./threads/ct.c -7 +3 cannam@95: cannam@95: Sun Dec 24 11:48:01 EST 2006 athena@fftw.org cannam@95: * Removed debugging leftovers. cannam@95: cannam@95: M ./dft/dftw-genericbuf.c -2 +2 cannam@95: cannam@95: Sun Dec 24 11:34:17 EST 2006 athena@fftw.org cannam@95: * Moved dftw-generic* to new dftw protocol. cannam@95: cannam@95: M ./dft/dftw-generic.c -30 +24 cannam@95: M ./dft/dftw-genericbuf.c -13 +13 cannam@95: cannam@95: Sun Dec 24 09:37:19 EST 2006 athena@fftw.org cannam@95: * Oops cannam@95: cannam@95: M ./tests/hook.c -1 +1 cannam@95: cannam@95: Sun Dec 24 09:31:46 EST 2006 athena@fftw.org cannam@95: * Fixed wrong verification of rank-1 rdft2 cannam@95: cannam@95: M ./tests/hook.c -4 +6 cannam@95: cannam@95: Sat Dec 23 20:11:29 EST 2006 athena@fftw.org cannam@95: * minor tweaks cannam@95: cannam@95: M ./dft/dftw-direct.c -2 +4 cannam@95: cannam@95: Sat Dec 23 19:50:03 EST 2006 athena@fftw.org cannam@95: * Removed obsolete items. cannam@95: cannam@95: M ./TODO -13 cannam@95: cannam@95: Sat Dec 23 17:56:37 EST 2006 athena@fftw.org cannam@95: * Modified the problem_dftw invocation protocol. cannam@95: cannam@95: apply() now requires pointers to the beginning of the full array. cannam@95: Each thread processes a slice mb <= m < me. This protocol is cannam@95: consistent with the one used in hc2hc, where there is no other choice. cannam@95: cannam@95: cannam@95: cannam@95: M ./dft/dftw-direct.c -27 +31 cannam@95: M ./dft/dftw-directsq.c -1 +3 cannam@95: M ./threads/ct.c -7 +2 cannam@95: cannam@95: Sat Dec 23 16:18:25 EST 2006 athena@fftw.org cannam@95: * typo cannam@95: cannam@95: M ./simd/simd-altivec.h -1 +1 cannam@95: cannam@95: Sat Dec 23 16:06:56 EST 2006 athena@fftw.org cannam@95: * changed hc2hc twiddle storage to be the same as hc2c cannam@95: cannam@95: M ./genfft/gen_hc2hc.ml -2 +3 cannam@95: M ./rdft/hc2hc-direct.c -14 +11 cannam@95: cannam@95: Sat Dec 23 15:16:36 EST 2006 athena@fftw.org cannam@95: * Allowed extra_iter in dftw-direct. Rationalized twiddle factors in hc2c. cannam@95: cannam@95: M ./dft/dftw-direct.c -10 +38 cannam@95: M ./dft/simd/Makefile.am -1 +1 cannam@95: M ./dft/simd/codelets/Makefile.am -1 +12 cannam@95: M ./dft/simd/t.c -2 +45 cannam@95: A ./dft/simd/t1bu.h cannam@95: A ./dft/simd/t1fu.h cannam@95: M ./genfft/gen_hc2c.ml -2 +3 cannam@95: M ./genfft/gen_hc2cdft.ml -2 +3 cannam@95: M ./genfft/gen_hc2cdft_c.ml -1 +1 cannam@95: M ./genfft/gen_hc2hc.ml -1 +1 cannam@95: M ./genfft/gen_twiddle.ml -1 +1 cannam@95: M ./genfft/gen_twiddle_c.ml -1 +1 cannam@95: M ./genfft/gen_twidsq.ml -1 +1 cannam@95: M ./genfft/gen_twidsq_c.ml -1 +1 cannam@95: M ./genfft/twiddle.ml -12 +10 cannam@95: M ./genfft/twiddle.mli -1 +1 cannam@95: M ./kernel/twiddle.c -4 +8 cannam@95: M ./rdft/ct-hc2c-direct.c -4 +7 cannam@95: M ./rdft/simd/hc2cbv.h -2 +2 cannam@95: M ./rdft/simd/hc2cfv.h -2 +2 cannam@95: M ./simd/simd-altivec.h -12 +9 cannam@95: M ./simd/simd-sse.h -12 +9 cannam@95: M ./simd/simd-sse2.h -11 +6 cannam@95: cannam@95: Sat Dec 23 10:37:11 EST 2006 athena@fftw.org cannam@95: * Implemented unmentionable hack to use 4-way SIMD with an odd number of cannam@95: iterations. cannam@95: cannam@95: M ./dft/direct.c -9 +34 cannam@95: M ./rdft/ct-hc2c-direct.c -43 +88 cannam@95: M ./simd/simd-altivec.h -1 +3 cannam@95: M ./simd/simd-sse.h -1 +3 cannam@95: cannam@95: Fri Dec 22 22:13:30 EST 2006 athena@fftw.org cannam@95: * altivec support for new codelets cannam@95: cannam@95: M ./simd/simd-altivec.h -7 +28 cannam@95: cannam@95: Fri Dec 22 19:09:15 EST 2006 athena@fftw.org cannam@95: * fixed incorrect computation of W cannam@95: cannam@95: M ./genfft/gen_hc2cdft_c.ml -1 +1 cannam@95: cannam@95: Fri Dec 22 18:51:22 EST 2006 athena@fftw.org cannam@95: * Implemented 4-way simd hc2cdftv cannam@95: cannam@95: Also eliminated the twiddle_shift hack. A zillion changes dictated cannam@95: by this choice, which was in turn necessary for the hc2cdftv thing cannam@95: to work. cannam@95: cannam@95: cannam@95: M ./Makefile.am -13 +8 cannam@95: M ./configure.ac -3 cannam@95: M ./dft/codelet-dft.h -9 +9 cannam@95: M ./dft/codelets/t.c -4 +4 cannam@95: M ./dft/dftw-direct.c -35 +30 cannam@95: M ./dft/dftw-directsq.c -7 +4 cannam@95: M ./dft/simd/q1b.c -6 +6 cannam@95: M ./dft/simd/q1f.c -6 +6 cannam@95: M ./dft/simd/t.c -14 +14 cannam@95: M ./dft/simd/t3b.h +1 cannam@95: M ./dft/simd/t3f.h +1 cannam@95: M ./dft/simd/ts.c -6 +6 cannam@95: M ./dft/simd/ts.h +1 cannam@95: M ./genfft/Makefile.am -17 +10 cannam@95: R ./genfft/gen_conv.ml cannam@95: M ./genfft/gen_hc2c.ml -25 +19 cannam@95: M ./genfft/gen_hc2cdft.ml -19 +19 cannam@95: M ./genfft/gen_hc2cdft_c.ml -20 +24 cannam@95: M ./genfft/gen_hc2hc.ml -23 +17 cannam@95: M ./genfft/gen_mdct.ml -9 +9 cannam@95: M ./genfft/gen_notw.ml -8 +8 cannam@95: M ./genfft/gen_notw_c.ml -6 +6 cannam@95: M ./genfft/gen_r2cb.ml -9 +9 cannam@95: M ./genfft/gen_r2cf.ml -9 +9 cannam@95: M ./genfft/gen_r2r.ml -6 +6 cannam@95: M ./genfft/gen_twiddle.ml -37 +39 cannam@95: M ./genfft/gen_twiddle_c.ml -38 +42 cannam@95: M ./genfft/gen_twidsq.ml -42 +44 cannam@95: M ./genfft/gen_twidsq_c.ml -44 +47 cannam@95: M ./genfft/genutil.ml -6 +6 cannam@95: M ./genfft/simd.ml -6 +7 cannam@95: M ./genfft/simd.mli -2 +1 cannam@95: M ./genfft/variable.ml -7 +12 cannam@95: M ./genfft/variable.mli -1 +3 cannam@95: M ./kernel/ifftw.h -7 cannam@95: M ./kernel/twiddle.c -14 cannam@95: M ./rdft/codelet-rdft.h -5 +5 cannam@95: M ./rdft/codelets/hc2c.c -2 +2 cannam@95: M ./rdft/conf.c -1 +1 cannam@95: M ./rdft/ct-hc2c-direct.c -47 +36 cannam@95: M ./rdft/ct-hc2c.c -1 +3 cannam@95: M ./rdft/hc2hc-direct.c -39 +31 cannam@95: M ./rdft/simd/codelets/Makefile.am -2 +2 cannam@95: M ./rdft/simd/hc2cbv.c -7 +8 cannam@95: M ./rdft/simd/hc2cbv.h -4 +3 cannam@95: M ./rdft/simd/hc2cfv.c -7 +8 cannam@95: M ./rdft/simd/hc2cfv.h -4 +3 cannam@95: M ./simd/simd-sse.h +4 cannam@95: M ./simd/simd-sse2.h -3 +5 cannam@95: M ./simd/simd.h -2 cannam@95: M ./tests/Makefile.am +9 cannam@95: cannam@95: Fri Dec 22 08:45:46 EST 2006 athena@fftw.org cannam@95: * Fixed verification of rdft2 problems with new format. cannam@95: cannam@95: M ./tests/hook.c -1 +5 cannam@95: cannam@95: Fri Dec 22 00:05:59 EST 2006 athena@fftw.org cannam@95: * Added file cannam@95: cannam@95: A ./rdft/simd/Makefile.am cannam@95: cannam@95: Fri Dec 22 00:02:50 EST 2006 athena@fftw.org cannam@95: * Hmm, previous commit did not work cannam@95: cannam@95: A ./rdft/simd/ cannam@95: A ./rdft/simd/codelets/ cannam@95: A ./rdft/simd/codelets/Makefile.am cannam@95: A ./rdft/simd/hc2hcv.h cannam@95: ./rdft/simd/hc2hcv.h -> ./rdft/simd/hc2cfv.h cannam@95: A ./rdft/simd/hc2cbv.h cannam@95: A ./rdft/simd/hc2cv.c cannam@95: ./rdft/simd/hc2cv.c -> ./rdft/simd/hc2cfv.c cannam@95: M ./rdft/simd/codelets/Makefile.am +51 cannam@95: A ./rdft/simd/hc2cbv.c cannam@95: M ./rdft/simd/hc2cbv.h +29 cannam@95: M ./rdft/simd/hc2cfv.c +41 cannam@95: M ./rdft/simd/hc2cfv.h +29 cannam@95: cannam@95: Thu Dec 21 23:58:33 EST 2006 athena@fftw.org cannam@95: * Added SIMD r2cdft codelets. cannam@95: cannam@95: A ./genfft/gen_hc2cdft_c.ml cannam@95: M ./Makefile.am -14 +22 cannam@95: M ./configure.ac +5 cannam@95: M ./dft/simd/n1b.h -1 cannam@95: M ./dft/simd/n1f.h -1 cannam@95: M ./dft/simd/n2b.h -1 cannam@95: M ./dft/simd/n2f.h -1 cannam@95: M ./dft/simd/n2s.h -1 cannam@95: M ./genfft/Makefile.am -8 +14 cannam@95: M ./genfft/algsimp.ml -2 +12 cannam@95: M ./genfft/c.ml +1 cannam@95: M ./genfft/expr.ml -1 +3 cannam@95: M ./genfft/expr.mli -1 +1 cannam@95: M ./genfft/gen_hc2cdft_c.ml +217 cannam@95: M ./genfft/simd.ml -2 +9 cannam@95: M ./genfft/to_alist.ml +1 cannam@95: M ./kernel/ifftw.h +6 cannam@95: M ./rdft/Makefile.am -1 +1 cannam@95: M ./rdft/codelet-rdft.h +5 cannam@95: M ./rdft/codelets/hc2c.c -2 +12 cannam@95: M ./rdft/conf.c +4 cannam@95: M ./rdft/ct-hc2c-direct.c -5 +60 cannam@95: M ./rdft/hc2hc-direct.c -5 cannam@95: M ./rdft/hc2hc.h -2 cannam@95: M ./rdft/khc2hc.c -1 cannam@95: M ./simd/nonportable/sse.c -9 +4 cannam@95: M ./simd/nonportable/sse2.c -9 +4 cannam@95: M ./simd/simd-sse.h -4 +4 cannam@95: M ./simd/simd-sse2.h -4 +7 cannam@95: M ./simd/sse.c -2 +2 cannam@95: M ./simd/sse2.c -2 +2 cannam@95: M ./support/Makefile.codelets +1 cannam@95: cannam@95: Thu Dec 21 21:19:21 EST 2006 athena@fftw.org cannam@95: * Bug in buffering, grrr... cannam@95: cannam@95: M ./rdft/ct-hc2c-direct.c -1 +1 cannam@95: cannam@95: Thu Dec 21 20:58:14 EST 2006 athena@fftw.org cannam@95: * Oops, memory leak. cannam@95: cannam@95: M ./rdft/hc2hc-direct.c +1 cannam@95: cannam@95: Thu Dec 21 17:12:31 EST 2006 athena@fftw.org cannam@95: * minor changes, cleanup. cannam@95: cannam@95: M ./rdft/ct-hc2c.c -2 cannam@95: M ./rdft/hc2hc.c -16 +10 cannam@95: M ./threads/hc2hc.c -17 +12 cannam@95: cannam@95: Wed Dec 20 22:09:28 EST 2006 athena@fftw.org cannam@95: * Unified hc2hc-direct, hc2hc-directbuf. Cleanup. cannam@95: cannam@95: M ./rdft/Makefile.am -4 +3 cannam@95: M ./rdft/ct-hc2c-direct.c -8 +8 cannam@95: R ./rdft/hc2hc-common.c cannam@95: M ./rdft/hc2hc-direct.c -44 +153 cannam@95: R ./rdft/hc2hc-directbuf.c cannam@95: M ./rdft/hc2hc.h -4 cannam@95: cannam@95: Wed Dec 20 17:55:56 EST 2006 athena@fftw.org cannam@95: * removed obsolete rdft2-radix2 cannam@95: cannam@95: M ./rdft/Makefile.am -1 +1 cannam@95: M ./rdft/conf.c -1 cannam@95: M ./rdft/rdft.h -1 cannam@95: R ./rdft/rdft2-radix2.c cannam@95: cannam@95: Wed Dec 20 17:51:20 EST 2006 athena@fftw.org cannam@95: * implemented reduction rdft2->dft cannam@95: cannam@95: A ./genfft/gen_hc2cdft.ml cannam@95: M ./genfft/Makefile.am -9 +16 cannam@95: M ./genfft/gen_hc2c.ml -1 +1 cannam@95: M ./genfft/gen_hc2cdft.ml +207 cannam@95: M ./rdft/codelet-rdft.h -2 +7 cannam@95: M ./rdft/codelets/r2cb/Makefile.am -1 +12 cannam@95: M ./rdft/codelets/r2cf/Makefile.am -1 +12 cannam@95: M ./rdft/ct-hc2c-direct.c -5 +8 cannam@95: M ./rdft/ct-hc2c.c -24 +95 cannam@95: M ./rdft/ct-hc2c.h -2 +6 cannam@95: M ./rdft/khc2c.c -2 +3 cannam@95: M ./support/Makefile.codelets +1 cannam@95: cannam@95: Wed Dec 20 09:29:39 EST 2006 athena@fftw.org cannam@95: * Implemented buffered direct-r2c, direct-hc2c. cannam@95: Also, removed some old cruft: cannam@95: cannam@95: * okp() functions were never used and a pain to maintain---now they cannam@95: are gone. cannam@95: cannam@95: * ``m'' in hc2hc and hc2c codelets is now the number of iterations, cannam@95: not the ``logical'' m. cannam@95: cannam@95: M ./genfft/gen_hc2c.ml -5 +3 cannam@95: M ./genfft/gen_hc2hc.ml -9 +9 cannam@95: M ./genfft/gen_r2cb.ml -4 +1 cannam@95: M ./genfft/gen_r2cf.ml -4 +1 cannam@95: M ./genfft/gen_r2r.ml -5 +1 cannam@95: M ./rdft/codelet-rdft.h -30 +4 cannam@95: M ./rdft/codelets/hc2c.c -14 +2 cannam@95: M ./rdft/codelets/hfb.c -14 +2 cannam@95: M ./rdft/codelets/r2c.c -18 +4 cannam@95: M ./rdft/codelets/r2r.c -15 +1 cannam@95: M ./rdft/ct-hc2c-direct.c -37 +127 cannam@95: M ./rdft/direct-r2c.c -41 +189 cannam@95: M ./rdft/direct-r2r.c -15 +6 cannam@95: M ./rdft/direct2.c -20 +11 cannam@95: M ./rdft/hc2hc-direct.c -23 +9 cannam@95: M ./rdft/hc2hc-directbuf.c -79 +42 cannam@95: M ./rdft/kr2c.c +1 cannam@95: M ./rdft/rdft.h +1 cannam@95: cannam@95: Tue Dec 19 17:07:04 EST 2006 stevenj@fftw.org cannam@95: * added memcpy-loop rank0 solver (it makes a 5-20% difference for transposes of large tuples) cannam@95: cannam@95: M ./rdft/rank0.c +33 cannam@95: cannam@95: Tue Dec 19 16:15:54 EST 2006 stevenj@fftw.org cannam@95: * new variable to disable libbench2's problem allocation during speed benchmarking (to benchmark MPI transforms where the array does not fit into the memory of a single process) cannam@95: cannam@95: M ./libbench2/bench-user.h +2 cannam@95: M ./libbench2/speed.c -3 +8 cannam@95: M ./mpi/mpi-bench.c -5 +15 cannam@95: cannam@95: Tue Dec 19 14:55:08 EST 2006 stevenj@fftw.org cannam@95: * allow transpose-inplace to use input as scratch for DESTROY_INPUT plans (to avoid non-square in-place transpositions) ... on supersgj, the planner often prefers transpose-inplace to transpose-alltoall in this case (apparently MPI_Alltoall in LAM MPI isn't that great) cannam@95: cannam@95: M ./mpi/transpose-inplace.c -52 +99 cannam@95: cannam@95: Tue Dec 19 17:07:14 EST 2006 athena@fftw.org cannam@95: * For some reason HB2 codelets were not generated. cannam@95: cannam@95: M ./rdft/codelets/r2cb/Makefile.am -2 +2 cannam@95: M ./rdft/codelets/r2cf/Makefile.am -1 +1 cannam@95: cannam@95: Tue Dec 19 15:12:39 EST 2006 athena@fftw.org cannam@95: * split rdft/direct.c into direct-r2r and direct-r2c, since the file was getting out of control. cannam@95: cannam@95: A ./rdft/direct-r2c.c cannam@95: A ./rdft/direct-r2r.c cannam@95: R ./rdft/direct.c cannam@95: M ./rdft/Makefile.am -4 +5 cannam@95: M ./rdft/direct-r2c.c +195 cannam@95: M ./rdft/direct-r2r.c +154 cannam@95: cannam@95: Tue Dec 19 02:59:35 EST 2006 stevenj@fftw.org cannam@95: * added dft-rank1 solver - MPI now supports 1d complex DFTs! cannam@95: cannam@95: M ./mpi/Makefile.am -2 +2 cannam@95: M ./mpi/api.c -9 +54 cannam@95: A ./mpi/choose-radix.c cannam@95: M ./mpi/conf.c +1 cannam@95: M ./mpi/dft-problem.c -2 +2 cannam@95: A ./mpi/dft-rank1.c cannam@95: M ./mpi/fftw3-mpi.h -2 +8 cannam@95: M ./mpi/ifftw-mpi.h +4 cannam@95: M ./mpi/mpi-bench.c -25 +24 cannam@95: M ./mpi/mpi-dft.h -4 +5 cannam@95: M ./mpi/mpi-transpose.h -3 +3 cannam@95: M ./mpi/transpose-problem.c -1 +1 cannam@95: M ./tests/bench.c -8 +8 cannam@95: M ./tests/fftw-bench.h -1 +1 cannam@95: cannam@95: Tue Dec 19 01:27:20 EST 2006 stevenj@fftw.org cannam@95: * fftw_flops must call cost_hook directly; iestimate_cost always uses COST_MAX cannam@95: cannam@95: M ./api/flops.c -1 +7 cannam@95: M ./kernel/ifftw.h -2 +2 cannam@95: M ./kernel/planner.c -4 +3 cannam@95: cannam@95: Tue Dec 19 01:16:54 EST 2006 stevenj@fftw.org cannam@95: * fftw_flops and fftw_estimate_cost must now be called from every process, to prevent deadlocks in the MPI code (since they sum/max the cost over all processes) cannam@95: cannam@95: M ./tests/fftw-bench.c -7 +11 cannam@95: cannam@95: Tue Dec 19 00:55:34 EST 2006 stevenj@fftw.org cannam@95: * whoops, typo in assert cannam@95: cannam@95: M ./mpi/block.c -1 +1 cannam@95: cannam@95: Tue Dec 19 00:51:07 EST 2006 stevenj@fftw.org cannam@95: * remove multiplication by FFT_SIGN cannam@95: cannam@95: M ./kernel/trig.c -2 +7 cannam@95: cannam@95: Mon Dec 18 22:51:45 EST 2006 stevenj@fftw.org cannam@95: * need to synchronize ESTIMATE costs in MPI planner, and sum ESTIMATE costs for flop reporting: generalize measure_hook to cost_hook(..., {COST_SUM, COST_MAX}) cannam@95: cannam@95: M ./api/flops.c -1 +1 cannam@95: M ./kernel/ifftw.h -3 +6 cannam@95: M ./kernel/planner.c -8 +12 cannam@95: M ./mpi/api.c -6 +7 cannam@95: cannam@95: Mon Dec 18 15:36:15 EST 2006 stevenj@fftw.org cannam@95: * previous patch slowed down transpose-alltoall when TRANSPOSED_IN and DESTROY_INPUT; now allow planner to choose old behavior in this case cannam@95: cannam@95: M ./mpi/transpose-alltoall.c -11 +31 cannam@95: cannam@95: Mon Dec 18 15:15:39 EST 2006 stevenj@fftw.org cannam@95: * transpose-alltoall doesn't require input to be destroyed if TRANSPOSED_IN is set cannam@95: cannam@95: M ./mpi/transpose-alltoall.c -26 +41 cannam@95: cannam@95: Mon Dec 18 17:41:25 EST 2006 athena@fftw.org cannam@95: * Added t2-style hc2c codelets, fixed typos. cannam@95: cannam@95: M ./api/plan-guru-dft-c2r.h -1 +1 cannam@95: M ./api/plan-guru-dft-r2c.h -1 +1 cannam@95: M ./api/plan-guru-dft.h -2 +2 cannam@95: M ./rdft/codelets/r2cb/Makefile.am -1 +9 cannam@95: M ./rdft/codelets/r2cf/Makefile.am -1 +9 cannam@95: cannam@95: Mon Dec 18 16:02:42 EST 2006 athena@fftw.org cannam@95: * Renamed certain variables to avoid calling an output stride `is'. cannam@95: cannam@95: M ./rdft/codelet-rdft.h -1 +1 cannam@95: M ./rdft/direct.c -17 +17 cannam@95: M ./rdft/direct2.c -7 +7 cannam@95: cannam@95: Mon Dec 18 15:54:43 EST 2006 athena@fftw.org cannam@95: * Oops---wrong stride cannam@95: cannam@95: M ./rdft/ct-hc2c.c -1 +1 cannam@95: cannam@95: Mon Dec 18 14:59:16 EST 2006 athena@fftw.org cannam@95: * Oops, forgot to add file cannam@95: cannam@95: A ./rdft/kr2c.c cannam@95: cannam@95: Mon Dec 18 14:17:02 EST 2006 athena@fftw.org cannam@95: * Renamed r2hc/hc2r codelets to r2c cannam@95: After the recent changes, r2hc/hc2r codelets became rdft2 cannam@95: problems, so I renamed them accordingly to r2cf/r2cb. cannam@95: Codelet parameters are now a real array and a complex array, instead cannam@95: of an input array and an output array, and forward and backward cannam@95: codelets have the same type, which removes some clutter from the rdft cannam@95: code. cannam@95: cannam@95: ./genfft/gen_hc2r.ml -> ./genfft/gen_r2cb.ml cannam@95: ./genfft/gen_r2hc.ml -> ./genfft/gen_r2cf.ml cannam@95: ./rdft/codelets/hc2r -> ./rdft/codelets/r2cb cannam@95: ./rdft/codelets/r2hc -> ./rdft/codelets/r2cf cannam@95: A ./rdft/codelets/r2c.c cannam@95: A ./rdft/codelets/r2cb.h cannam@95: A ./rdft/codelets/r2cbIII.h cannam@95: A ./rdft/codelets/r2cf.h cannam@95: A ./rdft/codelets/r2cfII.h cannam@95: M ./Makefile.am -2 +2 cannam@95: M ./configure.ac -2 +2 cannam@95: M ./doc/fftw3.texi -1 +1 cannam@95: M ./genfft/Makefile.am -9 +9 cannam@95: M ./genfft/gen_r2cb.ml -58 +51 cannam@95: M ./genfft/gen_r2cf.ml -58 +51 cannam@95: M ./rdft/Makefile.am -3 +3 cannam@95: M ./rdft/codelet-rdft.h -59 +16 cannam@95: M ./rdft/codelets/Makefile.am -3 +3 cannam@95: R ./rdft/codelets/hc2r.c cannam@95: R ./rdft/codelets/hc2r.h cannam@95: R ./rdft/codelets/hc2rIII.h cannam@95: M ./rdft/codelets/r2c.c +51 cannam@95: M ./rdft/codelets/r2cb/Makefile.am -21 +21 cannam@95: M ./rdft/codelets/r2cb.h +23 cannam@95: M ./rdft/codelets/r2cbIII.h +23 cannam@95: M ./rdft/codelets/r2cf/Makefile.am -20 +20 cannam@95: M ./rdft/codelets/r2cf.h +23 cannam@95: M ./rdft/codelets/r2cfII.h +23 cannam@95: R ./rdft/codelets/r2hc.c cannam@95: R ./rdft/codelets/r2hc.h cannam@95: R ./rdft/codelets/r2hcII.h cannam@95: M ./rdft/conf.c -2 +2 cannam@95: M ./rdft/ct-hc2c.c -2 +2 cannam@95: M ./rdft/direct.c -46 +28 cannam@95: M ./rdft/direct2.c -72 +23 cannam@95: R ./rdft/khc2r.c cannam@95: R ./rdft/kr2hc.c cannam@95: M ./rdft/rdft.h -4 +2 cannam@95: M ./support/Makefile.codelets -2 +2 cannam@95: cannam@95: Mon Dec 18 10:48:07 EST 2006 athena@fftw.org cannam@95: * Implemented backward radix-2k rdft2. cannam@95: cannam@95: A ./rdft/codelets/hc2cf.h cannam@95: A ./rdft/codelets/hc2cb.h cannam@95: M ./genfft/gen_hc2c.ml -26 +30 cannam@95: M ./rdft/codelets/Makefile.am -1 +1 cannam@95: M ./rdft/codelets/hc2c.c -1 +5 cannam@95: R ./rdft/codelets/hc2c.h cannam@95: M ./rdft/codelets/hc2cb.h +23 cannam@95: M ./rdft/codelets/hc2cf.h +23 cannam@95: M ./rdft/codelets/hc2r/Makefile.am -1 +11 cannam@95: M ./rdft/codelets/r2hc/Makefile.am -7 +7 cannam@95: M ./rdft/ct-hc2c.c -3 +25 cannam@95: M ./rdft/rdft2-radix2.c -72 +1 cannam@95: cannam@95: Mon Dec 18 13:56:09 EST 2006 stevenj@fftw.org cannam@95: * move extract_reim into kernel, since it is used by internal MPI stuff and not just in the API code any more cannam@95: cannam@95: ./api/extract-reim.c -> ./kernel/extract-reim.c cannam@95: M ./api/Makefile.am -1 +1 cannam@95: M ./api/api.h -1 +1 cannam@95: M ./api/plan-many-dft-c2r.c -1 +1 cannam@95: M ./api/plan-many-dft-r2c.c -1 +1 cannam@95: M ./api/plan-many-dft.c -2 +2 cannam@95: M ./kernel/Makefile.am -5 +6 cannam@95: M ./kernel/extract-reim.c -6 +6 cannam@95: M ./kernel/ifftw.h +1 cannam@95: M ./mpi/dft-rank-geq2.c -3 cannam@95: M ./mpi/dft-rank1-bigvec.c -3 cannam@95: M ./mpi/dft-serial.c -3 cannam@95: cannam@95: Mon Dec 18 08:40:14 EST 2006 athena@fftw.org cannam@95: * Do not check r1==cr unless rnk>0 cannam@95: cannam@95: M ./rdft/problem2.c -1 +5 cannam@95: cannam@95: Sun Dec 17 21:03:50 EST 2006 athena@fftw.org cannam@95: * Implemented radix-2k RDFT2, forward only for now cannam@95: cannam@95: M ./api/execute-dft-c2r.c -1 +2 cannam@95: M ./api/execute-dft-r2c.c -1 +2 cannam@95: M ./api/execute-split-dft-c2r.c -1 +2 cannam@95: M ./api/execute-split-dft-r2c.c -1 +2 cannam@95: M ./api/f77funcs.h -4 +8 cannam@95: M ./api/plan-guru-dft-c2r.h -8 +8 cannam@95: M ./api/plan-guru-dft-r2c.h -9 +9 cannam@95: M ./api/plan-guru-split-dft-c2r.h -6 +6 cannam@95: M ./api/plan-guru-split-dft-r2c.h -9 +9 cannam@95: M ./api/plan-many-dft-c2r.c -3 +3 cannam@95: M ./api/plan-many-dft-r2c.c -1 +1 cannam@95: M ./dft/problem.c -3 +5 cannam@95: M ./genfft/Makefile.am -5 +11 cannam@95: A ./genfft/gen_hc2c.ml cannam@95: M ./genfft/gen_hc2hc.ml -30 +30 cannam@95: M ./genfft/gen_hc2r.ml -7 +16 cannam@95: M ./genfft/gen_r2hc.ml -8 +18 cannam@95: M ./rdft/Makefile.am -1 +2 cannam@95: M ./rdft/buffered2.c -31 +45 cannam@95: M ./rdft/codelet-rdft.h -11 +41 cannam@95: M ./rdft/codelets/Makefile.am -1 +1 cannam@95: A ./rdft/codelets/hc2c.c cannam@95: A ./rdft/codelets/hc2c.h cannam@95: M ./rdft/codelets/hc2r.c -4 +5 cannam@95: M ./rdft/codelets/hfb.c -4 +4 cannam@95: M ./rdft/codelets/r2hc.c -4 +5 cannam@95: M ./rdft/codelets/r2hc/Makefile.am -1 +11 cannam@95: A ./rdft/ct-hc2c-direct.c cannam@95: A ./rdft/ct-hc2c.c cannam@95: A ./rdft/ct-hc2c.h cannam@95: M ./rdft/direct.c -16 +21 cannam@95: M ./rdft/direct2.c -15 +32 cannam@95: M ./rdft/hc2hc-common.c -2 +2 cannam@95: M ./rdft/hc2hc-direct.c -4 +4 cannam@95: M ./rdft/hc2hc-directbuf.c -8 +8 cannam@95: M ./rdft/hc2hc-generic.c -4 +3 cannam@95: M ./rdft/hc2hc.c -4 +4 cannam@95: A ./rdft/khc2c.c cannam@95: M ./rdft/nop2.c -7 +10 cannam@95: M ./rdft/problem.c -2 +1 cannam@95: M ./rdft/problem2.c -38 +117 cannam@95: M ./rdft/rank-geq2-rdft2.c -13 +17 cannam@95: M ./rdft/rank0-rdft2.c -35 +44 cannam@95: M ./rdft/rdft.h -8 +27 cannam@95: M ./rdft/rdft2-inplace-strides.c -2 +6 cannam@95: M ./rdft/rdft2-radix2.c -131 +37 cannam@95: M ./rdft/solve2.c -1 +3 cannam@95: M ./rdft/vrank-geq1-rdft2.c -7 +7 cannam@95: M ./support/Makefile.codelets +1 cannam@95: M ./tests/hook.c -8 +23 cannam@95: M ./threads/vrank-geq1-rdft2.c -10 +10 cannam@95: cannam@95: Mon Dec 18 01:23:45 EST 2006 stevenj@fftw.org cannam@95: * separate TRANSPOSED/SCRAMBLED flags internally (this is required so that dft-rank1-bigvec and the future dft-rank1 won't have incompatible SCRAMBLED formats) cannam@95: cannam@95: M ./mpi/api.c -16 +2 cannam@95: M ./mpi/dft-rank-geq2.c -7 +7 cannam@95: M ./mpi/dft-rank1-bigvec.c -8 +9 cannam@95: M ./mpi/dft-serial.c -1 +1 cannam@95: M ./mpi/ifftw-mpi.h -1 +8 cannam@95: M ./mpi/mpi-dft.h -1 +2 cannam@95: M ./mpi/mpi-transpose.h -2 +2 cannam@95: M ./mpi/transpose-alltoall.c -3 +4 cannam@95: M ./mpi/transpose-inplace.c -6 +7 cannam@95: M ./mpi/transpose-radix2.c -5 +6 cannam@95: cannam@95: Mon Dec 18 01:02:27 EST 2006 stevenj@fftw.org cannam@95: * ops_add -> ops_add2 where possible, to shrink code cannam@95: cannam@95: M ./mpi/dft-rank-geq2.c -8 +4 cannam@95: M ./mpi/dft-rank1-bigvec.c -4 +2 cannam@95: M ./mpi/transpose-alltoall.c -6 +3 cannam@95: M ./mpi/transpose-inplace.c -8 +4 cannam@95: M ./mpi/transpose-radix2.c -10 +5 cannam@95: M ./rdft/vrank3-transpose.c -9 +5 cannam@95: cannam@95: Mon Dec 18 00:43:02 EST 2006 stevenj@fftw.org cannam@95: * added dft-rank1-bigvec solver (easy case for 1d parallel transforms) cannam@95: cannam@95: M ./mpi/Makefile.am -1 +1 cannam@95: M ./mpi/api.c -2 +14 cannam@95: M ./mpi/conf.c +1 cannam@95: A ./mpi/dft-rank1-bigvec.c cannam@95: M ./mpi/mpi-dft.h +1 cannam@95: cannam@95: Sun Dec 17 20:42:21 EST 2006 stevenj@fftw.org cannam@95: * rewrote MPI stuff to use dtensor data structure cannam@95: cannam@95: A dtensor is an ordered tuple of triplets (n, ib, ob) giving the size of cannam@95: a dimension (n) and its input and output block sizes of a distributed cannam@95: row-major multi-dimensional array. An MPI DFT (etc.) is now specified cannam@95: in terms of dtensors, which provide a much more flexible data layout. cannam@95: cannam@95: For example, we can now describe multidimensional block distributions, cannam@95: which are important if the number of processors is greater than the cannam@95: size of any given dimension. Currently, we only have solvers for cannam@95: 1d slab distributions, and this is all that is supported in the basic cannam@95: and advanced APIs. The guru API allows one to specify more general cannam@95: distributions, however, which will be useful when/if we have solvers cannam@95: for this case. cannam@95: cannam@95: We now also don't need a TRANSPOSED flag, at least internally, since cannam@95: TRANSPOSED multi-dimensional DFT plans just correspond to dtensors cannam@95: where the input and output block distributions are different. cannam@95: cannam@95: Other changes include the use of the XM(foo) macro for X(mpi_foo). cannam@95: cannam@95: cannam@95: M ./mpi/Makefile.am -1 +1 cannam@95: M ./mpi/any-true.c -1 +1 cannam@95: M ./mpi/api.c -129 +335 cannam@95: M ./mpi/block.c -32 +95 cannam@95: M ./mpi/conf.c -6 +6 cannam@95: M ./mpi/dft-problem.c -46 +33 cannam@95: M ./mpi/dft-rank-geq2.c -60 +72 cannam@95: M ./mpi/dft-serial.c -26 +20 cannam@95: M ./mpi/dft-solve.c -1 +1 cannam@95: A ./mpi/dtensor.c cannam@95: M ./mpi/fftw3-mpi.h -24 +38 cannam@95: M ./mpi/ifftw-mpi.h -22 +60 cannam@95: M ./mpi/mpi-bench.c -115 +289 cannam@95: M ./mpi/mpi-dft.h -12 +13 cannam@95: M ./mpi/mpi-transpose.h -10 +11 cannam@95: M ./mpi/transpose-alltoall.c -17 +16 cannam@95: M ./mpi/transpose-inplace.c -24 +21 cannam@95: M ./mpi/transpose-problem.c -19 +18 cannam@95: M ./mpi/transpose-radix2.c -13 +11 cannam@95: M ./mpi/transpose-solve.c -1 +1 cannam@95: M ./tests/bench.c +4 cannam@95: M ./tests/fftw-bench.c +2 cannam@95: M ./tests/fftw-bench.h +1 cannam@95: cannam@95: Fri Dec 15 16:01:23 EST 2006 athena@fftw.org cannam@95: * Distinguished mutexes from semaphores. cannam@95: The distinction is useful because the linux implementation of cannam@95: sem_post() in unnecessarily slow when semaphores are used for mutual cannam@95: exclusion. This change made spinlocks messier to implement, so I cannam@95: excised them. cannam@95: cannam@95: M ./tests/fftw-bench.c -8 cannam@95: M ./threads/threads.c -105 +40 cannam@95: cannam@95: Fri Dec 15 12:46:11 EST 2006 athena@fftw.org cannam@95: * Use posix semaphores where available. cannam@95: Paranoid declaration of all shared variables as ``volatile''. Paranoid cannam@95: initialization of all shared variables within locks. cannam@95: cannam@95: M ./threads/threads.c -36 +68 cannam@95: cannam@95: Wed Dec 13 20:12:13 EST 2006 Matteo Frigo cannam@95: * paranoia cannam@95: cannam@95: M ./threads/threads.c -2 +2 cannam@95: cannam@95: Tue Dec 12 17:28:13 EST 2006 stevenj@fftw.org cannam@95: * punt on detecting unsolvable rdft2 problems; make r==iio rdft2 problems unsolvable, since it doesn't look like we've consistently checked for this case and it's not clear why we would want to support it (it was also not documented in the manual) cannam@95: cannam@95: M ./doc/fftw3.texi -1 +2 cannam@95: M ./rdft/problem2.c -30 +2 cannam@95: cannam@95: Sat Dec 9 12:14:13 EST 2006 athena@fftw.org cannam@95: * Obey stupid const rules cannam@95: cannam@95: M ./api/api.h -2 +2 cannam@95: M ./api/apiplan.c -1 +1 cannam@95: M ./dft/dft.h -3 +3 cannam@95: M ./dft/indirect.c -4 +4 cannam@95: M ./dft/problem.c -9 +8 cannam@95: M ./kernel/ifftw.h -5 +5 cannam@95: M ./kernel/planner.c -2 +2 cannam@95: M ./kernel/problem.c -4 +4 cannam@95: M ./rdft/buffered2.c -1 +1 cannam@95: M ./rdft/dht-rader.c -1 +1 cannam@95: M ./rdft/indirect.c -4 +4 cannam@95: M ./rdft/problem.c -14 +13 cannam@95: M ./rdft/problem2.c -9 +8 cannam@95: M ./rdft/rank-geq2-rdft2.c -1 +1 cannam@95: M ./rdft/rdft-dht.c -1 +1 cannam@95: M ./rdft/rdft.h -13 +13 cannam@95: M ./rdft/rdft2-radix2.c -5 +5 cannam@95: cannam@95: Fri Dec 8 18:21:50 EST 2006 stevenj@fftw.org cannam@95: * added unsolvable check for rdft2 problem cannam@95: cannam@95: An in-place rdft2 problem is ill-formed if the real data, including the cannam@95: extra "padding" elements, do not coincide with the complex data. cannam@95: cannam@95: CHANGE: the new code considers all in-place split r2c and c2r problems cannam@95: to be ill-formed. Previously, these could be done, but only if the cannam@95: entire multi-dimensional array fit into the buffer, which is kind of cannam@95: stupid. I'm not sure it's worth it to even try to support the cannam@95: split in-place r2c case. cannam@95: cannam@95: M ./rdft/problem2.c -1 +33 cannam@95: cannam@95: Fri Dec 8 13:47:53 EST 2006 stevenj@fftw.org cannam@95: * check in-placeness after joining taints cannam@95: cannam@95: M ./dft/problem.c -9 +9 cannam@95: cannam@95: Fri Dec 8 13:43:44 EST 2006 athena@fftw.org cannam@95: * Grrr... paranoid-check was not testing in exhaustive mode cannam@95: cannam@95: M ./tests/Makefile.am -5 +5 cannam@95: cannam@95: Fri Dec 8 10:00:30 EST 2006 Matteo Frigo cannam@95: * Implemented PROBLEM_UNSOLVABLE. cannam@95: In-place DFT and RDFT problems with inconsistent I/O strides are cannam@95: now unsolvable, and we don't check for them any longer in solvers. cannam@95: cannam@95: While I was at it, declared all problem pointers to be ``const'' cannam@95: for extra safety. cannam@95: cannam@95: M ./api/api.h -2 +2 cannam@95: M ./api/apiplan.c -3 +5 cannam@95: M ./dft/buffered.c -4 +1 cannam@95: M ./dft/dft.h -4 +4 cannam@95: M ./dft/direct.c -8 +6 cannam@95: M ./dft/indirect.c -4 +4 cannam@95: M ./dft/problem.c -11 +15 cannam@95: M ./dft/vrank-geq1.c -3 cannam@95: M ./kernel/ifftw.h -6 +10 cannam@95: M ./kernel/planner.c -6 +6 cannam@95: M ./kernel/problem.c -1 +40 cannam@95: M ./rdft/buffered.c -4 +1 cannam@95: M ./rdft/buffered2.c -1 +1 cannam@95: M ./rdft/dht-rader.c -1 +1 cannam@95: M ./rdft/direct.c -4 +3 cannam@95: M ./rdft/indirect.c -4 +4 cannam@95: M ./rdft/problem.c -14 +17 cannam@95: M ./rdft/problem2.c -10 +10 cannam@95: M ./rdft/rank-geq2-rdft2.c -1 +1 cannam@95: M ./rdft/rdft-dht.c -1 +1 cannam@95: M ./rdft/rdft.h -13 +13 cannam@95: M ./rdft/rdft2-radix2.c -5 +5 cannam@95: M ./rdft/vrank-geq1.c -2 cannam@95: cannam@95: Thu Dec 7 20:13:46 EST 2006 Matteo Frigo cannam@95: * Avoid qsort'ing one element. cannam@95: cannam@95: M ./kernel/tensor7.c -6 +2 cannam@95: cannam@95: Thu Dec 7 18:25:47 EST 2006 Matteo Frigo cannam@95: * In-place vrank>=1 is now applicable only if the problem is really in-place. cannam@95: cannam@95: M ./dft/buffered.c -7 +12 cannam@95: M ./dft/direct.c -11 +11 cannam@95: M ./dft/vrank-geq1.c +3 cannam@95: M ./kernel/ifftw.h +2 cannam@95: M ./kernel/tensor7.c +41 cannam@95: M ./rdft/buffered.c -7 +12 cannam@95: M ./rdft/direct.c -7 +5 cannam@95: M ./rdft/vrank-geq1.c +2 cannam@95: cannam@95: Tue Dec 5 12:21:38 EST 2006 athena@fftw.org cannam@95: * unused variable cannam@95: cannam@95: M ./kernel/align.c +1 cannam@95: cannam@95: Sun Dec 3 19:16:33 EST 2006 Matteo Frigo cannam@95: * Removed CVS $Id$ everywhere, since darcs does not update them. cannam@95: cannam@95: M ./api/fftw3.h -1 cannam@95: M ./api/version.c -1 cannam@95: M ./commercialize.sh +1 cannam@95: M ./dft/buffered.c -1 cannam@95: M ./dft/codelet-dft.h -1 cannam@95: M ./dft/codelets/standard/Makefile.am +1 cannam@95: M ./dft/conf.c -1 cannam@95: M ./dft/ct.c -1 cannam@95: M ./dft/dft.h -1 cannam@95: M ./dft/dftw-direct.c -1 cannam@95: M ./dft/dftw-directsq.c -1 cannam@95: M ./dft/direct.c -1 cannam@95: M ./dft/indirect.c -1 cannam@95: M ./dft/kdft-dif.c -1 cannam@95: M ./dft/kdft-difsq.c -1 cannam@95: M ./dft/kdft-dit.c -1 cannam@95: M ./dft/kdft.c -1 cannam@95: M ./dft/nop.c -1 cannam@95: M ./dft/plan.c -1 cannam@95: M ./dft/problem.c -1 cannam@95: M ./dft/rank-geq2.c -1 cannam@95: M ./dft/solve.c -1 cannam@95: M ./dft/vrank-geq1.c -1 cannam@95: M ./dft/zero.c -1 cannam@95: M ./doc/fftw3.texi -1 cannam@95: M ./genfft/algsimp.ml -2 cannam@95: M ./genfft/algsimp.mli -2 cannam@95: M ./genfft/annotate.ml -2 cannam@95: M ./genfft/annotate.mli -1 cannam@95: M ./genfft/assoctable.ml -1 cannam@95: M ./genfft/assoctable.mli -1 cannam@95: M ./genfft/c.ml -1 cannam@95: M ./genfft/c.mli -1 cannam@95: M ./genfft/complex.ml -1 cannam@95: M ./genfft/complex.mli -1 cannam@95: M ./genfft/conv.mli -1 cannam@95: M ./genfft/dag.ml -1 cannam@95: M ./genfft/dag.mli -1 cannam@95: M ./genfft/expr.ml -1 cannam@95: M ./genfft/expr.mli -1 cannam@95: M ./genfft/fft.ml -2 cannam@95: M ./genfft/fft.mli -2 cannam@95: M ./genfft/gen_conv.ml -3 +1 cannam@95: M ./genfft/gen_hc2hc.ml -3 +1 cannam@95: M ./genfft/gen_hc2r.ml -3 +1 cannam@95: M ./genfft/gen_mdct.ml -3 +1 cannam@95: M ./genfft/gen_notw.ml -3 +1 cannam@95: M ./genfft/gen_notw_c.ml -3 +1 cannam@95: M ./genfft/gen_r2hc.ml -3 +1 cannam@95: M ./genfft/gen_r2r.ml -3 +1 cannam@95: M ./genfft/gen_twiddle.ml -3 +1 cannam@95: M ./genfft/gen_twiddle_c.ml -3 +1 cannam@95: M ./genfft/gen_twidsq.ml -3 +1 cannam@95: M ./genfft/gen_twidsq_c.ml -3 +1 cannam@95: M ./genfft/genutil.ml -8 +1 cannam@95: M ./genfft/littlesimp.ml -1 cannam@95: M ./genfft/littlesimp.mli -1 cannam@95: M ./genfft/magic.ml -1 cannam@95: M ./genfft/monads.ml -1 cannam@95: M ./genfft/number.ml -1 cannam@95: M ./genfft/number.mli -1 cannam@95: M ./genfft/oracle.ml -1 cannam@95: M ./genfft/oracle.mli -1 cannam@95: M ./genfft/schedule.ml -1 cannam@95: M ./genfft/schedule.mli -1 cannam@95: M ./genfft/simd.ml -1 cannam@95: M ./genfft/simd.mli -1 cannam@95: M ./genfft/simdmagic.ml -1 cannam@95: M ./genfft/to_alist.ml -1 cannam@95: M ./genfft/to_alist.mli -1 cannam@95: M ./genfft/trig.ml -1 cannam@95: M ./genfft/trig.mli -1 cannam@95: M ./genfft/twiddle.ml -1 cannam@95: M ./genfft/twiddle.mli -1 cannam@95: M ./genfft/unique.ml -1 cannam@95: M ./genfft/unique.mli -1 cannam@95: M ./genfft/util.ml -1 cannam@95: M ./genfft/util.mli -1 cannam@95: M ./genfft/variable.ml -1 cannam@95: M ./genfft/variable.mli -1 cannam@95: M ./kernel/align.c -1 cannam@95: M ./kernel/alloc.c -1 cannam@95: M ./kernel/assert.c -1 cannam@95: M ./kernel/awake.c -1 cannam@95: M ./kernel/cycle.h -1 cannam@95: M ./kernel/debug.c -1 cannam@95: M ./kernel/iabs.c -1 cannam@95: M ./kernel/ifftw.h -1 cannam@95: M ./kernel/kalloc.c -1 cannam@95: M ./kernel/minmax.c -1 cannam@95: M ./kernel/ops.c -1 cannam@95: M ./kernel/pickdim.c -1 cannam@95: M ./kernel/plan.c -1 cannam@95: M ./kernel/planner.c -1 cannam@95: M ./kernel/primes.c -1 cannam@95: M ./kernel/print.c -1 cannam@95: M ./kernel/problem.c -1 cannam@95: M ./kernel/scan.c -1 cannam@95: M ./kernel/solver.c -1 cannam@95: M ./kernel/solvtab.c -1 cannam@95: M ./kernel/stride.c -1 cannam@95: M ./kernel/tensor.c -1 cannam@95: M ./kernel/tensor1.c -1 cannam@95: M ./kernel/tensor2.c -1 cannam@95: M ./kernel/tensor4.c -1 cannam@95: M ./kernel/tensor5.c -1 cannam@95: M ./kernel/tensor7.c -1 cannam@95: M ./kernel/tensor8.c -1 cannam@95: M ./kernel/tensor9.c -1 cannam@95: M ./kernel/timer.c -1 cannam@95: M ./kernel/trig.c -1 cannam@95: M ./kernel/twiddle.c -1 cannam@95: M ./libbench/accopy-from.c -1 cannam@95: M ./libbench/accopy-to.c -1 cannam@95: M ./libbench/allocate.c -1 cannam@95: M ./libbench/bench-main.c -1 cannam@95: M ./libbench/bench-user.h -1 cannam@95: M ./libbench/bench.h -1 cannam@95: M ./libbench/can-do.c -1 cannam@95: M ./libbench/ccopy-from.c -1 cannam@95: M ./libbench/ccopy-to.c -1 cannam@95: M ./libbench/deallocate.c -1 cannam@95: M ./libbench/getopt-utils.c -1 cannam@95: M ./libbench/info.c -1 cannam@95: M ./libbench/main.c -1 cannam@95: M ./libbench/prime.c -1 cannam@95: M ./libbench/problem.c -1 cannam@95: M ./libbench/report.c -1 cannam@95: M ./libbench/speed.c -1 cannam@95: M ./libbench/timer.c -1 cannam@95: M ./libbench/verify.c -1 cannam@95: M ./libbench/zero.c -1 cannam@95: M ./libbench2/aligned-main.c -1 cannam@95: M ./libbench2/allocate.c -1 cannam@95: M ./libbench2/bench-main.c -1 cannam@95: M ./libbench2/bench-user.h -1 cannam@95: M ./libbench2/bench.h -1 cannam@95: M ./libbench2/can-do.c -1 cannam@95: M ./libbench2/dotens2.c -1 cannam@95: M ./libbench2/info.c -1 cannam@95: M ./libbench2/main.c -1 cannam@95: M ./libbench2/problem.c -1 cannam@95: M ./libbench2/report.c -1 cannam@95: M ./libbench2/speed.c -1 cannam@95: M ./libbench2/tensor.c -1 cannam@95: M ./libbench2/timer.c -1 cannam@95: M ./libbench2/useropt.c -1 cannam@95: M ./libbench2/verify-dft.c -1 cannam@95: M ./libbench2/verify-lib.c -1 cannam@95: M ./libbench2/verify-rdft2.c -1 cannam@95: M ./libbench2/verify.c -1 cannam@95: M ./libbench2/zero.c -1 cannam@95: M ./mpi/conf.c -1 cannam@95: M ./rdft/buffered.c -1 cannam@95: M ./rdft/buffered2.c -1 cannam@95: M ./rdft/codelets/hc2r/Makefile.am +1 cannam@95: M ./rdft/codelets/r2hc/Makefile.am +1 cannam@95: M ./rdft/conf.c -1 cannam@95: M ./rdft/dft-r2hc.c -1 cannam@95: M ./rdft/dht-r2hc.c -1 cannam@95: M ./rdft/direct.c -1 cannam@95: M ./rdft/direct2.c -1 cannam@95: M ./rdft/hc2hc-direct.c -1 cannam@95: M ./rdft/hc2hc-directbuf.c -1 cannam@95: M ./rdft/indirect.c -1 cannam@95: M ./rdft/khc2hc.c -1 cannam@95: M ./rdft/khc2r.c -1 cannam@95: M ./rdft/kr2hc.c -1 cannam@95: M ./rdft/kr2r.c -1 cannam@95: M ./rdft/nop.c -1 cannam@95: M ./rdft/nop2.c -1 cannam@95: M ./rdft/plan.c -1 cannam@95: M ./rdft/plan2.c -1 cannam@95: M ./rdft/problem.c -1 cannam@95: M ./rdft/problem2.c -1 cannam@95: M ./rdft/rank-geq2-rdft2.c -1 cannam@95: M ./rdft/rank-geq2.c -1 cannam@95: M ./rdft/rank0-rdft2.c -1 cannam@95: M ./rdft/rank0.c -1 cannam@95: M ./rdft/rdft-dht.c -1 cannam@95: M ./rdft/rdft2-inplace-strides.c -1 cannam@95: M ./rdft/rdft2-radix2.c -1 cannam@95: M ./rdft/rdft2-tensor-max-index.c -1 cannam@95: M ./rdft/solve.c -1 cannam@95: M ./rdft/solve2.c -1 cannam@95: M ./rdft/vrank-geq1-rdft2.c -1 cannam@95: M ./rdft/vrank-geq1.c -1 cannam@95: M ./rdft/vrank3-transpose.c -1 cannam@95: M ./reodft/conf.c -1 cannam@95: M ./reodft/redft00e-r2hc-pad.c -1 cannam@95: M ./reodft/redft00e-r2hc.c -1 cannam@95: M ./reodft/reodft00e-splitradix.c -1 cannam@95: M ./reodft/reodft010e-r2hc.c -1 cannam@95: M ./reodft/reodft11e-r2hc-odd.c -1 cannam@95: M ./reodft/reodft11e-r2hc.c -1 cannam@95: M ./reodft/reodft11e-radix2.c -1 cannam@95: M ./reodft/rodft00e-r2hc-pad.c -1 cannam@95: M ./reodft/rodft00e-r2hc.c -1 cannam@95: M ./simd/altivec.c -1 cannam@95: M ./simd/nonportable/sse.c -1 cannam@95: M ./simd/nonportable/sse2.c -1 cannam@95: M ./simd/sse.c -1 cannam@95: M ./simd/sse2.c -1 cannam@95: M ./simd/taint.c -1 cannam@95: M ./simd/x86-cpuid.h -1 cannam@95: M ./threads/conf.c -1 cannam@95: M ./threads/ct.c -1 cannam@95: M ./threads/dft-vrank-geq1.c -1 cannam@95: M ./threads/rdft-vrank-geq1.c -1 cannam@95: M ./threads/vrank-geq1-rdft2.c -1 cannam@95: cannam@95: Sun Dec 3 16:11:17 EST 2006 Matteo Frigo cannam@95: * generalized dftw to encompass q codelets. As a side effect, q codelets are now threaded. cannam@95: cannam@95: R ./dft/ctsq.c cannam@95: A ./dft/dftw-directsq.c cannam@95: M ./dft/Makefile.am -2 +2 cannam@95: M ./dft/ct.c -25 +55 cannam@95: M ./dft/ct.h -6 +9 cannam@95: M ./dft/dftw-direct.c -59 +72 cannam@95: M ./dft/dftw-directsq.c +163 cannam@95: M ./dft/dftw-generic.c -24 +32 cannam@95: M ./dft/dftw-genericbuf.c -30 +37 cannam@95: M ./dft/indirect-transpose.c -1 +1 cannam@95: M ./dft/kdft-difsq.c -1 +1 cannam@95: M ./libbench2/my-getopt.c -3 +3 cannam@95: M ./threads/ct.c -28 +53 cannam@95: cannam@95: Sat Nov 25 16:34:38 EST 2006 stevenj@fftw.org cannam@95: * add missing __declspec attribute to threads API functions when compiling for Windows (thanks to Robert O. Morris for the bug report) cannam@95: cannam@95: M ./api/f77api.c -29 +1 cannam@95: M ./api/x77.h +28 cannam@95: M ./threads/f77api.c +13 cannam@95: M ./threads/f77funcs.h -3 +3 cannam@95: cannam@95: Mon Nov 20 17:39:20 EST 2006 stevenj@fftw.org cannam@95: * add AC_SUBST to AX_OPENMP, thanks to Sebastien Maret for the suggestion cannam@95: cannam@95: M ./m4/ax_openmp.m4 -1 +2 cannam@95: cannam@95: Mon Oct 16 23:02:29 EDT 2006 stevenj@fftw.org cannam@95: * not gcc bug for MIPS (thanks to Jonathan Day) cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn +4 cannam@95: cannam@95: Sat Sep 23 17:52:36 EDT 2006 stevenj@fftw.org cannam@95: * in maintainer/debug mode, don't modify CFLAGS if they were explicitly set (-pedantic seems to cause problems with LAM's mpicc, so I need a way to override) cannam@95: cannam@95: M ./configure.ac +2 cannam@95: cannam@95: Wed Nov 22 18:14:47 EST 2006 athena@fftw.org cannam@95: * Removed obsolete comment. cannam@95: cannam@95: M ./dft/dftw-generic.c -1 cannam@95: cannam@95: Sun Nov 19 11:21:44 EST 2006 athena@fftw.org cannam@95: * Use p->v when comparing TW_FULL fields. cannam@95: cannam@95: M ./kernel/twiddle.c -2 +1 cannam@95: cannam@95: Thu Nov 16 14:49:05 EST 2006 athena@fftw.org cannam@95: * removed useless definition cannam@95: cannam@95: M ./simd/simd.h -1 cannam@95: cannam@95: Mon Nov 13 09:18:32 EST 2006 athena@fftw.org cannam@95: * paranoid avoidance of integer overflows cannam@95: cannam@95: M ./libbench2/timer.c -3 +3 cannam@95: cannam@95: Mon Nov 13 09:00:11 EST 2006 athena@fftw.org cannam@95: * avoid potential overflows in cycle counters cannam@95: At the suggestion of Alex Cichowski, convert all ticks cannam@95: to double before operating on them, to avoid potential cannam@95: signed/unsigned confusion and integer overflow. cannam@95: cannam@95: M ./kernel/cycle.h -5 +8 cannam@95: cannam@95: Sun Nov 5 09:00:52 EST 2006 Matteo Frigo cannam@95: * Removed unused struct field cannam@95: cannam@95: M ./threads/threads.c -1 cannam@95: cannam@95: Sat Nov 4 09:43:13 EST 2006 Matteo Frigo cannam@95: * use pthread condition variables instead of semaphores cannam@95: Condition variables are more likely to be portable everywhere, and cannam@95: somehow they appear to introduce less overhead at least on my cannam@95: linux box. cannam@95: cannam@95: M ./threads/threads.c -6 +19 cannam@95: cannam@95: Tue Oct 31 20:45:24 EST 2006 Matteo Frigo cannam@95: * Imprecise help message. cannam@95: cannam@95: M ./libbench2/my-getopt.c -1 +1 cannam@95: cannam@95: Mon Oct 30 20:13:35 EST 2006 athena@fftw.org cannam@95: * Experimental implementation of spinlocks. cannam@95: This patch implements spinlocks via a semi-portable hack, and adds cannam@95: the -ospinlocks option to the bench program so that we can play with them. cannam@95: cannam@95: M ./tests/fftw-bench.c +8 cannam@95: M ./threads/threads.c -28 +104 cannam@95: cannam@95: Thu Oct 26 22:29:18 EDT 2006 athena@fftw.org cannam@95: * Updated manual for new openmp configure options. cannam@95: cannam@95: M ./doc/fftw3.texi -20 +17 cannam@95: cannam@95: Thu Oct 26 21:52:39 EDT 2006 athena@fftw.org cannam@95: * Added back openmp. cannam@95: Rationalized threads naming conventions: cannam@95: cannam@95: * threads explicitly managed by us are enabled by --enable-threads, cannam@95: predicated on HAVE_THREADS, etc. cannam@95: * openmp is enabled by --enable-openmp, predicated on HAVE_OPENMP, etc. cannam@95: * SMP denotes either THREADS or OPENMP. cannam@95: cannam@95: M ./Makefile.am -1 +1 cannam@95: M ./configure.ac -10 +27 cannam@95: M ./kernel/alloc.c -1 +1 cannam@95: M ./kernel/ifftw.h -1 +1 cannam@95: M ./mpi/Makefile.am -1 +1 cannam@95: M ./tests/Makefile.am -5 +5 cannam@95: M ./tests/fftw-bench.c -3 +7 cannam@95: M ./threads/Makefile.am -2 +2 cannam@95: M ./threads/api.c -1 +16 cannam@95: M ./threads/conf.c -2 +2 cannam@95: A ./threads/openmp.c cannam@95: M ./threads/threads.c -13 +19 cannam@95: M ./threads/threads.h -2 +2 cannam@95: M ./tools/Makefile.am -1 +1 cannam@95: M ./tools/fftw-wisdom.c -5 +9 cannam@95: cannam@95: Mon Oct 23 20:14:31 EDT 2006 athena@fftw.org cannam@95: * different thread protocols cannam@95: cannam@95: M ./threads/threads.c -80 +71 cannam@95: cannam@95: Sun Oct 22 14:49:32 EDT 2006 athena@fftw.org cannam@95: * fix memory leak cannam@95: Added pthread_attr_destroy to avoid memory leak. cannam@95: cannam@95: M ./threads/threads.c +1 cannam@95: cannam@95: Sun Oct 22 14:23:30 EDT 2006 athena@fftw.org cannam@95: * Experimental new pthread implementation that recycles threads. cannam@95: cannam@95: M ./threads/threads.c -528 +155 cannam@95: cannam@95: Tue Oct 24 23:28:10 EDT 2006 athena@fftw.org cannam@95: * switched buddies cannam@95: Switched order of buddies in rdft2 rank-geq2 for consistency cannam@95: with analogous dft and rdft solvers. Furthermore, this change reduces cannam@95: the MEASURE planning time for rank == 3. cannam@95: cannam@95: M ./rdft/rank-geq2-rdft2.c -4 +1 cannam@95: cannam@95: Sat Oct 21 10:10:00 EDT 2006 Matteo Frigo cannam@95: * Typo. cannam@95: cannam@95: M ./tools/fftw_wisdom.1.in -1 +1 cannam@95: cannam@95: Sat Oct 21 09:56:50 EDT 2006 Matteo Frigo cannam@95: * Out of place is the default. Thanks to Kirk Kern for pointing this out. cannam@95: cannam@95: M ./tests/README -2 +2 cannam@95: cannam@95: Fri Sep 29 01:36:11 EDT 2006 stevenj@fftw.org cannam@95: * rename "test" to "tst", since a user (Igor Levicki) reports that "test" is a reserved words in some x86 assemblers cannam@95: cannam@95: M ./simd/x86-cpuid.h -10 +10 cannam@95: cannam@95: Tue Sep 26 09:01:08 EDT 2006 Matteo Frigo cannam@95: * Stylistic change. cannam@95: cannam@95: M ./kernel/timer.c -2 +1 cannam@95: cannam@95: Tue Sep 26 08:45:37 EDT 2006 athena@fftw.org cannam@95: * Do not set tmin=1e10, since a large FFT may take longer than that. cannam@95: cannam@95: M ./kernel/timer.c -5 +4 cannam@95: cannam@95: Sat Sep 23 22:07:10 EDT 2006 athena@fftw.org cannam@95: * Disable certain gcc optimizations. cannam@95: When PRECOMPUTE_ARRAY_INDICES is #define'd, array indices have the cannam@95: form array[stride[k]] for compile-time constant k. Apparently new cannam@95: gcc's copy stride[k] onto the stack before the codelet loop, which is cannam@95: an idiotic optimization if ever there was one. This patch confuses cannam@95: gcc enough to prevent this optimization. cannam@95: cannam@95: M ./kernel/ifftw.h -3 +6 cannam@95: M ./kernel/stride.c -1 +1 cannam@95: cannam@95: Sat Sep 23 13:02:58 EDT 2006 stevenj@fftw.org cannam@95: * re-enable TOMS algorithm - it is the best for large vector lengths, since for such sizes the cache line is not an issue and the bookkeeping overhead is negligible cannam@95: cannam@95: M ./rdft/vrank3-transpose.c -18 +15 cannam@95: cannam@95: Thu Sep 21 15:40:15 EDT 2006 stevenj@fftw.org cannam@95: * add measure_hook so that MPI can synchronize timing measurements (otherwise different processors might end up with different MPI plans, yikes!) cannam@95: cannam@95: M ./kernel/ifftw.h +2 cannam@95: M ./kernel/planner.c +5 cannam@95: M ./mpi/api.c -2 +26 cannam@95: cannam@95: Tue Sep 19 21:26:19 EDT 2006 stevenj@fftw.org cannam@95: * added O(p log p) transpose algorithm (radix 2) cannam@95: cannam@95: M ./mpi/Makefile.am -1 +1 cannam@95: M ./mpi/conf.c +1 cannam@95: M ./mpi/mpi-transpose.h +1 cannam@95: A ./mpi/transpose-radix2.c cannam@95: cannam@95: Tue Sep 19 21:05:09 EDT 2006 stevenj@fftw.org cannam@95: * comments cannam@95: cannam@95: M ./mpi/testsched.c -8 +10 cannam@95: cannam@95: Tue Sep 19 19:54:58 EDT 2006 stevenj@fftw.org cannam@95: * whoops cannam@95: cannam@95: ./mpi/any_true.c -> ./mpi/any-true.c cannam@95: M ./mpi/Makefile.am -1 +1 cannam@95: cannam@95: Tue Sep 19 19:49:01 EDT 2006 stevenj@fftw.org cannam@95: * synchronize planning so that if one process fails to create a plan then all of them do. cannam@95: cannam@95: M ./mpi/Makefile.am -1 +1 cannam@95: A ./mpi/any_true.c cannam@95: M ./mpi/dft-rank-geq2.c -4 +4 cannam@95: M ./mpi/dft-serial.c -1 +1 cannam@95: M ./mpi/ifftw-mpi.h +3 cannam@95: M ./mpi/transpose-alltoall.c -5 +5 cannam@95: M ./mpi/transpose-inplace.c -5 +5 cannam@95: cannam@95: Tue Sep 19 18:17:38 EDT 2006 stevenj@fftw.org cannam@95: * call MPI_Alltoall instead of MPI_Alltoallv for equal-blocks case, in case MPI implementation has special optimizations for the common case of equal sizes cannam@95: cannam@95: M ./mpi/transpose-alltoall.c -6 +19 cannam@95: cannam@95: Tue Sep 19 12:07:35 EDT 2006 stevenj@fftw.org cannam@95: * whoops cannam@95: cannam@95: M ./tools/Makefile.am -1 +1 cannam@95: cannam@95: Tue Sep 19 02:20:06 EDT 2006 stevenj@fftw.org cannam@95: * typo in comment cannam@95: cannam@95: M ./mpi/testsched.c -3 +3 cannam@95: cannam@95: Tue Sep 19 01:58:55 EDT 2006 stevenj@fftw.org cannam@95: * more filename simplifications cannam@95: cannam@95: ./mpi/mpi-dft-rank-geq2.c -> ./mpi/dft-rank-geq2.c cannam@95: ./mpi/mpi-dft-serial.c -> ./mpi/dft-serial.c cannam@95: ./mpi/mpi-transpose-alltoall.c -> ./mpi/transpose-alltoall.c cannam@95: ./mpi/mpi-transpose-inplace.c -> ./mpi/transpose-inplace.c cannam@95: ./mpi/problem.c -> ./mpi/dft-problem.c cannam@95: ./mpi/solve.c -> ./mpi/dft-solve.c cannam@95: ./mpi/tproblem.c -> ./mpi/transpose-problem.c cannam@95: ./mpi/tsolve.c -> ./mpi/transpose-solve.c cannam@95: M ./mpi/Makefile.am -3 +3 cannam@95: cannam@95: Tue Sep 19 01:49:52 EDT 2006 stevenj@fftw.org cannam@95: * canonicalize file names (hyphens, not underscores) cannam@95: cannam@95: ./mpi/fftw3_mpi.h -> ./mpi/fftw3-mpi.h cannam@95: ./mpi/ifftw_mpi.h -> ./mpi/ifftw-mpi.h cannam@95: ./mpi/mpi_bench.c -> ./mpi/mpi-bench.c cannam@95: ./mpi/mpi_dft.h -> ./mpi/mpi-dft.h cannam@95: ./mpi/mpi_transpose.h -> ./mpi/mpi-transpose.h cannam@95: ./mpi/mpi_transpose_alltoall.c -> ./mpi/mpi-transpose-alltoall.c cannam@95: ./mpi/mpi_transpose_inplace.c -> ./mpi/mpi-transpose-inplace.c cannam@95: ./tests/fftw_bench.h -> ./tests/fftw-bench.h cannam@95: ./tests/fftw_bench_common.c -> ./tests/fftw-bench.c cannam@95: M ./mpi/Makefile.am -5 +5 cannam@95: M ./mpi/api.c -4 +4 cannam@95: M ./mpi/block.c -1 +1 cannam@95: M ./mpi/conf.c -2 +2 cannam@95: M ./mpi/mpi-bench.c -2 +2 cannam@95: M ./mpi/mpi-dft-rank-geq2.c -2 +2 cannam@95: M ./mpi/mpi-dft-serial.c -1 +1 cannam@95: M ./mpi/mpi-dft.h -1 +1 cannam@95: M ./mpi/mpi-transpose-alltoall.c -1 +1 cannam@95: M ./mpi/mpi-transpose-inplace.c -1 +1 cannam@95: M ./mpi/mpi-transpose.h -1 +1 cannam@95: M ./mpi/problem.c -1 +1 cannam@95: M ./mpi/solve.c -1 +1 cannam@95: M ./mpi/tproblem.c -1 +1 cannam@95: M ./mpi/tsolve.c -1 +1 cannam@95: M ./tests/Makefile.am -1 +1 cannam@95: M ./tests/bench.c -2 +2 cannam@95: M ./tests/fftw-bench.c -1 +1 cannam@95: cannam@95: Tue Sep 19 01:34:07 EDT 2006 stevenj@fftw.org cannam@95: * add mpi-dft-serial cannam@95: cannam@95: M ./mpi/Makefile.am -1 +1 cannam@95: M ./mpi/conf.c +1 cannam@95: M ./mpi/mpi-dft-rank-geq2.c -3 +9 cannam@95: A ./mpi/mpi-dft-serial.c cannam@95: M ./mpi/mpi_dft.h +1 cannam@95: M ./mpi/mpi_transpose_alltoall.c -1 +1 cannam@95: M ./mpi/mpi_transpose_inplace.c -1 +1 cannam@95: cannam@95: Tue Sep 19 01:21:47 EDT 2006 stevenj@fftw.org cannam@95: * silence warnings cannam@95: cannam@95: M ./dft/indirect-transpose.c -1 cannam@95: M ./libbench2/problem.c -2 cannam@95: cannam@95: Tue Sep 19 00:31:59 EDT 2006 stevenj@fftw.org cannam@95: * make "t" problem semantics match FFTW_MPI_TRANSPOSED cannam@95: cannam@95: M ./libbench2/problem.c -4 +3 cannam@95: M ./mpi/mpi_bench.c -8 +18 cannam@95: cannam@95: Mon Sep 18 23:50:43 EDT 2006 stevenj@fftw.org cannam@95: * whoops, fixed backwards mpi_dft cannam@95: cannam@95: M ./mpi/mpi-dft-rank-geq2.c -7 +7 cannam@95: cannam@95: Mon Sep 18 22:26:31 EDT 2006 stevenj@fftw.org cannam@95: * initial stab at rank-geq2 mpi-dft; seems to be mostly working cannam@95: cannam@95: M ./libbench2/bench-user.h -1 +1 cannam@95: M ./mpi/Makefile.am -1 +4 cannam@95: M ./mpi/api.c -26 +180 cannam@95: M ./mpi/conf.c +2 cannam@95: M ./mpi/fftw3_mpi.h -11 +51 cannam@95: M ./mpi/ifftw_mpi.h -4 +1 cannam@95: A ./mpi/mpi-dft-rank-geq2.c cannam@95: M ./mpi/mpi_bench.c -55 +166 cannam@95: M ./mpi/mpi_dft.h -2 +2 cannam@95: M ./mpi/mpi_transpose.h -6 +1 cannam@95: M ./mpi/mpi_transpose_alltoall.c -7 +10 cannam@95: M ./mpi/mpi_transpose_inplace.c -7 +8 cannam@95: M ./mpi/problem.c -2 +7 cannam@95: A ./mpi/solve.c cannam@95: M ./mpi/tproblem.c -2 +2 cannam@95: cannam@95: Sun Sep 17 13:41:32 EDT 2006 stevenj@fftw.org cannam@95: * support SCRAMBLED_OUT in alltoall transpose cannam@95: cannam@95: M ./kernel/ifftw.h +4 cannam@95: M ./kernel/tensor2.c +20 cannam@95: M ./mpi/ifftw_mpi.h -1 +16 cannam@95: M ./mpi/mpi_dft.h -1 +1 cannam@95: M ./mpi/mpi_transpose_alltoall.c -23 +48 cannam@95: M ./mpi/mpi_transpose_inplace.c -23 +2 cannam@95: M ./mpi/problem.c -2 +3 cannam@95: cannam@95: Sun Sep 17 12:34:30 EDT 2006 stevenj@fftw.org cannam@95: * skeleton of future support for block-cyclic cannam@95: cannam@95: M ./mpi/block.c -7 +27 cannam@95: M ./mpi/ifftw_mpi.h +1 cannam@95: M ./mpi/mpi_transpose_alltoall.c +2 cannam@95: M ./mpi/mpi_transpose_inplace.c -2 +5 cannam@95: cannam@95: Sun Sep 17 12:11:19 EDT 2006 stevenj@fftw.org cannam@95: * test program now checks scrambled in/out via -obflag=28/29 cannam@95: cannam@95: M ./mpi/api.c -1 +1 cannam@95: M ./mpi/fftw3_mpi.h -1 +1 cannam@95: M ./mpi/mpi_bench.c -13 +59 cannam@95: M ./mpi/problem.c -1 cannam@95: cannam@95: Sun Sep 17 11:58:36 EDT 2006 stevenj@fftw.org cannam@95: * added -obflag to make it easier to set high-order bits cannam@95: cannam@95: M ./tests/fftw_bench_common.c +1 cannam@95: cannam@95: Sun Sep 17 01:30:51 EDT 2006 stevenj@fftw.org cannam@95: * use proper child plans for 2nd transpose in transpose_alltoall; implement opcount in transpose_inplace cannam@95: cannam@95: M ./mpi/mpi_transpose_alltoall.c -47 +47 cannam@95: M ./mpi/mpi_transpose_inplace.c -1 +10 cannam@95: cannam@95: Sun Sep 17 01:08:01 EDT 2006 stevenj@fftw.org cannam@95: * fix in test program for transposes of vectors -- transpose routines seem to completely work now (except for scrambled in/out, which is untested) cannam@95: cannam@95: M ./mpi/mpi_bench.c -1 +1 cannam@95: cannam@95: Sun Sep 17 01:01:16 EDT 2006 stevenj@fftw.org cannam@95: * fixed bug in transpose_alltoall for unequal blocks cannam@95: cannam@95: M ./mpi/mpi_transpose_alltoall.c -2 +2 cannam@95: M ./mpi/mpi_transpose_inplace.c -2 +2 cannam@95: cannam@95: Sat Sep 16 15:29:46 EDT 2006 stevenj@fftw.org cannam@95: * correctly handle cld2rest cannam@95: cannam@95: M ./mpi/mpi_transpose_inplace.c -24 +28 cannam@95: cannam@95: Sat Sep 16 15:29:31 EDT 2006 stevenj@fftw.org cannam@95: * some debugging code and other fixes cannam@95: cannam@95: M ./mpi/mpi_bench.c -13 +34 cannam@95: cannam@95: Sat Sep 16 14:54:30 EDT 2006 stevenj@fftw.org cannam@95: * whoops, forgot to check in mpi_bench.c file cannam@95: cannam@95: A ./mpi/mpi_bench.c cannam@95: cannam@95: Sat Sep 16 14:54:02 EDT 2006 stevenj@fftw.org cannam@95: * added bench_exit routine so that it can be overridden (by MPI_Abort) if needed cannam@95: cannam@95: M ./libbench2/Makefile.am -5 +5 cannam@95: A ./libbench2/bench-exit.c cannam@95: M ./libbench2/bench-user.h +1 cannam@95: M ./libbench2/util.c -1 +1 cannam@95: M ./libbench2/verify-lib.c -1 +1 cannam@95: M ./libbench2/verify-r2r.c -1 +1 cannam@95: cannam@95: Sat Sep 16 14:52:56 EDT 2006 stevenj@fftw.org cannam@95: * bug fix in mpi_transpose_inplace for case where some processors are idle cannam@95: cannam@95: M ./mpi/mpi_transpose_inplace.c -3 +4 cannam@95: cannam@95: Fri Sep 15 18:47:13 EDT 2006 stevenj@fftw.org cannam@95: * allow vecloop for sz->rnk==0 in exceptional (SLOW) cases, e.g. it is necessary for loops of non-square transposes (otherwise e.g. ik1v5:200:200x10:20:1x20:1:10 planning fails) cannam@95: cannam@95: M ./rdft/vrank-geq1.c -2 +6 cannam@95: cannam@95: Thu Sep 14 23:36:48 EDT 2006 stevenj@fftw.org cannam@95: * first pass at working mpi_bench test program; transpose seems to work iff dimensions are divisible by #processors cannam@95: cannam@95: M ./libbench2/Makefile.am -1 +3 cannam@95: A ./libbench2/after-ccopy-from.c cannam@95: A ./libbench2/after-ccopy-to.c cannam@95: A ./libbench2/after-hccopy-from.c cannam@95: A ./libbench2/after-hccopy-to.c cannam@95: A ./libbench2/after-rcopy-from.c cannam@95: A ./libbench2/after-rcopy-to.c cannam@95: M ./libbench2/bench-main.c -4 +6 cannam@95: M ./libbench2/bench-user.h +10 cannam@95: M ./libbench2/verify-dft.c +2 cannam@95: M ./libbench2/verify-r2r.c +4 cannam@95: M ./libbench2/verify-rdft2.c +4 cannam@95: M ./mpi/api.c -2 +6 cannam@95: cannam@95: Wed Sep 13 17:28:07 EDT 2006 stevenj@fftw.org cannam@95: * whoops cannam@95: cannam@95: M ./mpi/mpi_transpose_inplace.c -2 +2 cannam@95: cannam@95: Tue Sep 12 22:27:03 EDT 2006 stevenj@fftw.org cannam@95: * initial stub for mpi_bench cannam@95: cannam@95: M ./Makefile.am -2 +2 cannam@95: M ./libbench2/bench-main.c +3 cannam@95: M ./libbench2/bench-user.h +1 cannam@95: M ./libbench2/verify-lib.c -1 +1 cannam@95: M ./mpi/Makefile.am -1 +14 cannam@95: M ./tests/bench.c +9 cannam@95: M ./tests/fftw_bench.h +1 cannam@95: M ./tests/fftw_bench_common.c +2 cannam@95: cannam@95: Tue Sep 12 21:54:31 EDT 2006 stevenj@fftw.org cannam@95: * do no output at all if verbose < 0 (for use with MPI, where we only want output from process 0) cannam@95: cannam@95: M ./libbench2/bench-main.c -1 +1 cannam@95: M ./libbench2/bench.h +1 cannam@95: M ./libbench2/ovtpvt.c -1 +15 cannam@95: M ./libbench2/useropt.c -1 +1 cannam@95: M ./libbench2/util.c -2 +1 cannam@95: M ./libbench2/verify-lib.c -5 +4 cannam@95: M ./libbench2/verify-r2r.c -4 +4 cannam@95: cannam@95: Tue Sep 12 21:39:15 EDT 2006 stevenj@fftw.org cannam@95: * whoops cannam@95: cannam@95: M ./tests/fftw_bench_common.c +3 cannam@95: cannam@95: Tue Sep 12 21:31:40 EDT 2006 stevenj@fftw.org cannam@95: * split bench.c into bench.c and fftw_bench_common.c so that we can re-use some of the code in the MPI test program cannam@95: cannam@95: M ./tests/Makefile.am -1 +1 cannam@95: M ./tests/bench.c -210 +6 cannam@95: A ./tests/fftw_bench.h cannam@95: A ./tests/fftw_bench_common.c cannam@95: M ./tools/Makefile.am -1 +2 cannam@95: cannam@95: Tue Sep 12 21:00:36 EDT 2006 stevenj@fftw.org cannam@95: * MPI stuff at least compiles now cannam@95: cannam@95: M ./configure.ac -9 +11 cannam@95: M ./kernel/ifftw.h +7 cannam@95: M ./mpi/api.c -3 +4 cannam@95: M ./mpi/fftw3_mpi.h +1 cannam@95: M ./mpi/mpi_transpose_alltoall.c -10 +12 cannam@95: M ./mpi/mpi_transpose_inplace.c -11 +14 cannam@95: M ./mpi/problem.c -5 +4 cannam@95: M ./mpi/tproblem.c -4 +3 cannam@95: cannam@95: Mon Sep 11 22:26:36 EDT 2006 stevenj@fftw.org cannam@95: * initial (nonfunctional) start at MPI support (similar to FFTW 2.x in spirit, but mostly rewritten) cannam@95: cannam@95: A ./mpi/ cannam@95: M ./Makefile.am -1 +1 cannam@95: M ./configure.ac +11 cannam@95: A ./m4/acx_mpi.m4 cannam@95: A ./mpi/Makefile.am cannam@95: A ./mpi/api.c cannam@95: A ./mpi/block.c cannam@95: A ./mpi/conf.c cannam@95: A ./mpi/fftw3_mpi.h cannam@95: A ./mpi/ifftw_mpi.h cannam@95: A ./mpi/mpi_dft.h cannam@95: A ./mpi/mpi_transpose.h cannam@95: A ./mpi/mpi_transpose_alltoall.c cannam@95: A ./mpi/mpi_transpose_inplace.c cannam@95: A ./mpi/problem.c cannam@95: A ./mpi/testsched.c cannam@95: A ./mpi/tproblem.c cannam@95: A ./mpi/tsolve.c cannam@95: cannam@95: Mon Sep 11 22:25:38 EDT 2006 stevenj@fftw.org cannam@95: * make X(plan_awake) work for NULL argument to reduce code size cannam@95: cannam@95: M ./kernel/plan.c -3 +5 cannam@95: M ./rdft/vrank3-transpose.c -3 +3 cannam@95: cannam@95: Fri Sep 15 23:47:08 EDT 2006 stevenj@fftw.org cannam@95: * -mt should go before -mthreads to avoid spurious warnings on HPUX (thanks to Peter O'Gorman for the bug report) cannam@95: cannam@95: M ./m4/acx_pthread.m4 -2 +3 cannam@95: cannam@95: Mon Sep 11 13:53:44 EDT 2006 stevenj@fftw.org cannam@95: * Fortran init_threads wrapper didn't return result; thanks to Markus Wetzstein for the bug report cannam@95: cannam@95: M ./threads/f77funcs.h -2 +2 cannam@95: cannam@95: Thu Sep 7 18:43:55 EDT 2006 stevenj@fftw.org cannam@95: * make sure wrappers are included even if Fortran compiler was not detected (unless --disable-fortran was specified explicitly) ... this was supposed to be done before, but the definition was in the wrong place, grr cannam@95: cannam@95: M ./api/f77api.c -21 +21 cannam@95: cannam@95: Thu Aug 31 19:33:29 EDT 2006 stevenj@fftw.org cannam@95: * I'm sick of answering this question about non-deterministic results cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn +13 cannam@95: cannam@95: Tue Aug 22 21:27:29 EDT 2006 Matteo Frigo cannam@95: * Add --tag=CC flag to libtool. cannam@95: cannam@95: This change is consistent with the libtool invocation in the latest cannam@95: automake, and is required to compile with (some version of) xlc. cannam@95: cannam@95: cannam@95: cannam@95: M ./support/Makefile.codelets -2 +3 cannam@95: cannam@95: Mon Aug 21 21:40:36 EDT 2006 athena@fftw.org cannam@95: * avoid ``fma'' because it is defined in c99. cannam@95: cannam@95: M ./tests/bench.c -3 +3 cannam@95: cannam@95: Sun Aug 20 11:40:53 EDT 2006 Matteo Frigo cannam@95: * Obey -standalone flag. cannam@95: cannam@95: M ./genfft/gen_twiddle.ml -2 +4 cannam@95: cannam@95: Sat Aug 19 13:34:27 EDT 2006 Matteo Frigo cannam@95: * obey -standalone when generating simd codelets cannam@95: cannam@95: M ./genfft/gen_notw_c.ml -2 +1 cannam@95: M ./genfft/gen_twiddle_c.ml -2 +3 cannam@95: cannam@95: Sat Aug 19 13:33:43 EDT 2006 Matteo Frigo cannam@95: * removed obsolete athfft cannam@95: cannam@95: M ./genfft/Makefile.am -10 +9 cannam@95: R ./genfft/gen_athnotw.ml cannam@95: R ./genfft/gen_athtw.ml cannam@95: cannam@95: Thu Aug 17 21:50:50 EDT 2006 stevenj@fftw.org cannam@95: * updated citation to Proc. IEEE paper cannam@95: cannam@95: M ./doc/fftw3.texi -8 +5 cannam@95: cannam@95: Thu Aug 17 21:47:05 EDT 2006 stevenj@fftw.org cannam@95: * use darcs changes --summary to make nice changelog; emacs fill-region hack is obsolete cannam@95: cannam@95: M ./mkdist.sh -3 +1 cannam@95: cannam@95: Mon Aug 14 17:53:19 EDT 2006 athena@fftw.org cannam@95: * removed timer calibration cannam@95: Timer calibration seems not to work any longer on recent processors--- cannam@95: too much noise. I have remove it completely. cannam@95: cannam@95: M ./libbench2/Makefile.am -1 +1 cannam@95: M ./libbench2/bench.h -2 cannam@95: M ./libbench2/timer.c -70 +3 cannam@95: R ./libbench2/timer2.c cannam@95: cannam@95: Mon Aug 14 10:47:15 EDT 2006 Matteo Frigo cannam@95: * removed k7 cannam@95: Removed obsolete k7 support. cannam@95: cannam@95: M ./Makefile.am -10 +2 cannam@95: M ./NEWS +3 cannam@95: M ./TODO -5 cannam@95: M ./api/version.c -3 cannam@95: M ./bootstrap.sh -2 cannam@95: M ./commercialize.sh -8 +1 cannam@95: M ./configure.ac -13 cannam@95: M ./dft/Makefile.am -1 +1 cannam@95: M ./dft/codelet-dft.h -4 cannam@95: M ./dft/conf.c -3 cannam@95: R ./dft/k7/Makefile.am cannam@95: R ./dft/k7/codelets/Makefile.am cannam@95: R ./dft/k7/codelets/ cannam@95: R ./dft/k7/k7.c cannam@95: R ./dft/k7/ cannam@95: M ./doc/fftw3.texi -31 +22 cannam@95: R ./genfft-k7/Makefile.am cannam@95: R ./genfft-k7/algsimp.ml cannam@95: R ./genfft-k7/algsimp.mli cannam@95: R ./genfft-k7/assignmentsToVfpinstrs.ml cannam@95: R ./genfft-k7/assignmentsToVfpinstrs.mli cannam@95: R ./genfft-k7/assoctable.ml cannam@95: R ./genfft-k7/assoctable.mli cannam@95: R ./genfft-k7/balanceVfpinstrs.ml cannam@95: R ./genfft-k7/balanceVfpinstrs.mli cannam@95: R ./genfft-k7/complex.ml cannam@95: R ./genfft-k7/complex.mli cannam@95: R ./genfft-k7/expr.ml cannam@95: R ./genfft-k7/expr.mli cannam@95: R ./genfft-k7/fft.ml cannam@95: R ./genfft-k7/fft.mli cannam@95: R ./genfft-k7/genUtil.ml cannam@95: R ./genfft-k7/gen_notw.ml cannam@95: R ./genfft-k7/gen_twiddle.ml cannam@95: R ./genfft-k7/id.ml cannam@95: R ./genfft-k7/id.mli cannam@95: R ./genfft-k7/k7Basics.ml cannam@95: R ./genfft-k7/k7Basics.mli cannam@95: R ./genfft-k7/k7ExecutionModel.ml cannam@95: R ./genfft-k7/k7ExecutionModel.mli cannam@95: R ./genfft-k7/k7FlatInstructionScheduling.ml cannam@95: R ./genfft-k7/k7FlatInstructionScheduling.mli cannam@95: R ./genfft-k7/k7InstructionSchedulingBasics.ml cannam@95: R ./genfft-k7/k7InstructionSchedulingBasics.mli cannam@95: R ./genfft-k7/k7RegisterAllocationBasics.ml cannam@95: R ./genfft-k7/k7RegisterAllocationBasics.mli cannam@95: R ./genfft-k7/k7RegisterAllocator.ml cannam@95: R ./genfft-k7/k7RegisterAllocator.mli cannam@95: R ./genfft-k7/k7RegisterAllocatorEATranslation.ml cannam@95: R ./genfft-k7/k7RegisterAllocatorEATranslation.mli cannam@95: R ./genfft-k7/k7RegisterAllocatorInit.ml cannam@95: R ./genfft-k7/k7RegisterAllocatorInit.mli cannam@95: R ./genfft-k7/k7RegisterReallocation.ml cannam@95: R ./genfft-k7/k7RegisterReallocation.mli cannam@95: R ./genfft-k7/k7Translate.ml cannam@95: R ./genfft-k7/k7Translate.mli cannam@95: R ./genfft-k7/k7Unparsing.ml cannam@95: R ./genfft-k7/k7Unparsing.mli cannam@95: R ./genfft-k7/k7Vectorization.ml cannam@95: R ./genfft-k7/k7Vectorization.mli cannam@95: R ./genfft-k7/littlesimp.ml cannam@95: R ./genfft-k7/littlesimp.mli cannam@95: R ./genfft-k7/magic.ml cannam@95: R ./genfft-k7/memoMonad.ml cannam@95: R ./genfft-k7/memoMonad.mli cannam@95: R ./genfft-k7/monads.ml cannam@95: R ./genfft-k7/nonDetMonad.ml cannam@95: R ./genfft-k7/nonDetMonad.mli cannam@95: R ./genfft-k7/nullVectorization.ml cannam@95: R ./genfft-k7/nullVectorization.mli cannam@95: R ./genfft-k7/number.ml cannam@95: R ./genfft-k7/number.mli cannam@95: R ./genfft-k7/oracle.ml cannam@95: R ./genfft-k7/oracle.mli cannam@95: R ./genfft-k7/stateMonad.ml cannam@95: R ./genfft-k7/stateMonad.mli cannam@95: R ./genfft-k7/to_alist.ml cannam@95: R ./genfft-k7/to_alist.mli cannam@95: R ./genfft-k7/twiddle.ml cannam@95: R ./genfft-k7/twiddle.mli cannam@95: R ./genfft-k7/util.ml cannam@95: R ./genfft-k7/util.mli cannam@95: R ./genfft-k7/vAnnotatedScheduler.ml cannam@95: R ./genfft-k7/vAnnotatedScheduler.mli cannam@95: R ./genfft-k7/vDag.ml cannam@95: R ./genfft-k7/vDag.mli cannam@95: R ./genfft-k7/vFpBasics.ml cannam@95: R ./genfft-k7/vFpBasics.mli cannam@95: R ./genfft-k7/vFpUnparsing.ml cannam@95: R ./genfft-k7/vFpUnparsing.mli cannam@95: R ./genfft-k7/vImproveSchedule.ml cannam@95: R ./genfft-k7/vImproveSchedule.mli cannam@95: R ./genfft-k7/vK7Optimization.ml cannam@95: R ./genfft-k7/vK7Optimization.mli cannam@95: R ./genfft-k7/vScheduler.ml cannam@95: R ./genfft-k7/vScheduler.mli cannam@95: R ./genfft-k7/vSimdBasics.ml cannam@95: R ./genfft-k7/vSimdBasics.mli cannam@95: R ./genfft-k7/vSimdIndexing.ml cannam@95: R ./genfft-k7/vSimdIndexing.mli cannam@95: R ./genfft-k7/vSimdUnparsing.ml cannam@95: R ./genfft-k7/vSimdUnparsing.mli cannam@95: R ./genfft-k7/variable.ml cannam@95: R ./genfft-k7/variable.mli cannam@95: R ./genfft-k7/ cannam@95: M ./kernel/align.c -2 cannam@95: M ./kernel/ifftw.h -5 +1 cannam@95: M ./mkdist.sh -1 +1 cannam@95: M ./support/Makefile.am -1 +1 cannam@95: M ./support/Makefile.codelets -6 +1 cannam@95: R ./support/codelet_asmprelude cannam@95: cannam@95: Sun Aug 13 11:02:11 EDT 2006 athena@fftw.org cannam@95: * Use darcs instead of cvs. cannam@95: cannam@95: M ./mkdist.sh -5 +2 cannam@95: cannam@95: Wed Jul 19 08:52:15 EDT 2006 athena cannam@95: * [project @ 2006-07-19 12:52:15 by athena] cannam@95: Treat a the string "-" as a nonoption. cannam@95: cannam@95: M ./libbench2/my-getopt.c -1 +4 cannam@95: cannam@95: Tue Jul 4 17:10:47 EDT 2006 stevenj cannam@95: * [project @ 2006-07-04 21:10:47 by stevenj] cannam@95: comment out pkginclude dir for now cannam@95: cannam@95: M ./Makefile.am -2 +2 cannam@95: M ./api/Makefile.am -2 +2 cannam@95: M ./dft/Makefile.am -2 +2 cannam@95: M ./kernel/Makefile.am -2 +2 cannam@95: M ./rdft/Makefile.am -2 +2 cannam@95: M ./reodft/Makefile.am -2 +2 cannam@95: M ./threads/Makefile.am -2 +2 cannam@95: cannam@95: Mon Jul 3 20:51:08 EDT 2006 stevenj cannam@95: * [project @ 2006-07-04 00:51:08 by stevenj] cannam@95: make sure CCAS = CC to avoid libtool confusion cannam@95: cannam@95: M ./configure.ac -1 +1 cannam@95: cannam@95: Fri Jun 23 04:07:31 EDT 2006 stevenj cannam@95: * [project @ 2006-06-23 08:07:31 by stevenj] cannam@95: install x77.h guru.h guru64.h in pkgincludedir cannam@95: cannam@95: M ./api/Makefile.am -1 +1 cannam@95: cannam@95: Fri Jun 23 04:03:42 EDT 2006 stevenj cannam@95: * [project @ 2006-06-23 08:03:42 by stevenj] cannam@95: whitespace cannam@95: cannam@95: M ./configure.ac -3 cannam@95: cannam@95: Fri Jun 23 02:33:45 EDT 2006 stevenj cannam@95: * [project @ 2006-06-23 06:33:45 by stevenj] cannam@95: support cycle counter with xlc on Linux/ppc cannam@95: cannam@95: M ./kernel/cycle.h -2 +2 cannam@95: cannam@95: Tue Jun 20 08:16:08 EDT 2006 athena cannam@95: * [project @ 2006-06-20 12:16:08 by athena] cannam@95: Stylistic change. cannam@95: cannam@95: M ./tools/fftw-wisdom.c -1 +1 cannam@95: cannam@95: Tue Jun 20 02:20:34 EDT 2006 stevenj cannam@95: * [project @ 2006-06-20 06:20:34 by stevenj] cannam@95: bump date cannam@95: cannam@95: M ./m4/ax_cc_maxopt.m4 -1 +1 cannam@95: cannam@95: Tue Jun 20 02:20:06 EDT 2006 stevenj cannam@95: * [project @ 2006-06-20 06:20:06 by stevenj] cannam@95: correct bug reported by Andrew Salamon ... --enable-portable-binary was cannam@95: ignored (or rather, treated unpredictably) due to typo, grrr cannam@95: cannam@95: M ./m4/ax_cc_maxopt.m4 -1 +1 cannam@95: cannam@95: Thu Jun 1 20:30:06 EDT 2006 stevenj cannam@95: * [project @ 2006-06-02 00:30:06 by stevenj] cannam@95: install 'internal' header files into includedir/fftw3/, includedir/fftw3f/, etcetera....this will make it easier to write external libraries that plug into FFTW internals, e.g. to add new solvers cannam@95: cannam@95: M ./Makefile.am +3 cannam@95: M ./api/Makefile.am +3 cannam@95: M ./dft/Makefile.am +3 cannam@95: M ./kernel/Makefile.am +3 cannam@95: M ./rdft/Makefile.am +3 cannam@95: M ./reodft/Makefile.am +3 cannam@95: M ./threads/Makefile.am +3 cannam@95: cannam@95: Mon May 29 23:59:19 EDT 2006 stevenj cannam@95: * [project @ 2006-05-30 03:59:19 by stevenj] cannam@95: bug fix, thanks to James Donald for the bug report (only affects experimental semaphore stuff) cannam@95: cannam@95: M ./threads/threads.c +1 cannam@95: cannam@95: Mon May 29 23:58:16 EDT 2006 stevenj cannam@95: * [project @ 2006-05-30 03:58:16 by stevenj] cannam@95: comment cannam@95: cannam@95: M ./NEWS -3 +4 cannam@95: cannam@95: Mon May 29 21:02:50 EDT 2006 stevenj cannam@95: * [project @ 2006-05-30 01:02:50 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./m4/acx_pthread.m4 -1 +3 cannam@95: cannam@95: Sat May 27 19:36:15 EDT 2006 stevenj cannam@95: * [project @ 2006-05-27 23:36:15 by stevenj] cannam@95: version bump cannam@95: cannam@95: M ./m4/acx_pthread.m4 -1 +1 cannam@95: cannam@95: Sat May 27 14:54:47 EDT 2006 stevenj cannam@95: * [project @ 2006-05-27 18:54:47 by stevenj] cannam@95: only check for xlc_r/cc_r if we are not using gcc cannam@95: cannam@95: M ./m4/acx_pthread.m4 -2 +3 cannam@95: cannam@95: Fri May 26 15:00:38 EDT 2006 stevenj cannam@95: * [project @ 2006-05-26 19:00:38 by stevenj] cannam@95: use ptrdiff_t (it's C89 and standard C++, hooray) cannam@95: cannam@95: M ./api/fftw3.h -10 +5 cannam@95: cannam@95: Fri May 26 12:59:33 EDT 2006 stevenj cannam@95: * [project @ 2006-05-26 16:59:33 by stevenj] cannam@95: version bump cannam@95: cannam@95: M ./configure.ac -2 +2 cannam@95: cannam@95: Fri May 26 12:57:32 EDT 2006 stevenj cannam@95: * [project @ 2006-05-26 16:57:32 by stevenj] cannam@95: noted 64-bit guru API cannam@95: cannam@95: M ./NEWS +7 cannam@95: cannam@95: Fri May 26 12:53:09 EDT 2006 stevenj cannam@95: * [project @ 2006-05-26 16:53:09 by stevenj] cannam@95: note that newer versions of VC++ support long long cannam@95: cannam@95: M ./api/fftw3.h -2 +3 cannam@95: cannam@95: Fri May 26 12:46:09 EDT 2006 stevenj cannam@95: * [project @ 2006-05-26 16:46:09 by stevenj] cannam@95: try harder to get a portable 64-bit type cannam@95: cannam@95: M ./api/fftw3.h -2 +6 cannam@95: cannam@95: Thu May 25 22:04:18 EDT 2006 stevenj cannam@95: * [project @ 2006-05-26 02:04:18 by stevenj] cannam@95: added draft guru64 API cannam@95: cannam@95: A ./api/guru.h cannam@95: A ./api/guru64.h cannam@95: A ./api/mktensor-iodims.h cannam@95: A ./api/mktensor-iodims64.c cannam@95: A ./api/plan-guru-dft-c2r.h cannam@95: A ./api/plan-guru-dft-r2c.h cannam@95: A ./api/plan-guru-dft.h cannam@95: A ./api/plan-guru-r2r.h cannam@95: A ./api/plan-guru-split-dft-c2r.h cannam@95: A ./api/plan-guru-split-dft-r2c.h cannam@95: A ./api/plan-guru-split-dft.h cannam@95: A ./api/plan-guru64-dft-c2r.c cannam@95: A ./api/plan-guru64-dft-r2c.c cannam@95: A ./api/plan-guru64-dft.c cannam@95: A ./api/plan-guru64-r2r.c cannam@95: A ./api/plan-guru64-split-dft-c2r.c cannam@95: A ./api/plan-guru64-split-dft-r2c.c cannam@95: A ./api/plan-guru64-split-dft.c cannam@95: M ./api/Makefile.am -1 +7 cannam@95: M ./api/api.h +3 cannam@95: M ./api/fftw3.h -2 +56 cannam@95: M ./api/guru.h +4 cannam@95: M ./api/guru64.h +4 cannam@95: M ./api/mktensor-iodims.c -62 +2 cannam@95: M ./api/mktensor-iodims.h +62 cannam@95: M ./api/mktensor-iodims64.c +2 cannam@95: M ./api/plan-guru-dft-c2r.c -44 +2 cannam@95: M ./api/plan-guru-dft-c2r.h +44 cannam@95: M ./api/plan-guru-dft-r2c.c -43 +2 cannam@95: M ./api/plan-guru-dft-r2c.h +43 cannam@95: M ./api/plan-guru-dft.c -44 +2 cannam@95: M ./api/plan-guru-dft.h +44 cannam@95: M ./api/plan-guru-r2r.c -47 +2 cannam@95: M ./api/plan-guru-r2r.h +47 cannam@95: M ./api/plan-guru-split-dft-c2r.c -40 +2 cannam@95: M ./api/plan-guru-split-dft-c2r.h +40 cannam@95: M ./api/plan-guru-split-dft-r2c.c -39 +2 cannam@95: M ./api/plan-guru-split-dft-r2c.h +39 cannam@95: M ./api/plan-guru-split-dft.c -39 +2 cannam@95: M ./api/plan-guru-split-dft.h +39 cannam@95: M ./api/plan-guru64-dft-c2r.c +2 cannam@95: M ./api/plan-guru64-dft-r2c.c +2 cannam@95: M ./api/plan-guru64-dft.c +2 cannam@95: M ./api/plan-guru64-r2r.c +2 cannam@95: M ./api/plan-guru64-split-dft-c2r.c +2 cannam@95: M ./api/plan-guru64-split-dft-r2c.c +2 cannam@95: M ./api/plan-guru64-split-dft.c +2 cannam@95: cannam@95: Mon May 22 16:41:44 EDT 2006 stevenj cannam@95: * [project @ 2006-05-22 20:41:44 by stevenj] cannam@95: added FIXME note cannam@95: cannam@95: M ./m4/acx_pthread.m4 +1 cannam@95: cannam@95: Mon May 22 16:40:30 EDT 2006 stevenj cannam@95: * [project @ 2006-05-22 20:40:30 by stevenj] cannam@95: check for xlc_r in addition to cc_r; thanks to Guy Moebs for the bug report cannam@95: cannam@95: M ./m4/acx_pthread.m4 -3 +3 cannam@95: cannam@95: Fri Apr 21 12:35:25 EDT 2006 stevenj cannam@95: * [project @ 2006-04-21 16:35:25 by stevenj] cannam@95: added note about gcc 4.0.1 on MacOS/Intel cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -1 +4 cannam@95: cannam@95: Thu Apr 20 23:08:42 EDT 2006 stevenj cannam@95: * [project @ 2006-04-21 03:08:42 by stevenj] cannam@95: added code for Core Duo; thanks to Eric Branlund cannam@95: cannam@95: M ./m4/ax_gcc_archflag.m4 -2 +3 cannam@95: cannam@95: Thu Apr 20 20:21:03 EDT 2006 stevenj cannam@95: * [project @ 2006-04-21 00:21:03 by stevenj] cannam@95: fixed failure for -fPIC or for gcc-4 on Apple Intel machines; thanks to cannam@95: Eric Branlund for the bug report cannam@95: cannam@95: M ./m4/ax_gcc_x86_cpuid.m4 -4 +12 cannam@95: cannam@95: Tue Apr 11 20:00:31 EDT 2006 athena cannam@95: * [project @ 2006-04-12 00:00:31 by athena] cannam@95: Use -maltivec when checking for altivec.h. cannam@95: cannam@95: M ./configure.ac -1 +15 cannam@95: cannam@95: Mon Apr 3 15:52:44 EDT 2006 stevenj cannam@95: * [project @ 2006-04-03 19:52:44 by stevenj] cannam@95: note planner overwriting input in planner-flags reference cannam@95: cannam@95: M ./doc/fftw3.texi -1 +6 cannam@95: cannam@95: Tue Mar 28 09:05:26 EST 2006 athena cannam@95: * [project @ 2006-03-28 14:05:26 by athena] cannam@95: FAQ entry about --enable-k7 in 64-bit mode. cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn +11 cannam@95: cannam@95: Mon Mar 27 23:41:05 EST 2006 stevenj cannam@95: * [project @ 2006-03-28 04:41:05 by stevenj] cannam@95: sprintf -> snprintf, to avoid (harmless) complaints by users/compilers cannam@95: cannam@95: M ./configure.ac -1 +1 cannam@95: M ./libbench2/report.c -12 +19 cannam@95: M ./tools/fftw-wisdom.c +4 cannam@95: cannam@95: Mon Mar 27 23:30:22 EST 2006 stevenj cannam@95: * [project @ 2006-03-28 04:30:22 by stevenj] cannam@95: silence compiler warning cannam@95: cannam@95: M ./kernel/align.c -5 +6 cannam@95: cannam@95: Fri Mar 17 09:20:10 EST 2006 athena cannam@95: * [project @ 2006-03-17 14:20:10 by athena] cannam@95: Remove dft/codelets/inplace, add simd/nonportable to list of cannam@95: directories to be compiled on non-unix systems. cannam@95: cannam@95: M ./doc/fftw3.texi -13 +14 cannam@95: cannam@95: Sat Aug 12 23:52:49 EDT 2006 Unknown tagger cannam@95: tagged fftw-3-1-1 cannam@95: cannam@95: cannam@95: Sat Mar 4 16:17:56 EST 2006 stevenj cannam@95: * [project @ 2006-03-04 21:17:56 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Sat Mar 4 16:13:08 EST 2006 stevenj cannam@95: * [project @ 2006-03-04 21:13:08 by stevenj] cannam@95: note that we align the stack ourselves if necessary, with gcc and icc cannam@95: cannam@95: M ./doc/fftw3.texi -8 +10 cannam@95: cannam@95: Sat Mar 4 16:08:16 EST 2006 stevenj cannam@95: * [project @ 2006-03-04 21:08:16 by stevenj] cannam@95: clearer distinction between static and automatic storage in C cannam@95: cannam@95: M ./doc/fftw3.texi -29 +30 cannam@95: cannam@95: Sat Feb 25 20:27:01 EST 2006 stevenj cannam@95: * [project @ 2006-02-26 01:27:01 by stevenj] cannam@95: rm unused var cannam@95: cannam@95: M ./libbench2/verify-lib.c -2 +1 cannam@95: cannam@95: Sat Feb 25 17:30:28 EST 2006 athena cannam@95: * [project @ 2006-02-25 22:30:28 by athena] cannam@95: Improved usage of goto (Dijkstra miserere nostri) cannam@95: cannam@95: M ./libbench2/my-getopt.c -18 +17 cannam@95: cannam@95: Sat Feb 25 14:19:15 EST 2006 stevenj cannam@95: * [project @ 2006-02-25 19:19:15 by stevenj] cannam@95: boilerplate cannam@95: cannam@95: M ./libbench2/my-getopt.h +13 cannam@95: cannam@95: Sat Feb 25 14:14:40 EST 2006 stevenj cannam@95: * [project @ 2006-02-25 19:14:40 by stevenj] cannam@95: update for upcoming 3.1.1 cannam@95: cannam@95: M ./NEWS +17 cannam@95: cannam@95: Sat Feb 25 13:57:34 EST 2006 stevenj cannam@95: * [project @ 2006-02-25 18:57:34 by stevenj] cannam@95: replace obsolete IMPATIENT with MEASURE cannam@95: cannam@95: M ./tools/fftw-wisdom.c -1 +3 cannam@95: M ./tools/fftw_wisdom.1.in -2 +2 cannam@95: cannam@95: Sat Feb 25 13:52:25 EST 2006 stevenj cannam@95: * [project @ 2006-02-25 18:52:25 by stevenj] cannam@95: corrected comment cannam@95: cannam@95: M ./tools/fftw-wisdom.c -1 +1 cannam@95: cannam@95: Sat Feb 25 10:19:26 EST 2006 athena cannam@95: * [project @ 2006-02-25 15:19:26 by athena] cannam@95: -v does not take an argument. cannam@95: cannam@95: M ./tools/fftw-wisdom.c -6 +3 cannam@95: cannam@95: Sat Feb 25 10:17:18 EST 2006 athena cannam@95: * [project @ 2006-02-25 15:17:18 by athena] cannam@95: Obey the unix convention that -ab = -a -b cannam@95: cannam@95: M ./libbench2/my-getopt.c -15 +24 cannam@95: cannam@95: Fri Feb 24 23:13:49 EST 2006 stevenj cannam@95: * [project @ 2006-02-25 04:13:49 by stevenj] cannam@95: minor fixes (return error on unrecognized option) cannam@95: cannam@95: M ./libbench2/bench-main.c -2 +3 cannam@95: M ./libbench2/my-getopt.c +5 cannam@95: M ./tools/fftw-wisdom.c -3 +4 cannam@95: cannam@95: Fri Feb 24 22:46:12 EST 2006 stevenj cannam@95: * [project @ 2006-02-25 03:46:12 by stevenj] cannam@95: ugh cannam@95: cannam@95: M ./tools/fftw-wisdom.c -34 +30 cannam@95: cannam@95: Fri Feb 24 21:42:56 EST 2006 athena cannam@95: * [project @ 2006-02-25 02:42:56 by athena] cannam@95: require exact match for long options. cannam@95: cannam@95: M ./libbench2/my-getopt.c -1 +2 cannam@95: cannam@95: Fri Feb 24 21:38:02 EST 2006 athena cannam@95: * [project @ 2006-02-25 02:38:02 by athena] cannam@95: better fix cannam@95: cannam@95: M ./libbench2/my-getopt.c -1 +2 cannam@95: cannam@95: Fri Feb 24 21:37:06 EST 2006 athena cannam@95: * [project @ 2006-02-25 02:37:06 by athena] cannam@95: Fix cannam@95: cannam@95: M ./libbench2/my-getopt.c -1 +1 cannam@95: cannam@95: Fri Feb 24 21:25:48 EST 2006 athena cannam@95: * [project @ 2006-02-25 02:25:48 by athena] cannam@95: nothing cannam@95: cannam@95: A ./libbench2/my-getopt.c cannam@95: A ./libbench2/my-getopt.h cannam@95: M ./libbench2/Makefile.am -5 +4 cannam@95: M ./libbench2/bench-main.c -53 +48 cannam@95: M ./libbench2/bench.h -5 +1 cannam@95: R ./libbench2/getopt-utils.c cannam@95: R ./libbench2/getopt.c cannam@95: R ./libbench2/getopt.h cannam@95: R ./libbench2/getopt1.c cannam@95: M ./libbench2/my-getopt.c +153 cannam@95: M ./libbench2/my-getopt.h +33 cannam@95: cannam@95: Mon Feb 20 17:37:21 EST 2006 stevenj cannam@95: * [project @ 2006-02-20 22:37:21 by stevenj] cannam@95: rm transpose-indirect-inplace solver, which was buggy cannam@95: cannam@95: M ./dft/indirect-transpose.c -54 +11 cannam@95: cannam@95: Wed Feb 15 08:43:05 EST 2006 athena cannam@95: * [project @ 2006-02-15 13:43:05 by athena] cannam@95: Comment fix. cannam@95: cannam@95: M ./kernel/cycle.h -2 +2 cannam@95: cannam@95: Wed Feb 15 08:18:41 EST 2006 athena cannam@95: * [project @ 2006-02-15 13:18:41 by athena] cannam@95: Cycle counter for Visual C++ x86-64, courtesy of Dirk Michaelis cannam@95: cannam@95: M ./kernel/cycle.h -4 +5 cannam@95: cannam@95: Tue Feb 14 19:17:30 EST 2006 stevenj cannam@95: * [project @ 2006-02-15 00:17:30 by stevenj] cannam@95: rfftwnd.png is in builddir cannam@95: cannam@95: M ./doc/Makefile.am -1 +1 cannam@95: cannam@95: Tue Feb 14 19:03:27 EST 2006 stevenj cannam@95: * [project @ 2006-02-15 00:03:27 by stevenj] cannam@95: fixed typo: --enable-portable-binary, not --with cannam@95: cannam@95: M ./doc/fftw3.texi -3 +3 cannam@95: cannam@95: Mon Feb 13 07:59:06 EST 2006 athena cannam@95: * [project @ 2006-02-13 12:59:06 by athena] cannam@95: estimator tweaks. cannam@95: cannam@95: M ./dft/dftw-direct.c -2 +3 cannam@95: M ./rdft/hc2hc-direct.c -2 +2 cannam@95: cannam@95: Sun Feb 12 20:43:39 EST 2006 athena cannam@95: * [project @ 2006-02-13 01:43:39 by athena] cannam@95: sse/sse2 support for t3?v codelets cannam@95: cannam@95: M ./simd/simd-sse.h +22 cannam@95: M ./simd/simd-sse2.h -8 +20 cannam@95: cannam@95: Sun Feb 12 20:39:22 EST 2006 athena cannam@95: * [project @ 2006-02-13 01:39:22 by athena] cannam@95: Use CEXP instead of SIN/COS. cannam@95: cannam@95: M ./simd/simd-altivec.h -1 +1 cannam@95: cannam@95: Sun Feb 12 20:12:10 EST 2006 athena cannam@95: * [project @ 2006-02-13 01:12:10 by athena] cannam@95: bug in randomized cse eliminator. cannam@95: cannam@95: M ./genfft/oracle.ml -3 +5 cannam@95: cannam@95: Sun Feb 12 18:34:12 EST 2006 athena cannam@95: * [project @ 2006-02-12 23:34:12 by athena] cannam@95: Added support for t2-style simd codelets. This is altivec only for cannam@95: now; sse/sse2 don't even compile yet. cannam@95: cannam@95: A ./dft/simd/t3b.h cannam@95: A ./dft/simd/t3f.h cannam@95: M ./dft/simd/Makefile.am -2 +2 cannam@95: M ./dft/simd/codelets/Makefile.am -2 +11 cannam@95: M ./dft/simd/t3b.h +34 cannam@95: M ./dft/simd/t3f.h +34 cannam@95: M ./genfft/algsimp.ml -4 +26 cannam@95: M ./genfft/annotate.ml -5 +3 cannam@95: M ./genfft/c.ml -10 +9 cannam@95: M ./genfft/c.mli -2 +2 cannam@95: M ./genfft/complex.ml -93 +6 cannam@95: M ./genfft/complex.mli -5 +3 cannam@95: M ./genfft/expr.ml -6 +15 cannam@95: M ./genfft/expr.mli -2 +4 cannam@95: M ./genfft/gen_athtw.ml -3 +3 cannam@95: M ./genfft/gen_conv.ml -5 +3 cannam@95: M ./genfft/gen_hc2hc.ml -4 +3 cannam@95: M ./genfft/gen_hc2r.ml -5 +3 cannam@95: M ./genfft/gen_mdct.ml -7 +3 cannam@95: M ./genfft/gen_notw.ml -5 +2 cannam@95: M ./genfft/gen_notw_c.ml -5 +2 cannam@95: M ./genfft/gen_r2hc.ml -4 +3 cannam@95: M ./genfft/gen_r2r.ml -4 +3 cannam@95: M ./genfft/gen_twiddle.ml -3 +3 cannam@95: M ./genfft/gen_twiddle_c.ml -10 +4 cannam@95: M ./genfft/gen_twidsq.ml -7 +3 cannam@95: M ./genfft/gen_twidsq_c.ml -15 +5 cannam@95: M ./genfft/magic.ml -4 +4 cannam@95: M ./genfft/oracle.ml -1 +5 cannam@95: M ./genfft/schedule.ml -3 +3 cannam@95: M ./genfft/simd.ml -5 +11 cannam@95: M ./genfft/to_alist.ml -4 +16 cannam@95: M ./genfft/trig.ml -3 +2 cannam@95: M ./genfft/twiddle.ml -265 +35 cannam@95: M ./genfft/twiddle.mli -4 +4 cannam@95: M ./simd/simd-altivec.h +28 cannam@95: cannam@95: Sun Feb 12 15:30:27 EST 2006 athena cannam@95: * [project @ 2006-02-12 20:30:27 by athena] cannam@95: Added support for t2-style simd split-complex codelets. cannam@95: cannam@95: A ./dft/simd/ts.c cannam@95: A ./dft/simd/ts.h cannam@95: M ./dft/simd/Makefile.am -2 +2 cannam@95: M ./dft/simd/codelets/Makefile.am -2 +7 cannam@95: R ./dft/simd/t1s.c cannam@95: R ./dft/simd/t1s.h cannam@95: M ./dft/simd/ts.c +50 cannam@95: M ./dft/simd/ts.h +33 cannam@95: M ./genfft/twiddle.ml -5 +3 cannam@95: cannam@95: Fri Feb 10 18:21:28 EST 2006 stevenj cannam@95: * [project @ 2006-02-10 23:21:28 by stevenj] cannam@95: cannam@95: M ./m4/ax_openmp.m4 -1 +1 cannam@95: cannam@95: Fri Feb 10 18:19:46 EST 2006 stevenj cannam@95: * [project @ 2006-02-10 23:19:46 by stevenj] cannam@95: punctuation cannam@95: cannam@95: M ./m4/ax_openmp.m4 -2 +2 cannam@95: cannam@95: Fri Feb 10 18:00:35 EST 2006 stevenj cannam@95: * [project @ 2006-02-10 23:00:35 by stevenj] cannam@95: windows DLL stuff for Fortran interface cannam@95: cannam@95: M ./api/f77api.c +7 cannam@95: M ./api/f77funcs.h -43 +43 cannam@95: cannam@95: Fri Feb 10 09:48:52 EST 2006 athena cannam@95: * [project @ 2006-02-10 14:48:52 by athena] cannam@95: Bumped version to 3.1.1 cannam@95: cannam@95: M ./configure.ac -2 +2 cannam@95: cannam@95: Fri Feb 10 09:18:39 EST 2006 athena cannam@95: * [project @ 2006-02-10 14:18:39 by athena] cannam@95: Precompute array indices on x86-64. Speeds up Pentium IV and makes no cannam@95: appreciable difference on AMD. cannam@95: cannam@95: M ./kernel/ifftw.h -2 +2 cannam@95: cannam@95: Tue Feb 7 22:01:36 EST 2006 athena cannam@95: * [project @ 2006-02-08 03:01:36 by athena] cannam@95: Check whether the processor supports CPUID before issuing the cannam@95: instruction. (Grrr...) Code contributed by Eric J. Korpela. cannam@95: cannam@95: A ./simd/x86-cpuid.h cannam@95: M ./simd/Makefile.am -1 +1 cannam@95: M ./simd/sse.c -24 +6 cannam@95: M ./simd/sse2.c -26 +8 cannam@95: M ./simd/x86-cpuid.h +127 cannam@95: cannam@95: Tue Feb 7 21:36:47 EST 2006 athena cannam@95: * [project @ 2006-02-08 02:36:47 by athena] cannam@95: icc supports x86_64 these days. cannam@95: cannam@95: M ./kernel/cycle.h -2 +2 cannam@95: cannam@95: Sun Feb 5 18:19:55 EST 2006 athena cannam@95: * [project @ 2006-02-05 23:19:55 by athena] cannam@95: Paranoia. cannam@95: cannam@95: M ./kernel/primes.c -2 +2 cannam@95: cannam@95: Mon Jan 30 15:27:53 EST 2006 stevenj cannam@95: * [project @ 2006-01-30 20:27:53 by stevenj] cannam@95: whoops, fixed assert (y <= x) cannam@95: cannam@95: M ./kernel/primes.c -2 +2 cannam@95: cannam@95: Mon Jan 30 15:26:22 EST 2006 stevenj cannam@95: * [project @ 2006-01-30 20:26:22 by stevenj] cannam@95: note that safe_mulmod requires {x,y} < p (or at least < 2p), and added cannam@95: assert cannam@95: cannam@95: M ./kernel/primes.c -2 +4 cannam@95: cannam@95: Mon Jan 30 11:09:32 EST 2006 athena cannam@95: * [project @ 2006-01-30 16:09:32 by athena] cannam@95: fixed aix/xlc lossage cannam@95: cannam@95: M ./libbench2/bench-user.h -2 +2 cannam@95: M ./libbench2/timer.c -4 +4 cannam@95: cannam@95: Sun Jan 29 20:42:51 EST 2006 athena cannam@95: * [project @ 2006-01-30 01:42:51 by athena] cannam@95: In the impuse test, normalize the impulse so that the impulse and the cannam@95: random vectors have roughly the same L2 norm. This change reduces the cannam@95: number of bits that we lose because of floating-point cancellation, so cannam@95: that we can focus on the bits that we lose because of bugs. cannam@95: cannam@95: M ./libbench2/verify-lib.c -3 +3 cannam@95: cannam@95: Sun Jan 29 20:37:47 EST 2006 athena cannam@95: * [project @ 2006-01-30 01:37:47 by athena] cannam@95: Compute omega in trigreal precision, as opposed to R. cannam@95: cannam@95: M ./rdft/dht-rader.c -2 +2 cannam@95: cannam@95: Sat Aug 12 23:51:14 EDT 2006 Unknown tagger cannam@95: tagged fftw-3-1 cannam@95: cannam@95: cannam@95: Fri Jan 27 19:16:22 EST 2006 stevenj cannam@95: * [project @ 2006-01-28 00:16:22 by stevenj] cannam@95: add --with-combined-threads option as workaround to Windows inability to build shared libs with dependencies cannam@95: cannam@95: M ./Makefile.am -1 +7 cannam@95: M ./configure.ac -1 +3 cannam@95: M ./tests/Makefile.am +2 cannam@95: M ./threads/Makefile.am +4 cannam@95: M ./tools/Makefile.am +2 cannam@95: cannam@95: Fri Jan 27 17:20:45 EST 2006 stevenj cannam@95: * [project @ 2006-01-27 22:20:45 by stevenj] cannam@95: libfftw3_threads should *not* used -no-undefined because, in fact, it is not true -- this library depends on -lfftw3, and is not self-contained cannam@95: cannam@95: M ./threads/Makefile.am -1 +1 cannam@95: cannam@95: Thu Jan 26 22:04:34 EST 2006 stevenj cannam@95: * [project @ 2006-01-27 03:04:34 by stevenj] cannam@95: updated cannam@95: cannam@95: M ./NEWS -3 +10 cannam@95: cannam@95: Thu Jan 26 21:10:50 EST 2006 athena cannam@95: * [project @ 2006-01-27 02:10:50 by athena] cannam@95: Added paranoid stack alignment when awaking plans. While I was at it, cannam@95: removed obsolete, redundant AWAKE macro. cannam@95: cannam@95: M ./api/apiplan.c -3 +8 cannam@95: M ./dft/bluestein.c -1 +1 cannam@95: M ./dft/buffered.c -4 +4 cannam@95: M ./dft/ct.c -3 +3 cannam@95: M ./dft/ctsq.c -2 +2 cannam@95: M ./dft/dftw-generic.c -1 +1 cannam@95: M ./dft/dftw-genericbuf.c -1 +1 cannam@95: M ./dft/indirect-transpose.c -3 +3 cannam@95: M ./dft/indirect.c -3 +3 cannam@95: M ./dft/rader.c -3 +3 cannam@95: M ./dft/rank-geq2.c -3 +3 cannam@95: M ./dft/vrank-geq1.c -2 +2 cannam@95: M ./kernel/ifftw.h -2 +1 cannam@95: M ./kernel/timer.c -3 +3 cannam@95: M ./rdft/buffered.c -4 +4 cannam@95: M ./rdft/buffered2.c -3 +3 cannam@95: M ./rdft/dft-r2hc.c -2 +2 cannam@95: M ./rdft/dht-r2hc.c -2 +2 cannam@95: M ./rdft/dht-rader.c -3 +3 cannam@95: M ./rdft/hc2hc-direct.c -3 +3 cannam@95: M ./rdft/hc2hc-directbuf.c -3 +3 cannam@95: M ./rdft/hc2hc-generic.c -2 +2 cannam@95: M ./rdft/hc2hc.c -2 +2 cannam@95: M ./rdft/indirect.c -3 +3 cannam@95: M ./rdft/rank-geq2-rdft2.c -3 +3 cannam@95: M ./rdft/rank-geq2.c -3 +3 cannam@95: M ./rdft/rank0-rdft2.c -2 +2 cannam@95: M ./rdft/rdft-dht.c -2 +2 cannam@95: M ./rdft/rdft2-radix2.c -2 +2 cannam@95: M ./rdft/vrank-geq1-rdft2.c -2 +2 cannam@95: M ./rdft/vrank-geq1.c -2 +2 cannam@95: M ./rdft/vrank3-transpose.c -4 +4 cannam@95: M ./reodft/redft00e-r2hc-pad.c -3 +3 cannam@95: M ./reodft/redft00e-r2hc.c -2 +2 cannam@95: M ./reodft/reodft00e-splitradix.c -3 +3 cannam@95: M ./reodft/reodft010e-r2hc.c -2 +2 cannam@95: M ./reodft/reodft11e-r2hc-odd.c -2 +2 cannam@95: M ./reodft/reodft11e-r2hc.c -2 +2 cannam@95: M ./reodft/reodft11e-radix2.c -2 +2 cannam@95: M ./reodft/rodft00e-r2hc-pad.c -3 +3 cannam@95: M ./reodft/rodft00e-r2hc.c -2 +2 cannam@95: M ./tests/hook.c -2 +2 cannam@95: M ./threads/ct.c -3 +3 cannam@95: M ./threads/dft-vrank-geq1.c -2 +2 cannam@95: M ./threads/hc2hc.c -2 +2 cannam@95: M ./threads/rdft-vrank-geq1.c -2 +2 cannam@95: M ./threads/vrank-geq1-rdft2.c -2 +2 cannam@95: cannam@95: Thu Jan 26 20:54:39 EST 2006 athena cannam@95: * [project @ 2006-01-27 01:54:39 by athena] cannam@95: Updated for 3.1. cannam@95: cannam@95: M ./NEWS -1 +1 cannam@95: cannam@95: Thu Jan 26 19:15:12 EST 2006 athena cannam@95: * [project @ 2006-01-27 00:15:12 by athena] cannam@95: ditched one alignment check and noted that we should eliminate the rest as well cannam@95: cannam@95: M ./TODO +3 cannam@95: M ./libbench2/bench-main.c -13 +1 cannam@95: cannam@95: Wed Jan 25 23:02:19 EST 2006 athena cannam@95: * [project @ 2006-01-26 04:02:19 by athena] cannam@95: alignment hack cannam@95: cannam@95: M ./libbench2/bench-main.c -2 +4 cannam@95: cannam@95: Wed Jan 25 22:05:11 EST 2006 athena cannam@95: * [project @ 2006-01-26 03:05:11 by athena] cannam@95: detect pentium M cannam@95: cannam@95: M ./m4/ax_gcc_archflag.m4 -1 +2 cannam@95: cannam@95: Wed Jan 25 18:42:58 EST 2006 stevenj cannam@95: * [project @ 2006-01-25 23:42:58 by stevenj] cannam@95: don't trust host_cpu if it claims we are on i386/i486, and call cpuid anyway (if it fails we use no arch flag). This is needed on FreeBSD cannam@95: cannam@95: M ./m4/ax_gcc_archflag.m4 -3 +1 cannam@95: cannam@95: Wed Jan 25 18:00:04 EST 2006 stevenj cannam@95: * [project @ 2006-01-25 23:00:04 by stevenj] cannam@95: suggest --with-our-malloc16 in error message cannam@95: cannam@95: M ./kernel/kalloc.c -2 +2 cannam@95: cannam@95: Tue Jan 24 19:53:34 EST 2006 stevenj cannam@95: * [project @ 2006-01-25 00:53:34 by stevenj] cannam@95: ditto for -no-gcc cannam@95: cannam@95: M ./configure.ac -2 +4 cannam@95: cannam@95: Tue Jan 24 19:51:08 EST 2006 stevenj cannam@95: * [project @ 2006-01-25 00:51:08 by stevenj] cannam@95: flags required for successfull compilation should be added even if the cannam@95: user overrides CFLAGS cannam@95: cannam@95: M ./configure.ac -9 +12 cannam@95: cannam@95: Tue Jan 24 18:43:59 EST 2006 stevenj cannam@95: * [project @ 2006-01-24 23:43:59 by stevenj] cannam@95: upcoming gcc OpenMP support uses -fopenmp cannam@95: cannam@95: M ./m4/ax_openmp.m4 -4 +4 cannam@95: cannam@95: Tue Jan 24 18:26:59 EST 2006 stevenj cannam@95: * [project @ 2006-01-24 23:26:59 by stevenj] cannam@95: note that PGI uses -mp as well cannam@95: cannam@95: M ./m4/ax_openmp.m4 -1 +1 cannam@95: cannam@95: Mon Jan 23 15:31:24 EST 2006 athena cannam@95: * [project @ 2006-01-23 20:31:24 by athena] cannam@95: my best guess at how to fix the microsoft crap du jour cannam@95: cannam@95: M ./kernel/cycle.h -2 +2 cannam@95: M ./simd/sse.c -65 +70 cannam@95: M ./simd/sse2.c -62 +67 cannam@95: cannam@95: Mon Jan 23 14:05:14 EST 2006 stevenj cannam@95: * [project @ 2006-01-23 19:05:14 by stevenj] cannam@95: use -Masmkeyword for PGI cycle counter, grr cannam@95: cannam@95: M ./configure.ac +5 cannam@95: M ./kernel/cycle.h -2 +4 cannam@95: cannam@95: Sun Jan 22 18:09:06 EST 2006 athena cannam@95: * [project @ 2006-01-22 23:09:06 by athena] cannam@95: Bumped version number to 3.1. cannam@95: cannam@95: M ./configure.ac -1 +1 cannam@95: cannam@95: Sat Jan 21 10:03:59 EST 2006 athena cannam@95: * [project @ 2006-01-21 15:03:59 by athena] cannam@95: Report that --enable-k7 is incompatible with --enable-shared. cannam@95: cannam@95: M ./configure.ac +3 cannam@95: cannam@95: Sat Jan 21 09:17:54 EST 2006 athena cannam@95: * [project @ 2006-01-21 14:17:54 by athena] cannam@95: Do not use empty libraries in LIBADD, since otherwise the linker fails cannam@95: on Solaris. cannam@95: cannam@95: M ./Makefile.am -7 +20 cannam@95: cannam@95: Wed Jan 18 10:47:59 EST 2006 stevenj cannam@95: * [project @ 2006-01-18 15:47:59 by stevenj] cannam@95: warn end-users away from this file cannam@95: cannam@95: M ./bootstrap.sh -1 +10 cannam@95: cannam@95: Tue Jan 17 16:16:42 EST 2006 athena cannam@95: * [project @ 2006-01-17 21:16:42 by athena] cannam@95: Gcc sucks. cannam@95: cannam@95: M ./simd/simd-sse.h -1 +2 cannam@95: cannam@95: Tue Jan 17 11:48:55 EST 2006 athena cannam@95: * [project @ 2006-01-17 16:48:55 by athena] cannam@95: Disabled checks that may turn out to be too paranoid. cannam@95: cannam@95: M ./tests/hook.c -1 +2 cannam@95: cannam@95: Tue Jan 17 10:35:03 EST 2006 athena cannam@95: * [project @ 2006-01-17 15:35:03 by athena] cannam@95: Some paranoid checks. cannam@95: cannam@95: M ./tests/hook.c +16 cannam@95: cannam@95: Tue Jan 17 09:31:08 EST 2006 athena cannam@95: * [project @ 2006-01-17 14:31:08 by athena] cannam@95: Flush stdout after printing. cannam@95: cannam@95: M ./libbench2/ovtpvt.c +1 cannam@95: cannam@95: Tue Jan 17 08:28:18 EST 2006 athena cannam@95: * [project @ 2006-01-17 13:28:18 by athena] cannam@95: Run the leak detector in all cases, not just when verbose > 2. cannam@95: cannam@95: M ./kernel/alloc.c -2 +7 cannam@95: M ./tests/bench.c -2 +1 cannam@95: cannam@95: Tue Jan 17 08:11:41 EST 2006 athena cannam@95: * [project @ 2006-01-17 13:11:41 by athena] cannam@95: Eliminate calls to pow(), rint(). cannam@95: cannam@95: M ./api/mapflags.c -8 +3 cannam@95: cannam@95: Tue Jan 17 00:45:06 EST 2006 stevenj cannam@95: * [project @ 2006-01-17 05:45:06 by stevenj] cannam@95: put # in first column, for stylistic consistency cannam@95: cannam@95: M ./kernel/ifftw.h -3 +3 cannam@95: cannam@95: Tue Jan 17 00:17:27 EST 2006 athena cannam@95: * [project @ 2006-01-17 05:17:27 by athena] cannam@95: Made timeout part of impatience flags, in order to improve the cannam@95: usability of wisdom. Also, fixed bogus error recovery logic in cannam@95: planner.c:imprt(). cannam@95: cannam@95: M ./api/mapflags.c -1 +36 cannam@95: M ./kernel/ifftw.h -4 +7 cannam@95: M ./kernel/planner.c -23 +70 cannam@95: cannam@95: Mon Jan 16 23:03:34 EST 2006 stevenj cannam@95: * [project @ 2006-01-17 04:03:33 by stevenj] cannam@95: make timelimit < 0 .eq. FFTW_NO_TIMELIMIT cannam@95: cannam@95: M ./api/apiplan.c -1 +1 cannam@95: M ./api/fftw3.h -1 +3 cannam@95: M ./doc/fftw3.texi -5 +9 cannam@95: M ./kernel/planner.c -3 +3 cannam@95: cannam@95: Mon Jan 16 21:52:01 EST 2006 athena cannam@95: * [project @ 2006-01-17 02:52:01 by athena] cannam@95: Eliminated the FFTW_TIMELIMIT flag in favor of this simpler logic: cannam@95: fftw_set_timelimit(0) disables time limit. cannam@95: fftw_set_timelimit(X), X>0 sets the time limit to X. cannam@95: cannam@95: M ./api/apiplan.c -6 +1 cannam@95: M ./api/fftw3.h -2 +1 cannam@95: M ./api/the-planner.c +7 cannam@95: M ./doc/fftw3.texi -30 +27 cannam@95: M ./kernel/planner.c -2 +4 cannam@95: M ./tests/bench.c -1 cannam@95: cannam@95: Mon Jan 16 08:38:04 EST 2006 athena cannam@95: * [project @ 2006-01-16 13:38:04 by athena] cannam@95: Force the use of the estimator when wisdom fails because of md5 cannam@95: collisions, otherwise the planner takes forever. cannam@95: cannam@95: M ./api/apiplan.c -2 +10 cannam@95: cannam@95: Sun Jan 15 21:30:31 EST 2006 athena cannam@95: * [project @ 2006-01-16 02:30:31 by athena] cannam@95: Ranted about how broken gcc-4 is. cannam@95: cannam@95: M ./kernel/ifftw.h -5 +40 cannam@95: cannam@95: Sun Jan 15 19:59:38 EST 2006 stevenj cannam@95: * [project @ 2006-01-16 00:59:38 by stevenj] cannam@95: change fftw_timelimit global var to fftw_set_timelimit(double) function, for simpler usage with shared libraries and for consistency with e.g. set_numthreads cannam@95: cannam@95: M ./api/apiplan.c -2 +4 cannam@95: M ./api/fftw3.h -2 +3 cannam@95: M ./doc/fftw3.texi -11 +11 cannam@95: M ./tests/bench.c -1 +1 cannam@95: cannam@95: Sun Jan 15 19:32:27 EST 2006 athena cannam@95: * [project @ 2006-01-16 00:32:27 by athena] cannam@95: Minor tweaks. cannam@95: cannam@95: M ./doc/fftw3.texi -8 +11 cannam@95: cannam@95: Sun Jan 15 16:32:54 EST 2006 athena cannam@95: * [project @ 2006-01-15 21:32:54 by athena] cannam@95: tweaks to make sure that time_n() is always called from the same stack position. cannam@95: cannam@95: M ./libbench2/timer.c -12 +12 cannam@95: cannam@95: Sun Jan 15 16:09:53 EST 2006 athena cannam@95: * [project @ 2006-01-15 21:09:53 by athena] cannam@95: Major simplification of the timer calibration logic. Also, use an FFT cannam@95: as a unit of work instead of the old pointer chasing, because God cannam@95: knows how pointer chasing interacts with the idiotic cache-hit cannam@95: speculation on the Pentium IV. cannam@95: cannam@95: M ./libbench2/bench.h -2 +3 cannam@95: M ./libbench2/speed.c -5 +8 cannam@95: M ./libbench2/timer.c -96 +35 cannam@95: M ./libbench2/timer2.c -7 +99 cannam@95: cannam@95: Sun Jan 15 15:12:08 EST 2006 athena cannam@95: * [project @ 2006-01-15 20:12:08 by athena] cannam@95: Fixed broken aligment checks when sizeof(R)==12. cannam@95: cannam@95: M ./kernel/align.c -3 +8 cannam@95: cannam@95: Sun Jan 15 10:36:40 EST 2006 athena cannam@95: * [project @ 2006-01-15 15:36:40 by athena] cannam@95: Manual unrolling of loop. cannam@95: cannam@95: M ./libbench2/timer2.c -5 +2 cannam@95: cannam@95: Sun Jan 15 10:12:55 EST 2006 athena cannam@95: * [project @ 2006-01-15 15:12:55 by athena] cannam@95: Various improvements to timer calibration routines. cannam@95: cannam@95: A ./libbench2/timer2.c cannam@95: M ./libbench2/Makefile.am -1 +1 cannam@95: M ./libbench2/bench.h -1 +2 cannam@95: M ./libbench2/timer.c -18 +32 cannam@95: M ./libbench2/timer2.c +36 cannam@95: cannam@95: Sat Jan 14 22:16:09 EST 2006 athena cannam@95: * [project @ 2006-01-15 03:16:09 by athena] cannam@95: cygwin defines __CYGWIN__, not __WIN32__ etc. cannam@95: cannam@95: M ./libbench2/timer.c -16 +17 cannam@95: cannam@95: Sat Jan 14 20:40:12 EST 2006 athena cannam@95: * [project @ 2006-01-15 01:40:12 by athena] cannam@95: fixed confusion between libbench and user timers cannam@95: cannam@95: M ./libbench2/bench-user.h -3 +7 cannam@95: M ./libbench2/speed.c -5 +5 cannam@95: M ./libbench2/timer.c -6 +8 cannam@95: M ./tests/bench.c -8 +8 cannam@95: cannam@95: Sat Jan 14 12:32:44 EST 2006 stevenj cannam@95: * [project @ 2006-01-14 17:32:44 by stevenj] cannam@95: update cannam@95: cannam@95: M ./NEWS -2 +4 cannam@95: cannam@95: Sat Jan 14 10:24:11 EST 2006 athena cannam@95: * [project @ 2006-01-14 15:24:11 by athena] cannam@95: Comment. cannam@95: cannam@95: M ./simd/simd-sse.h +5 cannam@95: cannam@95: Sat Jan 14 10:19:28 EST 2006 athena cannam@95: * [project @ 2006-01-14 15:19:28 by athena] cannam@95: Workaround gcc bug. cannam@95: cannam@95: M ./simd/simd-sse.h -2 +3 cannam@95: cannam@95: Fri Jan 13 19:13:18 EST 2006 athena cannam@95: * [project @ 2006-01-14 00:13:18 by athena] cannam@95: Switched to -beta2. cannam@95: cannam@95: M ./configure.ac -1 +1 cannam@95: cannam@95: Thu Jan 12 22:21:57 EST 2006 athena cannam@95: * [project @ 2006-01-13 03:21:57 by athena] cannam@95: Fixed technically correct but highly obfuscated use of the enum tag cannam@95: R2HC as a null pointer. cannam@95: cannam@95: M ./rdft/buffered.c -5 +3 cannam@95: M ./rdft/indirect.c -4 +3 cannam@95: M ./rdft/problem.c -1 +8 cannam@95: M ./rdft/rank0-rdft2.c -4 +3 cannam@95: M ./rdft/rdft.h +1 cannam@95: M ./rdft/vrank3-transpose.c -25 +13 cannam@95: cannam@95: Thu Jan 12 19:25:20 EST 2006 stevenj cannam@95: * [project @ 2006-01-13 00:25:20 by stevenj] cannam@95: --enable-unsafe-mulmod is obsolete cannam@95: cannam@95: M ./configure.ac -4 cannam@95: cannam@95: Thu Jan 12 19:23:18 EST 2006 athena cannam@95: * [project @ 2006-01-13 00:23:18 by athena] cannam@95: More thoughts. cannam@95: cannam@95: M ./TODO +2 cannam@95: cannam@95: Thu Jan 12 19:17:57 EST 2006 athena cannam@95: * [project @ 2006-01-13 00:17:57 by athena] cannam@95: Removed loop unrolling because it slows things down on at least one cannam@95: powerpc and it generates clumsy x86 code. cannam@95: cannam@95: M ./rdft/buffered2.c -50 +12 cannam@95: cannam@95: Thu Jan 12 19:17:35 EST 2006 stevenj cannam@95: * [project @ 2006-01-13 00:17:35 by stevenj] cannam@95: tweaks cannam@95: cannam@95: M ./kernel/kalloc.c -6 +6 cannam@95: cannam@95: Thu Jan 12 15:55:52 EST 2006 stevenj cannam@95: * [project @ 2006-01-12 20:55:52 by stevenj] cannam@95: MacOSX x86 ABI specifies that the stack is kept 16-byte aligned cannam@95: cannam@95: M ./kernel/ifftw.h -2 +3 cannam@95: cannam@95: Thu Jan 12 12:46:49 EST 2006 athena cannam@95: * [project @ 2006-01-12 17:46:49 by athena] cannam@95: ``ret'' is a reserved word in the evil empire. cannam@95: cannam@95: M ./kernel/cycle.h -5 +5 cannam@95: cannam@95: Thu Jan 12 08:31:43 EST 2006 athena cannam@95: * [project @ 2006-01-12 13:31:09 by athena] cannam@95: Changed ret => result because ret ``is a reserved word'' in the evil cannam@95: empire. cannam@95: cannam@95: M ./simd/sse.c -4 +4 cannam@95: M ./simd/sse2.c -4 +4 cannam@95: cannam@95: Wed Jan 11 19:30:42 EST 2006 athena cannam@95: * [project @ 2006-01-12 00:30:42 by athena] cannam@95: Workaround Visual c++ lossage. cannam@95: cannam@95: M ./simd/simd-sse2.h -5 +4 cannam@95: cannam@95: Wed Jan 11 19:26:16 EST 2006 athena cannam@95: * [project @ 2006-01-12 00:26:16 by athena] cannam@95: Workaround visual c++ lossage. cannam@95: cannam@95: M ./simd/simd-sse.h +38 cannam@95: cannam@95: Wed Jan 11 19:10:52 EST 2006 athena cannam@95: * [project @ 2006-01-12 00:10:52 by athena] cannam@95: isprint() is guaranteed to work for unsigned char + EOF only. cannam@95: cannam@95: M ./libbench2/getopt-utils.c -2 +3 cannam@95: cannam@95: Wed Jan 11 13:47:49 EST 2006 stevenj cannam@95: * [project @ 2006-01-11 18:47:49 by stevenj] cannam@95: rm obsolete fixme cannam@95: cannam@95: M ./rdft/vrank3-transpose.c -2 +2 cannam@95: cannam@95: Wed Jan 11 13:38:46 EST 2006 stevenj cannam@95: * [project @ 2006-01-11 18:38:46 by stevenj] cannam@95: cannam@95: M ./rdft/vrank3-transpose.c -2 +2 cannam@95: cannam@95: Wed Jan 11 13:32:26 EST 2006 stevenj cannam@95: * [project @ 2006-01-11 18:32:26 by stevenj] cannam@95: fix comment cannam@95: cannam@95: M ./rdft/vrank3-transpose.c -2 +3 cannam@95: cannam@95: Wed Jan 11 12:27:05 EST 2006 athena cannam@95: * [project @ 2006-01-11 17:27:05 by athena] cannam@95: Paranoid use of K(x) for all constants x, to avoid runtime double->float conversions on sufficiently stupid compilers. cannam@95: cannam@95: M ./dft/bluestein.c -2 +2 cannam@95: M ./rdft/buffered2.c -2 +2 cannam@95: M ./rdft/dht-rader.c -1 +1 cannam@95: M ./rdft/rank0-rdft2.c -11 +11 cannam@95: M ./reodft/rodft00e-r2hc-pad.c -3 +3 cannam@95: cannam@95: Tue Jan 10 20:10:38 EST 2006 athena cannam@95: * [project @ 2006-01-11 01:10:38 by athena] cannam@95: Workaround to gcc nonsense. cannam@95: cannam@95: M ./simd/simd-sse.h -1 +15 cannam@95: cannam@95: Tue Jan 10 18:44:28 EST 2006 stevenj cannam@95: * [project @ 2006-01-10 23:44:28 by stevenj] cannam@95: bug fix: infinite loop in transpose-cut planning cannam@95: cannam@95: M ./rdft/vrank3-transpose.c -6 +14 cannam@95: cannam@95: Tue Jan 10 18:12:14 EST 2006 stevenj cannam@95: * [project @ 2006-01-10 23:12:14 by stevenj] cannam@95: clarified comment cannam@95: cannam@95: M ./api/fftw3.h -6 +6 cannam@95: cannam@95: Tue Jan 10 18:10:32 EST 2006 stevenj cannam@95: * [project @ 2006-01-10 23:10:32 by stevenj] cannam@95: more Windows decorations cannam@95: cannam@95: M ./tests/bench.c -1 +1 cannam@95: cannam@95: Tue Jan 10 17:57:45 EST 2006 stevenj cannam@95: * [project @ 2006-01-10 22:57:45 by stevenj] cannam@95: added FIXME comment cannam@95: cannam@95: M ./support/Makefile.codelets +1 cannam@95: cannam@95: Tue Jan 10 17:52:07 EST 2006 stevenj cannam@95: * [project @ 2006-01-10 22:52:07 by stevenj] cannam@95: 'make clean' should not delete codlist.c since it is included in the dist tarball cannam@95: cannam@95: M ./support/Makefile.codelets -2 +3 cannam@95: cannam@95: Tue Jan 10 17:50:12 EST 2006 athena cannam@95: * [project @ 2006-01-10 22:50:12 by athena] cannam@95: Change threshold for ``large'' Cooley-Tukey to 256K from 64K, since it cannam@95: seems to benefit the Pentium IV with sse and the planning cost is not cannam@95: too horrible. cannam@95: cannam@95: M ./dft/dftw-direct.c -2 +2 cannam@95: cannam@95: Tue Jan 10 17:45:11 EST 2006 stevenj cannam@95: * [project @ 2006-01-10 22:45:11 by stevenj] cannam@95: more missing Windows DLL decorations cannam@95: cannam@95: M ./kernel/ifftw.h -3 +3 cannam@95: cannam@95: Tue Jan 10 17:41:28 EST 2006 stevenj cannam@95: * [project @ 2006-01-10 22:41:28 by stevenj] cannam@95: remove unused var cannam@95: cannam@95: M ./rdft/dht-rader.c -2 cannam@95: cannam@95: Tue Jan 10 14:00:50 EST 2006 stevenj cannam@95: * [project @ 2006-01-10 19:00:50 by stevenj] cannam@95: allow compiler threads, if enabled, to take precedence over explicit threads cannam@95: cannam@95: M ./threads/threads.c -17 +17 cannam@95: cannam@95: Tue Jan 10 12:30:09 EST 2006 stevenj cannam@95: * [project @ 2006-01-10 17:30:09 by stevenj] cannam@95: cannam@95: M ./api/api.h -2 +2 cannam@95: cannam@95: Tue Jan 10 12:21:56 EST 2006 stevenj cannam@95: * [project @ 2006-01-10 17:21:56 by stevenj] cannam@95: cannam@95: M ./kernel/planner.c -2 +2 cannam@95: cannam@95: Tue Jan 10 09:13:20 EST 2006 athena cannam@95: * [project @ 2006-01-10 14:13:20 by athena] cannam@95: Fixed comment typo. cannam@95: cannam@95: M ./kernel/planner.c -2 +2 cannam@95: cannam@95: Tue Jan 10 08:59:22 EST 2006 athena cannam@95: * [project @ 2006-01-10 13:59:22 by athena] cannam@95: Rearranged timeout checks so as to eliminate one of them. cannam@95: cannam@95: M ./kernel/planner.c -9 +4 cannam@95: cannam@95: Tue Jan 10 08:56:55 EST 2006 athena cannam@95: * [project @ 2006-01-10 13:56:55 by athena] cannam@95: Converted residual CK() -> A(). cannam@95: cannam@95: M ./kernel/plan.c -2 +2 cannam@95: cannam@95: Tue Jan 10 08:36:13 EST 2006 athena cannam@95: * [project @ 2006-01-10 13:36:13 by athena] cannam@95: Maintain the invariant TIMED_OUT ==> NEED_TIMEOUT_CHECK. cannam@95: cannam@95: M ./kernel/planner.c -5 +10 cannam@95: cannam@95: Tue Jan 10 08:24:41 EST 2006 athena cannam@95: * [project @ 2006-01-10 13:24:41 by athena] cannam@95: silence some 64-bit warnings cannam@95: cannam@95: M ./api/mapflags.c -1 +1 cannam@95: M ./dft/rank-geq2.c -2 +2 cannam@95: M ./dft/vrank-geq1.c -2 +2 cannam@95: M ./kernel/buffered.c -1 +1 cannam@95: M ./kernel/md5.c -1 +1 cannam@95: M ./kernel/scan.c -2 +3 cannam@95: M ./rdft/rank-geq2-rdft2.c -2 +2 cannam@95: M ./rdft/rank-geq2.c -2 +2 cannam@95: M ./rdft/vrank-geq1-rdft2.c -2 +2 cannam@95: M ./rdft/vrank-geq1.c -2 +2 cannam@95: M ./threads/dft-vrank-geq1.c -2 +2 cannam@95: M ./threads/rdft-vrank-geq1.c -2 +2 cannam@95: M ./threads/vrank-geq1-rdft2.c -2 +2 cannam@95: cannam@95: Tue Jan 10 07:58:48 EST 2006 athena cannam@95: * [project @ 2006-01-10 12:58:48 by athena] cannam@95: Assertions. cannam@95: cannam@95: M ./tests/hook.c +3 cannam@95: cannam@95: Tue Jan 10 00:14:00 EST 2006 stevenj cannam@95: * [project @ 2006-01-10 05:14:00 by stevenj] cannam@95: some condensing cannam@95: cannam@95: M ./kernel/timer.c -10 +5 cannam@95: cannam@95: Tue Jan 10 00:03:32 EST 2006 stevenj cannam@95: * [project @ 2006-01-10 05:03:32 by stevenj] cannam@95: eliminate X(seconds) in favor of X(elapsed_since), in paranoia of clock wrap cannam@95: cannam@95: M ./api/apiplan.c -2 +2 cannam@95: M ./kernel/ifftw.h -3 +33 cannam@95: M ./kernel/planner.c -4 +4 cannam@95: M ./kernel/timer.c -58 +25 cannam@95: cannam@95: Mon Jan 9 23:21:21 EST 2006 stevenj cannam@95: * [project @ 2006-01-10 04:21:21 by stevenj] cannam@95: cannam@95: M ./kernel/timer.c -2 +2 cannam@95: cannam@95: Mon Jan 9 23:21:06 EST 2006 stevenj cannam@95: * [project @ 2006-01-10 04:21:06 by stevenj] cannam@95: hmm, a bit more pessimistic about clock wrapping cannam@95: cannam@95: M ./kernel/timer.c -4 +4 cannam@95: cannam@95: Mon Jan 9 23:20:26 EST 2006 athena cannam@95: * [project @ 2006-01-10 04:20:26 by athena] cannam@95: Revert to md5uint = unsigned int whenever possible, so as to cannam@95: avoid wasting space for unsigned long on 64-bit machines. cannam@95: cannam@95: M ./configure.ac +2 cannam@95: M ./kernel/ifftw.h -1 +5 cannam@95: cannam@95: Mon Jan 9 23:12:27 EST 2006 stevenj cannam@95: * [project @ 2006-01-10 04:12:27 by stevenj] cannam@95: note why clock() wrap should not be a concern cannam@95: cannam@95: M ./kernel/timer.c -1 +6 cannam@95: cannam@95: Mon Jan 9 22:57:16 EST 2006 stevenj cannam@95: * [project @ 2006-01-10 03:57:16 by stevenj] cannam@95: bugfix in recent timeout changes - check for case where last solver times out cannam@95: cannam@95: M ./kernel/planner.c -2 +8 cannam@95: cannam@95: Mon Jan 9 22:40:26 EST 2006 stevenj cannam@95: * [project @ 2006-01-10 03:40:26 by stevenj] cannam@95: started changes list from beta cannam@95: cannam@95: M ./NEWS -1 +11 cannam@95: cannam@95: Mon Jan 9 22:34:13 EST 2006 athena cannam@95: * [project @ 2006-01-10 03:34:13 by athena] cannam@95: Paranoia. cannam@95: cannam@95: M ./api/mapflags.c +4 cannam@95: cannam@95: Mon Jan 9 22:27:37 EST 2006 athena cannam@95: * [project @ 2006-01-10 03:27:37 by athena] cannam@95: Paranoid assertions. cannam@95: cannam@95: M ./kernel/planner.c -3 +7 cannam@95: cannam@95: Mon Jan 9 22:13:32 EST 2006 athena cannam@95: * [project @ 2006-01-10 03:13:32 by athena] cannam@95: Added FIXME comment stating the 64-bit uncleaniness of cannam@95: fftw_tensor_to_bench_tensor(). cannam@95: cannam@95: M ./tests/hook.c +1 cannam@95: cannam@95: Mon Jan 9 22:06:05 EST 2006 athena cannam@95: * [project @ 2006-01-10 03:06:05 by athena] cannam@95: Another 64-bit bug. cannam@95: cannam@95: M ./dft/simd/t.c -1 +1 cannam@95: cannam@95: Mon Jan 9 21:54:07 EST 2006 stevenj cannam@95: * [project @ 2006-01-10 02:54:07 by stevenj] cannam@95: more Windows DLL nonsense cannam@95: cannam@95: M ./api/api.h -5 +15 cannam@95: M ./kernel/ifftw.h -3 +5 cannam@95: M ./tests/hook.c +2 cannam@95: cannam@95: Mon Jan 9 21:18:25 EST 2006 stevenj cannam@95: * [project @ 2006-01-10 02:18:25 by stevenj] cannam@95: some additional dllexport tags required to build the test program, due to internal stuff called by hook.c cannam@95: cannam@95: M ./api/api.h -2 +3 cannam@95: M ./kernel/ifftw.h -6 +17 cannam@95: cannam@95: Mon Jan 9 20:31:15 EST 2006 stevenj cannam@95: * [project @ 2006-01-10 01:31:15 by stevenj] cannam@95: cannam@95: M ./api/fftw3.h -2 +2 cannam@95: cannam@95: Mon Jan 9 20:30:19 EST 2006 stevenj cannam@95: * [project @ 2006-01-10 01:30:19 by stevenj] cannam@95: comment cannam@95: cannam@95: M ./api/fftw3.h -4 +6 cannam@95: cannam@95: Mon Jan 9 20:20:28 EST 2006 stevenj cannam@95: * [project @ 2006-01-10 01:17:11 by stevenj] cannam@95: cannam@95: M ./api/api.h -7 +7 cannam@95: M ./api/fftw3.h -4 +3 cannam@95: cannam@95: Mon Jan 9 20:16:50 EST 2006 stevenj cannam@95: * [project @ 2006-01-10 01:16:50 by stevenj] cannam@95: clarification cannam@95: cannam@95: M ./api/fftw3.h -3 +3 cannam@95: cannam@95: Mon Jan 9 20:12:23 EST 2006 stevenj cannam@95: * [project @ 2006-01-10 01:12:23 by stevenj] cannam@95: define FFTW_DLL if DLL_EXPORT (defined by libtool) is supplied cannam@95: cannam@95: M ./api/api.h +10 cannam@95: cannam@95: Mon Jan 9 20:05:11 EST 2006 stevenj cannam@95: * [project @ 2006-01-10 01:05:11 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./api/fftw3.h -2 +2 cannam@95: cannam@95: Mon Jan 9 20:00:47 EST 2006 stevenj cannam@95: * [project @ 2006-01-10 01:00:47 by stevenj] cannam@95: another stab at Windows DLL mess cannam@95: cannam@95: M ./api/fftw3.h -172 +185 cannam@95: cannam@95: Mon Jan 9 19:23:42 EST 2006 athena cannam@95: * [project @ 2006-01-10 00:23:42 by athena] cannam@95: 64-bit clean SIMD header file. I missed those because sparse cannam@95: does not know vector types. Grrr... cannam@95: cannam@95: M ./simd/simd-altivec.h -11 +11 cannam@95: M ./simd/simd-sse.h -6 +6 cannam@95: M ./simd/simd-sse2.h -4 +4 cannam@95: cannam@95: Mon Jan 9 19:08:36 EST 2006 stevenj cannam@95: * [project @ 2006-01-10 00:08:36 by stevenj] cannam@95: this option is called AC_DISABLE_SHARED in the documentation cannam@95: cannam@95: M ./configure.ac -1 +1 cannam@95: cannam@95: Mon Jan 9 17:34:13 EST 2006 stevenj cannam@95: * [project @ 2006-01-09 22:34:13 by stevenj] cannam@95: fixed --with-gcc-arch to work when cross-compiling cannam@95: cannam@95: M ./m4/ax_gcc_archflag.m4 -3 +4 cannam@95: cannam@95: Mon Jan 9 12:04:04 EST 2006 athena cannam@95: * [project @ 2006-01-09 17:04:04 by athena] cannam@95: Moved the timeout check back into the search loop, sicut erat in cannam@95: principio. This gives us a precise control over the timeout. To cannam@95: avoid the overhead of X(seconds)(), only call X(seconds)() if some cannam@95: time measurement was taken since the last call to X(seconds)(). cannam@95: cannam@95: M ./api/apiplan.c -11 +1 cannam@95: M ./kernel/ifftw.h -1 +2 cannam@95: M ./kernel/planner.c -5 +36 cannam@95: cannam@95: Mon Jan 9 00:07:40 EST 2006 stevenj cannam@95: * [project @ 2006-01-09 05:07:40 by stevenj] cannam@95: comments cannam@95: cannam@95: M ./rdft/vrank3-transpose.c -5 +5 cannam@95: cannam@95: Sun Jan 8 23:58:23 EST 2006 stevenj cannam@95: * [project @ 2006-01-09 04:58:23 by stevenj] cannam@95: generalized transpose-cut routine to be able to call transpose-gcd recursivly; TOMS follow-the-cycles algorithm now seems to be completely superseded cannam@95: cannam@95: M ./rdft/vrank3-transpose.c -111 +142 cannam@95: cannam@95: Sun Jan 8 20:53:18 EST 2006 stevenj cannam@95: * [project @ 2006-01-09 01:53:18 by stevenj] cannam@95: cannam@95: M ./threads/threads.c -3 +2 cannam@95: cannam@95: Sun Jan 8 20:52:16 EST 2006 stevenj cannam@95: * [project @ 2006-01-09 01:52:16 by stevenj] cannam@95: ignore errors from setscope -- POSIX standard does not require PTHREAD_SCOPE_SYSTEM to be supported, and PTHREAD_SCOPE_PROCESS is usually okay in that case cannam@95: cannam@95: M ./threads/threads.c -7 +9 cannam@95: cannam@95: Sun Jan 8 15:58:40 EST 2006 stevenj cannam@95: * [project @ 2006-01-08 20:58:40 by stevenj] cannam@95: added TODO comment cannam@95: cannam@95: M ./rdft/vrank3-transpose.c -1 +7 cannam@95: cannam@95: Sun Jan 8 15:39:28 EST 2006 stevenj cannam@95: * [project @ 2006-01-08 20:39:28 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./rdft/vrank3-transpose.c -8 +18 cannam@95: cannam@95: Sun Jan 8 14:44:23 EST 2006 athena cannam@95: * [project @ 2006-01-08 19:44:23 by athena] cannam@95: Boasted ``much faster altivec performance''. cannam@95: cannam@95: M ./NEWS -1 +1 cannam@95: cannam@95: Sun Jan 8 11:44:52 EST 2006 athena cannam@95: * [project @ 2006-01-08 16:44:52 by athena] cannam@95: Added a new pass to the generator to schedule for the pipeline cannam@95: latency. (This schedule modifies the ``optimal'' cache-oblivious cannam@95: schedule and hence it uses more registers.) cannam@95: cannam@95: This pass is currently: cannam@95: cannam@95: * disabled for non-fma code, under the assumption that this will cannam@95: run on a register-starved fma. cannam@95: cannam@95: * enabled for non-simd fma code, under the assumption that this will cannam@95: run on a processor with 32 or more FP registers. The latency of 4 cannam@95: is conservative and does not introduce too much register pressure. cannam@95: cannam@95: * enabled for simd fma code, under the assumption that this will run cannam@95: on altivec. The latency of 8 seems to produce the best results. cannam@95: cannam@95: M ./configure.ac -1 +1 cannam@95: M ./dft/simd/codelets/Makefile.am -1 +1 cannam@95: M ./genfft/annotate.ml -3 +58 cannam@95: M ./genfft/magic.ml -3 +5 cannam@95: M ./genfft/schedule.ml -33 +5 cannam@95: M ./support/Makefile.codelets -2 +2 cannam@95: M ./support/twovers.sh -1 +1 cannam@95: cannam@95: Sun Jan 8 03:13:53 EST 2006 stevenj cannam@95: * [project @ 2006-01-08 08:13:53 by stevenj] cannam@95: fixed estimator for vrank3-transpose cannam@95: cannam@95: M ./rdft/vrank3-transpose.c -6 +22 cannam@95: cannam@95: Sun Jan 8 02:02:11 EST 2006 stevenj cannam@95: * [project @ 2006-01-08 07:02:11 by stevenj] cannam@95: more detail on VC++ workaround cannam@95: cannam@95: M ./NEWS -1 +2 cannam@95: cannam@95: Sun Jan 8 00:19:19 EST 2006 stevenj cannam@95: * [project @ 2006-01-08 05:19:19 by stevenj] cannam@95: typo cannam@95: cannam@95: M ./rdft/vrank3-transpose.c -2 +2 cannam@95: cannam@95: Sun Jan 8 00:16:20 EST 2006 stevenj cannam@95: * [project @ 2006-01-08 05:16:20 by stevenj] cannam@95: screw it, just use planner for all sub-transposes in vrank3-transpose (still just use memcpy for contiguous copies, though) cannam@95: cannam@95: M ./rdft/vrank3-transpose.c -95 +185 cannam@95: cannam@95: Sat Jan 7 23:13:45 EST 2006 stevenj cannam@95: * [project @ 2006-01-08 04:13:45 by stevenj] cannam@95: add an assert cannam@95: cannam@95: M ./kernel/tile2d.c +2 cannam@95: cannam@95: Sat Jan 7 21:57:34 EST 2006 stevenj cannam@95: * [project @ 2006-01-08 02:57:34 by stevenj] cannam@95: vrank3-transpose now uses planner to decide whether to use cpy2d, cpy2d_tiled, etc. cannam@95: cannam@95: M ./kernel/ifftw.h -1 +7 cannam@95: M ./rdft/rank0.c -7 +5 cannam@95: M ./rdft/vrank3-transpose.c -217 +231 cannam@95: cannam@95: Sat Jan 7 20:57:16 EST 2006 stevenj cannam@95: * [project @ 2006-01-08 01:57:16 by stevenj] cannam@95: too annoying to have isqrt unexpectedly fail for n==0 cannam@95: cannam@95: M ./kernel/primes.c -2 +4 cannam@95: cannam@95: Sat Jan 7 17:49:37 EST 2006 stevenj cannam@95: * [project @ 2006-01-07 22:49:37 by stevenj] cannam@95: clarifications cannam@95: cannam@95: M ./NEWS -1 +1 cannam@95: M ./doc/fftw3.texi -2 +5 cannam@95: cannam@95: Sat Jan 7 16:39:20 EST 2006 stevenj cannam@95: * [project @ 2006-01-07 21:39:20 by stevenj] cannam@95: comment fix cannam@95: cannam@95: M ./rdft/vrank3-transpose.c -4 +4 cannam@95: cannam@95: Sat Jan 7 15:16:22 EST 2006 stevenj cannam@95: * [project @ 2006-01-07 20:16:22 by stevenj] cannam@95: more faq updates cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -8 +9 cannam@95: cannam@95: Sat Jan 7 15:12:16 EST 2006 stevenj cannam@95: * [project @ 2006-01-07 20:12:16 by stevenj] cannam@95: enable fma on hppa, update FAQ entry cannam@95: cannam@95: M ./configure.ac +1 cannam@95: M ./doc/FAQ/fftw-faq.bfnn -3 +11 cannam@95: cannam@95: Sat Aug 12 23:47:06 EDT 2006 Unknown tagger cannam@95: tagged fftw-3-1-beta1 cannam@95: cannam@95: cannam@95: Sat Jan 7 14:06:31 EST 2006 athena cannam@95: * [project @ 2006-01-07 19:06:31 by athena] cannam@95: Accomodate different semantics of 'const' in C and C++ cannam@95: cannam@95: M ./dft/simd/t.c +4 cannam@95: cannam@95: Fri Jan 6 23:40:53 EST 2006 athena cannam@95: * [project @ 2006-01-07 04:40:53 by athena] cannam@95: Altivec is called VMX in IBM land. cannam@95: cannam@95: M ./NEWS -1 +1 cannam@95: cannam@95: Fri Jan 6 23:40:16 EST 2006 athena cannam@95: * [project @ 2006-01-07 04:40:16 by athena] cannam@95: Noted faster altivec support. cannam@95: cannam@95: M ./NEWS +2 cannam@95: cannam@95: Fri Jan 6 21:49:10 EST 2006 stevenj cannam@95: * [project @ 2006-01-07 02:49:10 by stevenj] cannam@95: updated icc flag detection cannam@95: cannam@95: M ./m4/ax_cc_maxopt.m4 -1 +1 cannam@95: cannam@95: Fri Jan 6 10:01:50 EST 2006 athena cannam@95: * [project @ 2006-01-06 15:01:50 by athena] cannam@95: Note ``memoize triggen''. cannam@95: cannam@95: M ./TODO -2 +1 cannam@95: cannam@95: Fri Jan 6 09:36:51 EST 2006 athena cannam@95: * [project @ 2006-01-06 14:36:51 by athena] cannam@95: Use --enable-threads to generate dependencies in the threads/ directory. cannam@95: cannam@95: M ./mkdist.sh -1 +1 cannam@95: cannam@95: Fri Jan 6 09:26:29 EST 2006 athena cannam@95: * [project @ 2006-01-06 14:26:29 by athena] cannam@95: Workaround to icc #defining __GNUC__. cannam@95: cannam@95: M ./kernel/ifftw.h -14 +14 cannam@95: cannam@95: Fri Jan 6 09:21:19 EST 2006 athena cannam@95: * [project @ 2006-01-06 14:21:19 by athena] cannam@95: Switched name to 3.1-beta1. cannam@95: cannam@95: M ./configure.ac -1 +1 cannam@95: cannam@95: Thu Jan 5 23:08:44 EST 2006 athena cannam@95: * [project @ 2006-01-06 04:08:44 by athena] cannam@95: More thoughts. cannam@95: cannam@95: M ./TODO +12 cannam@95: cannam@95: Thu Jan 5 22:30:51 EST 2006 athena cannam@95: * [project @ 2006-01-06 03:30:51 by athena] cannam@95: Note wish that (block_size % 4) == 0. cannam@95: cannam@95: M ./TODO +3 cannam@95: cannam@95: Thu Jan 5 22:19:09 EST 2006 athena cannam@95: * [project @ 2006-01-06 03:19:09 by athena] cannam@95: Check alignment of mstart, mcount in SIMD codelets. cannam@95: cannam@95: M ./dft/codelet-dft.h -2 +3 cannam@95: M ./dft/codelets/t.c -2 +3 cannam@95: M ./dft/ctsq.c -2 +3 cannam@95: M ./dft/dftw-direct.c -11 +17 cannam@95: M ./dft/k7/k7.c -3 +3 cannam@95: M ./dft/simd/q1b.c -1 +3 cannam@95: M ./dft/simd/q1f.c -1 +3 cannam@95: M ./dft/simd/t.c -9 +11 cannam@95: M ./dft/simd/t1s.c -3 +5 cannam@95: M ./threads/ct.c -3 +1 cannam@95: M ./threads/hc2hc.c -2 cannam@95: cannam@95: Thu Jan 5 21:56:19 EST 2006 athena cannam@95: * [project @ 2006-01-06 02:56:19 by athena] cannam@95: Enable threads at bootstrap time, so I get the compiler warnings that cannam@95: I would otherwise ignore. cannam@95: cannam@95: M ./bootstrap.sh -1 +1 cannam@95: cannam@95: Thu Jan 5 18:23:15 EST 2006 athena cannam@95: * [project @ 2006-01-05 23:23:15 by athena] cannam@95: made compilable by c++ cannam@95: cannam@95: M ./threads/dft-vrank-geq1.c -2 +2 cannam@95: M ./threads/rdft-vrank-geq1.c -2 +2 cannam@95: M ./threads/vrank-geq1-rdft2.c -2 +2 cannam@95: cannam@95: Thu Jan 5 17:39:02 EST 2006 athena cannam@95: * [project @ 2006-01-05 22:39:02 by athena] cannam@95: FIXED: incorrect twiddle_shift() cannam@95: cannam@95: M ./kernel/twiddle.c -13 +19 cannam@95: cannam@95: Thu Jan 5 16:01:51 EST 2006 athena cannam@95: * [project @ 2006-01-05 21:01:51 by athena] cannam@95: Replaced remnants of awake flag with the new enum wakefulness type. cannam@95: cannam@95: M ./reodft/redft00e-r2hc.c -4 +5 cannam@95: M ./reodft/reodft11e-r2hc.c -5 +7 cannam@95: M ./reodft/rodft00e-r2hc.c -4 +5 cannam@95: M ./threads/ct.c -4 +4 cannam@95: M ./threads/dft-vrank-geq1.c -3 +3 cannam@95: M ./threads/hc2hc.c -3 +3 cannam@95: M ./threads/rdft-vrank-geq1.c -3 +3 cannam@95: M ./threads/vrank-geq1-rdft2.c -3 +3 cannam@95: cannam@95: Thu Jan 5 11:20:59 EST 2006 athena cannam@95: * [project @ 2006-01-05 16:20:59 by athena] cannam@95: Oops---there is no need to find a free slot. cannam@95: cannam@95: M ./kernel/planner.c -10 +1 cannam@95: cannam@95: Thu Jan 5 09:41:58 EST 2006 athena cannam@95: * [project @ 2006-01-05 14:41:58 by athena] cannam@95: Assertions. cannam@95: cannam@95: M ./kernel/planner.c -1 +3 cannam@95: cannam@95: Thu Jan 5 09:29:55 EST 2006 athena cannam@95: * [project @ 2006-01-05 14:29:55 by athena] cannam@95: Commented the hash table lookup algorithm. cannam@95: cannam@95: M ./kernel/planner.c -10 +14 cannam@95: cannam@95: Thu Jan 5 09:12:00 EST 2006 athena cannam@95: * [project @ 2006-01-05 14:12:00 by athena] cannam@95: Fixed infinite loop in hashtable lookup/insert. Grrr... cannam@95: cannam@95: M ./kernel/planner.c -12 +31 cannam@95: cannam@95: Wed Jan 4 22:04:28 EST 2006 stevenj cannam@95: * [project @ 2006-01-05 03:04:26 by stevenj] cannam@95: updated copyright years to 2006 cannam@95: cannam@95: M ./COPYRIGHT -2 +2 cannam@95: M ./api/api.h -2 +2 cannam@95: M ./api/apiplan.c -2 +2 cannam@95: M ./api/configure.c -2 +2 cannam@95: M ./api/execute-dft-c2r.c -2 +2 cannam@95: M ./api/execute-dft-r2c.c -2 +2 cannam@95: M ./api/execute-dft.c -2 +2 cannam@95: M ./api/execute-r2r.c -2 +2 cannam@95: M ./api/execute-split-dft-c2r.c -2 +2 cannam@95: M ./api/execute-split-dft-r2c.c -2 +2 cannam@95: M ./api/execute-split-dft.c -2 +2 cannam@95: M ./api/execute.c -2 +2 cannam@95: M ./api/export-wisdom-to-file.c -2 +2 cannam@95: M ./api/export-wisdom-to-string.c -2 +2 cannam@95: M ./api/export-wisdom.c -2 +2 cannam@95: M ./api/extract-reim.c -2 +2 cannam@95: M ./api/f77api.c -2 +2 cannam@95: M ./api/f77funcs.h -2 +2 cannam@95: M ./api/fftw3.h -3 +3 cannam@95: M ./api/flops.c -2 +2 cannam@95: M ./api/forget-wisdom.c -2 +2 cannam@95: M ./api/import-system-wisdom.c -2 +2 cannam@95: M ./api/import-wisdom-from-file.c -2 +2 cannam@95: M ./api/import-wisdom-from-string.c -2 +2 cannam@95: M ./api/import-wisdom.c -2 +2 cannam@95: M ./api/malloc.c -2 +2 cannam@95: M ./api/map-r2r-kind.c -2 +2 cannam@95: M ./api/mapflags.c -2 +2 cannam@95: M ./api/mkprinter-file.c -2 +2 cannam@95: M ./api/mktensor-iodims.c -2 +2 cannam@95: M ./api/mktensor-rowmajor.c -2 +2 cannam@95: M ./api/plan-dft-1d.c -2 +2 cannam@95: M ./api/plan-dft-2d.c -2 +2 cannam@95: M ./api/plan-dft-3d.c -2 +2 cannam@95: M ./api/plan-dft-c2r-1d.c -2 +2 cannam@95: M ./api/plan-dft-c2r-2d.c -2 +2 cannam@95: M ./api/plan-dft-c2r-3d.c -2 +2 cannam@95: M ./api/plan-dft-c2r.c -2 +2 cannam@95: M ./api/plan-dft-r2c-1d.c -2 +2 cannam@95: M ./api/plan-dft-r2c-2d.c -2 +2 cannam@95: M ./api/plan-dft-r2c-3d.c -2 +2 cannam@95: M ./api/plan-dft-r2c.c -2 +2 cannam@95: M ./api/plan-dft.c -2 +2 cannam@95: M ./api/plan-guru-dft-c2r.c -2 +2 cannam@95: M ./api/plan-guru-dft-r2c.c -2 +2 cannam@95: M ./api/plan-guru-dft.c -2 +2 cannam@95: M ./api/plan-guru-r2r.c -2 +2 cannam@95: M ./api/plan-guru-split-dft-c2r.c -2 +2 cannam@95: M ./api/plan-guru-split-dft-r2c.c -2 +2 cannam@95: M ./api/plan-guru-split-dft.c -2 +2 cannam@95: M ./api/plan-many-dft-c2r.c -2 +2 cannam@95: M ./api/plan-many-dft-r2c.c -2 +2 cannam@95: M ./api/plan-many-dft.c -2 +2 cannam@95: M ./api/plan-many-r2r.c -2 +2 cannam@95: M ./api/plan-r2r-1d.c -2 +2 cannam@95: M ./api/plan-r2r-2d.c -2 +2 cannam@95: M ./api/plan-r2r-3d.c -2 +2 cannam@95: M ./api/plan-r2r.c -2 +2 cannam@95: M ./api/print-plan.c -2 +2 cannam@95: M ./api/rdft2-pad.c -2 +2 cannam@95: M ./api/the-planner.c -2 +2 cannam@95: M ./api/version.c -3 +3 cannam@95: M ./api/x77.h -2 +2 cannam@95: M ./dft/bluestein.c -2 +2 cannam@95: M ./dft/buffered.c -3 +3 cannam@95: M ./dft/codelet-dft.h -3 +3 cannam@95: M ./dft/codelets/n.c -2 +2 cannam@95: M ./dft/codelets/n.h -2 +2 cannam@95: M ./dft/codelets/t.c -2 +2 cannam@95: M ./dft/codelets/t.h -2 +2 cannam@95: M ./dft/conf.c -3 +3 cannam@95: M ./dft/ct.c -3 +3 cannam@95: M ./dft/ct.h -2 +2 cannam@95: M ./dft/ctsq.c -3 +3 cannam@95: M ./dft/dft.h -3 +3 cannam@95: M ./dft/dftw-direct.c -3 +3 cannam@95: M ./dft/dftw-generic.c -2 +2 cannam@95: M ./dft/dftw-genericbuf.c -2 +2 cannam@95: M ./dft/direct.c -3 +3 cannam@95: M ./dft/generic.c -2 +2 cannam@95: M ./dft/indirect-transpose.c -2 +2 cannam@95: M ./dft/indirect.c -3 +3 cannam@95: M ./dft/k7/k7.c -3 +3 cannam@95: M ./dft/kdft-dif.c -3 +3 cannam@95: M ./dft/kdft-difsq.c -3 +3 cannam@95: M ./dft/kdft-dit.c -3 +3 cannam@95: M ./dft/kdft.c -3 +3 cannam@95: M ./dft/nop.c -3 +3 cannam@95: M ./dft/plan.c -3 +3 cannam@95: M ./dft/problem.c -3 +3 cannam@95: M ./dft/rader.c -2 +2 cannam@95: M ./dft/rank-geq2.c -3 +3 cannam@95: M ./dft/simd/n1b.c -2 +2 cannam@95: M ./dft/simd/n1b.h -2 +2 cannam@95: M ./dft/simd/n1f.c -2 +2 cannam@95: M ./dft/simd/n1f.h -2 +2 cannam@95: M ./dft/simd/n2b.c -2 +2 cannam@95: M ./dft/simd/n2b.h -2 +2 cannam@95: M ./dft/simd/n2f.c -2 +2 cannam@95: M ./dft/simd/n2f.h -2 +2 cannam@95: M ./dft/simd/n2s.c -2 +2 cannam@95: M ./dft/simd/n2s.h -2 +2 cannam@95: M ./dft/simd/q1b.c -2 +2 cannam@95: M ./dft/simd/q1b.h -2 +2 cannam@95: M ./dft/simd/q1f.c -2 +2 cannam@95: M ./dft/simd/q1f.h -2 +2 cannam@95: M ./dft/simd/t.c -2 +2 cannam@95: M ./dft/simd/t1b.h -2 +2 cannam@95: M ./dft/simd/t1f.h -2 +2 cannam@95: M ./dft/simd/t1s.c -2 +2 cannam@95: M ./dft/simd/t1s.h -2 +2 cannam@95: M ./dft/simd/t2b.h -2 +2 cannam@95: M ./dft/simd/t2f.h -2 +2 cannam@95: M ./dft/solve.c -3 +3 cannam@95: M ./dft/vrank-geq1.c -3 +3 cannam@95: M ./dft/zero.c -3 +3 cannam@95: M ./doc/f77_wisdom.f -2 +2 cannam@95: M ./doc/fftw3.texi -1 +1 cannam@95: M ./genfft/algsimp.ml -4 +4 cannam@95: M ./genfft/algsimp.mli -3 +3 cannam@95: M ./genfft/annotate.ml -4 +4 cannam@95: M ./genfft/annotate.mli -3 +3 cannam@95: M ./genfft/assoctable.ml -3 +3 cannam@95: M ./genfft/assoctable.mli -3 +3 cannam@95: M ./genfft/c.ml -3 +3 cannam@95: M ./genfft/c.mli -3 +3 cannam@95: M ./genfft/complex.ml -3 +3 cannam@95: M ./genfft/complex.mli -3 +3 cannam@95: M ./genfft/conv.ml -2 +2 cannam@95: M ./genfft/conv.mli -3 +3 cannam@95: M ./genfft/dag.ml -3 +3 cannam@95: M ./genfft/dag.mli -3 +3 cannam@95: M ./genfft/expr.ml -3 +3 cannam@95: M ./genfft/expr.mli -3 +3 cannam@95: M ./genfft/fft.ml -4 +4 cannam@95: M ./genfft/fft.mli -3 +3 cannam@95: M ./genfft/gen_athnotw.ml -4 +4 cannam@95: M ./genfft/gen_athtw.ml -4 +4 cannam@95: M ./genfft/gen_conv.ml -4 +4 cannam@95: M ./genfft/gen_hc2hc.ml -4 +4 cannam@95: M ./genfft/gen_hc2r.ml -4 +4 cannam@95: M ./genfft/gen_mdct.ml -4 +4 cannam@95: M ./genfft/gen_notw.ml -4 +4 cannam@95: M ./genfft/gen_notw_c.ml -4 +4 cannam@95: M ./genfft/gen_r2hc.ml -4 +4 cannam@95: M ./genfft/gen_r2r.ml -4 +4 cannam@95: M ./genfft/gen_twiddle.ml -4 +4 cannam@95: M ./genfft/gen_twiddle_c.ml -4 +4 cannam@95: M ./genfft/gen_twidsq.ml -4 +4 cannam@95: M ./genfft/gen_twidsq_c.ml -4 +4 cannam@95: M ./genfft/genutil.ml -3 +3 cannam@95: M ./genfft/littlesimp.ml -3 +3 cannam@95: M ./genfft/littlesimp.mli -3 +3 cannam@95: M ./genfft/magic.ml -3 +3 cannam@95: M ./genfft/monads.ml -3 +3 cannam@95: M ./genfft/number.ml -3 +3 cannam@95: M ./genfft/number.mli -3 +3 cannam@95: M ./genfft/oracle.ml -3 +3 cannam@95: M ./genfft/oracle.mli -3 +3 cannam@95: M ./genfft/schedule.ml -3 +3 cannam@95: M ./genfft/schedule.mli -3 +3 cannam@95: M ./genfft/simd.ml -3 +3 cannam@95: M ./genfft/simd.mli -3 +3 cannam@95: M ./genfft/simdmagic.ml -3 +3 cannam@95: M ./genfft/to_alist.ml -3 +3 cannam@95: M ./genfft/to_alist.mli -3 +3 cannam@95: M ./genfft/trig.ml -3 +3 cannam@95: M ./genfft/trig.mli -3 +3 cannam@95: M ./genfft/twiddle.ml -3 +3 cannam@95: M ./genfft/twiddle.mli -3 +3 cannam@95: M ./genfft/unique.ml -3 +3 cannam@95: M ./genfft/unique.mli -3 +3 cannam@95: M ./genfft/util.ml -3 +3 cannam@95: M ./genfft/util.mli -3 +3 cannam@95: M ./genfft/variable.ml -3 +3 cannam@95: M ./genfft/variable.mli -3 +3 cannam@95: M ./genfft-k7/algsimp.ml -4 +4 cannam@95: M ./genfft-k7/algsimp.mli -3 +3 cannam@95: M ./genfft-k7/assoctable.ml -3 +3 cannam@95: M ./genfft-k7/assoctable.mli -3 +3 cannam@95: M ./genfft-k7/complex.ml -1 +1 cannam@95: M ./genfft-k7/complex.mli -1 +1 cannam@95: M ./genfft-k7/expr.ml -3 +3 cannam@95: M ./genfft-k7/expr.mli -3 +3 cannam@95: M ./genfft-k7/fft.ml -4 +4 cannam@95: M ./genfft-k7/gen_notw.ml -1 +1 cannam@95: M ./genfft-k7/littlesimp.ml -3 +3 cannam@95: M ./genfft-k7/littlesimp.mli -3 +3 cannam@95: M ./genfft-k7/monads.ml -3 +3 cannam@95: M ./genfft-k7/number.ml -3 +3 cannam@95: M ./genfft-k7/number.mli -3 +3 cannam@95: M ./genfft-k7/oracle.ml -3 +3 cannam@95: M ./genfft-k7/oracle.mli -3 +3 cannam@95: M ./genfft-k7/to_alist.ml -3 +3 cannam@95: M ./genfft-k7/to_alist.mli -3 +3 cannam@95: M ./genfft-k7/twiddle.ml -3 +3 cannam@95: M ./genfft-k7/twiddle.mli -3 +3 cannam@95: M ./genfft-k7/vScheduler.mli -1 +1 cannam@95: M ./kernel/align.c -3 +3 cannam@95: M ./kernel/alloc.c -3 +3 cannam@95: M ./kernel/assert.c -3 +3 cannam@95: M ./kernel/awake.c -3 +3 cannam@95: M ./kernel/buffered.c -2 +2 cannam@95: M ./kernel/cpy1d.c -2 +2 cannam@95: M ./kernel/cpy2d-pair.c -2 +2 cannam@95: M ./kernel/cpy2d.c -2 +2 cannam@95: M ./kernel/ct.c -2 +2 cannam@95: M ./kernel/cycle.h -3 +3 cannam@95: M ./kernel/debug.c -3 +3 cannam@95: M ./kernel/hash.c -2 +2 cannam@95: M ./kernel/iabs.c -3 +3 cannam@95: M ./kernel/ifftw.h -3 +3 cannam@95: M ./kernel/kalloc.c -3 +3 cannam@95: M ./kernel/md5-1.c -2 +2 cannam@95: M ./kernel/md5.c -2 +2 cannam@95: M ./kernel/minmax.c -3 +3 cannam@95: M ./kernel/ops.c -3 +3 cannam@95: M ./kernel/pickdim.c -3 +3 cannam@95: M ./kernel/plan.c -3 +3 cannam@95: M ./kernel/primes.c -3 +3 cannam@95: M ./kernel/print.c -3 +3 cannam@95: M ./kernel/problem.c -3 +3 cannam@95: M ./kernel/rader.c -2 +2 cannam@95: M ./kernel/scan.c -3 +3 cannam@95: M ./kernel/solver.c -3 +3 cannam@95: M ./kernel/solvtab.c -3 +3 cannam@95: M ./kernel/stride.c -3 +3 cannam@95: M ./kernel/tensor.c -3 +3 cannam@95: M ./kernel/tensor1.c -3 +3 cannam@95: M ./kernel/tensor2.c -3 +3 cannam@95: M ./kernel/tensor4.c -3 +3 cannam@95: M ./kernel/tensor5.c -3 +3 cannam@95: M ./kernel/tensor7.c -3 +3 cannam@95: M ./kernel/tensor8.c -3 +3 cannam@95: M ./kernel/tensor9.c -3 +3 cannam@95: M ./kernel/tile2d.c -2 +2 cannam@95: M ./kernel/timer.c -3 +3 cannam@95: M ./kernel/transpose.c -2 +2 cannam@95: M ./kernel/trig.c -3 +3 cannam@95: M ./kernel/twiddle.c -3 +3 cannam@95: M ./libbench/accopy-from.c -1 +1 cannam@95: M ./libbench/accopy-to.c -1 +1 cannam@95: M ./libbench/allocate.c -1 +1 cannam@95: M ./libbench/bench-main.c -1 +1 cannam@95: M ./libbench/bench-user.h -1 +1 cannam@95: M ./libbench/bench.h -1 +1 cannam@95: M ./libbench/can-do.c -1 +1 cannam@95: M ./libbench/ccopy-from.c -1 +1 cannam@95: M ./libbench/ccopy-to.c -1 +1 cannam@95: M ./libbench/deallocate.c -1 +1 cannam@95: M ./libbench/getopt-utils.c -1 +1 cannam@95: M ./libbench/info.c -1 +1 cannam@95: M ./libbench/main.c -1 +1 cannam@95: M ./libbench/prime.c -1 +1 cannam@95: M ./libbench/problem.c -1 +1 cannam@95: M ./libbench/report.c -1 +1 cannam@95: M ./libbench/speed.c -1 +1 cannam@95: M ./libbench/timer.c -1 +1 cannam@95: M ./libbench/verify.c -1 +1 cannam@95: M ./libbench/zero.c -1 +1 cannam@95: M ./libbench2/aligned-main.c -1 +1 cannam@95: M ./libbench2/allocate.c -1 +1 cannam@95: M ./libbench2/can-do.c -1 +1 cannam@95: M ./libbench2/dotens2.c -3 +3 cannam@95: M ./libbench2/getopt-utils.c -1 +1 cannam@95: M ./libbench2/info.c -1 +1 cannam@95: M ./libbench2/main.c -1 +1 cannam@95: M ./libbench2/report.c -1 +1 cannam@95: M ./libbench2/tensor.c -1 +1 cannam@95: M ./libbench2/useropt.c -1 +1 cannam@95: M ./libbench2/verify-dft.c -3 +3 cannam@95: M ./libbench2/verify-lib.c -3 +3 cannam@95: M ./libbench2/verify-r2r.c -2 +2 cannam@95: M ./libbench2/verify-rdft2.c -3 +3 cannam@95: M ./libbench2/verify.c -1 +1 cannam@95: M ./libbench2/verify.h -2 +2 cannam@95: M ./libbench2/zero.c -1 +1 cannam@95: M ./m4/ax_gcc_archflag.m4 -1 +1 cannam@95: M ./rdft/buffered.c -3 +3 cannam@95: M ./rdft/buffered2.c -3 +3 cannam@95: M ./rdft/codelet-rdft.h -2 +2 cannam@95: M ./rdft/codelets/hb.h -2 +2 cannam@95: M ./rdft/codelets/hc2r.c -2 +2 cannam@95: M ./rdft/codelets/hc2r.h -2 +2 cannam@95: M ./rdft/codelets/hc2rIII.h -2 +2 cannam@95: M ./rdft/codelets/hf.h -2 +2 cannam@95: M ./rdft/codelets/hfb.c -2 +2 cannam@95: M ./rdft/codelets/r2hc.c -2 +2 cannam@95: M ./rdft/codelets/r2hc.h -2 +2 cannam@95: M ./rdft/codelets/r2hcII.h -2 +2 cannam@95: M ./rdft/codelets/r2r.c -2 +2 cannam@95: M ./rdft/codelets/r2r.h -2 +2 cannam@95: M ./rdft/conf.c -3 +3 cannam@95: M ./rdft/dft-r2hc.c -3 +3 cannam@95: M ./rdft/dht-r2hc.c -3 +3 cannam@95: M ./rdft/dht-rader.c -2 +2 cannam@95: M ./rdft/direct.c -3 +3 cannam@95: M ./rdft/direct2.c -3 +3 cannam@95: M ./rdft/generic.c -2 +2 cannam@95: M ./rdft/hc2hc-common.c -2 +2 cannam@95: M ./rdft/hc2hc-direct.c -3 +3 cannam@95: M ./rdft/hc2hc-directbuf.c -3 +3 cannam@95: M ./rdft/hc2hc-generic.c -2 +2 cannam@95: M ./rdft/hc2hc.c -2 +2 cannam@95: M ./rdft/hc2hc.h -2 +2 cannam@95: M ./rdft/indirect.c -3 +3 cannam@95: M ./rdft/khc2hc.c -3 +3 cannam@95: M ./rdft/khc2r.c -3 +3 cannam@95: M ./rdft/kr2hc.c -3 +3 cannam@95: M ./rdft/kr2r.c -3 +3 cannam@95: M ./rdft/nop.c -3 +3 cannam@95: M ./rdft/nop2.c -3 +3 cannam@95: M ./rdft/plan.c -3 +3 cannam@95: M ./rdft/plan2.c -3 +3 cannam@95: M ./rdft/problem.c -3 +3 cannam@95: M ./rdft/problem2.c -3 +3 cannam@95: M ./rdft/rank-geq2-rdft2.c -3 +3 cannam@95: M ./rdft/rank-geq2.c -3 +3 cannam@95: M ./rdft/rank0-rdft2.c -3 +3 cannam@95: M ./rdft/rank0.c -3 +3 cannam@95: M ./rdft/rdft-dht.c -3 +3 cannam@95: M ./rdft/rdft.h -2 +2 cannam@95: M ./rdft/rdft2-inplace-strides.c -3 +3 cannam@95: M ./rdft/rdft2-radix2.c -3 +3 cannam@95: M ./rdft/rdft2-strides.c -2 +2 cannam@95: M ./rdft/rdft2-tensor-max-index.c -3 +3 cannam@95: M ./rdft/solve.c -3 +3 cannam@95: M ./rdft/solve2.c -3 +3 cannam@95: M ./rdft/vrank-geq1-rdft2.c -3 +3 cannam@95: M ./rdft/vrank-geq1.c -3 +3 cannam@95: M ./rdft/vrank3-transpose.c -3 +3 cannam@95: M ./reodft/conf.c -3 +3 cannam@95: M ./reodft/redft00e-r2hc-pad.c -3 +3 cannam@95: M ./reodft/redft00e-r2hc.c -3 +3 cannam@95: M ./reodft/reodft.h -2 +2 cannam@95: M ./reodft/reodft010e-r2hc.c -3 +3 cannam@95: M ./reodft/reodft11e-r2hc-odd.c -3 +3 cannam@95: M ./reodft/reodft11e-r2hc.c -3 +3 cannam@95: M ./reodft/reodft11e-radix2.c -3 +3 cannam@95: M ./reodft/rodft00e-r2hc-pad.c -3 +3 cannam@95: M ./reodft/rodft00e-r2hc.c -3 +3 cannam@95: M ./simd/altivec.c -3 +3 cannam@95: M ./simd/nonportable/sse.c -3 +3 cannam@95: M ./simd/nonportable/sse2.c -3 +3 cannam@95: M ./simd/simd-altivec.h -2 +2 cannam@95: M ./simd/simd-sse.h -2 +2 cannam@95: M ./simd/simd-sse2.h -2 +2 cannam@95: M ./simd/simd.h -2 +2 cannam@95: M ./simd/sse.c -3 +3 cannam@95: M ./simd/sse2.c -3 +3 cannam@95: M ./simd/taint.c -3 +3 cannam@95: M ./threads/api.c -2 +2 cannam@95: M ./threads/conf.c -3 +3 cannam@95: M ./threads/ct.c -3 +3 cannam@95: M ./threads/dft-vrank-geq1.c -3 +3 cannam@95: M ./threads/f77api.c -2 +2 cannam@95: M ./threads/f77funcs.h -2 +2 cannam@95: M ./threads/hc2hc.c -2 +2 cannam@95: M ./threads/rdft-vrank-geq1.c -3 +3 cannam@95: M ./threads/threads.c -2 +2 cannam@95: M ./threads/threads.h -2 +2 cannam@95: M ./threads/vrank-geq1-rdft2.c -3 +3 cannam@95: M ./tools/fftw-wisdom-to-conf.1 -4 +4 cannam@95: M ./tools/fftw-wisdom-to-conf.in -2 +2 cannam@95: M ./tools/fftw-wisdom.c -2 +2 cannam@95: M ./tools/fftw_wisdom.1.in -4 +4 cannam@95: cannam@95: Wed Jan 4 21:57:23 EST 2006 stevenj cannam@95: * [project @ 2006-01-05 02:57:23 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./m4/ax_gcc_archflag.m4 -7 +4 cannam@95: cannam@95: Wed Jan 4 21:52:18 EST 2006 stevenj cannam@95: * [project @ 2006-01-05 02:52:18 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./m4/ax_gcc_archflag.m4 -1 +1 cannam@95: cannam@95: Wed Jan 4 21:51:40 EST 2006 stevenj cannam@95: * [project @ 2006-01-05 02:51:40 by stevenj] cannam@95: more updates for recent pentia/amd cannam@95: cannam@95: M ./m4/ax_gcc_archflag.m4 -5 +12 cannam@95: cannam@95: Wed Jan 4 20:57:47 EST 2006 athena cannam@95: * [project @ 2006-01-05 01:57:47 by athena] cannam@95: Pruned TODO. cannam@95: cannam@95: M ./TODO -6 +2 cannam@95: cannam@95: Wed Jan 4 20:43:41 EST 2006 athena cannam@95: * [project @ 2006-01-05 01:43:41 by athena] cannam@95: Prototype of problem_destroy() cannam@95: cannam@95: M ./libbench2/bench-user.h -1 +2 cannam@95: M ./libbench2/bench.h -2 +1 cannam@95: cannam@95: Wed Jan 4 20:43:13 EST 2006 stevenj cannam@95: * [project @ 2006-01-05 01:43:13 by stevenj] cannam@95: rm obsoleted TODOs cannam@95: cannam@95: M ./TODO -5 cannam@95: cannam@95: Wed Jan 4 20:37:24 EST 2006 athena cannam@95: * [project @ 2006-01-05 01:37:24 by athena] cannam@95: Fallback to 970 if neither -mcpu=power5 nor -mcpu=power4 are supported. cannam@95: cannam@95: M ./m4/ax_gcc_archflag.m4 -3 +3 cannam@95: cannam@95: Wed Jan 4 20:29:07 EST 2006 stevenj cannam@95: * [project @ 2006-01-05 01:29:07 by stevenj] cannam@95: NEWS updates, clarifications, and reorganization cannam@95: cannam@95: M ./NEWS -6 +10 cannam@95: cannam@95: Wed Jan 4 19:54:41 EST 2006 stevenj cannam@95: * [project @ 2006-01-05 00:54:40 by stevenj] cannam@95: remove some compiler warnings, add an assert check, make estimator work properly for nop plans cannam@95: cannam@95: M ./dft/dftw-genericbuf.c +3 cannam@95: M ./kernel/planner.c -1 +2 cannam@95: M ./kernel/trig.c -1 +3 cannam@95: M ./m4/ax_gcc_x86_cpuid.m4 +2 cannam@95: M ./rdft/dft-r2hc.c -1 +2 cannam@95: cannam@95: Tue Jan 3 19:34:04 EST 2006 athena cannam@95: * [project @ 2006-01-04 00:34:03 by athena] cannam@95: Two big changes: cannam@95: cannam@95: 1) revised the twiddle generation machinery, to avoid generating cannam@95: twiddles when measuring, and to use a faster O(sqrt(N)) table cannam@95: when this entails no loss of precision. cannam@95: cannam@95: 2) implemented new ALLOW_PRUNING estimator hack. cannam@95: cannam@95: M ./api/apiplan.c -2 +10 cannam@95: M ./api/fftw3.h -1 +2 cannam@95: M ./api/mapflags.c -1 +4 cannam@95: M ./configure.ac -8 +9 cannam@95: M ./dft/bluestein.c -12 +18 cannam@95: M ./dft/buffered.c -5 +5 cannam@95: M ./dft/ct.c -4 +7 cannam@95: M ./dft/ctsq.c -4 +4 cannam@95: M ./dft/dftw-direct.c -3 +4 cannam@95: M ./dft/dftw-generic.c -131 +19 cannam@95: M ./dft/dftw-genericbuf.c -76 +36 cannam@95: M ./dft/direct.c -1 +2 cannam@95: M ./dft/generic.c -2 +2 cannam@95: M ./dft/indirect-transpose.c -4 +4 cannam@95: M ./dft/indirect.c -4 +4 cannam@95: M ./dft/rader.c -17 +22 cannam@95: M ./dft/rank-geq2.c -4 +4 cannam@95: M ./dft/vrank-geq1.c -3 +3 cannam@95: M ./genfft/twiddle.ml -22 +12 cannam@95: M ./kernel/awake.c -3 +3 cannam@95: M ./kernel/ifftw.h -23 +48 cannam@95: M ./kernel/plan.c -14 +9 cannam@95: M ./kernel/planner.c -1 +8 cannam@95: M ./kernel/timer.c -3 +3 cannam@95: M ./kernel/trig.c -41 +166 cannam@95: M ./kernel/twiddle.c -64 +103 cannam@95: M ./libbench2/bench-main.c -3 +8 cannam@95: M ./libbench2/bench.h -2 +2 cannam@95: M ./libbench2/problem.c -4 +3 cannam@95: M ./libbench2/speed.c -2 +8 cannam@95: M ./rdft/buffered.c -5 +5 cannam@95: M ./rdft/buffered2.c -4 +4 cannam@95: M ./rdft/dft-r2hc.c -3 +3 cannam@95: M ./rdft/dht-r2hc.c -3 +3 cannam@95: M ./rdft/dht-rader.c -16 +23 cannam@95: M ./rdft/direct.c -1 +3 cannam@95: M ./rdft/direct2.c -1 +2 cannam@95: M ./rdft/generic.c -2 +2 cannam@95: M ./rdft/hc2hc-direct.c -5 +7 cannam@95: M ./rdft/hc2hc-directbuf.c -5 +5 cannam@95: M ./rdft/hc2hc-generic.c -6 +7 cannam@95: M ./rdft/hc2hc.c -3 +7 cannam@95: M ./rdft/indirect.c -4 +4 cannam@95: M ./rdft/rank-geq2-rdft2.c -4 +4 cannam@95: M ./rdft/rank-geq2.c -4 +4 cannam@95: M ./rdft/rank0-rdft2.c -3 +3 cannam@95: M ./rdft/rdft-dht.c -3 +3 cannam@95: M ./rdft/rdft2-radix2.c -4 +5 cannam@95: M ./rdft/vrank-geq1-rdft2.c -3 +3 cannam@95: M ./rdft/vrank-geq1.c -3 +3 cannam@95: M ./reodft/redft00e-r2hc-pad.c -4 +4 cannam@95: M ./reodft/reodft00e-splitradix.c -5 +6 cannam@95: M ./reodft/reodft010e-r2hc.c -4 +5 cannam@95: M ./reodft/reodft11e-r2hc-odd.c -3 +3 cannam@95: M ./reodft/reodft11e-radix2.c -5 +7 cannam@95: M ./reodft/rodft00e-r2hc-pad.c -4 +4 cannam@95: M ./tests/hook.c -3 +6 cannam@95: cannam@95: Sat Dec 24 22:08:29 EST 2005 athena cannam@95: * [project @ 2005-12-25 03:08:29 by athena] cannam@95: Estimator tweaks, mostly to favor generic over rader for small n. cannam@95: cannam@95: M ./dft/generic.c +2 cannam@95: M ./rdft/generic.c +2 cannam@95: cannam@95: Sat Dec 24 17:55:47 EST 2005 athena cannam@95: * [project @ 2005-12-24 22:55:47 by athena] cannam@95: Grrr... missing break statement in switch. cannam@95: cannam@95: M ./tests/hook.c +1 cannam@95: cannam@95: Sat Dec 24 16:08:50 EST 2005 athena cannam@95: * [project @ 2005-12-24 21:08:49 by athena] cannam@95: Swapped fields TW and OPS in struct ct_desc_s, to make k7 asm cannam@95: code insensitive to -malign-double. For consistency, changed cannam@95: struct hc2hc_desc_s in the same way. cannam@95: cannam@95: M ./dft/codelet-dft.h -2 +2 cannam@95: M ./genfft/gen_hc2hc.ml -3 +3 cannam@95: M ./genfft/gen_twiddle.ml -3 +3 cannam@95: M ./genfft/gen_twiddle_c.ml -3 +3 cannam@95: M ./genfft/gen_twidsq.ml -3 +3 cannam@95: M ./genfft/gen_twidsq_c.ml -3 +3 cannam@95: M ./genfft-k7/gen_twiddle.ml -3 +2 cannam@95: M ./rdft/codelet-rdft.h -1 +1 cannam@95: cannam@95: Sat Dec 24 16:00:42 EST 2005 athena cannam@95: * [project @ 2005-12-24 21:00:42 by athena] cannam@95: Wrong check for infeasible slvndx in imprt(). cannam@95: cannam@95: M ./kernel/planner.c -4 +4 cannam@95: cannam@95: Sat Dec 24 15:56:59 EST 2005 athena cannam@95: * [project @ 2005-12-24 20:56:59 by athena] cannam@95: Removed obsolete function invoke_solver_if_correct_kind(). cannam@95: cannam@95: M ./kernel/planner.c -10 +1 cannam@95: cannam@95: Sat Dec 24 14:22:12 EST 2005 athena cannam@95: * [project @ 2005-12-24 19:22:12 by athena] cannam@95: Faster implementation of safe_mulmod(), avoiding divisions altogether. cannam@95: Works for 0 <= p <= INT_MAX. cannam@95: cannam@95: M ./kernel/primes.c -13 +22 cannam@95: cannam@95: Sat Dec 24 12:05:54 EST 2005 athena cannam@95: * [project @ 2005-12-24 17:05:54 by athena] cannam@95: FFTW_ALLOW_LARGE_GENERIC must belong to flags->l, it cannot be cannam@95: overridden by fftw. cannam@95: cannam@95: M ./api/mapflags.c -2 +2 cannam@95: cannam@95: Fri Dec 23 20:46:24 EST 2005 stevenj cannam@95: * [project @ 2005-12-24 01:46:24 by stevenj] cannam@95: no more need for limits.h, add some explanatory comments cannam@95: cannam@95: M ./kernel/primes.c -4 +9 cannam@95: cannam@95: Fri Dec 23 17:50:25 EST 2005 athena cannam@95: * [project @ 2005-12-23 22:50:25 by athena] cannam@95: Paranoia. cannam@95: cannam@95: M ./dft/k7/k7.c -1 +8 cannam@95: cannam@95: Fri Dec 23 17:40:41 EST 2005 athena cannam@95: * [project @ 2005-12-23 22:40:41 by athena] cannam@95: Fixed subtle bug involving overflow of the slvndx field in flags_t. cannam@95: cannam@95: M ./kernel/ifftw.h -5 +6 cannam@95: M ./kernel/planner.c -18 +26 cannam@95: cannam@95: Fri Dec 23 16:33:56 EST 2005 athena cannam@95: * [project @ 2005-12-23 21:33:56 by athena] cannam@95: Note 64-bit clean. cannam@95: cannam@95: M ./NEWS +2 cannam@95: cannam@95: Fri Dec 23 15:34:32 EST 2005 athena cannam@95: * [project @ 2005-12-23 20:34:32 by athena] cannam@95: Threads are now 64-bit clean cannam@95: cannam@95: M ./threads/ct.c -12 +13 cannam@95: M ./threads/dft-vrank-geq1.c -20 +17 cannam@95: M ./threads/hc2hc.c -7 +8 cannam@95: M ./threads/rdft-vrank-geq1.c -18 +15 cannam@95: M ./threads/threads.h -2 +2 cannam@95: M ./threads/vrank-geq1-rdft2.c -20 +20 cannam@95: cannam@95: Fri Dec 23 13:00:31 EST 2005 athena cannam@95: * [project @ 2005-12-23 18:00:31 by athena] cannam@95: Restored the old numbering TW_NEXT=3 etc, because the k7 code depends cannam@95: on it. cannam@95: cannam@95: M ./kernel/ifftw.h -3 +3 cannam@95: cannam@95: Fri Dec 23 11:58:00 EST 2005 athena cannam@95: * [project @ 2005-12-23 16:58:00 by athena] cannam@95: Portable implementation of MULMOD() and safe_mulmod(). cannam@95: Removed all unnecessary AC_CHECK_SIZEOF() from configure.ac. cannam@95: cannam@95: M ./configure.ac -5 cannam@95: M ./kernel/ifftw.h -27 +3 cannam@95: M ./kernel/primes.c -8 +3 cannam@95: cannam@95: Thu Dec 22 11:12:29 EST 2005 athena cannam@95: * [project @ 2005-12-22 16:12:29 by athena] cannam@95: Inline the loop body in r2r codelets like we do everywhere else. cannam@95: cannam@95: M ./genfft/gen_r2r.ml -44 +34 cannam@95: cannam@95: Thu Dec 22 10:48:53 EST 2005 athena cannam@95: * [project @ 2005-12-22 15:48:53 by athena] cannam@95: Oops. cannam@95: cannam@95: M ./dft/conf.c -3 +3 cannam@95: cannam@95: Thu Dec 22 10:25:15 EST 2005 athena cannam@95: * [project @ 2005-12-22 15:25:15 by athena] cannam@95: Renamed X(sin_and_cos)() to X(cexp)(). cannam@95: cannam@95: M ./dft/bluestein.c -1 +1 cannam@95: M ./dft/dftw-generic.c -2 +2 cannam@95: M ./dft/dftw-genericbuf.c -2 +2 cannam@95: M ./dft/rader.c -1 +1 cannam@95: M ./kernel/ifftw.h -2 +2 cannam@95: M ./kernel/trig.c -2 +2 cannam@95: M ./kernel/twiddle.c -4 +3 cannam@95: M ./rdft/dht-rader.c -1 +1 cannam@95: cannam@95: Wed Dec 21 22:49:58 EST 2005 athena cannam@95: * [project @ 2005-12-22 03:49:58 by athena] cannam@95: Somewhat faster generation of twiddle factors. cannam@95: cannam@95: M ./dft/bluestein.c -2 +1 cannam@95: M ./dft/conf.c -3 +3 cannam@95: M ./dft/dftw-generic.c -8 +5 cannam@95: M ./dft/dftw-genericbuf.c -8 +6 cannam@95: M ./dft/rader.c -3 +4 cannam@95: M ./kernel/Makefile.am -1 +1 cannam@95: M ./kernel/ifftw.h -4 +3 cannam@95: M ./kernel/trig.c -5 +67 cannam@95: R ./kernel/trig1.c cannam@95: M ./kernel/twiddle.c -11 +15 cannam@95: M ./rdft/dht-rader.c -2 +4 cannam@95: cannam@95: Tue Dec 20 23:50:01 EST 2005 athena cannam@95: * [project @ 2005-12-21 04:50:01 by athena] cannam@95: tweaks cannam@95: cannam@95: M ./kernel/md5.c -2 +1 cannam@95: cannam@95: Tue Dec 20 22:29:19 EST 2005 athena cannam@95: * [project @ 2005-12-21 03:29:19 by athena] cannam@95: Sped up planner, esp. in estimate mode. The planner now classifies cannam@95: all solvers into DFT, RDFT, and RDFT2, and it only invokes solvers cannam@95: appropriate for the problem being planned. Because we have several cannam@95: hundred solvers, the overhead of calling irrelevant solvers is cannam@95: significant, and this modification mitigates the issue somewhat. cannam@95: cannam@95: M ./dft/bluestein.c -14 +10 cannam@95: M ./dft/buffered.c -31 +30 cannam@95: M ./dft/ct.c -16 +13 cannam@95: M ./dft/ctsq.c -23 +20 cannam@95: M ./dft/dft.h -4 +1 cannam@95: M ./dft/direct.c -69 +62 cannam@95: M ./dft/generic.c -12 +8 cannam@95: M ./dft/indirect-transpose.c -23 +19 cannam@95: M ./dft/indirect.c -37 +33 cannam@95: M ./dft/nop.c -16 +15 cannam@95: M ./dft/problem.c -6 +2 cannam@95: M ./dft/rader.c -11 +7 cannam@95: M ./dft/rank-geq2.c -13 +9 cannam@95: M ./dft/vrank-geq1.c -16 +12 cannam@95: M ./kernel/ifftw.h -1 +25 cannam@95: M ./kernel/planner.c -5 +29 cannam@95: M ./rdft/buffered.c -31 +30 cannam@95: M ./rdft/buffered2.c -8 +5 cannam@95: M ./rdft/dft-r2hc.c -10 +6 cannam@95: M ./rdft/dht-r2hc.c -12 +9 cannam@95: M ./rdft/dht-rader.c -13 +9 cannam@95: M ./rdft/direct.c -49 +45 cannam@95: M ./rdft/direct2.c -45 +41 cannam@95: M ./rdft/generic.c -13 +9 cannam@95: M ./rdft/hc2hc.c -19 +15 cannam@95: M ./rdft/indirect.c -30 +26 cannam@95: M ./rdft/nop.c -16 +13 cannam@95: M ./rdft/nop2.c -17 +15 cannam@95: M ./rdft/problem.c -6 +2 cannam@95: M ./rdft/problem2.c -6 +2 cannam@95: M ./rdft/rank-geq2-rdft2.c -20 +16 cannam@95: M ./rdft/rank-geq2.c -13 +9 cannam@95: M ./rdft/rank0-rdft2.c -13 +10 cannam@95: M ./rdft/rank0.c -14 +10 cannam@95: M ./rdft/rdft-dht.c -15 +13 cannam@95: M ./rdft/rdft.h -6 cannam@95: M ./rdft/rdft2-radix2.c -25 +17 cannam@95: M ./rdft/vrank-geq1-rdft2.c -13 +12 cannam@95: M ./rdft/vrank-geq1.c -15 +11 cannam@95: M ./rdft/vrank3-transpose.c -27 +24 cannam@95: M ./reodft/redft00e-r2hc-pad.c -12 +9 cannam@95: M ./reodft/redft00e-r2hc.c -12 +9 cannam@95: M ./reodft/reodft00e-splitradix.c -17 +14 cannam@95: M ./reodft/reodft010e-r2hc.c -12 +9 cannam@95: M ./reodft/reodft11e-r2hc-odd.c -12 +9 cannam@95: M ./reodft/reodft11e-r2hc.c -11 +9 cannam@95: M ./reodft/reodft11e-radix2.c -12 +9 cannam@95: M ./reodft/rodft00e-r2hc-pad.c -12 +8 cannam@95: M ./reodft/rodft00e-r2hc.c -11 +8 cannam@95: M ./tests/hook.c -92 +98 cannam@95: cannam@95: Mon Dec 19 22:04:00 EST 2005 athena cannam@95: * [project @ 2005-12-20 03:04:00 by athena] cannam@95: Eliminated all calls to sprintf() in favor of own routines, so as not cannam@95: to force users to link stdio and the associated locale/pthreads crap. cannam@95: cannam@95: M ./kernel/print.c -35 +50 cannam@95: cannam@95: Mon Dec 19 21:27:25 EST 2005 athena cannam@95: * [project @ 2005-12-20 02:27:25 by athena] cannam@95: Implemented routine to print INT, removing the need for c99's cannam@95: %td format. cannam@95: cannam@95: M ./kernel/ifftw.h -2 +1 cannam@95: M ./kernel/print.c -7 +27 cannam@95: cannam@95: Mon Dec 19 12:06:33 EST 2005 athena cannam@95: * [project @ 2005-12-19 17:06:33 by athena] cannam@95: info->n is size_t cannam@95: cannam@95: M ./kernel/alloc.c -2 +2 cannam@95: cannam@95: Sun Dec 18 18:15:04 EST 2005 athena cannam@95: * [project @ 2005-12-18 23:15:04 by athena] cannam@95: Explicit casts in front of pointer difference in printf() context, cannam@95: just in case INT != ptrdiff_t. cannam@95: cannam@95: M ./configure.ac -1 +1 cannam@95: M ./dft/problem.c -3 +3 cannam@95: M ./rdft/problem.c -2 +2 cannam@95: M ./rdft/problem2.c -3 +3 cannam@95: cannam@95: Sun Dec 18 16:52:38 EST 2005 athena cannam@95: * [project @ 2005-12-18 21:52:38 by athena] cannam@95: Forgot to add %D to print.c cannam@95: cannam@95: M ./kernel/print.c -1 +6 cannam@95: cannam@95: Sun Dec 18 16:43:26 EST 2005 athena cannam@95: * [project @ 2005-12-18 21:43:26 by athena] cannam@95: Use %D as format character for type INT. cannam@95: cannam@95: M ./dft/bluestein.c -1 +1 cannam@95: M ./dft/buffered.c -2 +2 cannam@95: M ./dft/ct.c -2 +2 cannam@95: M ./dft/ctsq.c -2 +2 cannam@95: M ./dft/dftw-direct.c -3 +3 cannam@95: M ./dft/dftw-generic.c -1 +1 cannam@95: M ./dft/dftw-genericbuf.c -1 +1 cannam@95: M ./dft/direct.c -3 +3 cannam@95: M ./dft/generic.c -1 +1 cannam@95: M ./dft/problem.c -2 +2 cannam@95: M ./dft/rader.c -1 +1 cannam@95: M ./dft/vrank-geq1.c -2 +2 cannam@95: M ./kernel/print.c -9 +1 cannam@95: M ./kernel/tensor.c -2 +2 cannam@95: M ./rdft/buffered.c -2 +2 cannam@95: M ./rdft/buffered2.c -2 +2 cannam@95: M ./rdft/dft-r2hc.c -2 +2 cannam@95: M ./rdft/dht-r2hc.c -2 +2 cannam@95: M ./rdft/dht-rader.c -1 +1 cannam@95: M ./rdft/direct.c -2 +2 cannam@95: M ./rdft/direct2.c -2 +2 cannam@95: M ./rdft/generic.c -1 +1 cannam@95: M ./rdft/hc2hc-direct.c -2 +2 cannam@95: M ./rdft/hc2hc-directbuf.c -2 +2 cannam@95: M ./rdft/hc2hc-generic.c -1 +1 cannam@95: M ./rdft/hc2hc.c -1 +1 cannam@95: M ./rdft/problem.c -2 +2 cannam@95: M ./rdft/problem2.c -2 +2 cannam@95: M ./rdft/rank0.c -2 +2 cannam@95: M ./rdft/rdft-dht.c -2 +2 cannam@95: M ./rdft/rdft2-radix2.c -2 +2 cannam@95: M ./rdft/vrank-geq1-rdft2.c -2 +2 cannam@95: M ./rdft/vrank-geq1.c -2 +2 cannam@95: M ./rdft/vrank3-transpose.c -2 +2 cannam@95: M ./reodft/redft00e-r2hc-pad.c -2 +2 cannam@95: M ./reodft/redft00e-r2hc.c -2 +2 cannam@95: M ./reodft/reodft00e-splitradix.c -3 +3 cannam@95: M ./reodft/reodft010e-r2hc.c -2 +2 cannam@95: M ./reodft/reodft11e-r2hc-odd.c -2 +2 cannam@95: M ./reodft/reodft11e-r2hc.c -2 +2 cannam@95: M ./reodft/reodft11e-radix2.c -2 +2 cannam@95: M ./reodft/rodft00e-r2hc-pad.c -2 +2 cannam@95: M ./reodft/rodft00e-r2hc.c -2 +2 cannam@95: cannam@95: Sun Dec 18 15:14:03 EST 2005 athena cannam@95: * [project @ 2005-12-18 20:14:03 by athena] cannam@95: Changed type of an_int_guaranteed_to_be_zero. Changed name as well. cannam@95: cannam@95: M ./kernel/ifftw.h -3 +3 cannam@95: M ./kernel/stride.c -2 +2 cannam@95: cannam@95: Sun Dec 18 14:41:31 EST 2005 athena cannam@95: * [project @ 2005-12-18 19:41:31 by athena] cannam@95: converted %o -> INT cannam@95: cannam@95: M ./kernel/ifftw.h -1 +2 cannam@95: M ./kernel/planner.c -2 +2 cannam@95: M ./kernel/print.c -8 +6 cannam@95: cannam@95: Sat Dec 17 20:28:50 EST 2005 athena cannam@95: * [project @ 2005-12-18 01:28:50 by athena] cannam@95: Major 64-bit cleanup. cannam@95: cannam@95: M ./dft/bluestein.c -14 +15 cannam@95: M ./dft/buffered.c -20 +20 cannam@95: M ./dft/codelet-dft.h -17 +17 cannam@95: M ./dft/codelets/n.c -1 +1 cannam@95: M ./dft/codelets/t.c -1 +1 cannam@95: M ./dft/ct.c -6 +6 cannam@95: M ./dft/ct.h -5 +5 cannam@95: M ./dft/ctsq.c -6 +6 cannam@95: M ./dft/dftw-direct.c -17 +17 cannam@95: M ./dft/dftw-generic.c -28 +28 cannam@95: M ./dft/dftw-genericbuf.c -25 +25 cannam@95: M ./dft/direct.c -16 +14 cannam@95: M ./dft/generic.c -8 +8 cannam@95: M ./dft/indirect-transpose.c -5 +5 cannam@95: M ./dft/problem.c -3 +3 cannam@95: M ./dft/rader.c -10 +10 cannam@95: M ./dft/simd/n1b.c -1 +1 cannam@95: M ./dft/simd/n1f.c -1 +1 cannam@95: M ./dft/simd/n2b.c -1 +1 cannam@95: M ./dft/simd/n2f.c -1 +1 cannam@95: M ./dft/simd/n2s.c -1 +1 cannam@95: M ./dft/simd/q1b.c -1 +1 cannam@95: M ./dft/simd/q1f.c -1 +1 cannam@95: M ./dft/simd/t.c -5 +5 cannam@95: M ./dft/simd/t1s.c -1 +1 cannam@95: M ./dft/vrank-geq1.c -5 +5 cannam@95: M ./dft/zero.c -3 +3 cannam@95: M ./genfft/gen_hc2hc.ml -5 +5 cannam@95: M ./genfft/gen_hc2r.ml -6 +6 cannam@95: M ./genfft/gen_notw.ml -6 +6 cannam@95: M ./genfft/gen_notw_c.ml -6 +6 cannam@95: M ./genfft/gen_r2hc.ml -6 +6 cannam@95: M ./genfft/gen_r2r.ml -6 +6 cannam@95: M ./genfft/gen_twiddle.ml -5 +5 cannam@95: M ./genfft/gen_twiddle_c.ml -5 +5 cannam@95: M ./genfft/gen_twidsq.ml -5 +5 cannam@95: M ./genfft/gen_twidsq_c.ml -5 +5 cannam@95: M ./kernel/buffered.c -2 +2 cannam@95: M ./kernel/cpy1d.c -2 +2 cannam@95: M ./kernel/cpy2d-pair.c -9 +7 cannam@95: M ./kernel/cpy2d.c -20 +19 cannam@95: M ./kernel/ct.c -1 +1 cannam@95: M ./kernel/iabs.c -3 +3 cannam@95: M ./kernel/ifftw.h -81 +99 cannam@95: M ./kernel/md5-1.c -5 +5 cannam@95: M ./kernel/minmax.c -3 +3 cannam@95: M ./kernel/ops.c -4 +4 cannam@95: M ./kernel/planner.c -2 +2 cannam@95: M ./kernel/primes.c -27 +29 cannam@95: M ./kernel/rader.c -3 +3 cannam@95: M ./kernel/solvtab.c -1 +2 cannam@95: M ./kernel/stride.c -3 +3 cannam@95: M ./kernel/tensor.c -7 +8 cannam@95: M ./kernel/tensor1.c -2 +2 cannam@95: M ./kernel/tensor2.c -3 +21 cannam@95: M ./kernel/tensor4.c -8 +8 cannam@95: M ./kernel/tensor7.c -10 +16 cannam@95: M ./kernel/tile2d.c -31 +8 cannam@95: M ./kernel/transpose.c -12 +12 cannam@95: M ./kernel/trig.c -13 +3 cannam@95: M ./kernel/twiddle.c -23 +23 cannam@95: M ./rdft/buffered.c -15 +15 cannam@95: M ./rdft/buffered2.c -30 +30 cannam@95: M ./rdft/codelet-rdft.h -29 +29 cannam@95: M ./rdft/codelets/hc2r.c -1 +1 cannam@95: M ./rdft/codelets/hfb.c -1 +1 cannam@95: M ./rdft/codelets/r2hc.c -1 +1 cannam@95: M ./rdft/codelets/r2r.c -1 +1 cannam@95: M ./rdft/dft-r2hc.c -12 +12 cannam@95: M ./rdft/dht-r2hc.c -5 +5 cannam@95: M ./rdft/dht-rader.c -12 +12 cannam@95: M ./rdft/direct.c -11 +11 cannam@95: M ./rdft/direct2.c -8 +8 cannam@95: M ./rdft/generic.c -14 +14 cannam@95: M ./rdft/hc2hc-common.c -3 +3 cannam@95: M ./rdft/hc2hc-direct.c -13 +13 cannam@95: M ./rdft/hc2hc-directbuf.c -24 +24 cannam@95: M ./rdft/hc2hc-generic.c -48 +31 cannam@95: M ./rdft/hc2hc.c -5 +5 cannam@95: M ./rdft/hc2hc.h -7 +7 cannam@95: M ./rdft/problem.c -3 +3 cannam@95: M ./rdft/problem2.c -2 +2 cannam@95: M ./rdft/rank0-rdft2.c -7 +7 cannam@95: M ./rdft/rank0.c -11 +11 cannam@95: M ./rdft/rdft-dht.c -9 +9 cannam@95: M ./rdft/rdft.h -2 +2 cannam@95: M ./rdft/rdft2-inplace-strides.c -3 +3 cannam@95: M ./rdft/rdft2-radix2.c -23 +23 cannam@95: M ./rdft/rdft2-strides.c -1 +1 cannam@95: M ./rdft/rdft2-tensor-max-index.c -4 +4 cannam@95: M ./rdft/vrank-geq1-rdft2.c -6 +6 cannam@95: M ./rdft/vrank-geq1.c -5 +5 cannam@95: M ./rdft/vrank3-transpose.c -53 +54 cannam@95: M ./reodft/redft00e-r2hc-pad.c -11 +11 cannam@95: M ./reodft/redft00e-r2hc.c -11 +11 cannam@95: M ./reodft/reodft00e-splitradix.c -16 +17 cannam@95: M ./reodft/reodft010e-r2hc.c -26 +26 cannam@95: M ./reodft/reodft11e-r2hc-odd.c -19 +19 cannam@95: M ./reodft/reodft11e-r2hc.c -16 +16 cannam@95: M ./reodft/reodft11e-radix2.c -25 +25 cannam@95: M ./reodft/rodft00e-r2hc-pad.c -11 +11 cannam@95: M ./reodft/rodft00e-r2hc.c -11 +11 cannam@95: M ./simd/sse2.c -2 +3 cannam@95: M ./simd/taint.c -2 +2 cannam@95: cannam@95: Sat Aug 12 23:34:43 EDT 2006 Unknown tagger cannam@95: tagged before-64bit-rewrite cannam@95: cannam@95: cannam@95: Wed Dec 7 22:39:01 EST 2005 stevenj cannam@95: * [project @ 2005-12-08 03:39:01 by stevenj] cannam@95: PGI x86-64 cycle counter, courtesy Cristiano Calonaci cannam@95: cannam@95: M ./kernel/cycle.h -1 +12 cannam@95: cannam@95: Mon Dec 5 21:25:57 EST 2005 athena cannam@95: * [project @ 2005-12-06 02:25:57 by athena] cannam@95: Must insert into hash table when wisdom_state == WISDOM_ONLY, cannam@95: otherwise wisdom does not work. cannam@95: cannam@95: M ./kernel/planner.c -2 +3 cannam@95: cannam@95: Sat Oct 8 18:08:44 EDT 2005 stevenj cannam@95: * [project @ 2005-10-08 22:07:37 by stevenj] cannam@95: comment cannam@95: cannam@95: M ./m4/acx_pthread.m4 -1 +2 cannam@95: cannam@95: Sun Oct 2 11:49:13 EDT 2005 athena cannam@95: * [project @ 2005-10-02 15:49:13 by athena] cannam@95: Paranoia: made planner robust against MD5 collisions. cannam@95: cannam@95: M ./api/apiplan.c -6 +37 cannam@95: M ./kernel/ifftw.h -1 +19 cannam@95: M ./kernel/planner.c -25 +55 cannam@95: cannam@95: Tue Sep 27 22:33:18 EDT 2005 athena cannam@95: * [project @ 2005-09-28 02:33:18 by athena] cannam@95: Note that --enable-3dnow is unsupported. cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -2 +2 cannam@95: cannam@95: Tue Sep 27 22:31:04 EDT 2005 athena cannam@95: * [project @ 2005-09-28 02:31:04 by athena] cannam@95: * Removed --enable-3dnow support. cannam@95: cannam@95: * SIMD support for split complex arrays. cannam@95: cannam@95: M ./NEWS +4 cannam@95: cannam@95: Tue Sep 27 22:28:41 EDT 2005 athena cannam@95: * [project @ 2005-09-28 02:28:40 by athena] cannam@95: Removed --enabled-3dnow, since it is becoming useless as the world cannam@95: moves to x86-64, and it is a pain to maintain. (We should probably cannam@95: remove the k7 stuff as well.) cannam@95: cannam@95: M ./api/version.c -5 +1 cannam@95: M ./configure.ac -9 cannam@95: M ./genfft/gen_notw.ml -6 +4 cannam@95: M ./genfft/gen_notw_c.ml -4 +2 cannam@95: M ./genfft/gen_twiddle.ml -5 +3 cannam@95: M ./genfft/gen_twiddle_c.ml -4 +2 cannam@95: M ./genfft/gen_twidsq_c.ml -4 +2 cannam@95: M ./kernel/align.c -4 +2 cannam@95: M ./kernel/ifftw.h -2 +2 cannam@95: R ./simd/3dnow.c cannam@95: M ./simd/Makefile.am -2 +2 cannam@95: R ./simd/simd-3dnow.h cannam@95: M ./simd/simd-altivec.h -3 cannam@95: M ./simd/simd-sse.h -3 cannam@95: M ./simd/simd-sse2.h -3 cannam@95: M ./simd/simd.h -4 cannam@95: cannam@95: Tue Sep 27 21:59:16 EDT 2005 athena cannam@95: * [project @ 2005-09-28 01:59:16 by athena] cannam@95: Missing BEGIN_SIMD(), END_SIMD() statements. cannam@95: cannam@95: M ./genfft/gen_notw.ml -4 +6 cannam@95: M ./genfft/gen_twiddle.ml -3 +5 cannam@95: cannam@95: Tue Sep 27 12:16:08 EDT 2005 athena cannam@95: * [project @ 2005-09-27 16:16:08 by athena] cannam@95: Tweaks cannam@95: cannam@95: M ./simd/simd-sse.h -12 +9 cannam@95: cannam@95: Tue Sep 27 10:04:32 EDT 2005 athena cannam@95: * [project @ 2005-09-27 14:04:32 by athena] cannam@95: Fixed wrong opcount for simd codelets. cannam@95: cannam@95: M ./genfft/to_alist.ml -9 +9 cannam@95: cannam@95: Tue Sep 27 09:25:50 EDT 2005 athena cannam@95: * [project @ 2005-09-27 13:25:50 by athena] cannam@95: Fixed wrong opcount for simd codelets. cannam@95: cannam@95: M ./dft/dftw-direct.c -2 +2 cannam@95: cannam@95: Mon Sep 26 22:58:19 EDT 2005 athena cannam@95: * [project @ 2005-09-27 02:58:19 by athena] cannam@95: fixed flop counts cannam@95: cannam@95: M ./genfft/c.ml -23 +17 cannam@95: M ./simd/simd-altivec.h -2 +2 cannam@95: M ./simd/simd-sse2.h -9 +9 cannam@95: cannam@95: Mon Sep 26 22:34:40 EDT 2005 athena cannam@95: * [project @ 2005-09-27 02:34:40 by athena] cannam@95: Silence warnings cannam@95: cannam@95: M ./simd/simd-sse2.h +6 cannam@95: cannam@95: Mon Sep 26 20:52:36 EDT 2005 athena cannam@95: * [project @ 2005-09-27 00:52:36 by athena] cannam@95: Implemented split-complex SIMD codelets cannam@95: cannam@95: A ./dft/simd/n2s.c cannam@95: A ./dft/simd/n2s.h cannam@95: A ./dft/simd/t1s.c cannam@95: A ./dft/simd/t1s.h cannam@95: M ./dft/simd/Makefile.am -2 +3 cannam@95: M ./dft/simd/codelets/Makefile.am -6 +18 cannam@95: M ./dft/simd/n2s.c +50 cannam@95: M ./dft/simd/n2s.h +28 cannam@95: M ./dft/simd/t1s.c +48 cannam@95: M ./dft/simd/t1s.h +33 cannam@95: M ./genfft/annotate.ml -3 +3 cannam@95: M ./genfft/c.ml -1 +3 cannam@95: M ./genfft/gen_hc2hc.ml -3 +3 cannam@95: M ./genfft/gen_notw.ml -5 +26 cannam@95: M ./genfft/gen_twiddle.ml -4 +4 cannam@95: M ./genfft/gen_twiddle_c.ml -3 +3 cannam@95: M ./genfft/gen_twidsq_c.ml -3 +3 cannam@95: M ./genfft/genutil.ml -3 +3 cannam@95: M ./genfft/simd.ml -14 +8 cannam@95: M ./genfft/twiddle.ml -3 +3 cannam@95: M ./genfft/twiddle.mli -2 +2 cannam@95: M ./simd/simd-altivec.h -7 +33 cannam@95: M ./simd/simd-sse.h +33 cannam@95: M ./simd/simd-sse2.h -2 +38 cannam@95: M ./simd/simd.h +2 cannam@95: cannam@95: Sun Sep 25 22:25:35 EDT 2005 athena cannam@95: * [project @ 2005-09-26 02:25:35 by athena] cannam@95: Generalized the ``store pairs'' trick (now called ``store multiple''). cannam@95: cannam@95: M ./dft/simd/codelets/Makefile.am -2 +2 cannam@95: M ./genfft/annotate.ml -37 +25 cannam@95: M ./genfft/annotate.mli -4 +2 cannam@95: M ./genfft/expr.ml -6 +6 cannam@95: M ./genfft/expr.mli -2 +2 cannam@95: M ./genfft/gen_notw_c.ml -11 +13 cannam@95: M ./genfft/simd.ml -6 +10 cannam@95: M ./genfft/simdmagic.ml -3 +3 cannam@95: M ./simd/simd-3dnow.h -2 +2 cannam@95: M ./simd/simd-altivec.h -2 +4 cannam@95: M ./simd/simd-sse.h -3 +3 cannam@95: M ./simd/simd-sse2.h -2 +2 cannam@95: cannam@95: Sun Sep 25 18:58:20 EDT 2005 athena cannam@95: * [project @ 2005-09-25 22:58:20 by athena] cannam@95: Silence some warnings. cannam@95: cannam@95: M ./simd/simd-altivec.h -4 +4 cannam@95: cannam@95: Sat Sep 24 12:37:16 EDT 2005 athena cannam@95: * [project @ 2005-09-24 16:37:16 by athena] cannam@95: Removed obsolete cruft cannam@95: cannam@95: M ./simd/simd-altivec.h -6 +6 cannam@95: cannam@95: Mon Sep 19 22:55:19 EDT 2005 athena cannam@95: * [project @ 2005-09-20 02:55:19 by athena] cannam@95: Re-enabled check for because OSX requires it. cannam@95: cannam@95: M ./configure.ac -1 +1 cannam@95: M ./simd/simd-altivec.h +2 cannam@95: cannam@95: Sun Sep 11 11:03:03 EDT 2005 athena cannam@95: * [project @ 2005-09-11 15:03:03 by athena] cannam@95: Check for sizeof(unsigned int) unconditionally, because the cannam@95: result is used by ifftw.h. cannam@95: cannam@95: M ./configure.ac -2 +2 cannam@95: cannam@95: Sun Sep 11 10:59:40 EDT 2005 athena cannam@95: * [project @ 2005-09-11 14:59:40 by athena] cannam@95: Higher size limit for t2 codelets. cannam@95: cannam@95: M ./dft/simd/t.c -1 +1 cannam@95: cannam@95: Sun Sep 11 10:50:37 EDT 2005 athena cannam@95: * [project @ 2005-09-11 14:50:37 by athena] cannam@95: Heuristic: do not use t2 simd codelets for N>1024. cannam@95: cannam@95: A ./dft/simd/t.c cannam@95: M ./dft/simd/Makefile.am -2 +2 cannam@95: M ./dft/simd/t.c +98 cannam@95: R ./dft/simd/t1b.c cannam@95: R ./dft/simd/t1f.c cannam@95: M ./dft/simd/t2b.h -2 +1 cannam@95: M ./dft/simd/t2f.h -2 +1 cannam@95: cannam@95: Mon Sep 5 22:22:50 EDT 2005 athena cannam@95: * [project @ 2005-09-06 02:22:50 by athena] cannam@95: Larger tolerance in timer calibration routine. cannam@95: cannam@95: M ./libbench2/timer.c -2 +2 cannam@95: cannam@95: Mon Sep 5 16:03:33 EDT 2005 athena cannam@95: * [project @ 2005-09-05 20:03:33 by athena] cannam@95: #include unconditionally. (There is no point in checking.) cannam@95: cannam@95: M ./configure.ac -2 +1 cannam@95: M ./simd/simd-altivec.h -2 cannam@95: cannam@95: Mon Sep 5 15:23:27 EDT 2005 athena cannam@95: * [project @ 2005-09-05 19:23:27 by athena] cannam@95: Removed SSE and SSE2 asm because it was bitrotting. Use the Intel cannam@95: API instead, which seems to be supported by gcc >= 3.3. cannam@95: Moved files that require -msse, -msse2 to new directory. cannam@95: cannam@95: A ./simd/nonportable/ cannam@95: A ./simd/nonportable/Makefile.am cannam@95: A ./simd/nonportable/sse.c cannam@95: A ./simd/nonportable/sse2.c cannam@95: M ./Makefile.am +1 cannam@95: M ./configure.ac +1 cannam@95: M ./simd/Makefile.am -4 +3 cannam@95: M ./simd/nonportable/Makefile.am +8 cannam@95: M ./simd/nonportable/sse.c +43 cannam@95: M ./simd/nonportable/sse2.c +43 cannam@95: M ./simd/simd-sse.h -101 +19 cannam@95: M ./simd/simd-sse2.h -75 +7 cannam@95: R ./simd/sse-aux.c cannam@95: M ./simd/sse.c -5 +7 cannam@95: R ./simd/sse2-aux.c cannam@95: M ./simd/sse2.c -5 +7 cannam@95: cannam@95: Mon Sep 5 12:56:28 EDT 2005 athena cannam@95: * [project @ 2005-09-05 16:56:28 by athena] cannam@95: Parse cputypes of the form 7447A,altivecsupported cannam@95: cannam@95: M ./m4/ax_gcc_archflag.m4 -1 +1 cannam@95: cannam@95: Mon Sep 5 12:52:30 EDT 2005 athena cannam@95: * [project @ 2005-09-05 16:52:30 by athena] cannam@95: Distinguish powerpc 7400 from the 7450, which has a different cannam@95: pipeline. cannam@95: cannam@95: M ./m4/ax_gcc_archflag.m4 +2 cannam@95: cannam@95: Mon Sep 5 12:46:00 EDT 2005 athena cannam@95: * [project @ 2005-09-05 16:46:00 by athena] cannam@95: Paranoia: define RIGHT_CPU unconditionally. cannam@95: cannam@95: M ./simd/simd-altivec.h -3 +3 cannam@95: cannam@95: Thu Aug 11 20:56:41 EDT 2005 athena cannam@95: * [project @ 2005-08-12 00:56:41 by athena] cannam@95: Removed obsolete name fftw-wisdom2c. cannam@95: cannam@95: M ./tools/fftw-wisdom-to-conf.in -1 +1 cannam@95: cannam@95: Thu Aug 11 20:55:59 EDT 2005 athena cannam@95: * [project @ 2005-08-12 00:55:59 by athena] cannam@95: Avoid creation of temporary files---use cpp magic instead. cannam@95: This fix solves a security bug and avoids nonportable tempfile cannam@95: creation hacks. cannam@95: cannam@95: M ./tools/fftw-wisdom-to-conf.in -13 +20 cannam@95: cannam@95: Fri Aug 5 10:03:02 EDT 2005 athena cannam@95: * [project @ 2005-08-05 14:03:02 by athena] cannam@95: Workaround for with gcc-3.3 altivec bug. cannam@95: cannam@95: M ./configure.ac -1 +3 cannam@95: M ./simd/altivec.c -3 +2 cannam@95: M ./simd/simd-altivec.h -2 +2 cannam@95: cannam@95: Wed Jun 15 21:36:46 EDT 2005 stevenj cannam@95: * [project @ 2005-06-16 01:36:46 by stevenj] cannam@95: solaris fix: check -pthreads first since gcc does not like -pthread but chokes due to stubbed libc (grr) cannam@95: cannam@95: M ./m4/acx_pthread.m4 -2 +2 cannam@95: cannam@95: Fri Jun 3 17:19:56 EDT 2005 stevenj cannam@95: * [project @ 2005-06-03 21:19:56 by stevenj] cannam@95: note that VC++ bug was fixed in 2005 cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -4 +4 cannam@95: cannam@95: Mon May 30 16:30:45 EDT 2005 stevenj cannam@95: * [project @ 2005-05-30 20:30:32 by stevenj] cannam@95: generalized ax_cc_vendor to ax_compiler_vendor cannam@95: cannam@95: A ./m4/ax_compiler_vendor.m4 cannam@95: M ./configure.ac -2 +2 cannam@95: M ./m4/ax_cc_maxopt.m4 -5 +5 cannam@95: R ./m4/ax_cc_vendor.m4 cannam@95: M ./m4/ax_compiler_vendor.m4 +30 cannam@95: cannam@95: Mon May 30 15:55:07 EDT 2005 stevenj cannam@95: * [project @ 2005-05-30 19:55:07 by stevenj] cannam@95: updated message cannam@95: cannam@95: M ./m4/ax_cc_maxopt.m4 -2 +2 cannam@95: cannam@95: Mon May 30 15:45:14 EDT 2005 stevenj cannam@95: * [project @ 2005-05-30 19:45:14 by stevenj] cannam@95: update for new AC archive format cannam@95: cannam@95: M ./m4/acx_pthread.m4 -1 +4 cannam@95: M ./m4/ax_cc_maxopt.m4 -4 +9 cannam@95: M ./m4/ax_cc_vendor.m4 +6 cannam@95: M ./m4/ax_check_compiler_flags.m4 -1 +4 cannam@95: M ./m4/ax_gcc_aligns_stack.m4 -1 +4 cannam@95: M ./m4/ax_gcc_archflag.m4 -2 +5 cannam@95: M ./m4/ax_gcc_version.m4 -1 +4 cannam@95: M ./m4/ax_gcc_x86_cpuid.m4 -1 +4 cannam@95: M ./m4/ax_openmp.m4 -1 +4 cannam@95: cannam@95: Mon May 23 23:12:22 EDT 2005 stevenj cannam@95: * [project @ 2005-05-24 03:12:22 by stevenj] cannam@95: cannam@95: M ./api/fftw3.h -2 +2 cannam@95: cannam@95: Mon May 23 18:17:38 EDT 2005 stevenj cannam@95: * [project @ 2005-05-23 22:17:38 by stevenj] cannam@95: cannam@95: M ./NEWS -1 +1 cannam@95: cannam@95: Mon May 23 18:13:08 EDT 2005 stevenj cannam@95: * [project @ 2005-05-23 22:13:08 by stevenj] cannam@95: more notes cannam@95: cannam@95: M ./NEWS -1 +9 cannam@95: cannam@95: Sun May 22 23:37:08 EDT 2005 stevenj cannam@95: * [project @ 2005-05-23 03:37:08 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./m4/ax_cc_maxopt.m4 -8 +8 cannam@95: cannam@95: Sun May 22 22:37:50 EDT 2005 stevenj cannam@95: * [project @ 2005-05-23 02:37:50 by stevenj] cannam@95: note icc 8.x annoyance cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -1 +3 cannam@95: cannam@95: Sun May 22 22:36:04 EDT 2005 stevenj cannam@95: * [project @ 2005-05-23 02:36:04 by stevenj] cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -1 +1 cannam@95: cannam@95: Sun May 22 22:35:34 EDT 2005 stevenj cannam@95: * [project @ 2005-05-23 02:35:34 by stevenj] cannam@95: note gcc 3.4.[0123] bug, which is fixed in gcc 3.4.4 cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn +5 cannam@95: cannam@95: Sun May 22 22:21:26 EDT 2005 stevenj cannam@95: * [project @ 2005-05-23 02:21:26 by stevenj] cannam@95: added automatic detection of icc architecture flag cannam@95: cannam@95: M ./m4/ax_cc_maxopt.m4 -1 +34 cannam@95: cannam@95: Sun May 22 21:47:19 EDT 2005 stevenj cannam@95: * [project @ 2005-05-23 01:47:19 by stevenj] cannam@95: add -no-gcc to icc flags...even if it is Intel's fault, I'm sick of dealing with bug reports about this cannam@95: cannam@95: M ./configure.ac +2 cannam@95: cannam@95: Sun May 22 21:40:59 EDT 2005 stevenj cannam@95: * [project @ 2005-05-23 01:40:59 by stevenj] cannam@95: added @cindex portability cannam@95: cannam@95: M ./doc/fftw3.texi -2 +11 cannam@95: cannam@95: Sun May 22 21:34:10 EDT 2005 stevenj cannam@95: * [project @ 2005-05-23 01:34:10 by stevenj] cannam@95: note --without-gcc-arch cannam@95: cannam@95: M ./doc/fftw3.texi -2 +4 cannam@95: cannam@95: Sun May 22 20:54:54 EDT 2005 stevenj cannam@95: * [project @ 2005-05-23 00:54:54 by stevenj] cannam@95: bsd ppc detection; some odd 603 types cannam@95: cannam@95: M ./m4/ax_gcc_archflag.m4 -2 +3 cannam@95: cannam@95: Sun May 22 11:53:20 EDT 2005 stevenj cannam@95: * [project @ 2005-05-22 15:53:20 by stevenj] cannam@95: cannam@95: M ./m4/ax_gcc_archflag.m4 -2 +2 cannam@95: cannam@95: Sat May 21 20:34:52 EDT 2005 stevenj cannam@95: * [project @ 2005-05-22 00:34:52 by stevenj] cannam@95: ensure no spaces in cputype cannam@95: cannam@95: M ./m4/ax_gcc_archflag.m4 -2 +2 cannam@95: cannam@95: Sat May 21 20:31:41 EDT 2005 stevenj cannam@95: * [project @ 2005-05-22 00:31:41 by stevenj] cannam@95: nevermind cannam@95: cannam@95: M ./m4/ax_gcc_archflag.m4 -2 +2 cannam@95: cannam@95: Sat May 21 20:30:08 EDT 2005 stevenj cannam@95: * [project @ 2005-05-22 00:30:08 by stevenj] cannam@95: more bsd stuff cannam@95: cannam@95: M ./m4/ax_gcc_archflag.m4 -2 +2 cannam@95: cannam@95: Sat May 21 20:28:40 EDT 2005 stevenj cannam@95: * [project @ 2005-05-22 00:28:40 by stevenj] cannam@95: added BSD cpu detection for SPARC and better super/hypersparc detection cannam@95: cannam@95: M ./m4/ax_gcc_archflag.m4 -5 +5 cannam@95: cannam@95: Sat May 21 20:22:11 EDT 2005 stevenj cannam@95: * [project @ 2005-05-22 00:22:11 by stevenj] cannam@95: comment cannam@95: cannam@95: M ./m4/ax_gcc_archflag.m4 -1 +4 cannam@95: cannam@95: Fri May 20 19:40:09 EDT 2005 stevenj cannam@95: * [project @ 2005-05-20 23:40:09 by stevenj] cannam@95: "alternate" == "alternative" is US-centric cannam@95: cannam@95: M ./doc/fftw3.texi -3 +3 cannam@95: cannam@95: Fri May 20 19:36:26 EDT 2005 stevenj cannam@95: * [project @ 2005-05-20 23:36:26 by stevenj] cannam@95: typo cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Fri May 20 01:28:34 EDT 2005 stevenj cannam@95: * [project @ 2005-05-20 05:28:34 by stevenj] cannam@95: clarification cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -2 +2 cannam@95: cannam@95: Tue May 17 18:56:46 EDT 2005 stevenj cannam@95: * [project @ 2005-05-17 22:56:46 by stevenj] cannam@95: print out estimate-planner time from can_do in verbose>2 mode cannam@95: cannam@95: M ./tests/bench.c -1 +7 cannam@95: cannam@95: Mon May 9 00:47:19 EDT 2005 stevenj cannam@95: * [project @ 2005-05-09 04:47:19 by stevenj] cannam@95: comment cannam@95: cannam@95: M ./m4/ax_cc_vendor.m4 +1 cannam@95: cannam@95: Thu May 5 23:47:55 EDT 2005 stevenj cannam@95: * [project @ 2005-05-06 03:47:55 by stevenj] cannam@95: fixes for building Windows DLLs with Cygwin; thanks in part to Stephane Fillod cannam@95: cannam@95: M ./Makefile.am -1 +1 cannam@95: M ./api/api.h +2 cannam@95: M ./api/fftw3.h -2 +12 cannam@95: M ./configure.ac -2 +4 cannam@95: M ./threads/Makefile.am -1 +1 cannam@95: cannam@95: Fri Apr 22 19:47:43 EDT 2005 stevenj cannam@95: * [project @ 2005-04-22 23:47:43 by stevenj] cannam@95: -ffast-math seems to produce code that is either about the same speed or slightly faster (gcc 3.3 and 4.0, x86) cannam@95: cannam@95: M ./m4/ax_cc_maxopt.m4 -1 +4 cannam@95: cannam@95: Fri Apr 22 19:18:23 EDT 2005 stevenj cannam@95: * [project @ 2005-04-22 23:18:23 by stevenj] cannam@95: power5 fallback to power4 sched for older gcc's cannam@95: cannam@95: M ./m4/ax_gcc_archflag.m4 -2 +2 cannam@95: cannam@95: Fri Apr 22 19:14:53 EDT 2005 stevenj cannam@95: * [project @ 2005-04-22 23:14:53 by stevenj] cannam@95: check for power5 cannam@95: cannam@95: M ./m4/ax_gcc_archflag.m4 -1 +2 cannam@95: cannam@95: Tue Apr 19 21:55:13 EDT 2005 athena cannam@95: * [project @ 2005-04-20 01:55:13 by athena] cannam@95: Removed clause #3 cannam@95: cannam@95: M ./api/fftw3.h -5 +1 cannam@95: cannam@95: Tue Apr 19 21:44:57 EDT 2005 stevenj cannam@95: * [project @ 2005-04-20 01:44:57 by stevenj] cannam@95: license clarification cannam@95: cannam@95: M ./api/fftw3.h -1 +4 cannam@95: cannam@95: Tue Apr 19 21:42:51 EDT 2005 athena cannam@95: * [project @ 2005-04-20 01:42:51 by athena] cannam@95: Changed license of fftw3.h to X11. cannam@95: cannam@95: M ./api/fftw3.h -12 +24 cannam@95: cannam@95: Mon Apr 11 13:15:12 EDT 2005 stevenj cannam@95: * [project @ 2005-04-11 17:15:12 by stevenj] cannam@95: delete fixed-input code cannam@95: cannam@95: M ./genfft/gen_conv.ml -3 +2 cannam@95: cannam@95: Sun Apr 10 16:33:24 EDT 2005 athena cannam@95: * [project @ 2005-04-10 20:33:24 by athena] cannam@95: joned L-U-planner branch cannam@95: cannam@95: M ./api/apiplan.c -2 +2 cannam@95: M ./api/fftw3.h -4 +5 cannam@95: M ./api/mapflags.c -19 +35 cannam@95: M ./dft/bluestein.c -6 +7 cannam@95: M ./dft/buffered.c -18 +11 cannam@95: M ./dft/ct.c -7 +5 cannam@95: M ./dft/dftw-direct.c -1 +4 cannam@95: M ./dft/dftw-generic.c -2 +4 cannam@95: M ./dft/generic.c -1 +1 cannam@95: M ./dft/indirect-transpose.c -1 +1 cannam@95: M ./dft/indirect.c -8 +8 cannam@95: M ./dft/rader.c -14 +16 cannam@95: M ./dft/vrank-geq1.c -2 +2 cannam@95: M ./kernel/ifftw.h -49 +71 cannam@95: M ./kernel/planner.c -97 +146 cannam@95: M ./rdft/buffered.c -15 +9 cannam@95: M ./rdft/buffered2.c -9 +6 cannam@95: M ./rdft/dft-r2hc.c -7 +4 cannam@95: M ./rdft/dht-r2hc.c -7 +7 cannam@95: M ./rdft/dht-rader.c -13 +15 cannam@95: M ./rdft/generic.c -1 +1 cannam@95: M ./rdft/hc2hc-generic.c -1 +1 cannam@95: M ./rdft/hc2hc.c -3 +2 cannam@95: M ./rdft/indirect.c -6 +4 cannam@95: M ./rdft/rank-geq2-rdft2.c -2 +2 cannam@95: M ./rdft/rdft-dht.c -5 +5 cannam@95: M ./rdft/rdft2-radix2.c -2 +2 cannam@95: M ./rdft/vrank-geq1-rdft2.c -2 +2 cannam@95: M ./rdft/vrank-geq1.c -2 +2 cannam@95: M ./rdft/vrank3-transpose.c -6 +6 cannam@95: M ./reodft/redft00e-r2hc-pad.c -2 +2 cannam@95: M ./reodft/redft00e-r2hc.c -2 +2 cannam@95: M ./reodft/reodft00e-splitradix.c -34 +17 cannam@95: M ./reodft/reodft010e-r2hc.c -2 +2 cannam@95: M ./reodft/reodft11e-r2hc-odd.c -2 +2 cannam@95: M ./reodft/reodft11e-r2hc.c -2 +2 cannam@95: M ./reodft/reodft11e-radix2.c -2 +2 cannam@95: M ./reodft/rodft00e-r2hc-pad.c -2 +2 cannam@95: M ./reodft/rodft00e-r2hc.c -2 +2 cannam@95: M ./tests/bench.c -1 +1 cannam@95: M ./tests/hook.c -1 +1 cannam@95: cannam@95: Thu Apr 7 23:15:02 EDT 2005 stevenj cannam@95: * [project @ 2005-04-08 03:15:02 by stevenj] cannam@95: ref cannam@95: cannam@95: M ./reodft/reodft00e-splitradix.c -16 +33 cannam@95: cannam@95: Thu Apr 7 00:11:13 EDT 2005 stevenj cannam@95: * [project @ 2005-04-07 04:10:26 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./genfft/gen_r2r.ml -3 +3 cannam@95: cannam@95: Wed Apr 6 22:06:21 EDT 2005 stevenj cannam@95: * [project @ 2005-04-07 02:06:21 by stevenj] cannam@95: added (optional) new split-radix algorithm, enabled with -newsplit; also new -standalone option to omit desc; also -unitary, -normalization, and -normsqr options to generate r2r codelets with various normalization (to match lit. in DCT-II, use: -unitary -normsqr 2) cannam@95: cannam@95: M ./genfft/complex.ml -1 +22 cannam@95: M ./genfft/complex.mli -1 +7 cannam@95: M ./genfft/fft.ml -4 +86 cannam@95: M ./genfft/gen_athtw.ml -3 +3 cannam@95: M ./genfft/gen_hc2r.ml -4 +4 cannam@95: M ./genfft/gen_notw.ml -5 +4 cannam@95: M ./genfft/gen_notw_c.ml -3 +3 cannam@95: M ./genfft/gen_r2hc.ml -4 +4 cannam@95: M ./genfft/gen_r2r.ml -19 +43 cannam@95: M ./genfft/magic.ml -1 +6 cannam@95: M ./genfft/number.ml -4 +15 cannam@95: M ./genfft/number.mli -1 +2 cannam@95: cannam@95: Fri Mar 25 08:59:43 EST 2005 athena cannam@95: * [project @ 2005-03-25 13:59:43 by athena] cannam@95: Moved timeout check outside the search loop, because X(seconds) is cannam@95: expensive. cannam@95: cannam@95: M ./kernel/planner.c -9 +7 cannam@95: cannam@95: Sun Mar 20 18:35:53 EST 2005 athena cannam@95: * [project @ 2005-03-20 23:35:53 by athena] cannam@95: Enable vector recursion for in-place problems, otherwise cannam@95: dftw-genericbuf works only in PATIENT mode. cannam@95: cannam@95: M ./dft/ct.c -2 +4 cannam@95: cannam@95: Sun Mar 20 17:53:58 EST 2005 athena cannam@95: * [project @ 2005-03-20 22:53:58 by athena] cannam@95: oops cannam@95: cannam@95: M ./dft/dftw-genericbuf.c -1 +1 cannam@95: cannam@95: Sun Mar 20 17:49:13 EST 2005 athena cannam@95: * [project @ 2005-03-20 22:49:13 by athena] cannam@95: make solver UGLY for small N cannam@95: cannam@95: M ./dft/dftw-genericbuf.c -3 +13 cannam@95: cannam@95: Sun Mar 20 17:16:37 EST 2005 athena cannam@95: * [project @ 2005-03-20 22:16:37 by athena] cannam@95: new dftw-genericbuf solver cannam@95: cannam@95: A ./dft/dftw-genericbuf.c cannam@95: cannam@95: Sun Mar 20 16:12:44 EST 2005 athena cannam@95: * [project @ 2005-03-20 21:12:44 by athena] cannam@95: new dftw-genericbuf solver cannam@95: cannam@95: M ./dft/Makefile.am -4 +4 cannam@95: M ./dft/conf.c -1 +2 cannam@95: M ./dft/dft.h -1 +2 cannam@95: cannam@95: Thu Mar 17 21:48:19 EST 2005 athena cannam@95: * [project @ 2005-03-18 02:48:19 by athena] cannam@95: Hmm... what was I thinking? cannam@95: cannam@95: M ./simd/sse2-aux.c -2 +2 cannam@95: cannam@95: Thu Mar 17 19:20:54 EST 2005 athena cannam@95: * [project @ 2005-03-18 00:20:54 by athena] cannam@95: Workaround for a MSVC bug. cannam@95: cannam@95: M ./simd/simd-sse2.h -1 +6 cannam@95: M ./simd/sse2-aux.c -1 +8 cannam@95: cannam@95: Thu Mar 17 08:18:39 EST 2005 athena cannam@95: * [project @ 2005-03-17 13:18:39 by athena] cannam@95: Workaround for a MSVC bug that was reported by Eddie Yee. cannam@95: cannam@95: M ./simd/simd-sse.h -1 +6 cannam@95: M ./simd/sse-aux.c -1 +8 cannam@95: cannam@95: Tue Mar 15 13:25:53 EST 2005 athena cannam@95: * [project @ 2005-03-15 18:25:53 by athena] cannam@95: try both contiguous input and contiguous output when in doubt cannam@95: cannam@95: M ./rdft/rank0.c -3 +26 cannam@95: cannam@95: Tue Mar 15 08:44:41 EST 2005 athena cannam@95: * [project @ 2005-03-15 13:44:41 by athena] cannam@95: Added genfft flag -precompute-twiddles which moves the computation of cannam@95: the twiddle factors before the main schedule. This flag produces cannam@95: smaller code everywhere, and slightly faster code on powerpc. cannam@95: I observe no speed difference on x86. cannam@95: cannam@95: M ./dft/codelets/standard/Makefile.am -3 +3 cannam@95: M ./genfft/genutil.ml -3 +5 cannam@95: M ./genfft/magic.ml -1 +5 cannam@95: M ./genfft/schedule.ml -1 +53 cannam@95: M ./genfft/schedule.mli -1 +2 cannam@95: M ./rdft/codelets/hc2r/Makefile.am -1 +1 cannam@95: M ./rdft/codelets/r2hc/Makefile.am -1 +1 cannam@95: cannam@95: Mon Mar 14 21:43:53 EST 2005 stevenj cannam@95: * [project @ 2005-03-15 02:43:53 by stevenj] cannam@95: sp cannam@95: cannam@95: M ./kernel/kalloc.c -2 +2 cannam@95: cannam@95: Mon Mar 14 21:43:05 EST 2005 stevenj cannam@95: * [project @ 2005-03-15 02:43:05 by stevenj] cannam@95: whoops, spelling error (thanks to Steve Eddins for bug report) cannam@95: cannam@95: M ./kernel/alloc.c -2 +2 cannam@95: cannam@95: Sat Mar 12 15:03:47 EST 2005 athena cannam@95: * [project @ 2005-03-12 20:03:45 by athena] cannam@95: Do not approximate pcost = vl * child->pcost unless child is guaranteed cannam@95: not to be a simple codelet. cannam@95: cannam@95: M ./dft/vrank-geq1.c -2 +4 cannam@95: M ./rdft/vrank-geq1-rdft2.c -2 +4 cannam@95: M ./rdft/vrank-geq1.c -2 +4 cannam@95: cannam@95: Wed Mar 9 20:00:02 EST 2005 athena cannam@95: * [project @ 2005-03-10 01:00:02 by athena] cannam@95: Relaxed applicability conditions. cannam@95: cannam@95: M ./dft/direct.c -6 +3 cannam@95: cannam@95: Wed Mar 9 00:05:47 EST 2005 athena cannam@95: * [project @ 2005-03-09 05:05:47 by athena] cannam@95: Minor optimization cannam@95: cannam@95: M ./dft/dftw-generic.c -8 +8 cannam@95: cannam@95: Tue Mar 8 22:14:02 EST 2005 athena cannam@95: * [project @ 2005-03-09 03:14:02 by athena] cannam@95: Interpret K to mean *1024. Similarly for M. cannam@95: cannam@95: M ./libbench2/problem.c -1 +12 cannam@95: cannam@95: Tue Mar 8 20:44:25 EST 2005 athena cannam@95: * [project @ 2005-03-09 01:44:25 by athena] cannam@95: Hmm... somehow some previous commit got lost. cannam@95: cannam@95: M ./kernel/primes.c -2 +8 cannam@95: cannam@95: Tue Mar 8 20:30:42 EST 2005 athena cannam@95: * [project @ 2005-03-09 01:30:42 by athena] cannam@95: Paranoia cannam@95: cannam@95: M ./dft/ct.c -2 +2 cannam@95: cannam@95: Mon Mar 7 14:30:01 EST 2005 stevenj cannam@95: * [project @ 2005-03-07 19:30:01 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./configure.ac +1 cannam@95: cannam@95: Mon Mar 7 14:29:43 EST 2005 stevenj cannam@95: * [project @ 2005-03-07 19:29:43 by stevenj] cannam@95: move fftw-specific HP/UX tweak into configure.ac cannam@95: cannam@95: M ./configure.ac +3 cannam@95: M ./m4/ax_cc_maxopt.m4 -2 +2 cannam@95: cannam@95: Mon Mar 7 14:19:24 EST 2005 stevenj cannam@95: * [project @ 2005-03-07 19:19:24 by stevenj] cannam@95: ax_cc_family -> ax_cc_vendor (vendor names are easier to remember), add checks for many new compilers, use in ax_cc_maxopt cannam@95: cannam@95: A ./m4/ax_cc_vendor.m4 cannam@95: M ./configure.ac -3 +3 cannam@95: R ./m4/ax_cc_family.m4 cannam@95: M ./m4/ax_cc_maxopt.m4 -65 +42 cannam@95: M ./m4/ax_cc_vendor.m4 +23 cannam@95: cannam@95: Sun Mar 6 21:36:05 EST 2005 athena cannam@95: * [project @ 2005-03-07 02:36:05 by athena] cannam@95: Count FMA as one flop in estimator when HAVE_FMA cannam@95: cannam@95: M ./kernel/planner.c -1 +7 cannam@95: cannam@95: Sun Mar 6 19:16:06 EST 2005 athena cannam@95: * [project @ 2005-03-07 00:16:06 by athena] cannam@95: Do not try radix-2 generic. cannam@95: cannam@95: M ./dft/dftw-generic.c -1 +1 cannam@95: cannam@95: Sun Mar 6 13:04:23 EST 2005 athena cannam@95: * [project @ 2005-03-06 18:04:23 by athena] cannam@95: Use -O3 for xlc now that we use -O for CODELET_OPTIM cannam@95: cannam@95: M ./m4/ax_cc_maxopt.m4 -4 +4 cannam@95: cannam@95: Sun Mar 6 13:02:41 EST 2005 athena cannam@95: * [project @ 2005-03-06 18:02:41 by athena] cannam@95: New AX_CC_FAMILY macro, that detects the compiler based on symbols cannam@95: that it defines (as opposed to the name of the compiler). cannam@95: We need to start use this strategy everywhere else. cannam@95: cannam@95: A ./m4/ax_cc_family.m4 cannam@95: M ./configure.ac -25 +29 cannam@95: M ./m4/ax_cc_family.m4 +18 cannam@95: cannam@95: Sun Mar 6 11:33:15 EST 2005 athena cannam@95: * [project @ 2005-03-06 16:33:15 by athena] cannam@95: Runtime checks to guarantee small strides. cannam@95: cannam@95: M ./dft/direct.c -2 +18 cannam@95: cannam@95: Sat Mar 5 20:09:25 EST 2005 athena cannam@95: * [project @ 2005-03-06 01:09:25 by athena] cannam@95: Reduced the search space for rank-0 transforms cannam@95: cannam@95: M ./dft/vrank-geq1.c -6 +6 cannam@95: M ./kernel/tensor7.c -8 +20 cannam@95: M ./rdft/rank0.c -50 +48 cannam@95: M ./rdft/vrank-geq1.c -6 +5 cannam@95: cannam@95: Fri Mar 4 17:50:29 EST 2005 stevenj cannam@95: * [project @ 2005-03-04 22:50:29 by stevenj] cannam@95: little assert cannam@95: cannam@95: M ./kernel/primes.c -8 +3 cannam@95: cannam@95: Tue Mar 1 09:19:16 EST 2005 athena cannam@95: * [project @ 2005-03-01 14:19:16 by athena] cannam@95: Implemented directbuf, enabled for now. cannam@95: cannam@95: M ./dft/dft.h -1 +2 cannam@95: M ./dft/dftw-direct.c -10 +7 cannam@95: M ./dft/direct.c -11 +135 cannam@95: M ./dft/kdft.c -3 +3 cannam@95: cannam@95: Mon Feb 28 22:21:14 EST 2005 athena cannam@95: * [project @ 2005-03-01 03:21:14 by athena] cannam@95: Unified dftw-direct, dftw-directbuf in an attempt to tame code cannam@95: growth cannam@95: cannam@95: M ./dft/Makefile.am -4 +4 cannam@95: M ./dft/dftw-direct.c -15 +130 cannam@95: R ./dft/dftw-directbuf.c cannam@95: M ./dft/kdft-dif.c -2 +1 cannam@95: M ./dft/kdft-dit.c -2 +1 cannam@95: cannam@95: Sun Feb 27 13:51:24 EST 2005 stevenj cannam@95: * [project @ 2005-02-27 18:51:24 by stevenj] cannam@95: fixed copyright cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Sat Feb 26 22:21:03 EST 2005 athena cannam@95: * [project @ 2005-02-27 03:21:03 by athena] cannam@95: silence warnings cannam@95: cannam@95: M ./rdft/rank0.c -4 +5 cannam@95: cannam@95: Sat Feb 26 22:19:16 EST 2005 athena cannam@95: * [project @ 2005-02-27 03:19:16 by athena] cannam@95: oops cannam@95: cannam@95: M ./rdft/rank0.c -2 +2 cannam@95: cannam@95: Sat Feb 26 21:28:39 EST 2005 athena cannam@95: * [project @ 2005-02-27 02:28:39 by athena] cannam@95: Tweaking while thinking about a higher-rank transposer (bitreverser) cannam@95: cannam@95: M ./rdft/rank0.c -44 +63 cannam@95: cannam@95: Sat Feb 26 20:06:49 EST 2005 athena cannam@95: * [project @ 2005-02-27 01:06:49 by athena] cannam@95: Transposed the buffer, and skewed it. This allows for contiguous cannam@95: copy operations, and the codelet should not incur associativity cannam@95: conflicts if the buffer is large. cannam@95: cannam@95: M ./dft/dftw-directbuf.c -20 +26 cannam@95: cannam@95: Sat Feb 26 18:14:11 EST 2005 stevenj cannam@95: * [project @ 2005-02-26 23:14:11 by stevenj] cannam@95: make tensor_max_index more reasonable (take maximum of input and output cannam@95: max indices, computed separately) cannam@95: cannam@95: M ./kernel/tensor4.c -4 +5 cannam@95: cannam@95: Sat Feb 26 10:04:30 EST 2005 athena cannam@95: * [project @ 2005-02-26 15:04:30 by athena] cannam@95: Use cpy2d instead of cpy2d_tiled, because vl may be too large. cannam@95: cannam@95: M ./rdft/vrank3-transpose.c -5 +5 cannam@95: cannam@95: Sat Feb 26 00:31:52 EST 2005 athena cannam@95: * [project @ 2005-02-26 05:31:52 by athena] cannam@95: Fixed old bug that was introduced with yesterday's changes. cannam@95: cannam@95: M ./genfft/annotate.ml -4 +6 cannam@95: cannam@95: Fri Feb 25 21:54:23 EST 2005 athena cannam@95: * [project @ 2005-02-26 02:54:23 by athena] cannam@95: ``Interesting'' switch statement. cannam@95: cannam@95: M ./kernel/cpy1d.c -3 +23 cannam@95: cannam@95: Fri Feb 25 12:29:54 EST 2005 athena cannam@95: * [project @ 2005-02-25 17:29:54 by athena] cannam@95: Disabled -reorder-loads -reorder-stores, since they seem to do cannam@95: nothing. cannam@95: cannam@95: M ./support/Makefile.codelets -1 +1 cannam@95: cannam@95: Fri Feb 25 12:19:10 EST 2005 stevenj cannam@95: * [project @ 2005-02-25 17:19:03 by stevenj] cannam@95: Because of the recent changes to kernel/pickdim.c, splitrnk=0 is no cannam@95: longer equivalent to splitrnk=1 for rnk < 4, where the latter is the cannam@95: FFTW2 behavior. For small rnk, however, I observe the planner to pretty cannam@95: consistently choose the FFTW2 behavior (splitrnk=1), despite its not cannam@95: being asymptotically optimal in the cache oblivious sense. So, make cannam@95: splitrnk=1 instead of splitrnk=0 the default in FFTW_MEASURE and cannam@95: FFTW_ESTIMATE modes (rnk > 3 is pretty rare in practice anyway). cannam@95: cannam@95: M ./dft/rank-geq2.c -5 +2 cannam@95: M ./rdft/rank-geq2.c -5 +2 cannam@95: cannam@95: Fri Feb 25 00:33:27 EST 2005 stevenj cannam@95: * [project @ 2005-02-25 05:33:27 by stevenj] cannam@95: tweak cannam@95: cannam@95: M ./dft/indirect-transpose.c -3 +4 cannam@95: cannam@95: Fri Feb 25 00:29:09 EST 2005 stevenj cannam@95: * [project @ 2005-02-25 05:29:09 by stevenj] cannam@95: slight relaxation cannam@95: cannam@95: M ./dft/indirect-transpose.c -2 +4 cannam@95: cannam@95: Fri Feb 25 00:21:00 EST 2005 stevenj cannam@95: * [project @ 2005-02-25 05:21:00 by stevenj] cannam@95: cruft cannam@95: cannam@95: M ./dft/indirect-transpose.c -2 cannam@95: cannam@95: Fri Feb 25 00:03:14 EST 2005 stevenj cannam@95: * [project @ 2005-02-25 05:03:13 by stevenj] cannam@95: added experimental indirect-transpose solver: when transforming the columns of the matrix, allow us to do a transpose to make the DFTs contiguous cannam@95: cannam@95: A ./dft/indirect-transpose.c cannam@95: M ./dft/Makefile.am -3 +3 cannam@95: M ./dft/conf.c -1 +2 cannam@95: M ./dft/dft.h -1 +2 cannam@95: M ./dft/indirect-transpose.c +280 cannam@95: M ./dft/indirect.c -3 +10 cannam@95: M ./kernel/ifftw.h -1 +3 cannam@95: M ./kernel/tensor4.c -1 +32 cannam@95: cannam@95: Thu Feb 24 23:04:58 EST 2005 stevenj cannam@95: * [project @ 2005-02-25 04:04:58 by stevenj] cannam@95: check for abort() cannam@95: cannam@95: M ./configure.ac -1 +1 cannam@95: cannam@95: Thu Feb 24 23:04:43 EST 2005 stevenj cannam@95: * [project @ 2005-02-25 04:04:43 by stevenj] cannam@95: call abort() on failed assertion cannam@95: cannam@95: M ./kernel/assert.c -1 +5 cannam@95: cannam@95: Thu Feb 24 21:17:23 EST 2005 athena cannam@95: * [project @ 2005-02-25 02:17:23 by athena] cannam@95: Forgot to change X(isqrt) -> isqrt_maybe cannam@95: cannam@95: M ./kernel/primes.c -2 +2 cannam@95: cannam@95: Thu Feb 24 20:18:59 EST 2005 stevenj cannam@95: * [project @ 2005-02-25 01:17:59 by stevenj] cannam@95: require finite_rnk cannam@95: cannam@95: M ./dft/rank-geq2.c -1 +2 cannam@95: M ./rdft/rank-geq2-rdft2.c -1 +2 cannam@95: M ./rdft/rank-geq2.c -1 +2 cannam@95: cannam@95: Thu Feb 24 20:07:38 EST 2005 stevenj cannam@95: * [project @ 2005-02-25 01:07:38 by stevenj] cannam@95: #ifdef HAVE_STRING_H must come after rdft.h so that we get config.h cannam@95: cannam@95: M ./rdft/vrank3-transpose.c -3 +3 cannam@95: cannam@95: Thu Feb 24 18:59:40 EST 2005 athena cannam@95: * [project @ 2005-02-24 23:59:38 by athena] cannam@95: Implemented reordering of loads and stores so that the real and cannam@95: imaginary part are loaded/stored together. This should improve cannam@95: out-of-cache performance in the presence of associativity conflicts, cannam@95: and maybe worsen in-cache performance because of worse scheduling. cannam@95: Enabled for now, for experimental purposes. cannam@95: cannam@95: M ./genfft/annotate.ml -30 +76 cannam@95: M ./genfft/magic.ml -4 +10 cannam@95: M ./support/Makefile.codelets -1 +1 cannam@95: M ./support/twovers.sh -1 +1 cannam@95: cannam@95: Thu Feb 24 18:10:49 EST 2005 stevenj cannam@95: * [project @ 2005-02-24 23:10:49 by stevenj] cannam@95: fix comment cannam@95: cannam@95: M ./m4/ax_gcc_aligns_stack.m4 -2 +2 cannam@95: cannam@95: Thu Feb 24 18:10:23 EST 2005 stevenj cannam@95: * [project @ 2005-02-24 23:10:23 by stevenj] cannam@95: better message cannam@95: cannam@95: M ./m4/ax_gcc_aligns_stack.m4 -2 +2 cannam@95: cannam@95: Thu Feb 24 18:08:36 EST 2005 stevenj cannam@95: * [project @ 2005-02-24 23:08:36 by stevenj] cannam@95: use gcc version > 3.0 as fallback in check for alignment bug cannam@95: cannam@95: M ./m4/ax_gcc_aligns_stack.m4 -2 +2 cannam@95: cannam@95: Thu Feb 24 18:02:31 EST 2005 stevenj cannam@95: * [project @ 2005-02-24 23:02:31 by stevenj] cannam@95: don't use -malign-double unconditionally (it is only available on x86) cannam@95: cannam@95: M ./m4/ax_gcc_aligns_stack.m4 -2 +3 cannam@95: cannam@95: Thu Feb 24 12:03:30 EST 2005 athena cannam@95: * [project @ 2005-02-24 17:03:30 by athena] cannam@95: Subtler selection of tilesz. cannam@95: cannam@95: M ./kernel/transpose.c -3 +8 cannam@95: cannam@95: Thu Feb 24 11:52:25 EST 2005 athena cannam@95: * [project @ 2005-02-24 16:52:25 by athena] cannam@95: Call cpy2d_tiledbuf, not cpy2d_tiled. cannam@95: cannam@95: M ./rdft/rank0.c -5 +5 cannam@95: cannam@95: Thu Feb 24 11:29:28 EST 2005 athena cannam@95: * [project @ 2005-02-24 16:29:28 by athena] cannam@95: buffer sizes were wrong :-( cannam@95: cannam@95: M ./kernel/cpy2d.c -1 +1 cannam@95: M ./kernel/transpose.c -2 +2 cannam@95: cannam@95: Thu Feb 24 11:19:01 EST 2005 athena cannam@95: * [project @ 2005-02-24 16:19:01 by athena] cannam@95: Single function for computing tile size. Eliminate spurious assertions. cannam@95: cannam@95: M ./kernel/cpy2d.c -3 +8 cannam@95: M ./kernel/ifftw.h -1 +3 cannam@95: M ./kernel/tile2d.c +5 cannam@95: M ./kernel/transpose.c -4 +6 cannam@95: M ./rdft/rank0.c -3 +1 cannam@95: cannam@95: Thu Feb 24 10:00:02 EST 2005 athena cannam@95: * [project @ 2005-02-24 15:00:02 by athena] cannam@95: Do tiling recursively. cannam@95: cannam@95: M ./kernel/tile2d.c +28 cannam@95: cannam@95: Thu Feb 24 09:40:30 EST 2005 athena cannam@95: * [project @ 2005-02-24 14:40:30 by athena] cannam@95: Reworked tiled transposes; provide tiling with and without buffering. cannam@95: I can't believe that one has to waste his life with this @#$%. cannam@95: cannam@95: A ./kernel/tile2d.c cannam@95: M ./kernel/Makefile.am -1 +1 cannam@95: M ./kernel/cpy2d.c -36 +60 cannam@95: M ./kernel/ifftw.h -2 +10 cannam@95: M ./kernel/tile2d.c +41 cannam@95: M ./kernel/transpose.c -64 +122 cannam@95: M ./rdft/rank0.c -6 +41 cannam@95: M ./rdft/vrank3-transpose.c -2 +2 cannam@95: cannam@95: Wed Feb 23 22:21:19 EST 2005 athena cannam@95: * [project @ 2005-02-24 03:21:19 by athena] cannam@95: Clarified logic. I am not sure why the code was so confusing to begin cannam@95: with. The computation of *dp in the which_dim == 0 case was also cannam@95: wrong, returning e.g. *dp == -1 if sz->rnk == 1. cannam@95: cannam@95: M ./kernel/pickdim.c -6 +6 cannam@95: cannam@95: Wed Feb 23 22:00:15 EST 2005 athena cannam@95: * [project @ 2005-02-24 03:00:15 by athena] cannam@95: Enable aggressive inlining in codelets only, to avoid code bloat. cannam@95: cannam@95: M ./configure.ac -9 +10 cannam@95: cannam@95: Wed Feb 23 21:51:50 EST 2005 athena cannam@95: * [project @ 2005-02-24 02:51:50 by athena] cannam@95: Removed cache-oblivious copy/transpose algorithms in favor of cannam@95: explicitly blocked algorithms. The cache-oblivious algorithms fail if cannam@95: there are associativity conflicts, in which case buffering is cannam@95: necessary, as per Carter and Gatlin. Once you set the buffer size, cannam@95: there is no point whatsoever to do the algorithm recursively, and you cannam@95: may as well use blocking. cannam@95: cannam@95: M ./kernel/Makefile.am -2 +1 cannam@95: M ./kernel/cpy2d.c -62 +40 cannam@95: M ./kernel/ifftw.h -11 +11 cannam@95: M ./kernel/primes.c -4 +10 cannam@95: R ./kernel/transpose-rec.c cannam@95: M ./kernel/transpose.c +54 cannam@95: M ./rdft/rank0.c -49 +28 cannam@95: M ./rdft/vrank3-transpose.c -6 +6 cannam@95: cannam@95: Wed Feb 23 18:46:12 EST 2005 stevenj cannam@95: * [project @ 2005-02-23 23:46:12 by stevenj] cannam@95: --disable-fortran now differs from --enable-fortran that fails cannam@95: cannam@95: M ./configure.ac -2 +4 cannam@95: cannam@95: Wed Feb 23 18:42:21 EST 2005 stevenj cannam@95: * [project @ 2005-02-23 23:42:21 by stevenj] cannam@95: comment tweak cannam@95: cannam@95: M ./api/f77api.c -1 +1 cannam@95: cannam@95: Wed Feb 23 18:41:14 EST 2005 stevenj cannam@95: * [project @ 2005-02-23 23:41:14 by stevenj] cannam@95: If a Fortran compiler was not detected, just make our best guess at cannam@95: what wrappers to use...I'm sick of dealing with user complaints from cannam@95: cases where wrapper detection fails for whatever reason. cannam@95: cannam@95: M ./api/f77api.c +21 cannam@95: cannam@95: Wed Feb 23 18:10:40 EST 2005 stevenj cannam@95: * [project @ 2005-02-23 23:10:40 by stevenj] cannam@95: fflush(stdout) after print_plan, in case F77 doesn't cannam@95: cannam@95: M ./api/f77funcs.h +1 cannam@95: cannam@95: Tue Feb 22 22:54:42 EST 2005 athena cannam@95: * [project @ 2005-02-23 03:54:42 by athena] cannam@95: --enable-sse is necessary after all, to generate all dependencies cannam@95: correctly. cannam@95: cannam@95: M ./mkdist.sh -1 +1 cannam@95: cannam@95: Tue Feb 22 22:32:06 EST 2005 athena cannam@95: * [project @ 2005-02-23 03:32:06 by athena] cannam@95: Put cpy2d_pair into its own file, so that I can experiment with cannam@95: buffering of nontwiddle codelets. cannam@95: cannam@95: A ./kernel/cpy2d-pair.c cannam@95: M ./dft/dftw-directbuf.c -28 +7 cannam@95: M ./kernel/Makefile.am -5 +6 cannam@95: M ./kernel/cpy2d-pair.c +61 cannam@95: M ./kernel/ifftw.h -1 +11 cannam@95: cannam@95: Tue Feb 22 20:07:11 EST 2005 athena cannam@95: * [project @ 2005-02-23 01:07:11 by athena] cannam@95: Copy rfftwnd.png from ${srcdir}, not $PWD cannam@95: cannam@95: M ./doc/Makefile.am -1 +1 cannam@95: cannam@95: Tue Feb 22 17:08:48 EST 2005 athena cannam@95: * [project @ 2005-02-22 22:08:48 by athena] cannam@95: Do not bother memcpy-ing complex numbers. cannam@95: cannam@95: M ./rdft/rank0.c -2 +6 cannam@95: cannam@95: Tue Feb 22 16:20:46 EST 2005 athena cannam@95: * [project @ 2005-02-22 21:20:46 by athena] cannam@95: Tighther layout of buffers. I am not sure it matters, but just in case... cannam@95: cannam@95: M ./kernel/cpy2d.c -2 +2 cannam@95: M ./kernel/transpose-rec.c -4 +4 cannam@95: cannam@95: Tue Feb 22 10:13:02 EST 2005 athena cannam@95: * [project @ 2005-02-22 15:13:02 by athena] cannam@95: Usec cpy1d for rank-0 copies cannam@95: cannam@95: M ./rdft/rank0.c -3 +2 cannam@95: cannam@95: Tue Feb 22 10:06:13 EST 2005 athena cannam@95: * [project @ 2005-02-22 15:06:13 by athena] cannam@95: Implemented in-place transposes with buffering. Moved cannam@95: copy/transposition routines into own files, so that we can reuse them cannam@95: from multiple places. TODO: merge vrank3-transpose.c with rank0.c, or cannam@95: rename vrank3-transpose.c to rank0-fancy.c or something like that; cannam@95: decide whether square in-place transposes should be in rank0.c or cannam@95: vrank3-transpose.c; apply FIXME's in vrank3-transpose.c. cannam@95: cannam@95: A ./kernel/cpy1d.c cannam@95: A ./kernel/cpy2d.c cannam@95: A ./kernel/transpose-rec.c cannam@95: A ./kernel/transpose.c cannam@95: M ./kernel/Makefile.am -5 +5 cannam@95: M ./kernel/cpy1d.c +50 cannam@95: M ./kernel/cpy2d.c +154 cannam@95: M ./kernel/ifftw.h -1 +28 cannam@95: M ./kernel/transpose-rec.c +143 cannam@95: M ./kernel/transpose.c +72 cannam@95: M ./rdft/rank0.c -160 +85 cannam@95: M ./rdft/vrank3-transpose.c -231 +24 cannam@95: cannam@95: Mon Feb 21 23:29:52 EST 2005 athena cannam@95: * [project @ 2005-02-22 04:29:52 by athena] cannam@95: Indentation should be printed after newline, not at the beginning cannam@95: of print() cannam@95: cannam@95: M ./kernel/print.c -6 +11 cannam@95: cannam@95: Mon Feb 21 10:07:24 EST 2005 athena cannam@95: * [project @ 2005-02-21 15:07:24 by athena] cannam@95: generalized in anticipation of more complicated solvers. cannam@95: cannam@95: M ./rdft/rank0.c -24 +45 cannam@95: cannam@95: Sun Feb 20 22:18:59 EST 2005 athena cannam@95: * [project @ 2005-02-21 03:18:59 by athena] cannam@95: Implemented buffered recursive transpose cannam@95: cannam@95: M ./rdft/rank0.c -7 +85 cannam@95: cannam@95: Sun Feb 20 18:27:29 EST 2005 athena cannam@95: * [project @ 2005-02-20 23:27:29 by athena] cannam@95: Fixed comment cannam@95: cannam@95: M ./rdft/rank0.c -2 +2 cannam@95: cannam@95: Sun Feb 20 18:22:15 EST 2005 athena cannam@95: * [project @ 2005-02-20 23:22:15 by athena] cannam@95: grand unification of rank0 solvers cannam@95: cannam@95: M ./rdft/Makefile.am -3 +2 cannam@95: M ./rdft/conf.c -2 +1 cannam@95: R ./rdft/rank0-vrank2.c cannam@95: M ./rdft/rank0.c -93 +173 cannam@95: M ./rdft/rdft.h -1 cannam@95: cannam@95: Sun Feb 20 15:35:24 EST 2005 athena cannam@95: * [project @ 2005-02-20 20:35:24 by athena] cannam@95: manual tail-recursion optimization cannam@95: cannam@95: M ./rdft/vrank3-transpose.c -7 +11 cannam@95: cannam@95: Sat Feb 19 17:57:44 EST 2005 athena cannam@95: * [project @ 2005-02-19 22:57:44 by athena] cannam@95: implemented check for transpositions cannam@95: cannam@95: M ./libbench2/verify-lib.c -11 +11 cannam@95: M ./libbench2/verify-r2r.c -1 +1 cannam@95: M ./tests/check.pl -1 +19 cannam@95: cannam@95: Sat Feb 19 17:28:43 EST 2005 athena cannam@95: * [project @ 2005-02-19 22:28:43 by athena] cannam@95: Previous fix was wrong for rdft2 problems. cannam@95: cannam@95: M ./libbench2/verify-lib.c -3 +3 cannam@95: cannam@95: Sat Feb 19 17:23:36 EST 2005 athena cannam@95: * [project @ 2005-02-19 22:23:36 by athena] cannam@95: vecsz->rnk must be finite for this solver to apply. cannam@95: cannam@95: M ./rdft/dft-r2hc.c -2 +2 cannam@95: cannam@95: Sat Feb 19 17:15:19 EST 2005 athena cannam@95: * [project @ 2005-02-19 22:15:19 by athena] cannam@95: unified the various simple'' transposers cannam@95: cannam@95: M ./rdft/vrank3-transpose.c -101 +51 cannam@95: cannam@95: Sat Feb 19 16:55:29 EST 2005 athena cannam@95: * [project @ 2005-02-19 21:55:29 by athena] cannam@95: Fixed stupid bug in rec_transpose_swap. Fixed stupid verifier that did not catch the bug. cannam@95: cannam@95: M ./libbench2/verify-lib.c -17 +26 cannam@95: M ./libbench2/verify-r2r.c -2 +3 cannam@95: M ./rdft/vrank3-transpose.c -5 +5 cannam@95: cannam@95: Sat Feb 19 15:24:03 EST 2005 athena cannam@95: * [project @ 2005-02-19 20:24:03 by athena] cannam@95: Minor cleanup of transposition routines. cannam@95: cannam@95: M ./rdft/vrank3-transpose.c -188 +83 cannam@95: cannam@95: Sat Feb 19 09:31:14 EST 2005 athena cannam@95: * [project @ 2005-02-19 14:31:14 by athena] cannam@95: Make the batch size B=Theta(r) instead of B=Theta(1) in buffered cannam@95: twiddle solvers. Theory: for cache line size L, we want B = Omega(L) cannam@95: to utilize the cache line fully. We also want B*r =O(Z), where Z is cannam@95: the size of the cache. It is safe to assume that Z = Theta(L^2): cannam@95: cache designers will tend to make L as large as they can get away cannam@95: with, because they don't have to program the machines that they build, cannam@95: and Z < Theta(L^2) will screw up the little matrix transposition cannam@95: benchmarks that they use to design the cache. Hence, B=Theta(r) is cannam@95: the right number. cannam@95: cannam@95: M ./dft/dftw-directbuf.c -12 +17 cannam@95: M ./rdft/hc2hc-directbuf.c -16 +20 cannam@95: cannam@95: Fri Feb 18 23:47:22 EST 2005 stevenj cannam@95: * [project @ 2005-02-19 04:47:22 by stevenj] cannam@95: for --enable-portable-binary, only try -mcpu=$arch and -m$arch on x86, cannam@95: since these generate non-portable code on every other target (and cannam@95: some other targets, like Alpha, don't support -mtune=$arch). cannam@95: cannam@95: M ./m4/ax_gcc_archflag.m4 -2 +5 cannam@95: cannam@95: Thu Feb 17 21:15:42 EST 2005 athena cannam@95: * [project @ 2005-02-18 02:15:42 by athena] cannam@95: gcc/aix defines _POWER, not __powerpc__ like the rest of the world cannam@95: does. cannam@95: cannam@95: M ./kernel/ifftw.h -2 +2 cannam@95: cannam@95: Wed Feb 16 22:30:27 EST 2005 athena cannam@95: * [project @ 2005-02-17 03:30:27 by athena] cannam@95: enable fma for ia64, since it seems to help with the hpux compiler. cannam@95: cannam@95: M ./configure.ac +1 cannam@95: cannam@95: Wed Feb 16 21:47:48 EST 2005 athena cannam@95: * [project @ 2005-02-17 02:47:48 by athena] cannam@95: cannam@95: M ./TODO -1 +2 cannam@95: cannam@95: Wed Feb 16 15:27:18 EST 2005 athena cannam@95: * [project @ 2005-02-16 20:27:18 by athena] cannam@95: Fixes for darwin cannam@95: cannam@95: M ./simd/simd-altivec.h -2 +2 cannam@95: cannam@95: Wed Feb 16 14:27:42 EST 2005 athena cannam@95: * [project @ 2005-02-16 19:27:42 by athena] cannam@95: Made the correctness of the code more obvious. cannam@95: cannam@95: M ./api/apiplan.c -14 +16 cannam@95: cannam@95: Wed Feb 16 12:30:29 EST 2005 stevenj cannam@95: * [project @ 2005-02-16 17:30:29 by stevenj] cannam@95: s/with-portable-binary/enable-portable-binary/ to be GNUlly correct; I'm sticking with --with-gcc-arch=arch, however, as --enable-gcc-arch=arch has the wrong connotations for me cannam@95: cannam@95: M ./NEWS -2 +2 cannam@95: M ./m4/ax_cc_maxopt.m4 -3 +3 cannam@95: cannam@95: Wed Feb 16 11:44:48 EST 2005 stevenj cannam@95: * [project @ 2005-02-16 16:44:48 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./api/apiplan.c -2 +2 cannam@95: cannam@95: Wed Feb 16 11:23:38 EST 2005 stevenj cannam@95: * [project @ 2005-02-16 16:23:38 by stevenj] cannam@95: bless wisdom with patience used to create it cannam@95: cannam@95: M ./api/apiplan.c -1 +1 cannam@95: cannam@95: Wed Feb 16 11:18:56 EST 2005 stevenj cannam@95: * [project @ 2005-02-16 16:18:56 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./api/apiplan.c -7 +10 cannam@95: cannam@95: Wed Feb 16 10:50:28 EST 2005 stevenj cannam@95: * [project @ 2005-02-16 15:50:28 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./api/apiplan.c -1 +2 cannam@95: cannam@95: Tue Feb 15 23:53:53 EST 2005 stevenj cannam@95: * [project @ 2005-02-16 04:53:53 by stevenj] cannam@95: added 'timed' planner option cannam@95: cannam@95: M ./NEWS +3 cannam@95: M ./TODO -6 cannam@95: M ./api/apiplan.c -16 +56 cannam@95: M ./api/fftw3.h -1 +3 cannam@95: M ./doc/fftw3.texi -1 +27 cannam@95: M ./kernel/ifftw.h -1 +6 cannam@95: M ./kernel/planner.c -2 +14 cannam@95: M ./kernel/timer.c -1 +13 cannam@95: M ./tests/bench.c +5 cannam@95: cannam@95: Tue Feb 15 23:08:29 EST 2005 athena cannam@95: * [project @ 2005-02-16 04:08:27 by athena] cannam@95: Do not use SIMD_CFLAGS. The theory is that if taint.c is unsafe cannam@95: with SIMD_CFLAGS, then all files in this directory are as well. cannam@95: Conversely, if these files require SIMD_CFLAGS because they include cannam@95: "simd.h", then taint.c requires SIMD_CFLAGS as well, and thus we need cannam@95: some other hack. cannam@95: cannam@95: M ./dft/simd/Makefile.am -1 cannam@95: M ./simd/Makefile.am -8 +3 cannam@95: cannam@95: Tue Feb 15 22:49:05 EST 2005 athena cannam@95: * [project @ 2005-02-16 03:49:05 by athena] cannam@95: Do not override CFLAGS in Makefile.am. cannam@95: cannam@95: M ./dft/codelets/standard/Makefile.am -1 cannam@95: M ./dft/simd/Makefile.am -1 +1 cannam@95: M ./dft/simd/codelets/Makefile.am -1 +1 cannam@95: M ./rdft/codelets/hc2r/Makefile.am -1 cannam@95: M ./rdft/codelets/r2hc/Makefile.am -1 cannam@95: M ./rdft/codelets/r2r/Makefile.am -1 cannam@95: M ./support/Makefile.codelets -1 +6 cannam@95: cannam@95: Tue Feb 15 10:30:12 EST 2005 athena cannam@95: * [project @ 2005-02-15 15:30:12 by athena] cannam@95: Allow users to build long double version even if sizeof(long double) cannam@95: == sizeof(double) cannam@95: cannam@95: M ./configure.ac -3 cannam@95: cannam@95: Mon Feb 14 19:55:38 EST 2005 athena cannam@95: * [project @ 2005-02-15 00:55:38 by athena] cannam@95: Updated for 3.1 cannam@95: cannam@95: M ./commercialize.sh -12 +20 cannam@95: cannam@95: Mon Feb 14 19:07:14 EST 2005 athena cannam@95: * [project @ 2005-02-15 00:07:14 by athena] cannam@95: Oops, version.h is no longer used cannam@95: cannam@95: M ./api/version.c -2 +1 cannam@95: cannam@95: Mon Feb 14 18:51:05 EST 2005 athena cannam@95: * [project @ 2005-02-14 23:51:05 by athena] cannam@95: unified fma and non-fma versions cannam@95: cannam@95: A ./support/twovers.sh cannam@95: M ./api/Makefile.am -5 +2 cannam@95: M ./api/version.c -2 +6 cannam@95: M ./configure.ac -5 +10 cannam@95: M ./dft/codelets/standard/Makefile.am -7 +7 cannam@95: M ./dft/simd/codelets/Makefile.am -10 +10 cannam@95: M ./m4/ocaml.m4 -9 +3 cannam@95: M ./mkdist.sh -5 cannam@95: M ./rdft/codelets/hc2r/Makefile.am -4 +4 cannam@95: M ./rdft/codelets/r2hc/Makefile.am -4 +4 cannam@95: M ./rdft/codelets/r2r/Makefile.am -9 +9 cannam@95: M ./support/Makefile.am -2 +2 cannam@95: M ./support/Makefile.codelets -7 +2 cannam@95: M ./support/twovers.sh +17 cannam@95: cannam@95: Mon Feb 14 14:12:09 EST 2005 athena cannam@95: * [project @ 2005-02-14 19:12:09 by athena] cannam@95: forgot to remove inplace/Makefile from configure.ac cannam@95: cannam@95: M ./configure.ac -1 cannam@95: cannam@95: Mon Feb 14 12:08:52 EST 2005 athena cannam@95: * [project @ 2005-02-14 17:08:49 by athena] cannam@95: Merged dft/codelets/inplace with the main dft/codelets/standard cannam@95: directory. This step makes dft codelets consistent with the rest cannam@95: of the naming conventions, and will simplify the eventual merge cannam@95: of fma and non-fma codelets. cannam@95: cannam@95: M ./Makefile.am -1 cannam@95: M ./dft/codelet-dft.h -2 +1 cannam@95: M ./dft/codelets/Makefile.am -1 +1 cannam@95: R ./dft/codelets/inplace/Makefile.am cannam@95: R ./dft/codelets/inplace/ cannam@95: M ./dft/codelets/standard/Makefile.am -9 +31 cannam@95: M ./dft/conf.c -2 +1 cannam@95: cannam@95: Mon Feb 14 11:16:15 EST 2005 athena cannam@95: * [project @ 2005-02-14 16:16:15 by athena] cannam@95: inline altivec constants, since gcc seems to generate better code this way. cannam@95: cannam@95: M ./simd/altivec.c -21 +1 cannam@95: M ./simd/simd-altivec.h -16 +18 cannam@95: cannam@95: Sun Feb 13 18:17:32 EST 2005 athena cannam@95: * [project @ 2005-02-13 23:17:32 by athena] cannam@95: group altivec constants into a single array, for faster access cannam@95: cannam@95: M ./simd/altivec.c -11 +17 cannam@95: M ./simd/simd-altivec.h -20 +11 cannam@95: cannam@95: Sun Feb 13 18:15:37 EST 2005 athena cannam@95: * [project @ 2005-02-13 23:15:37 by athena] cannam@95: code cleanup cannam@95: cannam@95: M ./genfft/c.ml -23 +9 cannam@95: M ./genfft/c.mli -2 +1 cannam@95: M ./genfft/simd.ml -6 +6 cannam@95: cannam@95: Sun Feb 13 10:29:32 EST 2005 athena cannam@95: * [project @ 2005-02-13 15:29:32 by athena] cannam@95: removed some unused stuff cannam@95: cannam@95: M ./genfft/c.ml -3 +1 cannam@95: M ./genfft/c.mli -2 +1 cannam@95: cannam@95: Sat Feb 12 22:04:40 EST 2005 athena cannam@95: * [project @ 2005-02-13 03:04:40 by athena] cannam@95: New twiddle scheme for altivec, 3dnow cannam@95: cannam@95: M ./simd/simd-3dnow.h -4 +34 cannam@95: M ./simd/simd-altivec.h -4 +27 cannam@95: cannam@95: Sat Feb 12 20:17:35 EST 2005 athena cannam@95: * [project @ 2005-02-13 01:17:35 by athena] cannam@95: Implemented new twiddle scheme for sse2 cannam@95: cannam@95: M ./simd/simd-sse2.h -4 +26 cannam@95: cannam@95: Sat Feb 12 19:57:46 EST 2005 athena cannam@95: * [project @ 2005-02-13 00:57:40 by athena] cannam@95: Implemented experimental t2* codelets, which store twiddle factors cannam@95: in a more convenient format, at the expense of twice the storage. cannam@95: Currently only SSE works; I have to port SSE2, altivec, etc. to the cannam@95: new scheme. After this, we will decide whether these codelets cannam@95: are worth the price. cannam@95: cannam@95: A ./dft/simd/t2b.h cannam@95: A ./dft/simd/t2f.h cannam@95: M ./dft/simd/Makefile.am -1 +1 cannam@95: M ./dft/simd/codelets/Makefile.am -2 +14 cannam@95: M ./dft/simd/q1b.h +5 cannam@95: M ./dft/simd/q1f.h +5 cannam@95: M ./dft/simd/t1b.h +5 cannam@95: M ./dft/simd/t1f.h +5 cannam@95: M ./dft/simd/t2b.h +36 cannam@95: M ./dft/simd/t2f.h +36 cannam@95: M ./simd/simd-sse.h -36 +26 cannam@95: cannam@95: Fri Feb 11 08:07:12 EST 2005 athena cannam@95: * [project @ 2005-02-11 13:07:12 by athena] cannam@95: Forgot to define SIMD_STRIDE_OKPAIR cannam@95: cannam@95: M ./simd/simd-altivec.h +1 cannam@95: cannam@95: Thu Feb 10 22:20:00 EST 2005 athena cannam@95: * [project @ 2005-02-11 03:20:00 by athena] cannam@95: fixed sse2, 3dnow, and altivec, as promised cannam@95: cannam@95: M ./simd/simd-3dnow.h -5 +3 cannam@95: M ./simd/simd-altivec.h -1 +4 cannam@95: M ./simd/simd-sse.h +1 cannam@95: M ./simd/simd-sse2.h -5 +3 cannam@95: cannam@95: Thu Feb 10 21:47:40 EST 2005 athena cannam@95: * [project @ 2005-02-11 02:47:33 by athena] cannam@95: Generate n2?v_* codelets in such a way that we may or may not cannam@95: pair stores, depending on which mode happens to work best on cannam@95: a particular SIMD implementation. sse2, 3dnow, and altivec cannam@95: are currently broken---will fix soon. cannam@95: cannam@95: M ./dft/simd/n2b.c -1 +1 cannam@95: M ./dft/simd/n2f.c -1 +1 cannam@95: M ./genfft/annotate.ml -7 +9 cannam@95: M ./genfft/expr.ml -4 +6 cannam@95: M ./genfft/expr.mli -2 +2 cannam@95: M ./genfft/simd.ml -3 +7 cannam@95: M ./simd/simd-sse.h -1 +6 cannam@95: cannam@95: Thu Feb 10 08:53:22 EST 2005 athena cannam@95: * [project @ 2005-02-10 13:53:22 by athena] cannam@95: instantiate altivec constants only once cannam@95: cannam@95: M ./simd/altivec.c -1 +6 cannam@95: M ./simd/simd-altivec.h -4 +10 cannam@95: cannam@95: Thu Feb 10 06:37:56 EST 2005 athena cannam@95: * [project @ 2005-02-10 11:37:56 by athena] cannam@95: Fixed alignment checks for new SIMD scheme cannam@95: cannam@95: M ./dft/simd/n2b.c -3 +3 cannam@95: M ./dft/simd/n2f.c -3 +3 cannam@95: cannam@95: Wed Feb 9 21:35:01 EST 2005 athena cannam@95: * [project @ 2005-02-10 02:35:01 by athena] cannam@95: Change n2?v_* codelets to store pairs of vectors, with implicit cannam@95: 2x2 transposition. Works for 2-way SIMD as well. Tested with sse cannam@95: and sse2. I haven't tried altivec yet, but I observed a huge cannam@95: speedup when I transformed one codelet by hand. cannam@95: cannam@95: M ./dft/simd/codelets/Makefile.am -8 +6 cannam@95: M ./genfft/annotate.ml -3 +48 cannam@95: M ./genfft/annotate.mli -2 +5 cannam@95: M ./genfft/expr.ml -2 +5 cannam@95: M ./genfft/expr.mli -2 +2 cannam@95: M ./genfft/gen_notw_c.ml -3 +17 cannam@95: M ./genfft/genutil.ml -3 +2 cannam@95: M ./genfft/simd.ml -5 +9 cannam@95: M ./genfft/simdmagic.ml -1 +3 cannam@95: M ./simd/simd-3dnow.h +6 cannam@95: M ./simd/simd-altivec.h +10 cannam@95: M ./simd/simd-sse.h +6 cannam@95: M ./simd/simd-sse2.h +6 cannam@95: cannam@95: Tue Feb 8 21:28:38 EST 2005 athena cannam@95: * [project @ 2005-02-09 02:28:38 by athena] cannam@95: Resurrected old DIF codelets for experimental purposes. They cannam@95: are disabled for now, but I am keeping the setup around for cannam@95: future reference. cannam@95: cannam@95: M ./dft/codelets/standard/Makefile.am -1 +14 cannam@95: cannam@95: Tue Feb 8 20:10:19 EST 2005 stevenj cannam@95: * [project @ 2005-02-09 01:09:12 by stevenj] cannam@95: cannam@95: M ./doc/fftw3.texi -7 +7 cannam@95: cannam@95: Tue Feb 8 19:37:09 EST 2005 stevenj cannam@95: * [project @ 2005-02-09 00:37:09 by stevenj] cannam@95: clarifications, document --with-portable-binary and --with-gcc-arch cannam@95: cannam@95: M ./doc/fftw3.texi -17 +30 cannam@95: cannam@95: Tue Feb 8 19:23:41 EST 2005 stevenj cannam@95: * [project @ 2005-02-09 00:20:56 by stevenj] cannam@95: cannam@95: M ./NEWS -7 +7 cannam@95: cannam@95: Tue Feb 8 01:36:22 EST 2005 stevenj cannam@95: * [project @ 2005-02-08 06:36:22 by stevenj] cannam@95: more change comments cannam@95: cannam@95: M ./NEWS +61 cannam@95: cannam@95: Tue Feb 8 00:41:38 EST 2005 stevenj cannam@95: * [project @ 2005-02-08 05:41:38 by stevenj] cannam@95: fma is definitely beneficial on Itanium with the HP/UX compiler cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -1 +1 cannam@95: cannam@95: Mon Feb 7 22:58:47 EST 2005 athena cannam@95: * [project @ 2005-02-08 03:58:47 by athena] cannam@95: Silence warnings. cannam@95: cannam@95: M ./libbench2/bench-main.c -3 +4 cannam@95: cannam@95: Mon Feb 7 22:55:49 EST 2005 stevenj cannam@95: * [project @ 2005-02-08 03:55:49 by stevenj] cannam@95: when we compile our own getopt, change symbol names to avoid conflicts (e.g. avoid build failure on MacOS X with --enable-shared) cannam@95: cannam@95: M ./libbench2/getopt.h +26 cannam@95: cannam@95: Mon Feb 7 22:36:42 EST 2005 stevenj cannam@95: * [project @ 2005-02-08 03:36:42 by stevenj] cannam@95: grr, more bugfixes for in-place case cannam@95: cannam@95: M ./reodft/reodft00e-splitradix.c -9 +13 cannam@95: cannam@95: Mon Feb 7 22:29:35 EST 2005 athena cannam@95: * [project @ 2005-02-08 03:29:35 by athena] cannam@95: removed relics of FRANZ mode cannam@95: cannam@95: M ./dft/codelets/standard/Makefile.am -19 cannam@95: cannam@95: Mon Feb 7 18:48:36 EST 2005 athena cannam@95: * [project @ 2005-02-07 23:48:36 by athena] cannam@95: Somehow xlc does not like ``vector int dummy;'' cannam@95: cannam@95: M ./simd/altivec.c -2 +2 cannam@95: cannam@95: Mon Feb 7 13:59:47 EST 2005 athena cannam@95: * [project @ 2005-02-07 18:59:47 by athena] cannam@95: There is no need to enable sse to make the distribution. This might cannam@95: have been true in the past but not anymore. cannam@95: cannam@95: M ./mkdist.sh -6 +8 cannam@95: cannam@95: Mon Feb 7 13:55:17 EST 2005 athena cannam@95: * [project @ 2005-02-07 18:55:17 by athena] cannam@95: Oops---included fortran file in C sources cannam@95: cannam@95: M ./api/Makefile.am -1 +1 cannam@95: cannam@95: Mon Feb 7 13:42:45 EST 2005 athena cannam@95: * [project @ 2005-02-07 18:42:45 by athena] cannam@95: Set version string at ``make dist'' time, not at ``configure'' time, cannam@95: so we know whether a user is using the fma version or not. cannam@95: cannam@95: M ./api/Makefile.am -2 +9 cannam@95: M ./api/version.c -2 +3 cannam@95: cannam@95: Sun Feb 6 17:00:33 EST 2005 athena cannam@95: * [project @ 2005-02-06 22:00:33 by athena] cannam@95: Removed useless files cannam@95: cannam@95: R ./genfft/gen_hc2r_noinline.ml cannam@95: R ./genfft/gen_notw_noinline.ml cannam@95: R ./genfft/gen_notw_noinline_c.ml cannam@95: R ./genfft/gen_r2hc_noinline.ml cannam@95: cannam@95: Sun Feb 6 16:59:39 EST 2005 athena cannam@95: * [project @ 2005-02-06 21:59:39 by athena] cannam@95: Different (simpler?) way to prevent the compiler from optimizing loop cannam@95: inductive variables. We now explicitly corrupt stride variables by cannam@95: xor-ing them with another variable that happens to be zero (but the cannam@95: compiler does not know it). In this way, the compiler does not cannam@95: attempt to extract a zillion loop indices from codelets, which would cannam@95: overflow the register set. Set the -fno-loop-optimize flag to further cannam@95: help the process. cannam@95: cannam@95: Consequences: removed m* codelets. Smaller library size. Slightly cannam@95: faster code with gcc/powerpc (including altivec). Much faster code cannam@95: with xlc/powerpc. No changes for gcc/pentium. Maybe slightly faster cannam@95: with icc/pentium. cannam@95: cannam@95: M ./configure.ac -11 +9 cannam@95: M ./dft/codelets/standard/Makefile.am -52 +3 cannam@95: M ./dft/simd/codelets/Makefile.am -30 +6 cannam@95: M ./genfft/Makefile.am -36 +10 cannam@95: M ./genfft/c.ml -1 +4 cannam@95: M ./genfft/c.mli -1 +2 cannam@95: M ./genfft/gen_hc2hc.ml -3 +5 cannam@95: M ./genfft/gen_hc2r.ml -3 +7 cannam@95: M ./genfft/gen_notw.ml -3 +6 cannam@95: M ./genfft/gen_notw_c.ml -3 +6 cannam@95: M ./genfft/gen_r2hc.ml -3 +7 cannam@95: M ./genfft/gen_twiddle.ml -3 +5 cannam@95: M ./genfft/gen_twiddle_c.ml -3 +5 cannam@95: M ./genfft/gen_twidsq.ml -3 +6 cannam@95: M ./genfft/gen_twidsq_c.ml -3 +6 cannam@95: M ./genfft/genutil.ml -1 +2 cannam@95: M ./genfft/simd.ml -1 +2 cannam@95: M ./kernel/ifftw.h -3 +7 cannam@95: M ./kernel/stride.c -1 +3 cannam@95: M ./rdft/codelets/hc2r/Makefile.am -13 +3 cannam@95: M ./rdft/codelets/r2hc/Makefile.am -13 +3 cannam@95: M ./support/Makefile.codelets -4 cannam@95: cannam@95: Sat Feb 5 18:51:08 EST 2005 stevenj cannam@95: * [project @ 2005-02-05 23:51:08 by stevenj] cannam@95: paranoia about in-place rodft00 plans cannam@95: cannam@95: M ./reodft/reodft00e-splitradix.c -3 +17 cannam@95: cannam@95: Sat Feb 5 18:39:55 EST 2005 stevenj cannam@95: * [project @ 2005-02-05 23:39:55 by stevenj] cannam@95: don't believe pcost when using the estimator...there is no point, and cannam@95: it screws up estimator hacks to prefer in-codelet loops to vecloops cannam@95: cannam@95: M ./kernel/planner.c -2 +2 cannam@95: cannam@95: Sat Feb 5 18:34:25 EST 2005 athena cannam@95: * [project @ 2005-02-05 23:34:25 by athena] cannam@95: Reduced optimization level from -O3 to -O for xlc, since -O generates cannam@95: faster code. cannam@95: cannam@95: M ./m4/ax_cc_maxopt.m4 -4 +4 cannam@95: cannam@95: Sat Feb 5 16:26:58 EST 2005 stevenj cannam@95: * [project @ 2005-02-05 21:26:58 by stevenj] cannam@95: whoops, only applicable to redft00/rodft00 plans cannam@95: cannam@95: M ./reodft/reodft00e-splitradix.c -1 +2 cannam@95: cannam@95: Sat Feb 5 16:22:39 EST 2005 stevenj cannam@95: * [project @ 2005-02-05 21:22:39 by stevenj] cannam@95: fixed in-place operation, and don't create size-0 sub-plans cannam@95: cannam@95: M ./reodft/reodft00e-splitradix.c -17 +16 cannam@95: cannam@95: Fri Feb 4 11:30:30 EST 2005 athena cannam@95: * [project @ 2005-02-04 16:30:30 by athena] cannam@95: Autodetect altivec on linux. This code works with gcc-3.4 and cannam@95: -maltivec, with or without -mabi=altivec. The code *should* work with cannam@95: gcc-3.3 without -mabi=altivec. However, disabling -mabi=altivec on cannam@95: gcc-3.4 produces much worse code (I don't know why). cannam@95: cannam@95: M ./simd/altivec.c -5 +31 cannam@95: cannam@95: Fri Jan 28 00:04:58 EST 2005 stevenj cannam@95: * [project @ 2005-01-28 05:04:58 by stevenj] cannam@95: update reference cannam@95: cannam@95: M ./doc/fftw3.texi -3 +3 cannam@95: cannam@95: Thu Jan 27 15:48:28 EST 2005 stevenj cannam@95: * [project @ 2005-01-27 20:48:28 by stevenj] cannam@95: note that DCT-II/III are often called the'' DCT/DCT cannam@95: cannam@95: M ./doc/fftw3.texi -11 +17 cannam@95: cannam@95: Fri Jan 21 14:42:04 EST 2005 stevenj cannam@95: * [project @ 2005-01-21 19:42:04 by stevenj] cannam@95: added MSVC++ for ia64 (based on information at http://www.intel.com/cd/ids/developer/asmo-na/eng/19949.htm?prn=Y) cannam@95: cannam@95: M ./kernel/cycle.h -1 +21 cannam@95: cannam@95: Fri Jan 21 14:22:50 EST 2005 stevenj cannam@95: * [project @ 2005-01-21 19:22:50 by stevenj] cannam@95: vc++ defines _M_AMD64 on x86-64, apparently cannam@95: cannam@95: M ./kernel/cycle.h -3 +3 cannam@95: cannam@95: Tue Jan 18 22:30:27 EST 2005 stevenj cannam@95: * [project @ 2005-01-19 03:30:27 by stevenj] cannam@95: avoid gratuitous breakage with -Werror, requested by Simon Perreault cannam@95: cannam@95: M ./m4/acx_pthread.m4 -2 +2 cannam@95: cannam@95: Mon Jan 17 18:54:55 EST 2005 stevenj cannam@95: * [project @ 2005-01-17 23:54:55 by stevenj] cannam@95: comment typo cannam@95: cannam@95: M ./m4/ax_gcc_aligns_stack.m4 -2 +2 cannam@95: cannam@95: Sat Jan 15 16:56:23 EST 2005 stevenj cannam@95: * [project @ 2005-01-15 21:56:23 by stevenj] cannam@95: bumped shared-lib revision# cannam@95: cannam@95: M ./configure.ac -1 +1 cannam@95: cannam@95: Sat Jan 15 16:35:42 EST 2005 stevenj cannam@95: * [project @ 2005-01-15 21:35:42 by stevenj] cannam@95: add X(estimate_cost) to get estimator cost, and print from bench, to aid in tweaking estimator cannam@95: cannam@95: M ./api/fftw3.h -1 +2 cannam@95: M ./api/flops.c +5 cannam@95: M ./kernel/ifftw.h -1 +2 cannam@95: M ./kernel/planner.c -6 +11 cannam@95: M ./tests/bench.c -1 +1 cannam@95: cannam@95: Sat Jan 15 14:57:56 EST 2005 stevenj cannam@95: * [project @ 2005-01-15 19:57:56 by stevenj] cannam@95: cannam@95: M ./doc/fftw3.texi -3 +3 cannam@95: cannam@95: Sat Jan 15 12:57:07 EST 2005 stevenj cannam@95: * [project @ 2005-01-15 17:57:07 by stevenj] cannam@95: formatting fix cannam@95: cannam@95: M ./doc/fftw3.texi -3 +9 cannam@95: cannam@95: Sat Jan 15 12:31:28 EST 2005 stevenj cannam@95: * [project @ 2005-01-15 17:31:28 by stevenj] cannam@95: tweaks cannam@95: cannam@95: M ./doc/fftw3.texi -17 +17 cannam@95: M ./reodft/Makefile.am +1 cannam@95: M ./reodft/conf.c -4 +7 cannam@95: cannam@95: Sat Jan 15 12:03:24 EST 2005 stevenj cannam@95: * [project @ 2005-01-15 17:03:24 by stevenj] cannam@95: use less buffer space cannam@95: cannam@95: M ./reodft/reodft00e-splitradix.c -22 +23 cannam@95: cannam@95: Sat Jan 15 01:41:58 EST 2005 stevenj cannam@95: * [project @ 2005-01-15 06:41:58 by stevenj] cannam@95: added split-radix-based dct/dst I for odd n cannam@95: cannam@95: A ./reodft/reodft00e-splitradix.c cannam@95: M ./doc/fftw3.texi -16 +21 cannam@95: M ./reodft/Makefile.am -1 +1 cannam@95: M ./reodft/conf.c -1 +2 cannam@95: M ./reodft/redft00e-r2hc.c -2 +4 cannam@95: M ./reodft/reodft.h +1 cannam@95: M ./reodft/reodft00e-splitradix.c +337 cannam@95: M ./reodft/rodft00e-r2hc.c -2 +4 cannam@95: cannam@95: Fri Jan 14 21:50:08 EST 2005 stevenj cannam@95: * [project @ 2005-01-15 02:50:08 by stevenj] cannam@95: cannam@95: M ./api/fftw3.h -3 +3 cannam@95: cannam@95: Fri Jan 14 21:49:55 EST 2005 stevenj cannam@95: * [project @ 2005-01-15 02:49:55 by stevenj] cannam@95: warn silly users who confuse CVS id with FFTW version cannam@95: cannam@95: M ./api/fftw3.h -1 +3 cannam@95: cannam@95: Fri Jan 14 16:57:36 EST 2005 stevenj cannam@95: * [project @ 2005-01-14 21:57:36 by stevenj] cannam@95: get sparc cpu type on solaris as well as with linux cannam@95: cannam@95: M ./m4/ax_gcc_archflag.m4 -13 +11 cannam@95: cannam@95: Thu Jan 13 19:21:58 EST 2005 stevenj cannam@95: * [project @ 2005-01-14 00:21:58 by stevenj] cannam@95: detect prescott mobile (f37) cannam@95: cannam@95: M ./m4/ax_gcc_archflag.m4 -2 +2 cannam@95: cannam@95: Thu Jan 13 18:09:52 EST 2005 stevenj cannam@95: * [project @ 2005-01-13 23:09:52 by stevenj] cannam@95: use cpuid for x86_64 as well as i[56]86 cannam@95: cannam@95: M ./bootstrap.sh +1 cannam@95: M ./m4/ax_gcc_archflag.m4 -2 +2 cannam@95: cannam@95: Thu Jan 13 17:59:55 EST 2005 stevenj cannam@95: * [project @ 2005-01-13 22:59:55 by stevenj] cannam@95: update with x86info 1.7 and other sources (identify k8, nocona, etc), handle nonzero leading bytes in eax cannam@95: cannam@95: M ./m4/ax_gcc_archflag.m4 -23 +27 cannam@95: cannam@95: Thu Jan 13 16:30:33 EST 2005 stevenj cannam@95: * [project @ 2005-01-13 21:30:33 by stevenj] cannam@95: compactified check for JOINABLE; use AC_DEFINE_UNQUOTED instead of AC_DEFINE for PTHREAD_CREATE_JOINABLE (thanks to Oliver Niekrenz for the bug report) cannam@95: cannam@95: M ./m4/acx_pthread.m4 -23 +16 cannam@95: cannam@95: Wed Jan 12 12:22:13 EST 2005 athena cannam@95: * [project @ 2005-01-12 17:22:13 by athena] cannam@95: The scheduler hack was incorrect because it swapped instructions cannam@95: of the form A = *B and *B = C. Fixed. cannam@95: cannam@95: M ./genfft/annotate.ml -7 +13 cannam@95: cannam@95: Tue Jan 11 22:13:24 EST 2005 athena cannam@95: * [project @ 2005-01-12 03:13:24 by athena] cannam@95: Quote expressions such as ``if test $FOO = yes'' when $FOO may be cannam@95: empty. Also, $GCC is set to either ``yes'' or empty, never to ``no''. cannam@95: cannam@95: M ./m4/ax_cc_maxopt.m4 -2 +2 cannam@95: M ./m4/ax_gcc_archflag.m4 -2 +2 cannam@95: cannam@95: Tue Jan 11 19:30:47 EST 2005 athena cannam@95: * [project @ 2005-01-12 00:30:47 by athena] cannam@95: Hmm---somehow the previous commit did not work. cannam@95: cannam@95: M ./TODO -6 cannam@95: M ./configure.ac +9 cannam@95: M ./simd/altivec.c -3 +3 cannam@95: cannam@95: Tue Jan 11 16:54:45 EST 2005 athena cannam@95: * [project @ 2005-01-11 21:54:45 by athena] cannam@95: Fixed various gcc-related problems on powerpc: cannam@95: - gcc-3.4 becomes totally confused by expressions like cannam@95: vec_add(a, vec_add(b, vec_add(c, ...))) cannam@95: The compiler uses gigabytes of memory and then crashes, presumably cannam@95: because of the exponential-time search problem involved in typing the cannam@95: above expression (since vec_add can take either ints or floats). cannam@95: I changed VADD and similar macros to be inline functions, thus cannam@95: constraining the type system. cannam@95: cannam@95: - New flags cannam@95: --param inline-unit-growth=1000 --param large-function-growth=1000 cannam@95: to work around limitations of the gcc-3.4 inliner. cannam@95: cannam@95: M ./simd/simd-altivec.h -14 +12 cannam@95: cannam@95: Mon Jan 10 21:27:24 EST 2005 athena cannam@95: * [project @ 2005-01-11 02:27:24 by athena] cannam@95: Check for HAVE_ALTIVEC_H cannam@95: cannam@95: M ./simd/simd-altivec.h -1 +2 cannam@95: cannam@95: Mon Jan 10 21:09:30 EST 2005 athena cannam@95: * [project @ 2005-01-11 02:09:30 by athena] cannam@95: Remove support for altivec using gcc builtins, since these keep cannam@95: changing across gcc versions. These changes work on gcc-3.4/linux; I cannam@95: haven't tried MacOS X yet. (The altivec ``spec'' differs between cannam@95: Motorola/Apple and gcc, grrr...) cannam@95: cannam@95: M ./configure.ac -1 +2 cannam@95: M ./simd/altivec.c -8 +5 cannam@95: M ./simd/simd-altivec.h -69 +5 cannam@95: cannam@95: Mon Jan 10 18:57:30 EST 2005 athena cannam@95: * [project @ 2005-01-10 23:57:30 by athena] cannam@95: Stylistic changes cannam@95: cannam@95: M ./rdft/rank0-vrank2.c -21 +21 cannam@95: cannam@95: Mon Jan 10 17:34:41 EST 2005 athena cannam@95: * [project @ 2005-01-10 22:34:41 by athena] cannam@95: Changed incorrect ugliness condition. cannam@95: cannam@95: M ./rdft/dft-r2hc.c -4 +2 cannam@95: cannam@95: Mon Jan 10 16:09:43 EST 2005 stevenj cannam@95: * [project @ 2005-01-10 21:09:43 by stevenj] cannam@95: note x86info version number that was used, to make it easier to update cannam@95: the cpuid for changes in later versions cannam@95: cannam@95: M ./m4/ax_gcc_archflag.m4 -2 +2 cannam@95: cannam@95: Mon Jan 10 15:00:51 EST 2005 athena cannam@95: * [project @ 2005-01-10 20:00:51 by athena] cannam@95: Make dft-r2hc non-UGLY for rank-0 problems cannam@95: cannam@95: M ./rdft/dft-r2hc.c -1 +6 cannam@95: cannam@95: Mon Jan 10 14:50:23 EST 2005 athena cannam@95: * [project @ 2005-01-10 19:50:23 by athena] cannam@95: Do not use -mcpu=970 on power4 processors, because power4 does cannam@95: not have altivec. cannam@95: cannam@95: M ./m4/ax_gcc_archflag.m4 -2 +2 cannam@95: cannam@95: Mon Jan 10 14:48:47 EST 2005 athena cannam@95: * [project @ 2005-01-10 19:48:47 by athena] cannam@95: Note gcc-3.4 problem with inlining. cannam@95: cannam@95: M ./TODO +5 cannam@95: cannam@95: Mon Jan 10 13:51:08 EST 2005 athena cannam@95: * [project @ 2005-01-10 18:51:08 by athena] cannam@95: Oops, forgot to remove ``static'' from the declaration of noninlinable cannam@95: functions. cannam@95: cannam@95: M ./genfft/gen_hc2r_noinline.ml -3 +3 cannam@95: M ./genfft/gen_notw_noinline_c.ml -3 +3 cannam@95: cannam@95: Mon Jan 10 12:31:26 EST 2005 athena cannam@95: * [project @ 2005-01-10 17:31:26 by athena] cannam@95: Recognize power4. Use ``head -n COUNT'' instead of obsolete ``head cannam@95: -COUNT'' (which fails on gentoo). cannam@95: cannam@95: M ./m4/ax_gcc_archflag.m4 -3 +4 cannam@95: cannam@95: Sun Jan 9 22:12:16 EST 2005 athena cannam@95: * [project @ 2005-01-10 03:12:16 by athena] cannam@95: Remind to add FAQ entry concerning gcc-3.4.[1-3] crashes. cannam@95: cannam@95: M ./TODO +4 cannam@95: cannam@95: Sun Jan 9 21:53:08 EST 2005 stevenj cannam@95: * [project @ 2005-01-10 02:53:08 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./m4/ax_gcc_version.m4 -3 +3 cannam@95: cannam@95: Sun Jan 9 21:48:02 EST 2005 stevenj cannam@95: * [project @ 2005-01-10 02:48:02 by stevenj] cannam@95: support checking for major.minor.patchlevel cannam@95: cannam@95: M ./m4/ax_gcc_version.m4 -10 +11 cannam@95: cannam@95: Sun Jan 9 21:40:18 EST 2005 athena cannam@95: * [project @ 2005-01-10 02:40:18 by athena] cannam@95: Revert CODELET_OPTIM to -O on IA32, which is faster than -O2. cannam@95: cannam@95: M ./configure.ac -3 +4 cannam@95: cannam@95: Sun Jan 9 20:30:12 EST 2005 athena cannam@95: * [project @ 2005-01-10 01:30:12 by athena] cannam@95: /bin/sh allows no spaces in assignments. cannam@95: cannam@95: M ./configure.ac -1 +1 cannam@95: cannam@95: Sun Jan 9 20:05:55 EST 2005 athena cannam@95: * [project @ 2005-01-10 01:05:55 by athena] cannam@95: Make non-inlinable functions external, so that gcc becomes confused cannam@95: and does not try to inline them. cannam@95: cannam@95: M ./genfft/gen_hc2r_noinline.ml -5 +5 cannam@95: M ./genfft/gen_notw_noinline.ml -5 +5 cannam@95: M ./genfft/gen_notw_noinline_c.ml -5 +5 cannam@95: cannam@95: Sun Jan 9 13:44:25 EST 2005 athena cannam@95: * [project @ 2005-01-09 18:44:25 by athena] cannam@95: Add -fno-web to CFLAGS, because -fweb destroys FMAs. cannam@95: cannam@95: M ./configure.ac -13 +18 cannam@95: cannam@95: Sun Jan 9 10:31:47 EST 2005 athena cannam@95: * [project @ 2005-01-09 15:31:47 by athena] cannam@95: Allow -mcpu=970 besides -mcpu=G5 cannam@95: cannam@95: M ./m4/ax_gcc_archflag.m4 -2 +2 cannam@95: cannam@95: Sun Jan 9 10:26:20 EST 2005 athena cannam@95: * [project @ 2005-01-09 15:26:20 by athena] cannam@95: configure was not using -fno-schedule-insns :-( cannam@95: cannam@95: M ./configure.ac -2 +1 cannam@95: cannam@95: Sun Jan 9 08:52:40 EST 2005 athena cannam@95: * [project @ 2005-01-09 13:52:40 by athena] cannam@95: In mkplan() and elsewhere, use solver index instead of solver cannam@95: *pointer*, which looks marginally clearer. cannam@95: cannam@95: M ./kernel/planner.c -18 +22 cannam@95: cannam@95: Sun Jan 9 08:15:36 EST 2005 athena cannam@95: * [project @ 2005-01-09 13:15:36 by athena] cannam@95: Split planner hash table into two tables, for blessed and unblessed cannam@95: solutions respectively. Now an unblessed solution never overwrites a cannam@95: blessed solution, thus avoiding wisdom leakage by construction. cannam@95: Further, forget() is now a O(1) operation, which speeds up the cannam@95: estimator when the wisdom table is large. cannam@95: cannam@95: M ./TODO -6 cannam@95: M ./kernel/ifftw.h -14 +17 cannam@95: M ./kernel/planner.c -157 +126 cannam@95: cannam@95: Sat Jan 8 21:19:45 EST 2005 athena cannam@95: * [project @ 2005-01-09 02:19:45 by athena] cannam@95: New TODO idea. cannam@95: cannam@95: M ./TODO +7 cannam@95: cannam@95: Thu Jan 6 11:02:29 EST 2005 athena cannam@95: * [project @ 2005-01-06 16:02:29 by athena] cannam@95: Split search() into two routines to make the UGLY/NO_UGLY logic cannam@95: obvious. cannam@95: cannam@95: M ./kernel/planner.c -41 +48 cannam@95: cannam@95: Fri Dec 17 16:08:54 EST 2004 stevenj cannam@95: * [project @ 2004-12-17 21:08:54 by stevenj] cannam@95: push/pop 64-bit registers on ia64; thanks to Orion Poplawski for the fix cannam@95: cannam@95: M ./simd/3dnow.c -1 +13 cannam@95: M ./simd/sse.c -1 +7 cannam@95: M ./simd/sse2.c -1 +7 cannam@95: cannam@95: Thu Dec 9 21:41:09 EST 2004 stevenj cannam@95: * [project @ 2004-12-10 02:41:09 by stevenj] cannam@95: patch from FreeBSD ports - FreeBSD does not have memalign, but its cannam@95: malloc is 16-byte aligned cannam@95: cannam@95: M ./kernel/kalloc.c -1 +5 cannam@95: cannam@95: Tue Nov 23 17:06:47 EST 2004 stevenj cannam@95: * [project @ 2004-11-23 22:06:47 by stevenj] cannam@95: don't compile taint.c with SIMD_CFLAGS (fixed Debian bug #259612) cannam@95: cannam@95: M ./simd/Makefile.am -3 +10 cannam@95: cannam@95: Thu Nov 18 11:37:32 EST 2004 stevenj cannam@95: * [project @ 2004-11-18 16:37:32 by stevenj] cannam@95: revert incorrect change -- codlist.c should be rebuilt, but it is built in the build directory and not in the source directory cannam@95: cannam@95: M ./support/Makefile.codelets -1 +1 cannam@95: cannam@95: Wed Nov 17 22:53:53 EST 2004 stevenj cannam@95: * [project @ 2004-11-18 03:53:53 by stevenj] cannam@95: $(CODLIST) should be rebuilt only if Makefile.am changes, or cannam@95: alternatively only in maintainer mode, to prevent stomping in the cannam@95: source directory during user builds. (Thanks to Grant Cook for the cannam@95: bug report.) cannam@95: cannam@95: M ./support/Makefile.codelets -1 +1 cannam@95: cannam@95: Sat Nov 13 13:43:01 EST 2004 stevenj cannam@95: * [project @ 2004-11-13 18:43:01 by stevenj] cannam@95: corrected #ifdef for icc/ia64, thanks to Matt Boman cannam@95: cannam@95: M ./kernel/cycle.h -16 +18 cannam@95: cannam@95: Sat Nov 13 13:34:55 EST 2004 stevenj cannam@95: * [project @ 2004-11-13 18:34:55 by stevenj] cannam@95: spelling correction (Larsen, not Larson) cannam@95: cannam@95: M ./NEWS -1 +1 cannam@95: cannam@95: Mon Nov 8 22:12:39 EST 2004 stevenj cannam@95: * [project @ 2004-11-09 03:12:39 by stevenj] cannam@95: use standard withval cannam@95: cannam@95: M ./m4/ax_gcc_archflag.m4 -3 +3 cannam@95: cannam@95: Mon Nov 8 22:09:16 EST 2004 stevenj cannam@95: * [project @ 2004-11-09 03:09:16 by stevenj] cannam@95: match doc cannam@95: cannam@95: M ./m4/ax_gcc_x86_cpuid.m4 -2 +2 cannam@95: cannam@95: Mon Nov 8 22:00:34 EST 2004 stevenj cannam@95: * [project @ 2004-11-09 03:00:34 by stevenj] cannam@95: formatting cannam@95: cannam@95: M ./m4/ax_openmp.m4 -3 +3 cannam@95: cannam@95: Mon Nov 8 21:59:33 EST 2004 stevenj cannam@95: * [project @ 2004-11-09 02:59:33 by stevenj] cannam@95: make sure OPENMP_CFLAGS environment variable is used correctly cannam@95: cannam@95: M ./m4/ax_openmp.m4 -2 +6 cannam@95: cannam@95: Mon Nov 8 21:46:50 EST 2004 stevenj cannam@95: * [project @ 2004-11-09 02:46:50 by stevenj] cannam@95: replace ax_check_cc_flags with more generic ax_check_compiler_flags cannam@95: cannam@95: A ./m4/ax_check_compiler_flags.m4 cannam@95: M ./configure.ac -13 +8 cannam@95: M ./m4/ax_cc_maxopt.m4 -7 +6 cannam@95: R ./m4/ax_check_cc_flags.m4 cannam@95: M ./m4/ax_check_compiler_flags.m4 +37 cannam@95: M ./m4/ax_gcc_aligns_stack.m4 -5 +4 cannam@95: M ./m4/ax_gcc_archflag.m4 -3 +3 cannam@95: cannam@95: Mon Nov 8 17:49:42 EST 2004 stevenj cannam@95: * [project @ 2004-11-08 22:49:42 by stevenj] cannam@95: separate macro for OpenMP test cannam@95: cannam@95: A ./m4/ax_openmp.m4 cannam@95: M ./configure.ac -40 +6 cannam@95: M ./m4/ax_cc_maxopt.m4 -3 +5 cannam@95: M ./m4/ax_openmp.m4 +58 cannam@95: cannam@95: Fri Nov 5 16:24:22 EST 2004 stevenj cannam@95: * [project @ 2004-11-05 21:24:22 by stevenj] cannam@95: typo cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Fri Oct 29 00:48:13 EDT 2004 stevenj cannam@95: * [project @ 2004-10-29 04:48:13 by stevenj] cannam@95: cannam@95: M ./configure.ac -12 +19 cannam@95: cannam@95: Thu Oct 28 00:09:38 EDT 2004 stevenj cannam@95: * [project @ 2004-10-28 04:09:38 by stevenj] cannam@95: better guessing of sparc type on Linux cannam@95: cannam@95: M ./m4/ax_gcc_archflag.m4 -2 +15 cannam@95: cannam@95: Wed Oct 27 13:44:08 EDT 2004 stevenj cannam@95: * [project @ 2004-10-27 17:44:08 by stevenj] cannam@95: note default cannam@95: cannam@95: M ./m4/ax_gcc_archflag.m4 -2 +2 cannam@95: cannam@95: Wed Oct 27 13:41:57 EDT 2004 stevenj cannam@95: * [project @ 2004-10-27 17:41:57 by stevenj] cannam@95: tweak cannam@95: cannam@95: M ./m4/ax_gcc_archflag.m4 -2 +2 cannam@95: cannam@95: Wed Oct 27 13:34:25 EDT 2004 stevenj cannam@95: * [project @ 2004-10-27 17:34:25 by stevenj] cannam@95: comment cannam@95: cannam@95: M ./m4/ax_gcc_x86_cpuid.m4 -2 +3 cannam@95: cannam@95: Wed Oct 27 13:31:10 EDT 2004 stevenj cannam@95: * [project @ 2004-10-27 17:31:10 by stevenj] cannam@95: whoops, m4 is EXTRA_DIST, not SUBDIR, since it doesn't have a Makefile cannam@95: cannam@95: M ./Makefile.am -2 +2 cannam@95: cannam@95: Wed Oct 27 13:16:57 EDT 2004 stevenj cannam@95: * [project @ 2004-10-27 17:16:57 by stevenj] cannam@95: silence warnings cannam@95: cannam@95: M ./m4/ocaml.m4 -2 +2 cannam@95: cannam@95: Wed Oct 27 13:14:22 EDT 2004 stevenj cannam@95: * [project @ 2004-10-27 17:14:22 by stevenj] cannam@95: clean up m4 macros; try to detect correct gcc -march flag on x86; new --with-portable-binary, --with-gcc-arch= flags; use -O2 for codelets with gcc 3.4 to work around bug cannam@95: cannam@95: A ./m4/ cannam@95: A ./m4/acx_pthread.m4 cannam@95: A ./m4/amx_prog_as.m4 cannam@95: A ./m4/ax_cc_maxopt.m4 cannam@95: A ./m4/ax_check_cc_flags.m4 cannam@95: A ./m4/ax_gcc_aligns_stack.m4 cannam@95: A ./m4/ax_gcc_archflag.m4 cannam@95: A ./m4/ax_gcc_version.m4 cannam@95: A ./m4/ax_gcc_x86_cpuid.m4 cannam@95: A ./m4/ocaml.m4 cannam@95: M ./Makefile.am -2 +4 cannam@95: R ./acinclude.m4 cannam@95: R ./acx_pthread.m4 cannam@95: M ./bootstrap.sh -12 +10 cannam@95: M ./configure.ac -11 +26 cannam@95: M ./m4/acx_pthread.m4 +240 cannam@95: M ./m4/amx_prog_as.m4 +11 cannam@95: M ./m4/ax_cc_maxopt.m4 +126 cannam@95: M ./m4/ax_check_cc_flags.m4 +37 cannam@95: M ./m4/ax_gcc_aligns_stack.m4 +47 cannam@95: M ./m4/ax_gcc_archflag.m4 +147 cannam@95: M ./m4/ax_gcc_version.m4 +34 cannam@95: M ./m4/ax_gcc_x86_cpuid.m4 +37 cannam@95: M ./m4/ocaml.m4 +84 cannam@95: cannam@95: Tue Oct 26 16:46:14 EDT 2004 stevenj cannam@95: * [project @ 2004-10-26 20:46:14 by stevenj] cannam@95: rename cexp -> mcexp to avoid conflict with C99 builtin cannam@95: cannam@95: M ./libbench2/mp.c -3 +3 cannam@95: cannam@95: Mon Oct 25 16:58:23 EDT 2004 stevenj cannam@95: * [project @ 2004-10-25 20:58:23 by stevenj] cannam@95: use basename , w/o args, for compiler-name comparisons; also detect Compaq ccc on alpha-linus cannam@95: cannam@95: M ./acinclude.m4 -4 +10 cannam@95: cannam@95: Sun Oct 24 22:05:10 EDT 2004 stevenj cannam@95: * [project @ 2004-10-25 02:05:10 by stevenj] cannam@95: note recent icc problems cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -6 +9 cannam@95: cannam@95: Sun Oct 24 02:10:12 EDT 2004 stevenj cannam@95: * [project @ 2004-10-24 06:10:12 by stevenj] cannam@95: whoops, disable semaphores again (for now) cannam@95: cannam@95: M ./threads/threads.c -1 +1 cannam@95: cannam@95: Sun Oct 24 02:04:58 EDT 2004 stevenj cannam@95: * [project @ 2004-10-24 06:04:58 by stevenj] cannam@95: POSIX semaphores are *not* the same as SYSV semaphores cannam@95: cannam@95: M ./threads/threads.c -2 +2 cannam@95: cannam@95: Sun Oct 24 01:18:14 EDT 2004 stevenj cannam@95: * [project @ 2004-10-24 05:18:14 by stevenj] cannam@95: re-implement threaded stuff; dftw now takes parameters to indicate a portion of m loop cannam@95: cannam@95: A ./rdft/hc2hc.c cannam@95: A ./rdft/hc2hc.h cannam@95: A ./threads/ct.c cannam@95: A ./threads/hc2hc.c cannam@95: M ./dft/conf.c -2 +2 cannam@95: M ./dft/ct.c -8 +11 cannam@95: M ./dft/ct.h -8 +11 cannam@95: M ./dft/ctsq.c -2 +2 cannam@95: M ./dft/dft.h -5 +2 cannam@95: M ./dft/dftw-direct.c -15 +29 cannam@95: M ./dft/dftw-directbuf.c -15 +30 cannam@95: M ./dft/dftw-generic.c -14 +31 cannam@95: M ./dft/kdft-dif.c -7 +3 cannam@95: M ./dft/kdft-difsq.c -2 +2 cannam@95: M ./dft/kdft-dit.c -7 +3 cannam@95: M ./kernel/ifftw.h -1 +2 cannam@95: M ./kernel/twiddle.c -1 +10 cannam@95: M ./rdft/Makefile.am -1 +1 cannam@95: M ./rdft/conf.c -2 +2 cannam@95: R ./rdft/ct.c cannam@95: R ./rdft/ct.h cannam@95: M ./rdft/hc2hc-common.c -4 +33 cannam@95: M ./rdft/hc2hc-direct.c -26 +42 cannam@95: M ./rdft/hc2hc-directbuf.c -25 +44 cannam@95: M ./rdft/hc2hc-generic.c -37 +104 cannam@95: M ./rdft/hc2hc.c +221 cannam@95: M ./rdft/hc2hc.h +60 cannam@95: M ./rdft/khc2hc.c -8 +4 cannam@95: M ./rdft/rdft.h -3 +1 cannam@95: M ./threads/Makefile.am -2 +2 cannam@95: R ./threads/ct-dit.c cannam@95: M ./threads/ct.c +253 cannam@95: M ./threads/dft-vrank-geq1.c -2 +2 cannam@95: R ./threads/hc2hc-dif.c cannam@95: R ./threads/hc2hc-dit.c cannam@95: M ./threads/hc2hc.c +238 cannam@95: M ./threads/threads.c -27 +8 cannam@95: M ./threads/threads.h -5 +4 cannam@95: cannam@95: Thu Oct 21 20:44:51 EDT 2004 stevenj cannam@95: * [project @ 2004-10-22 00:44:51 by stevenj] cannam@95: more C++ notes cannam@95: cannam@95: M ./doc/fftw3.texi -4 +11 cannam@95: cannam@95: Thu Oct 14 09:50:38 EDT 2004 stevenj cannam@95: * [project @ 2004-10-14 13:50:38 by stevenj] cannam@95: note bug report for VC++ 6.0 from Dale Dickerhoof cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -1 +4 cannam@95: cannam@95: Fri Oct 1 16:06:59 EDT 2004 stevenj cannam@95: * [project @ 2004-10-01 20:06:59 by stevenj] cannam@95: fmt cannam@95: cannam@95: M ./api/fftw3.h -3 +3 cannam@95: cannam@95: Fri Oct 1 15:59:17 EDT 2004 stevenj cannam@95: * [project @ 2004-10-01 19:58:50 by stevenj] cannam@95: comment typo cannam@95: cannam@95: M ./rdft/vrank3-transpose.c -3 +3 cannam@95: cannam@95: Fri Oct 1 15:48:09 EDT 2004 stevenj cannam@95: * [project @ 2004-10-01 19:48:09 by stevenj] cannam@95: bug fix -- ishift/oshift only apply to execution of child plan cannam@95: cannam@95: M ./rdft/dft-r2hc.c -9 +4 cannam@95: cannam@95: Thu Sep 30 21:12:47 EDT 2004 athena cannam@95: * [project @ 2004-10-01 01:12:47 by athena] cannam@95: New planner that tries never to lose wisdom. cannam@95: cannam@95: M ./api/fftw3.h -2 +1 cannam@95: M ./api/mapflags.c -1 cannam@95: M ./kernel/ifftw.h -6 +4 cannam@95: M ./kernel/planner.c -137 +183 cannam@95: cannam@95: Thu Sep 30 13:36:43 EDT 2004 athena cannam@95: * [project @ 2004-09-30 17:36:43 by athena] cannam@95: Nested comment was triggering a warning. cannam@95: cannam@95: M ./api/fftw3.h -2 +2 cannam@95: cannam@95: Fri Sep 10 15:20:07 EDT 2004 stevenj cannam@95: * [project @ 2004-09-10 19:20:07 by stevenj] cannam@95: system "root" under dgjpp is /dev/env/DJDIR, not /dev/env/DJGPP, cannam@95: according to djgpp's libc.info; patch confirmed with J. M. Guerrero cannam@95: cannam@95: M ./api/import-system-wisdom.c -1 +1 cannam@95: cannam@95: Wed Sep 8 18:50:03 EDT 2004 stevenj cannam@95: * [project @ 2004-09-08 22:50:03 by stevenj] cannam@95: some minor portability fixes for djgpp; thanks to Juan Manuel Guerrero for the patch cannam@95: cannam@95: M ./api/import-system-wisdom.c -1 +8 cannam@95: M ./tests/Makefile.am -4 +4 cannam@95: M ./tools/fftw-wisdom-to-conf.in -1 +1 cannam@95: cannam@95: Thu Aug 19 12:41:23 EDT 2004 stevenj cannam@95: * [project @ 2004-08-19 16:41:23 by stevenj] cannam@95: pointer to tutorial for quick start cannam@95: cannam@95: M ./README +3 cannam@95: cannam@95: Thu Aug 19 12:39:50 EDT 2004 stevenj cannam@95: * [project @ 2004-08-19 16:39:50 by stevenj] cannam@95: point users to manual cannam@95: cannam@95: M ./api/fftw3.h -1 +16 cannam@95: cannam@95: Sat Aug 7 13:42:22 EDT 2004 stevenj cannam@95: * [project @ 2004-08-07 17:42:22 by stevenj] cannam@95: minor typo cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Sun Jul 18 18:54:18 EDT 2004 stevenj cannam@95: * [project @ 2004-07-18 22:54:18 by stevenj] cannam@95: use __DECCXX for Compaq cxx, not Linux-specific symbol cannam@95: cannam@95: M ./kernel/cycle.h -2 +2 cannam@95: cannam@95: Fri Jul 16 13:55:25 EDT 2004 stevenj cannam@95: * [project @ 2004-07-16 17:55:25 by stevenj] cannam@95: patch by John Bowman to make cycle counter work with DEC cxx under Linux cannam@95: cannam@95: M ./kernel/cycle.h -2 +2 cannam@95: cannam@95: Wed Jun 30 00:45:10 EDT 2004 stevenj cannam@95: * [project @ 2004-06-30 04:45:10 by stevenj] cannam@95: updated pruned FFT discussion, with link to further details on www.fftw.org/pruned.html cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -8 +16 cannam@95: M ./doc/FAQ/html.refs +1 cannam@95: cannam@95: Mon Jun 14 20:08:27 EDT 2004 stevenj cannam@95: * [project @ 2004-06-15 00:08:27 by stevenj] cannam@95: darwin is based on freebsd cannam@95: cannam@95: M ./acx_pthread.m4 -2 +2 cannam@95: cannam@95: Thu Jun 3 14:23:41 EDT 2004 stevenj cannam@95: * [project @ 2004-06-03 18:23:41 by stevenj] cannam@95: in --with-windows-f77-mangling, add lowercase + single underscore for Intel compilers, etc. (thanks to David Gomez for the bug report) cannam@95: cannam@95: M ./api/f77api.c +5 cannam@95: cannam@95: Wed Apr 7 00:46:07 EDT 2004 stevenj cannam@95: * [project @ 2004-04-07 04:46:07 by stevenj] cannam@95: whoops, extra alignment check cannam@95: cannam@95: M ./rdft/rank0-vrank2.c -1 +5 cannam@95: cannam@95: Wed Apr 7 00:16:49 EDT 2004 stevenj cannam@95: * [project @ 2004-04-07 04:16:49 by stevenj] cannam@95: disable most 2-float-as-double copying, add alignment check in one remaining place cannam@95: cannam@95: M ./kernel/ifftw.h -1 +11 cannam@95: M ./rdft/rank0-vrank2.c -22 +25 cannam@95: M ./rdft/vrank3-transpose.c -6 +6 cannam@95: cannam@95: Tue Apr 6 13:49:13 EDT 2004 stevenj cannam@95: * [project @ 2004-04-06 17:49:13 by stevenj] cannam@95: make sure it is clear that real-even/odd refers to symmetry, not size cannam@95: cannam@95: M ./doc/fftw3.texi -7 +7 cannam@95: cannam@95: Mon Apr 5 20:18:29 EDT 2004 stevenj cannam@95: * [project @ 2004-04-06 00:18:29 by stevenj] cannam@95: optimization cannam@95: cannam@95: M ./rdft/vrank3-transpose.c -11 +48 cannam@95: cannam@95: Fri Apr 2 21:31:00 EST 2004 stevenj cannam@95: * [project @ 2004-04-03 02:31:00 by stevenj] cannam@95: separate cutoff for ugliness...these cutoffs are still not ideal cannam@95: cannam@95: M ./rdft/vrank3-transpose.c -3 +4 cannam@95: cannam@95: Fri Apr 2 21:30:17 EST 2004 stevenj cannam@95: * [project @ 2004-04-03 02:30:17 by stevenj] cannam@95: transpose.c is gone cannam@95: cannam@95: M ./kernel/ifftw.h -15 +1 cannam@95: cannam@95: Fri Apr 2 21:18:27 EST 2004 stevenj cannam@95: * [project @ 2004-04-03 02:18:27 by stevenj] cannam@95: move all rank0 transforms to rdft cannam@95: cannam@95: A ./rdft/rank0-vrank2.c cannam@95: A ./rdft/vrank3-transpose.c cannam@95: M ./configure.ac +1 cannam@95: M ./dft/Makefile.am -2 +2 cannam@95: M ./dft/conf.c -4 +1 cannam@95: R ./dft/rank0.c cannam@95: R ./dft/vrank2-transpose.c cannam@95: R ./dft/vrank3-transpose.c cannam@95: M ./kernel/Makefile.am -2 +2 cannam@95: R ./kernel/transpose.c cannam@95: M ./rdft/Makefile.am -2 +3 cannam@95: M ./rdft/conf.c -1 +3 cannam@95: M ./rdft/dft-r2hc.c -20 +36 cannam@95: M ./rdft/rank0-vrank2.c +289 cannam@95: M ./rdft/rdft.h -2 +2 cannam@95: M ./rdft/vrank3-transpose.c +943 cannam@95: cannam@95: Fri Apr 2 20:35:35 EST 2004 stevenj cannam@95: * [project @ 2004-04-03 01:35:35 by stevenj] cannam@95: enable fp-moves/us comparison of rank-0 transforms cannam@95: cannam@95: M ./libbench2/mflops.c -2 +8 cannam@95: M ./libbench2/report.c -3 +7 cannam@95: cannam@95: Thu Apr 1 16:13:22 EST 2004 stevenj cannam@95: * [project @ 2004-04-01 21:13:22 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./kernel/transpose.c -1 +1 cannam@95: cannam@95: Thu Apr 1 15:25:30 EST 2004 stevenj cannam@95: * [project @ 2004-04-01 20:25:30 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./kernel/tensor7.c -3 +3 cannam@95: cannam@95: Wed Mar 31 18:11:02 EST 2004 stevenj cannam@95: * [project @ 2004-03-31 23:11:02 by stevenj] cannam@95: sort tensor dims by stride absolute values, not strides cannam@95: cannam@95: M ./kernel/tensor7.c -5 +7 cannam@95: cannam@95: Tue Mar 30 20:22:50 EST 2004 stevenj cannam@95: * [project @ 2004-03-31 01:22:50 by stevenj] cannam@95: cannam@95: M ./kernel/transpose.c -2 +1 cannam@95: cannam@95: Tue Mar 30 19:44:54 EST 2004 stevenj cannam@95: * [project @ 2004-03-31 00:44:54 by stevenj] cannam@95: added improved transpose algorithm for N x M where |N-M| is small cannam@95: cannam@95: M ./dft/dftw-generic.c +1 cannam@95: M ./dft/vrank2-transpose.c -7 +10 cannam@95: M ./dft/vrank3-transpose.c -6 +7 cannam@95: M ./kernel/ifftw.h -2 +2 cannam@95: M ./kernel/transpose.c -15 +73 cannam@95: M ./libbench2/problem.c -2 +4 cannam@95: cannam@95: Tue Mar 30 19:41:14 EST 2004 stevenj cannam@95: * [project @ 2004-03-31 00:41:14 by stevenj] cannam@95: check to make sure SIMD matches precision, and make sure user doesn't select both SSE and SSE2 cannam@95: cannam@95: M ./configure.ac -2 +7 cannam@95: cannam@95: Sun Mar 28 09:26:38 EST 2004 athena cannam@95: * [project @ 2004-03-28 14:26:38 by athena] cannam@95: Implemented hc2hc-generic hc2r. cannam@95: cannam@95: M ./rdft/hc2hc-generic.c -29 +79 cannam@95: cannam@95: Thu Mar 25 11:19:25 EST 2004 athena cannam@95: * [project @ 2004-03-25 16:19:25 by athena] cannam@95: Inverted loop for stride-1 access. cannam@95: cannam@95: M ./rdft/hc2hc-generic.c -12 +18 cannam@95: cannam@95: Thu Mar 25 11:18:49 EST 2004 athena cannam@95: * [project @ 2004-03-25 16:18:49 by athena] cannam@95: Swapped j <-> k for consistency cannam@95: cannam@95: M ./dft/dftw-generic.c -3 +3 cannam@95: cannam@95: Tue Mar 23 12:08:07 EST 2004 athena cannam@95: * [project @ 2004-03-23 17:08:07 by athena] cannam@95: Require that R be odd cannam@95: cannam@95: M ./rdft/hc2hc-generic.c -1 +1 cannam@95: cannam@95: Tue Mar 23 11:49:01 EST 2004 athena cannam@95: * [project @ 2004-03-23 16:49:01 by athena] cannam@95: Implemented hc2hc-generic (DIT only for now). cannam@95: cannam@95: A ./rdft/hc2hc-generic.c cannam@95: M ./rdft/Makefile.am -3 +3 cannam@95: M ./rdft/conf.c -5 +4 cannam@95: M ./rdft/dft-r2hc.c -2 +2 cannam@95: M ./rdft/hc2hc-generic.c +216 cannam@95: M ./rdft/rdft.h +1 cannam@95: cannam@95: Mon Mar 22 14:43:16 EST 2004 athena cannam@95: * [project @ 2004-03-22 19:43:16 by athena] cannam@95: Relax equality of twiddle description, since the `i' field cannam@95: is not used by TW_FULL or TW_HALF. cannam@95: cannam@95: M ./kernel/twiddle.c -5 +17 cannam@95: cannam@95: Mon Mar 22 13:22:44 EST 2004 athena cannam@95: * [project @ 2004-03-22 18:22:41 by athena] cannam@95: Do not allocate tw_instr's on the stack. Thus, the ``consistency check'' cannam@95: in twiddle.c becomes wrong. cannam@95: cannam@95: M ./dft/dftw-generic.c -8 +1 cannam@95: M ./kernel/twiddle.c -2 +1 cannam@95: cannam@95: Mon Mar 22 13:21:28 EST 2004 athena cannam@95: * [project @ 2004-03-22 18:21:28 by athena] cannam@95: Fixed incorrect malloc()/free() logic. cannam@95: cannam@95: M ./libbench2/mp.c -3 +6 cannam@95: cannam@95: Mon Mar 22 09:04:37 EST 2004 athena cannam@95: * [project @ 2004-03-22 14:04:37 by athena] cannam@95: Silence warnings cannam@95: cannam@95: M ./rdft/hc2hc-directbuf.c -2 +2 cannam@95: cannam@95: Mon Mar 22 09:02:55 EST 2004 athena cannam@95: * [project @ 2004-03-22 14:02:55 by athena] cannam@95: Separate file for hc2hc common routines cannam@95: cannam@95: A ./rdft/hc2hc-common.c cannam@95: M ./rdft/Makefile.am -4 +4 cannam@95: M ./rdft/ct.c -35 cannam@95: M ./rdft/hc2hc-common.c +58 cannam@95: cannam@95: Mon Mar 22 08:23:56 EST 2004 athena cannam@95: * [project @ 2004-03-22 13:23:56 by athena] cannam@95: (re)Implemented buffered hc2hc. Slight simplification of cannam@95: twiddle-factors management. cannam@95: cannam@95: A ./rdft/hc2hc-directbuf.c cannam@95: M ./dft/dftw-directbuf.c -2 +2 cannam@95: M ./rdft/Makefile.am -2 +3 cannam@95: M ./rdft/ct.h +2 cannam@95: M ./rdft/hc2hc-direct.c -16 +8 cannam@95: M ./rdft/hc2hc-directbuf.c +238 cannam@95: M ./rdft/khc2hc.c -1 +2 cannam@95: cannam@95: Sun Mar 21 19:53:05 EST 2004 athena cannam@95: * [project @ 2004-03-22 00:53:05 by athena] cannam@95: Incremented libtool revision number before we forget. cannam@95: cannam@95: M ./configure.ac -1 +1 cannam@95: cannam@95: Sun Mar 21 19:25:56 EST 2004 athena cannam@95: * [project @ 2004-03-22 00:25:56 by athena] cannam@95: Fixed opcnt cannam@95: cannam@95: M ./rdft/hc2hc-direct.c -2 +5 cannam@95: cannam@95: Sun Mar 21 17:56:15 EST 2004 athena cannam@95: * [project @ 2004-03-21 22:56:15 by athena] cannam@95: Renamed files. These solvers are not really cooley-tukey. cannam@95: cannam@95: A ./dft/dftw-direct.c cannam@95: A ./dft/dftw-directbuf.c cannam@95: A ./dft/dftw-generic.c cannam@95: M ./dft/Makefile.am -5 +5 cannam@95: R ./dft/ct-directw.c cannam@95: R ./dft/ct-directwbuf.c cannam@95: R ./dft/ct-generic.c cannam@95: M ./dft/dftw-direct.c +153 cannam@95: M ./dft/dftw-directbuf.c +213 cannam@95: M ./dft/dftw-generic.c +304 cannam@95: cannam@95: Sun Mar 21 12:38:45 EST 2004 athena cannam@95: * [project @ 2004-03-21 17:38:45 by athena] cannam@95: Started moving rdft/ to the new cooley-tukey ontology cannam@95: cannam@95: A ./rdft/ct.c cannam@95: A ./rdft/ct.h cannam@95: A ./rdft/hc2hc-direct.c cannam@95: A ./rdft/khc2hc.c cannam@95: M ./dft/ct.h -1 +1 cannam@95: M ./genfft/gen_hc2hc.ml -6 +3 cannam@95: M ./rdft/Makefile.am -5 +4 cannam@95: M ./rdft/codelet-rdft.h -3 +1 cannam@95: M ./rdft/ct.c +253 cannam@95: M ./rdft/ct.h +54 cannam@95: R ./rdft/hc2hc-buf.c cannam@95: R ./rdft/hc2hc-dif.c cannam@95: M ./rdft/hc2hc-direct.c +182 cannam@95: R ./rdft/hc2hc-dit.c cannam@95: R ./rdft/hc2hc.c cannam@95: R ./rdft/hc2hc.h cannam@95: R ./rdft/khc2hc-dif.c cannam@95: R ./rdft/khc2hc-dit.c cannam@95: M ./rdft/khc2hc.c +32 cannam@95: M ./rdft/rdft.h -6 +1 cannam@95: cannam@95: Sun Mar 21 10:59:42 EST 2004 athena cannam@95: * [project @ 2004-03-21 15:59:42 by athena] cannam@95: Plans in ct-*.c are subtypes of plan_dftw, not plan_dft cannam@95: cannam@95: M ./dft/ct-directw.c -2 +2 cannam@95: M ./dft/ct-directwbuf.c -2 +2 cannam@95: M ./dft/ct-generic.c -1 +1 cannam@95: cannam@95: Sun Mar 21 10:38:18 EST 2004 athena cannam@95: * [project @ 2004-03-21 15:38:18 by athena] cannam@95: Slight simplification cannam@95: cannam@95: M ./dft/ct-directw.c -2 +2 cannam@95: cannam@95: Sun Mar 21 10:20:06 EST 2004 athena cannam@95: * [project @ 2004-03-21 15:20:06 by athena] cannam@95: Minor simplification cannam@95: cannam@95: M ./dft/ct.c -4 +2 cannam@95: cannam@95: Sat Mar 20 08:43:57 EST 2004 athena cannam@95: * [project @ 2004-03-20 13:43:57 by athena] cannam@95: Workarounds for icc-8.0 nonsense. cannam@95: cannam@95: M ./simd/simd-sse.h +6 cannam@95: M ./simd/simd-sse2.h +6 cannam@95: cannam@95: Sun Mar 7 07:56:08 EST 2004 athena cannam@95: * [project @ 2004-03-07 12:56:08 by athena] cannam@95: FFTW_FORWARD is not technically an ``option''. cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Tue Feb 24 12:17:06 EST 2004 stevenj cannam@95: * [project @ 2004-02-24 17:17:06 by stevenj] cannam@95: Alejandro requested that his name be removed from @author cannam@95: cannam@95: M ./acx_pthread.m4 -4 +5 cannam@95: cannam@95: Mon Feb 23 17:42:56 EST 2004 stevenj cannam@95: * [project @ 2004-02-23 22:42:56 by stevenj] cannam@95: GNU Pth emulation library check cannam@95: cannam@95: M ./acx_pthread.m4 -3 +12 cannam@95: cannam@95: Sat Feb 21 17:51:13 EST 2004 stevenj cannam@95: * [project @ 2004-02-21 22:51:13 by stevenj] cannam@95: calling can-do calls the estimating-planner, which creates wisdom that we don't want ...we should be able to do all of the documented problems, anyway cannam@95: cannam@95: M ./tools/fftw-wisdom.c -1 +1 cannam@95: cannam@95: Sat Feb 21 17:46:06 EST 2004 stevenj cannam@95: * [project @ 2004-02-21 22:46:06 by stevenj] cannam@95: don't forget_wisdom because of side effects cannam@95: cannam@95: M ./tests/bench.c -1 cannam@95: cannam@95: Sat Feb 21 17:42:47 EST 2004 stevenj cannam@95: * [project @ 2004-02-21 22:42:47 by stevenj] cannam@95: forget wisdom from can_do cannam@95: cannam@95: M ./tests/bench.c +1 cannam@95: cannam@95: Thu Feb 19 14:11:14 EST 2004 stevenj cannam@95: * [project @ 2004-02-19 19:11:14 by stevenj] cannam@95: parenthesization cannam@95: cannam@95: M ./api/malloc.c -1 +1 cannam@95: cannam@95: Fri Feb 13 07:20:31 EST 2004 athena cannam@95: * [project @ 2004-02-13 12:20:31 by athena] cannam@95: Split malloc into kernel_malloc and API malloc cannam@95: cannam@95: A ./api/malloc.c cannam@95: A ./kernel/kalloc.c cannam@95: M ./api/Makefile.am -14 +14 cannam@95: M ./api/malloc.c +32 cannam@95: M ./kernel/Makefile.am -5 +5 cannam@95: M ./kernel/alloc.c -127 +5 cannam@95: M ./kernel/ifftw.h -1 +6 cannam@95: M ./kernel/kalloc.c +140 cannam@95: M ./tests/bench.c +4 cannam@95: cannam@95: Thu Feb 12 15:42:20 EST 2004 stevenj cannam@95: * [project @ 2004-02-12 20:42:20 by stevenj] cannam@95: X(malloc) must be extern "C" cannam@95: cannam@95: M ./kernel/alloc.c -1 +4 cannam@95: cannam@95: Thu Feb 12 15:41:44 EST 2004 stevenj cannam@95: * [project @ 2004-02-12 20:41:44 by stevenj] cannam@95: satsify C++ compiler cannam@95: cannam@95: M ./dft/bluestein.c -2 +2 cannam@95: cannam@95: Thu Feb 5 20:39:14 EST 2004 stevenj cannam@95: * [project @ 2004-02-06 01:39:14 by stevenj] cannam@95: with the new flags, fma is definitely beneficial on PA-RISC with HP/UX cc cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -6 +7 cannam@95: cannam@95: Thu Feb 5 19:52:17 EST 2004 stevenj cannam@95: * [project @ 2004-02-06 00:52:17 by stevenj] cannam@95: grr, Ofaster etcetera are not supported under older versions of the compiler. Note that +Ofltacc *disables* fp-reordering optimizations (which are enabled by +Oall). +Optrs_ansi is the older version of the aliasing stuff cannam@95: cannam@95: M ./acinclude.m4 -1 +1 cannam@95: cannam@95: Thu Feb 5 19:26:01 EST 2004 stevenj cannam@95: * [project @ 2004-02-06 00:26:01 by stevenj] cannam@95: +Otype_safety=ansi on hpux cannam@95: cannam@95: M ./acinclude.m4 -1 +1 cannam@95: cannam@95: Thu Feb 5 19:22:34 EST 2004 stevenj cannam@95: * [project @ 2004-02-06 00:22:34 by stevenj] cannam@95: just use +Ofaster on hpux (+O3 +Onolimit +Olibcalls +Ofltacc=relaxed -Wl,+mergeseg) cannam@95: cannam@95: M ./acinclude.m4 -1 +1 cannam@95: cannam@95: Fri Jan 30 14:17:15 EST 2004 stevenj cannam@95: * [project @ 2004-01-30 19:17:15 by stevenj] cannam@95: check for win32 threads for mingw32; thanks to Alessio Massaro cannam@95: cannam@95: M ./configure.ac +8 cannam@95: cannam@95: Thu Jan 29 15:23:33 EST 2004 stevenj cannam@95: * [project @ 2004-01-29 20:23:33 by stevenj] cannam@95: added missing 'static', thanks to Alessio Massaro cannam@95: cannam@95: M ./threads/threads.c -3 +3 cannam@95: cannam@95: Fri Jan 9 16:36:48 EST 2004 stevenj cannam@95: * [project @ 2004-01-09 21:36:48 by stevenj] cannam@95: print more like bluestein cannam@95: cannam@95: M ./rdft/dht-rader.c -1 +1 cannam@95: cannam@95: Fri Jan 9 15:45:22 EST 2004 stevenj cannam@95: * [project @ 2004-01-09 20:45:22 by stevenj] cannam@95: fixed op count for R2HC_ONLY_CONV cannam@95: cannam@95: M ./rdft/dht-rader.c -3 +2 cannam@95: cannam@95: Fri Jan 9 15:41:50 EST 2004 stevenj cannam@95: * [project @ 2004-01-09 20:41:50 by stevenj] cannam@95: include DESTROY_INPUT in buffered flags for in-place...otherwise in-place hc2r uses rdft-dhtcvs diff cannam@95: cannam@95: M ./dft/buffered.c -1 +10 cannam@95: M ./rdft/buffered.c -1 +10 cannam@95: M ./rdft/buffered2.c -1 +9 cannam@95: cannam@95: Fri Jan 9 15:41:09 EST 2004 stevenj cannam@95: * [project @ 2004-01-09 20:41:09 by stevenj] cannam@95: resurrected R2HC_ONLY_CONV option to share plans and save on planning time cannam@95: cannam@95: M ./rdft/dht-rader.c -3 +46 cannam@95: cannam@95: Fri Jan 9 14:47:00 EST 2004 stevenj cannam@95: * [project @ 2004-01-09 19:47:00 by stevenj] cannam@95: precompute folding for cyclic convolution cannam@95: cannam@95: M ./rdft/dht-rader.c -13 +9 cannam@95: cannam@95: Wed Jan 7 16:48:39 EST 2004 stevenj cannam@95: * [project @ 2004-01-07 21:48:39 by stevenj] cannam@95: minor cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -1 +1 cannam@95: cannam@95: Wed Jan 7 16:48:25 EST 2004 stevenj cannam@95: * [project @ 2004-01-07 21:48:25 by stevenj] cannam@95: note reports of successful compilation on Windows cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -5 +5 cannam@95: cannam@95: Wed Jan 7 14:16:16 EST 2004 stevenj cannam@95: * [project @ 2004-01-07 19:16:16 by stevenj] cannam@95: citation year cannam@95: cannam@95: M ./reodft/reodft010e-r2hc.c -2 +2 cannam@95: cannam@95: Tue Jan 6 01:07:36 EST 2004 stevenj cannam@95: * [project @ 2004-01-06 06:07:36 by stevenj] cannam@95: comment cannam@95: cannam@95: M ./rdft/dht-rader.c -1 +1 cannam@95: cannam@95: Tue Jan 6 01:07:08 EST 2004 stevenj cannam@95: * [project @ 2004-01-06 06:07:08 by stevenj] cannam@95: comment fix cannam@95: cannam@95: M ./rdft/dht-rader.c -1 +1 cannam@95: cannam@95: Tue Jan 6 01:06:57 EST 2004 stevenj cannam@95: * [project @ 2004-01-06 06:06:57 by stevenj] cannam@95: fixed naming cruft cannam@95: cannam@95: M ./rdft/dht-rader.c -18 +18 cannam@95: cannam@95: Tue Jan 6 00:56:16 EST 2004 stevenj cannam@95: * [project @ 2004-01-06 05:56:16 by stevenj] cannam@95: space cannam@95: cannam@95: M ./rdft/dht-rader.c -1 cannam@95: cannam@95: Tue Jan 6 00:55:53 EST 2004 stevenj cannam@95: * [project @ 2004-01-06 05:55:45 by stevenj] cannam@95: comment cannam@95: cannam@95: M ./rdft/dht-rader.c -1 +1 cannam@95: cannam@95: Tue Jan 6 00:54:07 EST 2004 stevenj cannam@95: * [project @ 2004-01-06 05:54:07 by stevenj] cannam@95: moved assert cannam@95: cannam@95: M ./rdft/dht-rader.c -1 +1 cannam@95: cannam@95: Tue Jan 6 00:49:16 EST 2004 stevenj cannam@95: * [project @ 2004-01-06 05:49:16 by stevenj] cannam@95: comment cannam@95: cannam@95: M ./rdft/dht-rader.c +5 cannam@95: cannam@95: Tue Jan 6 00:41:06 EST 2004 stevenj cannam@95: * [project @ 2004-01-06 05:41:06 by stevenj] cannam@95: delete old R2HC_ONLY_CONV hack, now defunct cannam@95: cannam@95: M ./rdft/dht-rader.c -46 cannam@95: cannam@95: Tue Jan 6 00:32:58 EST 2004 stevenj cannam@95: * [project @ 2004-01-06 05:32:58 by stevenj] cannam@95: added padded real rader cannam@95: cannam@95: M ./rdft/dht-rader.c -56 +95 cannam@95: cannam@95: Mon Jan 5 22:56:58 EST 2004 stevenj cannam@95: * [project @ 2004-01-06 03:56:58 by stevenj] cannam@95: removed unused var cannam@95: cannam@95: M ./rdft/generic.c -1 +1 cannam@95: cannam@95: Mon Jan 5 21:20:29 EST 2004 stevenj cannam@95: * [project @ 2004-01-06 02:20:29 by stevenj] cannam@95: handle both FFT_SIGN values cannam@95: cannam@95: M ./rdft/generic.c -2 +14 cannam@95: cannam@95: Fri Jan 2 06:07:51 EST 2004 athena cannam@95: * [project @ 2004-01-02 11:07:51 by athena] cannam@95: Oops: d->ros ==> d->ios cannam@95: cannam@95: M ./rdft/codelets/r2hc.c -1 +1 cannam@95: cannam@95: Fri Jan 2 06:05:10 EST 2004 athena cannam@95: * [project @ 2004-01-02 11:05:10 by athena] cannam@95: Oops: d->ris should have been d->iis cannam@95: cannam@95: M ./rdft/codelets/hc2r.c -1 +1 cannam@95: cannam@95: Thu Jan 1 16:00:07 EST 2004 athena cannam@95: * [project @ 2004-01-01 21:00:07 by athena] cannam@95: Removed rdft rader cooley-tukey, to be superseded by a generic cannam@95: reduction of rdft twiddle problems to dft + pre/post processing cannam@95: cannam@95: M ./dft/Makefile.am -2 +2 cannam@95: M ./dft/dft.h -5 +1 cannam@95: R ./dft/rader-omega.c cannam@95: M ./dft/rader.c -2 +38 cannam@95: M ./rdft/Makefile.am -2 +2 cannam@95: M ./rdft/conf.c -2 +1 cannam@95: R ./rdft/rader-hc2hc.c cannam@95: cannam@95: Thu Jan 1 15:44:09 EST 2004 athena cannam@95: * [project @ 2004-01-01 20:44:09 by athena] cannam@95: In anticipation of the upcoming revision of rdft, removed rdft generic cannam@95: dit/dif cooley-tukey, in favor of generic rh2c and hc2r solvers. cannam@95: Cleaned up stuff that became unused after this change, such as cannam@95: TW_GENERIC. cannam@95: cannam@95: M ./dft/ct.c -33 +3 cannam@95: M ./dft/generic.c -1 +2 cannam@95: M ./kernel/ifftw.h -2 +3 cannam@95: M ./kernel/primes.c -1 +32 cannam@95: M ./kernel/twiddle.c -18 +1 cannam@95: M ./rdft/generic.c -258 +103 cannam@95: cannam@95: Thu Jan 1 12:59:30 EST 2004 athena cannam@95: * [project @ 2004-01-01 17:59:30 by athena] cannam@95: Removed useless file cannam@95: cannam@95: M ./kernel/Makefile.am -3 +3 cannam@95: M ./kernel/ifftw.h -2 +1 cannam@95: R ./kernel/square.c cannam@95: cannam@95: Fri Dec 26 13:54:00 EST 2003 stevenj cannam@95: * [project @ 2003-12-26 18:54:00 by stevenj] cannam@95: whoops, don't call AC_F77_DUMMY_MAIN if no Fortran compiler is found; thanks to Charles Radley for the bug report. cannam@95: cannam@95: M ./configure.ac -2 +3 cannam@95: cannam@95: Fri Dec 19 13:58:05 EST 2003 stevenj cannam@95: * [project @ 2003-12-19 18:58:05 by stevenj] cannam@95: guess good flags for Solaris/intel, suggested by J. Gregory Wright cannam@95: cannam@95: M ./acinclude.m4 -1 +1 cannam@95: cannam@95: Fri Dec 5 19:55:13 EST 2003 stevenj cannam@95: * [project @ 2003-12-06 00:55:13 by stevenj] cannam@95: blah cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -1 +1 cannam@95: M ./doc/FAQ/html.refs -1 +1 cannam@95: cannam@95: Sun Nov 30 06:59:41 EST 2003 athena cannam@95: * [project @ 2003-11-30 11:59:41 by athena] cannam@95: DIF generic solver was destroying the input. cannam@95: cannam@95: M ./rdft/generic.c -2 +16 cannam@95: cannam@95: Sat Nov 29 19:28:39 EST 2003 athena cannam@95: * [project @ 2003-11-30 00:28:39 by athena] cannam@95: Fixed bug that caused HC2R transforms to destroy the input in cannam@95: certain cases, even if the user specified FFTW_PRESERVE_INPUT. cannam@95: cannam@95: M ./NEWS +5 cannam@95: M ./rdft/rader-hc2hc.c -2 +14 cannam@95: cannam@95: Sat Nov 29 16:49:01 EST 2003 athena cannam@95: * [project @ 2003-11-29 21:49:01 by athena] cannam@95: Implemented swap_io hack for r2r verifier. cannam@95: cannam@95: M ./libbench2/verify-r2r.c -1 +8 cannam@95: cannam@95: Thu Nov 20 22:00:53 EST 2003 stevenj cannam@95: * [project @ 2003-11-21 03:00:53 by stevenj] cannam@95: citation cannam@95: cannam@95: M ./reodft/reodft010e-r2hc.c -2 +6 cannam@95: cannam@95: Fri Nov 14 20:57:55 EST 2003 athena cannam@95: * [project @ 2003-11-15 01:57:55 by athena] cannam@95: Trying to get ``make paranoid-check'' to work. (Still broken.) cannam@95: cannam@95: M ./kernel/ifftw.h -2 +3 cannam@95: M ./kernel/planner.c -2 +2 cannam@95: M ./tests/hook.c -9 +8 cannam@95: cannam@95: Fri Nov 14 20:05:54 EST 2003 stevenj cannam@95: * [project @ 2003-11-15 01:05:54 by stevenj] cannam@95: fixes for input-preservation tests cannam@95: cannam@95: M ./libbench2/bench-user.h -1 +2 cannam@95: M ./libbench2/tensor.c -1 +15 cannam@95: M ./libbench2/verify-dft.c -3 +12 cannam@95: M ./libbench2/verify-lib.c -1 +4 cannam@95: M ./libbench2/verify-r2r.c -13 +33 cannam@95: M ./libbench2/verify-rdft2.c -6 +22 cannam@95: M ./libbench2/verify.h -1 +1 cannam@95: cannam@95: Fri Nov 14 19:19:31 EST 2003 athena cannam@95: * [project @ 2003-11-15 00:19:31 by athena] cannam@95: Assume FFTW_PRESERVE_INPUT unless either the `d' flag is given in the cannam@95: problem, or the problem is multidimensional c2r (which fftw3 cannot cannam@95: without destroying the input). With this change, we can at least test cannam@95: that FFTW_PRESERVE_INPUT works in the c2r 1d case. cannam@95: cannam@95: M ./tests/bench.c -12 +21 cannam@95: cannam@95: Fri Nov 14 19:14:40 EST 2003 stevenj cannam@95: * [project @ 2003-11-15 00:14:40 by stevenj] cannam@95: apply should copy back input for input-preservation check cannam@95: cannam@95: M ./libbench2/verify-dft.c -1 +3 cannam@95: M ./libbench2/verify-r2r.c +13 cannam@95: M ./libbench2/verify-rdft2.c -1 +4 cannam@95: cannam@95: Fri Nov 14 19:01:36 EST 2003 athena cannam@95: * [project @ 2003-11-15 00:01:14 by athena] cannam@95: Undone previous bogus changes cannam@95: cannam@95: M ./rdft/rank-geq2-rdft2.c -1 +10 cannam@95: M ./tests/bench.c -14 +10 cannam@95: M ./tests/check.pl -1 cannam@95: cannam@95: Fri Nov 14 18:27:12 EST 2003 athena cannam@95: * [project @ 2003-11-14 23:27:12 by athena] cannam@95: Check dr[fb] in addition to r[fb] cannam@95: cannam@95: M ./tests/check.pl +1 cannam@95: cannam@95: Fri Nov 14 17:33:44 EST 2003 athena cannam@95: * [project @ 2003-11-14 22:33:42 by athena] cannam@95: Fixed conditions under which the rank-geq2-rdft2 solver is applicable. cannam@95: cannam@95: The old solver was not applicable for out-of-place problems cannam@95: unless DESTROY_INPUT. This is bogus. As long as the subsolvers cannam@95: honor !DESTROY_INPUT, the solver is always applicable. cannam@95: cannam@95: Changed semantics of test program, so that PRESERVE_INPUT is always cannam@95: true unless the problem specifies destroy_input explicitly. Without cannam@95: this change, there is no way to test the new solver. cannam@95: cannam@95: M ./rdft/rank-geq2-rdft2.c -10 +1 cannam@95: M ./tests/bench.c -10 +14 cannam@95: cannam@95: Thu Oct 30 15:10:42 EST 2003 stevenj cannam@95: * [project @ 2003-10-30 20:10:42 by stevenj] cannam@95: added AIX OpenMP (-qsmp=omp) support; thanks to Greg Bauer cannam@95: cannam@95: M ./configure.ac +5 cannam@95: cannam@95: Thu Oct 30 10:11:39 EST 2003 athena cannam@95: * [project @ 2003-10-30 15:11:39 by athena] cannam@95: G5 CFLAGS cannam@95: cannam@95: M ./acinclude.m4 +6 cannam@95: cannam@95: Fri Oct 24 04:17:39 EDT 2003 stevenj cannam@95: * [project @ 2003-10-24 08:17:39 by stevenj] cannam@95: western FAQ cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn +5 cannam@95: cannam@95: Thu Oct 23 11:34:11 EDT 2003 athena cannam@95: * [project @ 2003-10-23 15:34:11 by athena] cannam@95: Oops. cannam@95: cannam@95: M ./simd/altivec.c -1 +2 cannam@95: cannam@95: Thu Oct 23 11:28:28 EDT 2003 athena cannam@95: * [project @ 2003-10-23 15:28:27 by athena] cannam@95: Autodetect altivec cannam@95: cannam@95: M ./configure.ac -2 +3 cannam@95: M ./simd/altivec.c -1 +28 cannam@95: M ./simd/simd-altivec.h -1 +2 cannam@95: cannam@95: Wed Oct 22 01:14:10 EDT 2003 stevenj cannam@95: * [project @ 2003-10-22 05:14:10 by stevenj] cannam@95: MinGW gets confused by a single / cannam@95: cannam@95: M ./tests/check.pl -4 +4 cannam@95: cannam@95: Fri Oct 17 10:46:41 EDT 2003 athena cannam@95: * [project @ 2003-10-17 14:46:41 by athena] cannam@95: Paranoid portability fix cannam@95: cannam@95: M ./libbench2/mp.c -2 +2 cannam@95: cannam@95: Thu Oct 16 11:07:46 EDT 2003 athena cannam@95: * [project @ 2003-10-16 15:07:46 by athena] cannam@95: size -> length, which should make clear that we are not talking cannam@95: about arbitrary precision. cannam@95: cannam@95: M ./doc/fftw3.texi -4 +4 cannam@95: cannam@95: Wed Oct 15 15:01:40 EDT 2003 stevenj cannam@95: * [project @ 2003-10-15 19:01:40 by stevenj] cannam@95: pruned transforms are a FAQ cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn +11 cannam@95: cannam@95: Wed Oct 8 23:54:17 EDT 2003 stevenj cannam@95: * [project @ 2003-10-09 03:54:17 by stevenj] cannam@95: NO_SEARCH has already been mapped to FFTW_WISDOM_ONLY cannam@95: cannam@95: M ./TODO -1 +1 cannam@95: cannam@95: Wed Oct 8 23:53:19 EDT 2003 stevenj cannam@95: * [project @ 2003-10-09 03:53:19 by stevenj] cannam@95: newline cannam@95: cannam@95: M ./TODO +1 cannam@95: cannam@95: Sat Sep 27 20:27:32 EDT 2003 stevenj cannam@95: * [project @ 2003-09-28 00:27:32 by stevenj] cannam@95: fix cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Sat Sep 27 20:24:39 EDT 2003 stevenj cannam@95: * [project @ 2003-09-28 00:24:39 by stevenj] cannam@95: clarification cannam@95: cannam@95: M ./doc/fftw3.texi -2 +4 cannam@95: cannam@95: Sat Sep 27 17:43:57 EDT 2003 stevenj cannam@95: * [project @ 2003-09-27 21:43:57 by stevenj] cannam@95: minor fix cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Sat Sep 27 17:42:30 EDT 2003 stevenj cannam@95: * [project @ 2003-09-27 21:42:30 by stevenj] cannam@95: grammar cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Sat Sep 27 17:29:04 EDT 2003 stevenj cannam@95: * [project @ 2003-09-27 21:29:04 by stevenj] cannam@95: html output fix cannam@95: cannam@95: M ./doc/fftw3.texi -5 +5 cannam@95: cannam@95: Sat Sep 27 17:22:48 EDT 2003 stevenj cannam@95: * [project @ 2003-09-27 21:22:48 by stevenj] cannam@95: mentioned sqrt(2) factors for DCT/DST cannam@95: cannam@95: M ./doc/fftw3.texi -5 +48 cannam@95: cannam@95: Sat Sep 27 17:07:18 EDT 2003 stevenj cannam@95: * [project @ 2003-09-27 21:07:18 by stevenj] cannam@95: FFTW_WISDOM_ONLY flag (undocumented for now), suggested by Phil Dumont cannam@95: cannam@95: M ./api/fftw3.h -1 +2 cannam@95: M ./api/mapflags.c +1 cannam@95: cannam@95: Tue Sep 23 23:36:19 EDT 2003 stevenj cannam@95: * [project @ 2003-09-24 03:36:19 by stevenj] cannam@95: removed UpTime code cannam@95: cannam@95: M ./kernel/cycle.h -20 +3 cannam@95: cannam@95: Tue Sep 23 23:27:29 EDT 2003 stevenj cannam@95: * [project @ 2003-09-24 03:27:29 by stevenj] cannam@95: updated documentation for mach_absolute_time cannam@95: cannam@95: M ./kernel/cycle.h -3 +3 cannam@95: cannam@95: Tue Sep 23 23:25:52 EDT 2003 stevenj cannam@95: * [project @ 2003-09-24 03:25:52 by stevenj] cannam@95: use mach_absolute_time on MacOS/Darwin, as a fallback; don't bother checking for UpTime since it requires extra libs cannam@95: cannam@95: M ./configure.ac -3 +2 cannam@95: M ./kernel/cycle.h -1 +11 cannam@95: cannam@95: Tue Sep 23 22:59:29 EDT 2003 stevenj cannam@95: * [project @ 2003-09-24 02:59:29 by stevenj] cannam@95: support Apple UpTime function for asm-less xlc, grrr... cannam@95: cannam@95: M ./configure.ac -2 +3 cannam@95: M ./kernel/cycle.h -1 +19 cannam@95: cannam@95: Tue Sep 23 15:42:29 EDT 2003 stevenj cannam@95: * [project @ 2003-09-23 19:42:29 by stevenj] cannam@95: additional paranoia for xlc etc. cannam@95: cannam@95: M ./api/api.h -1 +1 cannam@95: M ./api/fftw3.h -2 +2 cannam@95: cannam@95: Mon Sep 22 15:28:56 EDT 2003 stevenj cannam@95: * [project @ 2003-09-22 19:28:56 by stevenj] cannam@95: work around _Complex_I weirdness in xlc, reported by Greg Allen cannam@95: cannam@95: M ./api/api.h -2 +4 cannam@95: M ./api/fftw3.h -2 +2 cannam@95: cannam@95: Fri Sep 5 18:03:11 EDT 2003 stevenj cannam@95: * [project @ 2003-09-05 22:03:11 by stevenj] cannam@95: typo cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -2 +2 cannam@95: cannam@95: Fri Sep 5 13:11:40 EDT 2003 athena cannam@95: * [project @ 2003-09-05 17:11:40 by athena] cannam@95: New script that produces commercial version. cannam@95: cannam@95: A ./commercialize.sh cannam@95: cannam@95: Fri Sep 5 07:27:06 EDT 2003 athena cannam@95: * [project @ 2003-09-05 11:27:06 by athena] cannam@95: Noted that VC++ is buggy. Noted that we know nothing about Windows. cannam@95: Noted that the sky is blue as well. cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -10 +22 cannam@95: cannam@95: Tue Sep 2 09:04:19 EDT 2003 athena cannam@95: * [project @ 2003-09-02 13:04:19 by athena] cannam@95: Noted that certain arrays are no longer used after the planner has cannam@95: completed. cannam@95: cannam@95: M ./doc/fftw3.texi -2 +12 cannam@95: cannam@95: Tue Aug 26 08:22:38 EDT 2003 athena cannam@95: * [project @ 2003-08-26 12:22:38 by athena] cannam@95: Typo cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Mon Aug 25 21:27:43 EDT 2003 athena cannam@95: * [project @ 2003-08-26 01:27:43 by athena] cannam@95: New item cannam@95: cannam@95: M ./TODO -2 +1 cannam@95: cannam@95: Thu Aug 21 17:36:08 EDT 2003 stevenj cannam@95: * [project @ 2003-08-21 21:36:08 by stevenj] cannam@95: try creating output file before planning (thanks to Phil Dumont for the suggestion) cannam@95: cannam@95: M ./tools/fftw-wisdom.c -10 +10 cannam@95: cannam@95: Tue Aug 19 10:08:07 EDT 2003 athena cannam@95: * [project @ 2003-08-19 14:08:07 by athena] cannam@95: Clarified fftw_cleanup() cannam@95: cannam@95: M ./doc/fftw3.texi -4 +7 cannam@95: cannam@95: Sat Aug 16 03:13:41 EDT 2003 stevenj cannam@95: * [project @ 2003-08-16 07:13:41 by stevenj] cannam@95: typo cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Mon Jul 28 18:01:13 EDT 2003 stevenj cannam@95: * [project @ 2003-07-28 22:01:13 by stevenj] cannam@95: use time() instead of clock() (FIXME: what to do for non-POSIX systems?) ...thanks to JP Sugarbroad and James A. Treacy for the bug report cannam@95: cannam@95: M ./tools/fftw-wisdom.c -4 +4 cannam@95: cannam@95: Thu Jul 24 18:58:10 EDT 2003 athena cannam@95: * [project @ 2003-07-24 22:58:10 by athena] cannam@95: Need __volatile__ in sparc cycle counter. This is why the debian cannam@95: port hangs. cannam@95: cannam@95: M ./kernel/cycle.h -2 +2 cannam@95: cannam@95: Sun Jul 20 16:02:43 EDT 2003 stevenj cannam@95: * [project @ 2003-07-20 20:02:43 by stevenj] cannam@95: merged 3.0.1 notes cannam@95: cannam@95: M ./NEWS +30 cannam@95: cannam@95: Sun Jul 13 20:57:34 EDT 2003 stevenj cannam@95: * [project @ 2003-07-14 00:57:34 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./libbench2/bench-main.c -2 +2 cannam@95: cannam@95: Thu Jul 10 11:48:50 EDT 2003 athena cannam@95: * [project @ 2003-07-10 15:48:50 by athena] cannam@95: Dealing with constants in a way that seems to confuse gcc less. cannam@95: cannam@95: M ./simd/simd-sse.h -5 +2 cannam@95: M ./simd/simd-sse2.h -6 +3 cannam@95: cannam@95: Wed Jul 9 17:39:23 EDT 2003 athena cannam@95: * [project @ 2003-07-09 21:39:16 by athena] cannam@95: Enabled scheduler hack for FMA, where it seems to help. cannam@95: cannam@95: M ./genfft/annotate.ml -7 +3 cannam@95: M ./genfft/magic.ml -1 +5 cannam@95: M ./support/Makefile.codelets -1 +1 cannam@95: cannam@95: Wed Jul 9 12:57:38 EDT 2003 athena cannam@95: * [project @ 2003-07-09 16:57:38 by athena] cannam@95: Hmm---the new scheduler seems make things worse for gcc/x86, better cannam@95: for gcc/ppc, and about the same for icc/x86. Disabled for now. cannam@95: cannam@95: M ./genfft/annotate.ml -3 +6 cannam@95: cannam@95: Wed Jul 9 08:09:53 EDT 2003 athena cannam@95: * [project @ 2003-07-09 12:09:53 by athena] cannam@95: New scheduling pass that keeps ``x = a + b'' and ``y = a - b'' close cannam@95: together. This property was no longer automatic for the dags cannam@95: generated in SIMD mode. cannam@95: cannam@95: I cannot measure any speed difference due to this change. However, cannam@95: the change is justified by a minimal-screwup argument. Moreover, the cannam@95: sse2 fftw library is now 1% smaller than it was before. cannam@95: cannam@95: M ./genfft/annotate.ml -32 +70 cannam@95: cannam@95: Tue Jul 8 20:42:22 EDT 2003 athena cannam@95: * [project @ 2003-07-09 00:42:22 by athena] cannam@95: -(FNMS()) => FMS() cannam@95: cannam@95: M ./genfft/c.ml -1 +2 cannam@95: cannam@95: Sun Jul 6 13:53:23 EDT 2003 stevenj cannam@95: * [project @ 2003-07-06 17:53:23 by stevenj] cannam@95: added more convenient target name cannam@95: cannam@95: M ./doc/FAQ/Makefile.am +2 cannam@95: cannam@95: Sat Jul 5 13:30:10 EDT 2003 stevenj cannam@95: * [project @ 2003-07-05 17:30:10 by stevenj] cannam@95: typo cannam@95: cannam@95: M ./kernel/ifftw.h -2 +2 cannam@95: cannam@95: Sat Jul 5 13:19:36 EDT 2003 athena cannam@95: * [project @ 2003-07-05 17:19:36 by athena] cannam@95: Consistent naming cannam@95: cannam@95: M ./dft/ct-generic.c -2 +2 cannam@95: cannam@95: Sat Jul 5 13:05:51 EDT 2003 athena cannam@95: * [project @ 2003-07-05 17:05:51 by athena] cannam@95: Got rid of problemw. cannam@95: cannam@95: A ./dft/ct-directw.c cannam@95: A ./dft/ct-directwbuf.c cannam@95: A ./dft/ct-generic.c cannam@95: A ./dft/ct.h cannam@95: M ./dft/Makefile.am -5 +4 cannam@95: M ./dft/conf.c -3 +2 cannam@95: M ./dft/ct-directw.c +153 cannam@95: M ./dft/ct-directwbuf.c +213 cannam@95: M ./dft/ct-generic.c +304 cannam@95: M ./dft/ct.c -62 +24 cannam@95: M ./dft/ct.h +56 cannam@95: M ./dft/ctsq.c -3 +3 cannam@95: M ./dft/dft.h -39 +2 cannam@95: R ./dft/dftw-dft.c cannam@95: M ./dft/direct.c -2 +1 cannam@95: R ./dft/directw.c cannam@95: R ./dft/directwbuf.c cannam@95: M ./dft/generic.c -2 cannam@95: M ./dft/kdft-dif.c -4 +4 cannam@95: M ./dft/kdft-difsq.c -3 +3 cannam@95: M ./dft/kdft-dit.c -4 +4 cannam@95: M ./dft/plan.c -11 +1 cannam@95: R ./dft/problemw.c cannam@95: M ./dft/rader.c -2 cannam@95: M ./dft/solve.c -8 +1 cannam@95: cannam@95: Fri Jul 4 06:56:26 EDT 2003 athena cannam@95: * [project @ 2003-07-04 10:56:26 by athena] cannam@95: Increase TIME_MIN on intel only cannam@95: cannam@95: M ./kernel/cycle.h -1 +3 cannam@95: M ./kernel/timer.c -2 +2 cannam@95: cannam@95: Fri Jul 4 06:36:02 EDT 2003 athena cannam@95: * [project @ 2003-07-04 10:36:02 by athena] cannam@95: A little hack to get more consistent scheduling. cannam@95: cannam@95: M ./genfft/schedule.ml -7 +8 cannam@95: cannam@95: Thu Jul 3 16:47:42 EDT 2003 athena cannam@95: * [project @ 2003-07-03 20:47:42 by athena] cannam@95: New experimental scheduler (currently disabled). cannam@95: cannam@95: The old scheduler is ``optimal'' in the sense that it minimizes cannam@95: register pressure. The only way to reduce register pressure is to cannam@95: schedule dependent instructions as closely as possible, so as to cannam@95: minimize the life time of registers. This strategy maximizes the cannam@95: number of pipeline stalls, however. With enough registers and short cannam@95: enough pipelines, this tradeoff is fine. This is no longer the case cannam@95: for the devilish pipeline of the Pentium IV or (probably) the PowerPC cannam@95: 970. cannam@95: cannam@95: The new scheduler switches to a ``list scheduler'' for dags smaller cannam@95: than a specified size. The list scheduler executes a butterfly left cannam@95: to right one column at the time. This amounts to the best possible cannam@95: pipeline utilization, and the worst possible register pressure. cannam@95: cannam@95: The ``specified size'' defaults to 0, i.e., no change from fftw2 and cannam@95: fftw-3.0. It seems like a value of 7--10 produces the best results cannam@95: for Pentium IV (probably screwing the G3/G4 powerpcs and sparc, but I cannam@95: haven't tried.) As time goes by, we may want to increase this number cannam@95: to favor newer processors over older processors. cannam@95: cannam@95: M ./genfft/magic.ml -1 +3 cannam@95: M ./genfft/schedule.ml -9 +35 cannam@95: cannam@95: Wed Jun 25 17:43:59 EDT 2003 stevenj cannam@95: * [project @ 2003-06-25 21:43:59 by stevenj] cannam@95: remove non-portable use of tempfile; thanks to Nicolas Decoster for the patch cannam@95: cannam@95: M ./tools/fftw-wisdom-to-conf.in -1 +1 cannam@95: cannam@95: Wed Jun 25 17:14:03 EDT 2003 stevenj cannam@95: * [project @ 2003-06-25 21:14:03 by stevenj] cannam@95: increase stupid HP preprocessor limits cannam@95: cannam@95: M ./acinclude.m4 -1 +1 cannam@95: cannam@95: Thu Jun 19 15:21:52 EDT 2003 athena cannam@95: * [project @ 2003-06-19 19:21:52 by athena] cannam@95: Distribute gen_mdct.ml cannam@95: cannam@95: M ./genfft/Makefile.am -8 +8 cannam@95: cannam@95: Wed Jun 11 06:55:21 EDT 2003 athena cannam@95: * [project @ 2003-06-11 10:55:21 by athena] cannam@95: Cleared int/ptrdiff_t confusions cannam@95: cannam@95: M ./rdft/buffered2.c -2 +3 cannam@95: M ./rdft/rdft2-radix2.c -5 +5 cannam@95: cannam@95: Tue Jun 10 22:15:42 EDT 2003 athena cannam@95: * [project @ 2003-06-11 02:15:41 by athena] cannam@95: Cleared int/ptrdiff_t confusion cannam@95: cannam@95: M ./dft/dftw-dft.c -1 +1 cannam@95: M ./dft/directwbuf.c -2 +2 cannam@95: M ./dft/rank0.c -2 +2 cannam@95: M ./dft/vrank2-transpose.c -2 +2 cannam@95: M ./dft/vrank3-transpose.c -2 +2 cannam@95: M ./kernel/planner.c -2 +2 cannam@95: cannam@95: Sun Jun 8 09:52:57 EDT 2003 athena cannam@95: * [project @ 2003-06-08 13:52:57 by athena] cannam@95: Increased TIME_MIN. This seems to produce more reliable plans cannam@95: on Pentium IV. cannam@95: cannam@95: M ./kernel/timer.c -2 +2 cannam@95: cannam@95: Sat Jun 7 21:43:00 EDT 2003 athena cannam@95: * [project @ 2003-06-08 01:43:00 by athena] cannam@95: Removed relic -trivial-stores, which dates back to Franz's early cannam@95: experiments. Speed improved on SSE2, both with gcc and icc. cannam@95: cannam@95: M ./dft/simd/codelets/Makefile.am -1 +1 cannam@95: cannam@95: Thu Jun 5 22:29:52 EDT 2003 stevenj cannam@95: * [project @ 2003-06-06 02:29:52 by stevenj] cannam@95: fix direntry cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Thu Jun 5 13:41:34 EDT 2003 stevenj cannam@95: * [project @ 2003-06-05 17:41:34 by stevenj] cannam@95: added imdct cannam@95: cannam@95: M ./genfft/gen_mdct.ml -13 +66 cannam@95: cannam@95: Wed Jun 4 19:54:38 EDT 2003 athena cannam@95: * [project @ 2003-06-04 23:54:38 by athena] cannam@95: Collect pattern (a * b) +- (c * d) in generic-arith, because this cannam@95: operation can usually be computed with one rounding in fixed-point cannam@95: (and it possibly exposes a FMA instruction) cannam@95: cannam@95: M ./genfft/c.ml -5 +15 cannam@95: cannam@95: Wed Jun 4 15:11:29 EDT 2003 athena cannam@95: * [project @ 2003-06-04 19:11:29 by athena] cannam@95: Generic-arithmetic unparser cannam@95: cannam@95: M ./genfft/c.ml -8 +31 cannam@95: M ./genfft/magic.ml -1 +5 cannam@95: cannam@95: Sun Jun 1 09:05:30 EDT 2003 athena cannam@95: * [project @ 2003-06-01 13:05:30 by athena] cannam@95: Oops---randomized CSE was using the same random numbers cannam@95: over and over cannam@95: cannam@95: M ./genfft/oracle.ml -8 +6 cannam@95: M ./genfft-k7/oracle.ml -11 +11 cannam@95: cannam@95: Sun Jun 1 07:01:17 EDT 2003 athena cannam@95: * [project @ 2003-06-01 11:01:17 by athena] cannam@95: Paranoia. cannam@95: cannam@95: M ./genfft/c.ml -3 +3 cannam@95: cannam@95: Sun Jun 1 07:00:54 EDT 2003 athena cannam@95: * [project @ 2003-06-01 11:00:54 by athena] cannam@95: Use relative error instead of absolute error, to avoid problems cannam@95: when normalization factors are used. cannam@95: cannam@95: M ./genfft/oracle.ml -3 +5 cannam@95: cannam@95: Sat May 31 22:11:28 EDT 2003 stevenj cannam@95: * [project @ 2003-06-01 02:11:28 by stevenj] cannam@95: slight opt cannam@95: cannam@95: M ./reodft/reodft11e-radix2.c -3 +3 cannam@95: cannam@95: Sat May 31 22:10:45 EDT 2003 stevenj cannam@95: * [project @ 2003-06-01 02:10:45 by stevenj] cannam@95: slight optimization cannam@95: cannam@95: M ./reodft/reodft11e-radix2.c -3 +3 cannam@95: cannam@95: Sat May 31 20:43:31 EDT 2003 stevenj cannam@95: * [project @ 2003-06-01 00:43:31 by stevenj] cannam@95: *W is const cannam@95: cannam@95: M ./genfft/gen_mdct.ml -3 +3 cannam@95: cannam@95: Sat May 31 20:41:15 EDT 2003 stevenj cannam@95: * [project @ 2003-06-01 00:41:15 by stevenj] cannam@95: comment cannam@95: cannam@95: M ./genfft/gen_mdct.ml -2 +4 cannam@95: cannam@95: Thu May 29 21:31:31 EDT 2003 stevenj cannam@95: * [project @ 2003-05-30 01:31:31 by stevenj] cannam@95: added experimental MDCT cannam@95: cannam@95: A ./genfft/gen_mdct.ml cannam@95: M ./genfft/Makefile.am -1 +7 cannam@95: M ./genfft/gen_mdct.ml +208 cannam@95: cannam@95: Wed May 28 22:01:37 EDT 2003 stevenj cannam@95: * [project @ 2003-05-29 02:01:37 by stevenj] cannam@95: altivec (fma) needs simd codlist.c too cannam@95: cannam@95: M ./mkdist.sh -1 +1 cannam@95: cannam@95: Wed May 28 22:00:49 EDT 2003 stevenj cannam@95: * [project @ 2003-05-29 02:00:49 by stevenj] cannam@95: make sure we include SIMD codlist.c for non-Unix folks cannam@95: cannam@95: M ./mkdist.sh -1 +1 cannam@95: cannam@95: Tue May 27 20:31:25 EDT 2003 stevenj cannam@95: * [project @ 2003-05-28 00:31:25 by stevenj] cannam@95: noted howmany_rank == 0 is a single transform cannam@95: cannam@95: M ./doc/fftw3.texi -1 +3 cannam@95: cannam@95: Tue May 27 20:02:31 EDT 2003 stevenj cannam@95: * [project @ 2003-05-28 00:02:31 by stevenj] cannam@95: further stride clarification cannam@95: cannam@95: M ./doc/fftw3.texi -4 +4 cannam@95: cannam@95: Mon May 26 10:21:22 EDT 2003 athena cannam@95: * [project @ 2003-05-26 14:21:22 by athena] cannam@95: Removed transposed dftw problems. cannam@95: cannam@95: I now consider transposed dftw a Bad Idea, since it does not cannam@95: apply to the case that it was originally meant for (speed up four-step) cannam@95: and it complicates the implementation of the other thing I want to try cannam@95: (dftw m-slices). cannam@95: cannam@95: M ./dft/Makefile.am -3 +3 cannam@95: M ./dft/conf.c -2 +1 cannam@95: M ./dft/ct.c -5 +3 cannam@95: M ./dft/ctsq.c -74 +57 cannam@95: M ./dft/dft.h -9 +5 cannam@95: M ./dft/dftw-dft.c -83 cannam@95: M ./dft/directw.c -5 +1 cannam@95: M ./dft/directwbuf.c -5 +1 cannam@95: R ./dft/directwsq.c cannam@95: M ./dft/kdft-difsq.c -2 +2 cannam@95: M ./dft/problemw.c -13 +9 cannam@95: cannam@95: Mon May 26 07:22:59 EDT 2003 athena cannam@95: * [project @ 2003-05-26 11:22:59 by athena] cannam@95: Obsolete comment cannam@95: cannam@95: M ./dft/buffered.c -6 +1 cannam@95: cannam@95: Sat May 24 15:00:53 EDT 2003 athena cannam@95: * [project @ 2003-05-24 19:00:53 by athena] cannam@95: comment cannam@95: cannam@95: M ./dft/ct.c -2 +2 cannam@95: cannam@95: Sat May 24 07:20:35 EDT 2003 athena cannam@95: * [project @ 2003-05-24 11:20:35 by athena] cannam@95: Oops---wrong test NO_UGLYP instead of !NO_UGLYP cannam@95: cannam@95: M ./dft/dftw-dft.c -3 +3 cannam@95: cannam@95: Sat May 24 07:05:34 EDT 2003 athena cannam@95: * [project @ 2003-05-24 11:05:34 by athena] cannam@95: Implemented radix r, where n=r^2 * p cannam@95: cannam@95: M ./dft/ct.c -19 +53 cannam@95: cannam@95: Wed May 21 01:54:32 EDT 2003 stevenj cannam@95: * [project @ 2003-05-21 05:54:32 by stevenj] cannam@95: xlc seems to properly use fma as well cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -2 +2 cannam@95: cannam@95: Tue May 20 23:07:43 EDT 2003 stevenj cannam@95: * [project @ 2003-05-21 03:07:43 by stevenj] cannam@95: print warning if there is no cycle counter cannam@95: cannam@95: M ./configure.ac +18 cannam@95: M ./doc/fftw3.texi -2 +3 cannam@95: cannam@95: Tue May 20 17:32:04 EDT 2003 stevenj cannam@95: * [project @ 2003-05-20 21:32:04 by stevenj] cannam@95: updated Funda reference cannam@95: cannam@95: M ./libbench2/verify-lib.c -1 +4 cannam@95: M ./libbench2/verify-r2r.c +3 cannam@95: cannam@95: Mon May 19 20:12:36 EDT 2003 athena cannam@95: * [project @ 2003-05-20 00:12:36 by athena] cannam@95: const cannam@95: cannam@95: M ./dft/ct.c -2 +2 cannam@95: cannam@95: Mon May 19 15:41:09 EDT 2003 athena cannam@95: * [project @ 2003-05-19 19:41:09 by athena] cannam@95: Implemented generic dif square transposed (q-style) solver. cannam@95: cannam@95: M ./dft/dftw-dft.c -8 +88 cannam@95: M ./dft/directwsq.c -2 +2 cannam@95: cannam@95: Mon May 19 07:00:36 EDT 2003 athena cannam@95: * [project @ 2003-05-19 11:00:36 by athena] cannam@95: applicable() is now a property of the solver (in anticipation of cannam@95: transposed solvers) cannam@95: cannam@95: M ./dft/dftw-dft.c -35 +35 cannam@95: cannam@95: Mon May 19 06:33:40 EDT 2003 athena cannam@95: * [project @ 2003-05-19 10:33:40 by athena] cannam@95: Slight cleanup cannam@95: cannam@95: M ./dft/dftw-dft.c -7 +7 cannam@95: cannam@95: Sun May 18 13:05:51 EDT 2003 athena cannam@95: * [project @ 2003-05-18 17:05:51 by athena] cannam@95: Nothing, really cannam@95: cannam@95: M ./dft/bluestein.c -9 +3 cannam@95: M ./kernel/ifftw.h -1 +2 cannam@95: M ./kernel/primes.c -1 +9 cannam@95: cannam@95: Sun May 18 09:05:20 EDT 2003 athena cannam@95: * [project @ 2003-05-18 13:05:20 by athena] cannam@95: Moved vector loop inside bytwiddle(), in anticipation of cannam@95: a q-style dftw-dit transposed solver. cannam@95: cannam@95: M ./dft/dftw-dft.c -38 +41 cannam@95: cannam@95: Sun May 18 08:52:02 EDT 2003 athena cannam@95: * [project @ 2003-05-18 12:52:02 by athena] cannam@95: Fixed flops count cannam@95: cannam@95: M ./dft/dftw-dft.c -1 +1 cannam@95: cannam@95: Sun May 18 08:47:20 EDT 2003 athena cannam@95: * [project @ 2003-05-18 12:47:20 by athena] cannam@95: style cannam@95: cannam@95: M ./dft/dftw-dft.c -5 +3 cannam@95: cannam@95: Sun May 18 07:16:34 EDT 2003 athena cannam@95: * [project @ 2003-05-18 11:16:34 by athena] cannam@95: Faster inner loop. cannam@95: cannam@95: M ./dft/dftw-dft.c -13 +25 cannam@95: cannam@95: Sat May 17 08:02:38 EDT 2003 athena cannam@95: * [project @ 2003-05-17 12:02:38 by athena] cannam@95: Print vector length cannam@95: cannam@95: M ./dft/dftw-dft.c -2 +2 cannam@95: cannam@95: Sat May 17 07:55:33 EDT 2003 athena cannam@95: * [project @ 2003-05-17 11:55:33 by athena] cannam@95: Oops cannam@95: cannam@95: M ./dft/dftw-dft.c -7 +7 cannam@95: cannam@95: Sat May 17 07:50:35 EDT 2003 athena cannam@95: * [project @ 2003-05-17 11:50:35 by athena] cannam@95: Allow vl > 1 cannam@95: cannam@95: M ./dft/dftw-dft.c -7 +16 cannam@95: cannam@95: Sat May 17 07:01:42 EDT 2003 athena cannam@95: * [project @ 2003-05-17 11:01:42 by athena] cannam@95: Radix can be derived from problem---no need to pre-specify it. cannam@95: cannam@95: M ./dft/ctsq.c -21 +12 cannam@95: cannam@95: Fri May 16 22:50:50 EDT 2003 stevenj cannam@95: * [project @ 2003-05-17 02:50:50 by stevenj] cannam@95: fixed comment cannam@95: cannam@95: M ./kernel/transpose.c -1 +1 cannam@95: cannam@95: Fri May 16 22:48:20 EDT 2003 stevenj cannam@95: * [project @ 2003-05-17 02:48:20 by stevenj] cannam@95: whoops, gcd should be static cannam@95: cannam@95: M ./kernel/transpose.c -1 +1 cannam@95: cannam@95: Fri May 16 22:40:32 EDT 2003 stevenj cannam@95: * [project @ 2003-05-17 02:40:32 by stevenj] cannam@95: more unrolling cannam@95: cannam@95: M ./kernel/transpose.c -7 +31 cannam@95: cannam@95: Fri May 16 20:03:48 EDT 2003 athena cannam@95: * [project @ 2003-05-17 00:03:48 by athena] cannam@95: Hack to avoid infinite recursion. cannam@95: cannam@95: M ./dft/bluestein.c +3 cannam@95: cannam@95: Fri May 16 19:52:43 EDT 2003 stevenj cannam@95: * [project @ 2003-05-16 23:52:43 by stevenj] cannam@95: consistency cannam@95: cannam@95: M ./dft/codelet-dft.h -4 +4 cannam@95: cannam@95: Fri May 16 19:45:15 EDT 2003 athena cannam@95: * [project @ 2003-05-16 23:45:15 by athena] cannam@95: Wrong comment. cannam@95: cannam@95: M ./dft/bluestein.c -1 +1 cannam@95: cannam@95: Fri May 16 19:45:03 EDT 2003 athena cannam@95: * [project @ 2003-05-16 23:45:03 by athena] cannam@95: Style. cannam@95: cannam@95: M ./dft/bluestein.c -3 +3 cannam@95: cannam@95: Fri May 16 18:35:27 EDT 2003 stevenj cannam@95: * [project @ 2003-05-16 22:35:27 by stevenj] cannam@95: punctuation cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -2 +2 cannam@95: cannam@95: Fri May 16 18:33:45 EDT 2003 stevenj cannam@95: * [project @ 2003-05-16 22:33:45 by stevenj] cannam@95: added allzero FAQ cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn +7 cannam@95: cannam@95: Fri May 16 18:22:45 EDT 2003 stevenj cannam@95: * [project @ 2003-05-16 22:22:45 by stevenj] cannam@95: simplification: instead of cldb, just use cldf with inputs/output values swapped cannam@95: cannam@95: M ./dft/bluestein.c -22 +11 cannam@95: cannam@95: Fri May 16 15:47:17 EDT 2003 athena cannam@95: * [project @ 2003-05-16 19:47:17 by athena] cannam@95: Allow more general transform sizes. cannam@95: cannam@95: M ./dft/bluestein.c -15 +21 cannam@95: cannam@95: Fri May 16 14:22:37 EDT 2003 stevenj cannam@95: * [project @ 2003-05-16 18:22:37 by stevenj] cannam@95: slight change cannam@95: cannam@95: M ./kernel/ifftw.h -4 +4 cannam@95: cannam@95: Fri May 16 14:22:05 EDT 2003 stevenj cannam@95: * [project @ 2003-05-16 18:22:05 by stevenj] cannam@95: MS has __int64 type, not long long (grr) cannam@95: cannam@95: M ./kernel/ifftw.h -1 +4 cannam@95: cannam@95: Fri May 16 13:34:16 EDT 2003 athena cannam@95: * [project @ 2003-05-16 17:34:16 by athena] cannam@95: Fixed printout cannam@95: cannam@95: M ./dft/ct.c -2 +4 cannam@95: cannam@95: Fri May 16 13:23:00 EDT 2003 athena cannam@95: * [project @ 2003-05-16 17:23:00 by athena] cannam@95: Fixed flop count cannam@95: cannam@95: M ./dft/bluestein.c +5 cannam@95: cannam@95: Fri May 16 13:02:06 EDT 2003 athena cannam@95: * [project @ 2003-05-16 17:02:06 by athena] cannam@95: New bluestein solver cannam@95: cannam@95: A ./dft/bluestein.c cannam@95: M ./dft/Makefile.am -5 +6 cannam@95: M ./dft/bluestein.c +255 cannam@95: M ./dft/conf.c -1 +2 cannam@95: M ./dft/dft.h -1 +2 cannam@95: cannam@95: Fri May 16 09:51:05 EDT 2003 athena cannam@95: * [project @ 2003-05-16 13:51:05 by athena] cannam@95: Implemented generic radix. cannam@95: cannam@95: M ./dft/ct.c -14 +35 cannam@95: cannam@95: Fri May 16 08:19:38 EDT 2003 athena cannam@95: * [project @ 2003-05-16 12:19:38 by athena] cannam@95: Removed conditional branch from inner loop in generic.c cannam@95: cannam@95: M ./dft/generic.c -18 +20 cannam@95: M ./kernel/ifftw.h -2 +2 cannam@95: M ./kernel/twiddle.c -1 +18 cannam@95: cannam@95: Fri May 16 07:48:28 EDT 2003 athena cannam@95: * [project @ 2003-05-16 11:48:28 by athena] cannam@95: Simplified indexing cannam@95: cannam@95: M ./dft/generic.c -29 +20 cannam@95: cannam@95: Fri May 16 06:53:56 EDT 2003 athena cannam@95: * [project @ 2003-05-16 10:53:56 by athena] cannam@95: Better still. cannam@95: cannam@95: M ./dft/generic.c -26 +37 cannam@95: cannam@95: Fri May 16 06:24:31 EDT 2003 athena cannam@95: * [project @ 2003-05-16 10:24:31 by athena] cannam@95: Further improvement of generic solver cannam@95: cannam@95: M ./dft/generic.c -8 +8 cannam@95: cannam@95: Fri May 16 05:57:07 EDT 2003 athena cannam@95: * [project @ 2003-05-16 09:57:07 by athena] cannam@95: Cleanup cannam@95: cannam@95: M ./dft/rader.c -54 +51 cannam@95: cannam@95: Fri May 16 05:42:57 EDT 2003 athena cannam@95: * [project @ 2003-05-16 09:42:57 by athena] cannam@95: Cleanup cannam@95: cannam@95: M ./dft/dftw-dft.c -30 +25 cannam@95: cannam@95: Fri May 16 05:31:40 EDT 2003 athena cannam@95: * [project @ 2003-05-16 09:31:40 by athena] cannam@95: Generic now only works for odd sized. Added check. cannam@95: cannam@95: M ./dft/generic.c -1 +2 cannam@95: cannam@95: Thu May 15 21:53:25 EDT 2003 athena cannam@95: * [project @ 2003-05-16 01:53:25 by athena] cannam@95: Increased GENERIC_MIN_BAD because of new algorithm. cannam@95: cannam@95: M ./kernel/ifftw.h -2 +2 cannam@95: cannam@95: Thu May 15 21:40:27 EDT 2003 athena cannam@95: * [project @ 2003-05-16 01:40:27 by athena] cannam@95: Much, much better. cannam@95: cannam@95: M ./dft/generic.c -9 +30 cannam@95: cannam@95: Thu May 15 21:25:00 EDT 2003 athena cannam@95: * [project @ 2003-05-16 01:25:00 by athena] cannam@95: Still trying to understand why rdft-generic-dit is faster cannam@95: then dft-generic... cannam@95: cannam@95: M ./dft/generic.c -9 +8 cannam@95: cannam@95: Thu May 15 21:04:33 EDT 2003 athena cannam@95: * [project @ 2003-05-16 01:04:33 by athena] cannam@95: Nothing, really cannam@95: cannam@95: M ./dft/generic.c -2 +2 cannam@95: cannam@95: Thu May 15 20:59:45 EDT 2003 athena cannam@95: * [project @ 2003-05-16 00:59:45 by athena] cannam@95: Never be clever for the sake of being clever. cannam@95: cannam@95: M ./dft/generic.c -1 +1 cannam@95: cannam@95: Thu May 15 20:58:06 EDT 2003 athena cannam@95: * [project @ 2003-05-16 00:58:06 by athena] cannam@95: Simplified. generic-dit is gone. The solver is now out-of-place cannam@95: only---buffering is done by the buffered solver. cannam@95: cannam@95: M ./dft/generic.c -94 +45 cannam@95: cannam@95: Thu May 15 19:18:18 EDT 2003 athena cannam@95: * [project @ 2003-05-15 23:18:18 by athena] cannam@95: rader-dit is gone. cannam@95: cannam@95: M ./dft/rader.c -200 cannam@95: cannam@95: Thu May 15 19:13:03 EDT 2003 athena cannam@95: * [project @ 2003-05-15 23:13:03 by athena] cannam@95: Cast cannam@95: cannam@95: M ./dft/plan.c -2 +2 cannam@95: cannam@95: Thu May 15 19:09:07 EDT 2003 athena cannam@95: * [project @ 2003-05-15 23:09:07 by athena] cannam@95: Introduced twiddle problem ``dftw''. Changed most other things cannam@95: to deal with this change. cannam@95: cannam@95: A ./dft/ctsq.c cannam@95: A ./dft/dftw-dft.c cannam@95: A ./dft/directw.c cannam@95: A ./dft/directwbuf.c cannam@95: A ./dft/directwsq.c cannam@95: A ./dft/problemw.c cannam@95: M ./configure.ac -1 +1 cannam@95: M ./dft/Makefile.am -5 +5 cannam@95: M ./dft/buffered.c -2 +1 cannam@95: M ./dft/codelet-dft.h -13 +9 cannam@95: M ./dft/conf.c -1 +4 cannam@95: R ./dft/ct-dif.c cannam@95: R ./dft/ct-dit.c cannam@95: R ./dft/ct-ditbuf.c cannam@95: R ./dft/ct-ditf.c cannam@95: M ./dft/ct.c -91 +157 cannam@95: R ./dft/ct.h cannam@95: M ./dft/ctsq.c +213 cannam@95: M ./dft/dft.h -7 +45 cannam@95: M ./dft/dftw-dft.c +297 cannam@95: M ./dft/directw.c +172 cannam@95: M ./dft/directwbuf.c +230 cannam@95: M ./dft/directwsq.c +148 cannam@95: M ./dft/kdft-dif.c -4 +5 cannam@95: M ./dft/kdft-difsq.c -3 +3 cannam@95: M ./dft/kdft-dit.c -5 +5 cannam@95: M ./dft/plan.c -1 +11 cannam@95: M ./dft/problemw.c +103 cannam@95: M ./dft/rader.c -1 +1 cannam@95: M ./dft/solve.c -1 +8 cannam@95: cannam@95: Sat Aug 12 23:17:14 EDT 2006 Unknown tagger cannam@95: tagged fftw-3-0-1-branch cannam@95: cannam@95: cannam@95: Thu May 15 18:47:18 EDT 2003 stevenj cannam@95: * [project @ 2003-05-15 22:47:18 by stevenj] cannam@95: whoops, X(safe_mulmod) not fftw_safe_mulmod cannam@95: cannam@95: M ./kernel/primes.c -3 +3 cannam@95: cannam@95: Thu May 15 16:53:16 EDT 2003 stevenj cannam@95: * [project @ 2003-05-15 20:53:16 by stevenj] cannam@95: add VC++ versions of asm cannam@95: cannam@95: M ./simd/sse.c -1 +17 cannam@95: M ./simd/sse2.c -1 +17 cannam@95: cannam@95: Thu May 15 15:03:06 EDT 2003 stevenj cannam@95: * [project @ 2003-05-15 19:03:05 by stevenj] cannam@95: VC++ reportedly supports the intel intrinsics, but requires __inline instead of __inline__ cannam@95: cannam@95: M ./simd/simd-sse.h -1 +4 cannam@95: M ./simd/simd-sse2.h -1 +5 cannam@95: cannam@95: Thu May 15 14:32:06 EDT 2003 stevenj cannam@95: * [project @ 2003-05-15 18:32:06 by stevenj] cannam@95: precompute array indices with VC++ cannam@95: cannam@95: M ./kernel/ifftw.h -2 +2 cannam@95: cannam@95: Wed May 14 21:57:39 EDT 2003 stevenj cannam@95: * [project @ 2003-05-15 01:57:39 by stevenj] cannam@95: added doc note cannam@95: cannam@95: M ./acx_pthread.m4 -1 +5 cannam@95: cannam@95: Wed May 14 19:45:54 EDT 2003 stevenj cannam@95: * [project @ 2003-05-14 23:45:54 by stevenj] cannam@95: autodetect windows cannam@95: cannam@95: M ./threads/threads.c -3 +4 cannam@95: cannam@95: Wed May 14 15:08:49 EDT 2003 stevenj cannam@95: * [project @ 2003-05-14 19:08:49 by stevenj] cannam@95: don't bother with #ifdef HAVE_CONFIG_H, since non-Unix users always forget to define it cannam@95: cannam@95: M ./libbench2/getopt.c -2 cannam@95: cannam@95: Tue May 13 16:58:07 EDT 2003 stevenj cannam@95: * [project @ 2003-05-13 20:58:07 by stevenj] cannam@95: VC++ uses __inline cannam@95: cannam@95: M ./kernel/cycle.h -2 +2 cannam@95: cannam@95: Tue May 13 14:51:26 EDT 2003 stevenj cannam@95: * [project @ 2003-05-13 18:51:26 by stevenj] cannam@95: added leak question cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn +8 cannam@95: cannam@95: Mon May 12 18:26:51 EDT 2003 stevenj cannam@95: * [project @ 2003-05-12 22:26:51 by stevenj] cannam@95: LARGE_INTEGER needs windows.h (supposedly, there is some problem converting _itnt64 to double...damn MS and their nonstandard types) cannam@95: cannam@95: M ./kernel/cycle.h -1 +2 cannam@95: cannam@95: Mon May 12 18:22:16 EDT 2003 stevenj cannam@95: * [project @ 2003-05-12 22:22:16 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./libbench2/timer.c -2 +2 cannam@95: cannam@95: Mon May 12 17:16:19 EDT 2003 stevenj cannam@95: * [project @ 2003-05-12 21:16:19 by stevenj] cannam@95: added 256x256 to canonical list cannam@95: cannam@95: M ./tools/fftw-wisdom.c -1 +1 cannam@95: cannam@95: Mon May 12 07:02:06 EDT 2003 athena cannam@95: * [project @ 2003-05-12 11:02:06 by athena] cannam@95: Oops... cannam@95: cannam@95: M ./kernel/transpose.c +1 cannam@95: cannam@95: Sun May 11 11:04:46 EDT 2003 athena cannam@95: * [project @ 2003-05-11 15:04:46 by athena] cannam@95: Unrolled loops, changed cutoff cannam@95: cannam@95: M ./kernel/transpose.c -23 +80 cannam@95: cannam@95: Sun May 11 10:20:04 EDT 2003 athena cannam@95: * [project @ 2003-05-11 14:20:04 by athena] cannam@95: Do not multiply strides by 2 twice. cannam@95: cannam@95: M ./tests/bench.c -16 +15 cannam@95: cannam@95: Wed May 7 21:09:43 EDT 2003 stevenj cannam@95: * [project @ 2003-05-08 01:09:43 by stevenj] cannam@95: added 'make smallcheck' cannam@95: cannam@95: M ./tests/Makefile.am +13 cannam@95: cannam@95: Wed May 7 20:46:10 EDT 2003 stevenj cannam@95: * [project @ 2003-05-08 00:46:10 by stevenj] cannam@95: --without-cycle-counter becomes --with-slow-timer, updated docs cannam@95: cannam@95: M ./configure.ac -3 +3 cannam@95: M ./doc/fftw3.texi -15 +15 cannam@95: M ./kernel/timer.c -3 +3 cannam@95: cannam@95: Wed May 7 18:05:29 EDT 2003 stevenj cannam@95: * [project @ 2003-05-07 22:05:29 by stevenj] cannam@95: remove duplicate -openmp check; Sun requires -xopenmp cannam@95: cannam@95: M ./configure.ac -5 +5 cannam@95: cannam@95: Wed May 7 17:59:23 EDT 2003 stevenj cannam@95: * [project @ 2003-05-07 21:59:23 by stevenj] cannam@95: fixed compilation under Sun C++ cannam@95: cannam@95: M ./dft/ct-ditbuf.c -3 +3 cannam@95: M ./rdft/hc2hc-buf.c -3 +3 cannam@95: cannam@95: Wed May 7 14:24:46 EDT 2003 athena cannam@95: * [project @ 2003-05-07 18:24:46 by athena] cannam@95: Use estimator if cycle counter is unavailable, regardless cannam@95: of the FFTW_MEASURE/ESTIMATE setting. cannam@95: cannam@95: M ./kernel/planner.c -3 +13 cannam@95: M ./kernel/timer.c -57 +67 cannam@95: cannam@95: Tue May 6 23:15:34 EDT 2003 stevenj cannam@95: * [project @ 2003-05-07 03:15:34 by stevenj] cannam@95: _WIN32 (not __WIN32__) is always defined cannam@95: cannam@95: M ./kernel/cycle.h -3 +2 cannam@95: cannam@95: Tue May 6 23:11:52 EDT 2003 stevenj cannam@95: * [project @ 2003-05-07 03:11:52 by stevenj] cannam@95: minor cleanup cannam@95: cannam@95: M ./kernel/cycle.h -2 +2 cannam@95: cannam@95: Tue May 6 22:50:07 EDT 2003 stevenj cannam@95: * [project @ 2003-05-07 02:50:07 by stevenj] cannam@95: tentative VC++ stuff, some consolidation cannam@95: cannam@95: M ./kernel/cycle.h -51 +56 cannam@95: cannam@95: Tue May 6 12:17:56 EDT 2003 stevenj cannam@95: * [project @ 2003-05-06 16:17:56 by stevenj] cannam@95: made cycle.h more self-contained cannam@95: cannam@95: M ./kernel/cycle.h -3 +13 cannam@95: M ./kernel/timer.c -15 +18 cannam@95: cannam@95: Tue May 6 08:30:39 EDT 2003 athena cannam@95: * [project @ 2003-05-06 12:30:39 by athena] cannam@95: Use ``%'' flag to denote commutative operations. cannam@95: cannam@95: M ./simd/simd-3dnow.h -3 +3 cannam@95: M ./simd/simd-sse.h -3 +3 cannam@95: M ./simd/simd-sse2.h -3 +3 cannam@95: cannam@95: Mon May 5 20:42:30 EDT 2003 stevenj cannam@95: * [project @ 2003-05-06 00:42:30 by stevenj] cannam@95: MIT license, brief documentation cannam@95: cannam@95: M ./kernel/cycle.h -12 +41 cannam@95: cannam@95: Mon May 5 20:31:16 EDT 2003 stevenj cannam@95: * [project @ 2003-05-06 00:31:16 by stevenj] cannam@95: whoops, forgot f77_wisdom.f cannam@95: cannam@95: M ./doc/Makefile.am -1 +1 cannam@95: cannam@95: Sun May 4 19:37:09 EDT 2003 athena cannam@95: * [project @ 2003-05-04 23:37:09 by athena] cannam@95: Improved speed of accuracy test. cannam@95: cannam@95: M ./dft/problem.c -6 +3 cannam@95: M ./libbench/mp.c -16 +26 cannam@95: M ./libbench2/bench.h -1 +2 cannam@95: M ./libbench2/mp.c -46 +92 cannam@95: M ./libbench2/verify-lib.c -1 +3 cannam@95: M ./rdft/problem2.c -3 +2 cannam@95: cannam@95: Tue Apr 29 11:45:34 EDT 2003 athena cannam@95: * [project @ 2003-04-29 15:45:34 by athena] cannam@95: s390 cycle counter cannam@95: cannam@95: M ./kernel/cycle.h -1 +21 cannam@95: cannam@95: Sat Apr 26 12:26:15 EDT 2003 stevenj cannam@95: * [project @ 2003-04-26 16:26:15 by stevenj] cannam@95: forgot r2r directory cannam@95: cannam@95: M ./doc/fftw3.texi -10 +10 cannam@95: cannam@95: Fri Apr 25 20:52:23 EDT 2003 stevenj cannam@95: * [project @ 2003-04-26 00:52:23 by stevenj] cannam@95: delete unused files, since they don't compile any more cannam@95: cannam@95: M ./rdft/Makefile.am -2 cannam@95: R ./rdft/vrank2-transpose.c cannam@95: R ./rdft/vrank3-transpose.c cannam@95: cannam@95: Thu Apr 24 06:37:41 EDT 2003 athena cannam@95: * [project @ 2003-04-24 10:37:41 by athena] cannam@95: Better gcc code generation cannam@95: cannam@95: M ./simd/simd-sse2.h -3 +5 cannam@95: cannam@95: Wed Apr 23 15:30:50 EDT 2003 stevenj cannam@95: * [project @ 2003-04-23 19:30:50 by stevenj] cannam@95: ccc is the Compaq C compiler on Linux/alpha cannam@95: cannam@95: M ./acinclude.m4 -1 +1 cannam@95: cannam@95: Wed Apr 23 00:06:03 EDT 2003 stevenj cannam@95: * [project @ 2003-04-23 04:06:03 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./doc/fftw3.texi -3 +3 cannam@95: cannam@95: Sat Aug 12 23:16:25 EDT 2006 Unknown tagger cannam@95: tagged fftw-3-0 cannam@95: cannam@95: cannam@95: Sat Apr 19 09:18:25 EDT 2003 athena cannam@95: * [project @ 2003-04-19 13:18:25 by athena] cannam@95: ia64 cycle counter with intel compiler. cannam@95: cannam@95: M ./kernel/cycle.h -1 +19 cannam@95: cannam@95: Fri Apr 18 18:27:30 EDT 2003 athena cannam@95: * [project @ 2003-04-18 22:27:30 by athena] cannam@95: More gcc bugs. Sigh. cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -4 +12 cannam@95: cannam@95: Fri Apr 18 18:01:49 EDT 2003 athena cannam@95: * [project @ 2003-04-18 22:01:49 by athena] cannam@95: touch ChangeLog to observe GNU standards cannam@95: cannam@95: M ./bootstrap.sh +1 cannam@95: cannam@95: Fri Apr 18 18:01:12 EDT 2003 athena cannam@95: * [project @ 2003-04-18 22:01:12 by athena] cannam@95: We now build ChangeLog automatically at distribution time cannam@95: cannam@95: R ./ChangeLog cannam@95: cannam@95: Fri Apr 18 18:00:17 EDT 2003 athena cannam@95: * [project @ 2003-04-18 22:00:17 by athena] cannam@95: Automatic ChangeLog hackery cannam@95: cannam@95: M ./mkdist.sh +7 cannam@95: cannam@95: Fri Apr 18 13:25:26 EDT 2003 stevenj cannam@95: * [project @ 2003-04-18 17:25:26 by stevenj] cannam@95: plural cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -1 +1 cannam@95: cannam@95: Fri Apr 18 13:25:01 EDT 2003 stevenj cannam@95: * [project @ 2003-04-18 17:25:01 by stevenj] cannam@95: updated cannam@95: cannam@95: M ./NEWS +12 cannam@95: cannam@95: Fri Apr 18 12:59:41 EDT 2003 athena cannam@95: * [project @ 2003-04-18 16:59:41 by athena] cannam@95: Updated cannam@95: cannam@95: M ./ChangeLog -6789 +3415 cannam@95: cannam@95: Fri Apr 18 11:48:39 EDT 2003 stevenj cannam@95: * [project @ 2003-04-18 15:48:39 by stevenj] cannam@95: a -> an cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -1 +1 cannam@95: cannam@95: Fri Apr 18 11:47:56 EDT 2003 stevenj cannam@95: * [project @ 2003-04-18 15:47:56 by stevenj] cannam@95: hyphen cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -1 +1 cannam@95: cannam@95: Fri Apr 18 11:47:28 EDT 2003 stevenj cannam@95: * [project @ 2003-04-18 15:47:28 by stevenj] cannam@95: comma cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -1 +1 cannam@95: cannam@95: Fri Apr 18 11:46:59 EDT 2003 stevenj cannam@95: * [project @ 2003-04-18 15:46:59 by stevenj] cannam@95: minor cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -4 +4 cannam@95: cannam@95: Fri Apr 18 10:37:31 EDT 2003 athena cannam@95: * [project @ 2003-04-18 14:37:31 by athena] cannam@95: Updated cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -9 +36 cannam@95: cannam@95: Fri Apr 18 10:14:59 EDT 2003 athena cannam@95: * [project @ 2003-04-18 14:14:59 by athena] cannam@95: New script that builds the distributions cannam@95: cannam@95: A ./mkdist.sh cannam@95: cannam@95: Fri Apr 18 08:51:07 EDT 2003 athena cannam@95: * [project @ 2003-04-18 12:51:07 by athena] cannam@95: Oops again cannam@95: cannam@95: M ./dft/simd/codelets/Makefile.am -4 +4 cannam@95: cannam@95: Fri Apr 18 08:39:05 EDT 2003 athena cannam@95: * [project @ 2003-04-18 12:39:05 by athena] cannam@95: Oops, forgot -sign 1 cannam@95: cannam@95: M ./dft/simd/codelets/Makefile.am -4 +4 cannam@95: cannam@95: Fri Apr 18 08:28:25 EDT 2003 athena cannam@95: * [project @ 2003-04-18 12:28:25 by athena] cannam@95: Reorganization of simd codelets cannam@95: cannam@95: M ./configure.ac -4 cannam@95: M ./dft/simd/codelets/Makefile.am -136 +22 cannam@95: M ./dft/simd/n1b.c -3 +3 cannam@95: M ./dft/simd/n1b.h -3 cannam@95: M ./dft/simd/n1f.c -3 +3 cannam@95: M ./dft/simd/n1f.h -3 cannam@95: M ./dft/simd/n2b.c -3 +3 cannam@95: M ./dft/simd/n2b.h +3 cannam@95: M ./dft/simd/n2f.c -3 +3 cannam@95: M ./dft/simd/n2f.h +3 cannam@95: cannam@95: Thu Apr 17 21:21:45 EDT 2003 athena cannam@95: * [project @ 2003-04-18 01:21:45 by athena] cannam@95: k7 assembly was not updated after conversion of opcnt from cannam@95: int to double cannam@95: cannam@95: M ./genfft-k7/gen_notw.ml -5 +5 cannam@95: M ./genfft-k7/gen_twiddle.ml -5 +6 cannam@95: cannam@95: Thu Apr 17 19:15:53 EDT 2003 athena cannam@95: * [project @ 2003-04-17 23:15:53 by athena] cannam@95: Capital `X' looks bad in all-lowercase plans cannam@95: cannam@95: M ./dft/vrank2-transpose.c -2 +2 cannam@95: M ./dft/vrank3-transpose.c -2 +2 cannam@95: cannam@95: Thu Apr 17 18:53:29 EDT 2003 athena cannam@95: * [project @ 2003-04-17 22:53:29 by athena] cannam@95: Removed redundant inline/noinline codelets cannam@95: cannam@95: M ./dft/codelets/standard/Makefile.am -3 +2 cannam@95: M ./dft/simd/codelets/Makefile.am -27 +12 cannam@95: M ./rdft/codelets/hc2r/Makefile.am -8 +4 cannam@95: M ./rdft/codelets/r2hc/Makefile.am -8 +4 cannam@95: cannam@95: Thu Apr 17 15:25:50 EDT 2003 athena cannam@95: * [project @ 2003-04-17 19:25:50 by athena] cannam@95: New noinline cannam@95: Noinline real codelets cannam@95: cannam@95: A ./genfft/gen_hc2r_noinline.ml cannam@95: A ./genfft/gen_r2hc_noinline.ml cannam@95: M ./genfft/Makefile.am -10 +24 cannam@95: M ./genfft/gen_hc2hc.ml -5 +5 cannam@95: M ./genfft/gen_hc2r.ml -48 +31 cannam@95: M ./genfft/gen_hc2r_noinline.ml +185 cannam@95: M ./genfft/gen_r2hc.ml -49 +31 cannam@95: M ./genfft/gen_r2hc_noinline.ml +181 cannam@95: M ./genfft/gen_r2r.ml -3 +3 cannam@95: M ./kernel/ifftw.h -4 +1 cannam@95: M ./rdft/codelets/hc2r/Makefile.am -1 +15 cannam@95: M ./rdft/codelets/r2hc/Makefile.am -1 +15 cannam@95: M ./support/Makefile.codelets -1 +3 cannam@95: cannam@95: Thu Apr 17 15:23:03 EDT 2003 stevenj cannam@95: * [project @ 2003-04-17 19:23:03 by stevenj] cannam@95: more ideas cannam@95: cannam@95: M ./TODO +6 cannam@95: cannam@95: Thu Apr 17 13:18:45 EDT 2003 athena cannam@95: * [project @ 2003-04-17 17:18:45 by athena] cannam@95: Removed duplicate rules. cannam@95: cannam@95: M ./dft/simd/codelets/Makefile.am -6 cannam@95: cannam@95: Thu Apr 17 10:51:09 EDT 2003 athena cannam@95: * [project @ 2003-04-17 14:51:09 by athena] cannam@95: acx_pthread.m4 was not distributed cannam@95: cannam@95: M ./Makefile.am -1 +1 cannam@95: cannam@95: Thu Apr 17 07:21:17 EDT 2003 athena cannam@95: * [project @ 2003-04-17 11:21:17 by athena] cannam@95: Oops cannam@95: cannam@95: M ./support/Makefile.codelets -1 +1 cannam@95: cannam@95: Thu Apr 17 07:07:19 EDT 2003 athena cannam@95: * [project @ 2003-04-17 11:07:19 by athena] cannam@95: Both inlined and non-inlined notw codelets. cannam@95: cannam@95: A ./genfft/gen_notw_noinline.ml cannam@95: M ./dft/codelets/standard/Makefile.am -1 +10 cannam@95: M ./genfft/Makefile.am -11 +17 cannam@95: M ./genfft/gen_notw.ml -51 +37 cannam@95: M ./genfft/gen_notw_noinline.ml +164 cannam@95: M ./support/Makefile.codelets +1 cannam@95: cannam@95: Thu Apr 17 06:44:21 EDT 2003 athena cannam@95: * [project @ 2003-04-17 10:44:21 by athena] cannam@95: Initial experiment with both inlined and non-inlined simd codelets. cannam@95: Both are included for now. cannam@95: cannam@95: A ./genfft/gen_notw_noinline_c.ml cannam@95: M ./dft/simd/codelets/Makefile.am -3 +68 cannam@95: M ./genfft/Makefile.am -9 +17 cannam@95: M ./genfft/gen_notw_noinline_c.ml +160 cannam@95: M ./support/Makefile.codelets -1 +2 cannam@95: cannam@95: Thu Apr 17 05:57:36 EDT 2003 athena cannam@95: * [project @ 2003-04-17 09:57:36 by athena] cannam@95: --enable-fma to build FMA distribution cannam@95: cannam@95: M ./configure.ac +7 cannam@95: M ./support/Makefile.codelets -1 +7 cannam@95: cannam@95: Wed Apr 16 17:21:53 EDT 2003 athena cannam@95: * [project @ 2003-04-16 21:21:53 by athena] cannam@95: Inline SIMD nontwiddle codelets cannam@95: cannam@95: M ./genfft/gen_notw_c.ml -47 +41 cannam@95: cannam@95: Wed Apr 16 16:18:29 EDT 2003 athena cannam@95: * [project @ 2003-04-16 20:18:29 by athena] cannam@95: Pathetic attempt at saving a couple of registers... cannam@95: cannam@95: M ./simd/simd-sse.h -6 +9 cannam@95: M ./simd/simd-sse2.h -6 +9 cannam@95: cannam@95: Wed Apr 16 15:51:27 EDT 2003 athena cannam@95: * [project @ 2003-04-16 19:51:27 by athena] cannam@95: for (i = 0; i < m; ++i) ==> for (i = m; i > 0; --i) cannam@95: No proof of evidence that this is any faster, but just in case... cannam@95: cannam@95: M ./genfft/gen_hc2r.ml -3 +3 cannam@95: M ./genfft/gen_notw.ml -3 +3 cannam@95: M ./genfft/gen_notw_c.ml -3 +3 cannam@95: M ./genfft/gen_r2hc.ml -3 +3 cannam@95: M ./genfft/gen_r2r.ml -3 +3 cannam@95: M ./genfft/gen_twiddle.ml -5 +5 cannam@95: M ./genfft/gen_twiddle_c.ml -5 +5 cannam@95: cannam@95: Tue Apr 15 15:03:20 EDT 2003 stevenj cannam@95: * [project @ 2003-04-15 19:03:20 by stevenj] cannam@95: added hack to make sure that codelet loops are preferred to vecloop solvers in the estimator cannam@95: cannam@95: M ./dft/vrank-geq1.c -1 +2 cannam@95: M ./rdft/vrank-geq1-rdft2.c -1 +2 cannam@95: M ./rdft/vrank-geq1.c -1 +2 cannam@95: cannam@95: Tue Apr 15 14:53:44 EDT 2003 stevenj cannam@95: * [project @ 2003-04-15 18:53:44 by stevenj] cannam@95: use double for flops cannam@95: cannam@95: M ./api/f77funcs.h -1 +1 cannam@95: M ./api/fftw3.h -2 +2 cannam@95: M ./api/flops.c -1 +1 cannam@95: M ./doc/fftw3.texi -3 +6 cannam@95: M ./kernel/ifftw.h -5 +5 cannam@95: M ./tests/bench.c -2 +2 cannam@95: cannam@95: Tue Apr 15 14:51:50 EDT 2003 stevenj cannam@95: * [project @ 2003-04-15 18:51:50 by stevenj] cannam@95: metrowerks reportedly supports gcc assembly extensions on ppc cannam@95: cannam@95: M ./kernel/cycle.h -2 +2 cannam@95: cannam@95: Mon Apr 14 15:00:50 EDT 2003 athena cannam@95: * [project @ 2003-04-14 19:00:48 by athena] cannam@95: foo_CFLAGS generates some automake junk that breaks the build cannam@95: on Redhat 7.3. Screw it. cannam@95: cannam@95: M ./dft/simd/Makefile.am -1 +1 cannam@95: M ./simd/Makefile.am -1 +1 cannam@95: cannam@95: Mon Apr 14 12:22:59 EDT 2003 athena cannam@95: * [project @ 2003-04-14 16:22:59 by athena] cannam@95: Carefully check return status cannam@95: cannam@95: M ./tests/check.pl -1 +21 cannam@95: cannam@95: Sun Apr 13 16:46:12 EDT 2003 athena cannam@95: * [project @ 2003-04-13 20:46:12 by athena] cannam@95: Removed annoying -FMA() expressions. cannam@95: cannam@95: M ./genfft/c.ml -2 +6 cannam@95: M ./genfft/simd.ml -2 +2 cannam@95: M ./kernel/ifftw.h -1 +9 cannam@95: M ./support/Makefile.codelets -1 +1 cannam@95: cannam@95: Sat Apr 12 14:32:22 EDT 2003 athena cannam@95: * [project @ 2003-04-12 18:32:22 by athena] cannam@95: Major fma hackery cannam@95: cannam@95: M ./kernel/ifftw.h -5 +17 cannam@95: cannam@95: Sat Apr 12 14:25:43 EDT 2003 athena cannam@95: * [project @ 2003-04-12 18:25:43 by athena] cannam@95: Slight cleanup cannam@95: cannam@95: M ./api/apiplan.c -8 +5 cannam@95: cannam@95: Sat Apr 12 10:04:51 EDT 2003 athena cannam@95: * [project @ 2003-04-12 14:04:51 by athena] cannam@95: Updated version number cannam@95: cannam@95: M ./configure.ac -1 +1 cannam@95: cannam@95: Sat Apr 12 08:03:07 EDT 2003 athena cannam@95: * [project @ 2003-04-12 12:03:07 by athena] cannam@95: Damn autoconf cannam@95: cannam@95: M ./acinclude.m4 -2 +2 cannam@95: cannam@95: Sat Apr 12 07:54:20 EDT 2003 athena cannam@95: * [project @ 2003-04-12 11:54:20 by athena] cannam@95: Recognize all 74xx processors cannam@95: cannam@95: M ./acinclude.m4 -4 +5 cannam@95: cannam@95: Sat Apr 12 07:35:17 EDT 2003 athena cannam@95: * [project @ 2003-04-12 11:35:17 by athena] cannam@95: Detect 7400 processor. cannam@95: cannam@95: M ./acinclude.m4 -3 +7 cannam@95: cannam@95: Fri Apr 11 20:42:11 EDT 2003 athena cannam@95: * [project @ 2003-04-12 00:42:11 by athena] cannam@95: No need to check for gcc-2.95 cannam@95: cannam@95: M ./acinclude.m4 -4 +2 cannam@95: cannam@95: Fri Apr 11 16:14:39 EDT 2003 stevenj cannam@95: * [project @ 2003-04-11 20:14:39 by stevenj] cannam@95: removed duplicate cannam@95: cannam@95: M ./NEWS -3 cannam@95: cannam@95: Fri Apr 11 08:45:37 EDT 2003 athena cannam@95: * [project @ 2003-04-11 12:45:37 by athena] cannam@95: mflops ==> ``mflops'' cannam@95: cannam@95: M ./libbench2/report.c -2 +2 cannam@95: cannam@95: Fri Apr 11 07:00:53 EDT 2003 athena cannam@95: * [project @ 2003-04-11 11:00:53 by athena] cannam@95: Print setup time as well cannam@95: cannam@95: M ./libbench2/report.c -3 +5 cannam@95: cannam@95: Sat Aug 12 23:14:42 EDT 2006 Unknown tagger cannam@95: tagged fftw-3-0-beta3 cannam@95: cannam@95: cannam@95: Thu Apr 10 15:36:18 EDT 2003 athena cannam@95: * [project @ 2003-04-10 19:36:15 by athena] cannam@95: Enforce pointer equality for in-place problems. cannam@95: cannam@95: M ./dft/problem.c -4 +8 cannam@95: M ./kernel/ifftw.h -1 +4 cannam@95: M ./rdft/problem.c -4 +3 cannam@95: M ./rdft/problem2.c -4 +7 cannam@95: M ./simd/taint.c -1 +8 cannam@95: cannam@95: Wed Apr 9 17:47:54 EDT 2003 stevenj cannam@95: * [project @ 2003-04-09 21:47:26 by stevenj] cannam@95: updated cannam@95: cannam@95: M ./ChangeLog +406 cannam@95: M ./NEWS +28 cannam@95: cannam@95: Wed Apr 9 14:53:38 EDT 2003 stevenj cannam@95: * [project @ 2003-04-09 18:53:38 by stevenj] cannam@95: cross-ref fftw-wisdom man page cannam@95: cannam@95: M ./tests/README -1 +1 cannam@95: cannam@95: Wed Apr 9 10:13:00 EDT 2003 athena cannam@95: * [project @ 2003-04-09 14:13:00 by athena] cannam@95: Undone previous change, committed by mistake. cannam@95: cannam@95: M ./kernel/planner.c -2 +1 cannam@95: cannam@95: Wed Apr 9 10:12:24 EDT 2003 athena cannam@95: * [project @ 2003-04-09 14:12:18 by athena] cannam@95: Quick and dirty README for bench cannam@95: cannam@95: A ./tests/README cannam@95: M ./kernel/planner.c -1 +2 cannam@95: M ./tests/Makefile.am -1 +1 cannam@95: M ./tests/README +57 cannam@95: cannam@95: Wed Apr 9 08:50:25 EDT 2003 athena cannam@95: * [project @ 2003-04-09 12:50:25 by athena] cannam@95: Consider additional command-line arguments as problems to be cannam@95: benchmarked. cannam@95: cannam@95: M ./libbench2/bench-main.c -7 +6 cannam@95: M ./libbench2/timer.c -1 +7 cannam@95: cannam@95: Wed Apr 9 08:44:13 EDT 2003 athena cannam@95: * [project @ 2003-04-09 12:44:13 by athena] cannam@95: Default report format is now human-readable. Removed cannam@95: unnecessary complexity in benchmark reporting. cannam@95: cannam@95: M ./libbench2/bench-main.c -18 +7 cannam@95: M ./libbench2/bench.h -5 +2 cannam@95: M ./libbench2/report.c -71 +39 cannam@95: cannam@95: Wed Apr 9 06:10:40 EDT 2003 athena cannam@95: * [project @ 2003-04-09 10:10:40 by athena] cannam@95: Updated for new interleaved/split api. cannam@95: cannam@95: M ./doc/fftw3.texi -100 +180 cannam@95: cannam@95: Wed Apr 9 03:01:03 EDT 2003 stevenj cannam@95: * [project @ 2003-04-09 07:01:03 by stevenj] cannam@95: updated citation cannam@95: cannam@95: M ./doc/fftw3.texi -3 +3 cannam@95: cannam@95: Tue Apr 8 19:35:59 EDT 2003 athena cannam@95: * [project @ 2003-04-08 23:35:59 by athena] cannam@95: Time for beta3 cannam@95: cannam@95: M ./configure.ac -1 +1 cannam@95: cannam@95: Tue Apr 8 17:40:59 EDT 2003 stevenj cannam@95: * [project @ 2003-04-08 21:40:59 by stevenj] cannam@95: whoops, added cannam@95: cannam@95: A ./reodft/redft00e-r2hc-pad.c cannam@95: cannam@95: Tue Apr 8 17:33:47 EDT 2003 stevenj cannam@95: * [project @ 2003-04-08 21:33:47 by stevenj] cannam@95: more comparison of different R*DFT types cannam@95: cannam@95: M ./doc/fftw3.texi -1 +32 cannam@95: cannam@95: Tue Apr 8 16:48:08 EDT 2003 stevenj cannam@95: * [project @ 2003-04-08 20:48:08 by stevenj] cannam@95: comments cannam@95: cannam@95: M ./reodft/redft00e-r2hc.c -5 +18 cannam@95: M ./reodft/rodft00e-r2hc.c -5 +18 cannam@95: cannam@95: Tue Apr 8 16:19:39 EDT 2003 stevenj cannam@95: * [project @ 2003-04-08 20:19:39 by stevenj] cannam@95: more accurate DCT-I and DST-I, at the expense of up to a factor of 2 in speed and memory cannam@95: cannam@95: A ./reodft/rodft00e-r2hc-pad.c cannam@95: M ./reodft/Makefile.am -4 +5 cannam@95: M ./reodft/conf.c -3 +5 cannam@95: M ./reodft/reodft.h +2 cannam@95: M ./reodft/rodft00e-r2hc-pad.c +200 cannam@95: cannam@95: Tue Apr 8 05:38:09 EDT 2003 athena cannam@95: * [project @ 2003-04-08 09:38:09 by athena] cannam@95: Workaround gcc/sparc bug cannam@95: cannam@95: M ./kernel/planner.c -2 +3 cannam@95: cannam@95: Tue Apr 8 01:34:12 EDT 2003 stevenj cannam@95: * [project @ 2003-04-08 05:34:12 by stevenj] cannam@95: rumors cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Mon Apr 7 18:54:11 EDT 2003 stevenj cannam@95: * [project @ 2003-04-07 22:54:11 by stevenj] cannam@95: added rdft2 paranoid mode cannam@95: cannam@95: M ./tests/hook.c +33 cannam@95: cannam@95: Mon Apr 7 18:47:37 EDT 2003 stevenj cannam@95: * [project @ 2003-04-07 22:47:37 by stevenj] cannam@95: added paranoid mode for r2r cannam@95: cannam@95: M ./tests/hook.c +51 cannam@95: cannam@95: Mon Apr 7 15:10:08 EDT 2003 stevenj cannam@95: * [project @ 2003-04-07 19:10:08 by stevenj] cannam@95: whoops, sincos is predefined on some systems cannam@95: cannam@95: M ./libbench2/verify-r2r.c -3 +3 cannam@95: cannam@95: Sat Apr 5 16:50:57 EST 2003 athena cannam@95: * [project @ 2003-04-05 21:50:57 by athena] cannam@95: bp->destroy_input was not initialized cannam@95: cannam@95: M ./tests/hook.c +1 cannam@95: cannam@95: Sat Apr 5 09:29:11 EST 2003 athena cannam@95: * [project @ 2003-04-05 14:29:11 by athena] cannam@95: Asserted correctness conditions for tainted pointers. cannam@95: cannam@95: (For now, use CK() while we test. They should be changed into cannam@95: A() at some point.) cannam@95: cannam@95: M ./dft/problem.c -7 +16 cannam@95: M ./kernel/ifftw.h -1 +3 cannam@95: M ./rdft/problem.c -3 +7 cannam@95: M ./rdft/problem2.c -6 +12 cannam@95: cannam@95: Sat Apr 5 08:18:23 EST 2003 athena cannam@95: * [project @ 2003-04-05 13:18:23 by athena] cannam@95: Untaint pointers before zero'ing arrays and before hashing cannam@95: cannam@95: M ./dft/problem.c -8 +10 cannam@95: M ./rdft/problem.c -4 +4 cannam@95: M ./rdft/problem2.c -9 +9 cannam@95: cannam@95: Sat Apr 5 07:11:56 EST 2003 athena cannam@95: * [project @ 2003-04-05 12:11:56 by athena] cannam@95: Alignment check did not work with icc, which seems to be cannam@95: confused by the fact that the variable is not used. cannam@95: cannam@95: M ./libbench2/bench-main.c -5 +4 cannam@95: cannam@95: Sat Apr 5 06:41:20 EST 2003 athena cannam@95: * [project @ 2003-04-05 11:41:20 by athena] cannam@95: More paranoid paranoid-check cannam@95: cannam@95: M ./tests/Makefile.am -1 +7 cannam@95: cannam@95: Sat Apr 5 06:19:25 EST 2003 athena cannam@95: * [project @ 2003-04-05 11:19:25 by athena] cannam@95: 0 == x & 7 parses as (0 == x) & 7, which is wrong cannam@95: cannam@95: M ./kernel/ifftw.h -2 +2 cannam@95: cannam@95: Fri Apr 4 21:35:49 EST 2003 stevenj cannam@95: * [project @ 2003-04-05 02:35:49 by stevenj] cannam@95: alignment checks cannam@95: cannam@95: M ./dft/direct.c -1 +2 cannam@95: M ./kernel/ifftw.h -1 +12 cannam@95: M ./kernel/planner.c -1 +2 cannam@95: M ./libbench2/bench-main.c -1 +3 cannam@95: M ./rdft/direct.c -1 +4 cannam@95: M ./rdft/direct2.c -1 +3 cannam@95: cannam@95: Fri Apr 4 21:04:14 EST 2003 stevenj cannam@95: * [project @ 2003-04-05 02:04:14 by stevenj] cannam@95: prevent infinite loops in exhaustive planning cannam@95: cannam@95: M ./rdft/rdft-dht.c -1 +6 cannam@95: cannam@95: Fri Apr 4 20:58:20 EST 2003 stevenj cannam@95: * [project @ 2003-04-05 01:58:20 by stevenj] cannam@95: split/unsplit guru interface cannam@95: cannam@95: A ./api/execute-split-dft-c2r.c cannam@95: A ./api/execute-split-dft-r2c.c cannam@95: A ./api/execute-split-dft.c cannam@95: A ./api/plan-guru-split-dft-c2r.c cannam@95: A ./api/plan-guru-split-dft-r2c.c cannam@95: A ./api/plan-guru-split-dft.c cannam@95: M ./api/Makefile.am -1 +3 cannam@95: M ./api/api.h -2 +3 cannam@95: M ./api/apiplan.c -3 +4 cannam@95: M ./api/execute-dft-c2r.c -2 +2 cannam@95: M ./api/execute-dft-r2c.c -2 +2 cannam@95: M ./api/execute-dft.c -3 +6 cannam@95: M ./api/execute-split-dft-c2r.c +29 cannam@95: M ./api/execute-split-dft-r2c.c +29 cannam@95: M ./api/execute-split-dft.c +29 cannam@95: M ./api/f77funcs.h -8 +73 cannam@95: M ./api/fftw3.h -5 +25 cannam@95: M ./api/mktensor-iodims.c -3 +3 cannam@95: M ./api/plan-guru-dft-c2r.c -4 +9 cannam@95: M ./api/plan-guru-dft-r2c.c -4 +9 cannam@95: M ./api/plan-guru-dft.c -4 +10 cannam@95: M ./api/plan-guru-r2r.c -3 +4 cannam@95: M ./api/plan-guru-split-dft-c2r.c +40 cannam@95: M ./api/plan-guru-split-dft-r2c.c +39 cannam@95: M ./api/plan-guru-split-dft.c +39 cannam@95: M ./api/plan-many-dft-c2r.c -1 +1 cannam@95: M ./api/plan-many-dft-r2c.c -1 +1 cannam@95: M ./api/plan-many-dft.c -1 +1 cannam@95: M ./api/plan-many-r2r.c -1 +1 cannam@95: M ./tests/bench.c -29 +15 cannam@95: cannam@95: Fri Apr 4 20:39:55 EST 2003 athena cannam@95: * [project @ 2003-04-05 01:39:55 by athena] cannam@95: Need UNTAINT in verifier too. cannam@95: cannam@95: M ./tests/hook.c -4 +4 cannam@95: cannam@95: Fri Apr 4 19:36:46 EST 2003 athena cannam@95: * [project @ 2003-04-05 00:36:46 by athena] cannam@95: Forgot #if HAVE_SIMD cannam@95: cannam@95: M ./simd/taint.c -1 +5 cannam@95: cannam@95: Fri Apr 4 19:30:37 EST 2003 athena cannam@95: * [project @ 2003-04-05 00:30:37 by athena] cannam@95: Keep track of two separate taint bits cannam@95: cannam@95: A ./simd/taint.c cannam@95: M ./api/fftw3.h -2 +2 cannam@95: M ./kernel/align.c -15 +1 cannam@95: M ./kernel/ifftw.h -2 +2 cannam@95: M ./simd/Makefile.am -1 +1 cannam@95: M ./simd/simd.h -2 +16 cannam@95: M ./simd/taint.c +33 cannam@95: cannam@95: Fri Apr 4 19:16:32 EST 2003 stevenj cannam@95: * [project @ 2003-04-05 00:16:32 by stevenj] cannam@95: added NO_SIMD problem flag, made UNALIGNED an API issue (taints input pointers) cannam@95: cannam@95: M ./api/api.h +2 cannam@95: M ./api/fftw3.h -1 +3 cannam@95: M ./api/mapflags.c -1 +1 cannam@95: M ./api/plan-guru-dft-c2r.c -1 +3 cannam@95: M ./api/plan-guru-dft-r2c.c -1 +3 cannam@95: M ./api/plan-guru-dft.c -1 +4 cannam@95: M ./api/plan-guru-r2r.c -1 +2 cannam@95: M ./api/plan-many-dft-c2r.c -1 +3 cannam@95: M ./api/plan-many-dft-r2c.c -1 +3 cannam@95: M ./api/plan-many-dft.c -1 +4 cannam@95: M ./api/plan-many-r2r.c -1 +2 cannam@95: M ./dft/k7/k7.c -5 +13 cannam@95: M ./dft/simd/n1b.c -1 +1 cannam@95: M ./dft/simd/n1f.c -1 +1 cannam@95: M ./dft/simd/n2b.c +1 cannam@95: M ./dft/simd/n2f.c +1 cannam@95: M ./dft/simd/q1b.c -1 +1 cannam@95: M ./dft/simd/q1f.c -1 +1 cannam@95: M ./dft/simd/t1b.c -1 +1 cannam@95: M ./dft/simd/t1f.c -1 +1 cannam@95: M ./kernel/ifftw.h -3 +3 cannam@95: M ./tests/bench.c +1 cannam@95: cannam@95: Fri Apr 4 18:14:14 EST 2003 stevenj cannam@95: * [project @ 2003-04-04 23:14:14 by stevenj] cannam@95: bugfix in buffered: wrong pointers passed for cldrest; also use TAINT instead of UNALIGNED in buffered2 cannam@95: cannam@95: M ./dft/buffered.c -6 +10 cannam@95: M ./rdft/buffered.c -6 +10 cannam@95: M ./rdft/buffered2.c -7 +5 cannam@95: cannam@95: Fri Apr 4 17:19:51 EST 2003 athena cannam@95: * [project @ 2003-04-04 22:19:51 by athena] cannam@95: Reverted previous change, committed accidentally cannam@95: cannam@95: M ./dft/vrank-geq1.c -3 +3 cannam@95: cannam@95: Fri Apr 4 17:18:39 EST 2003 athena cannam@95: * [project @ 2003-04-04 22:18:39 by athena] cannam@95: What was I thinking? cannam@95: cannam@95: M ./kernel/align.c -1 +2 cannam@95: cannam@95: Fri Apr 4 17:18:21 EST 2003 athena cannam@95: * [project @ 2003-04-04 22:18:21 by athena] cannam@95: cannam@95: M ./dft/vrank-geq1.c -3 +3 cannam@95: cannam@95: Fri Apr 4 16:48:32 EST 2003 stevenj cannam@95: * [project @ 2003-04-04 21:48:32 by stevenj] cannam@95: added --enable-debug-alignment cannam@95: cannam@95: M ./configure.ac +5 cannam@95: M ./libbench2/aligned-main.c -1 +7 cannam@95: cannam@95: Fri Apr 4 16:29:43 EST 2003 stevenj cannam@95: * [project @ 2003-04-04 21:29:43 by stevenj] cannam@95: X(taint) prototype, define corresponding function only if HAVE_SIMD cannam@95: cannam@95: M ./kernel/align.c -1 +3 cannam@95: M ./kernel/ifftw.h -1 +2 cannam@95: cannam@95: Fri Apr 4 16:15:53 EST 2003 athena cannam@95: * [project @ 2003-04-04 21:15:53 by athena] cannam@95: Initial checkin of tained pointers cannam@95: cannam@95: M ./dft/buffered.c -4 +9 cannam@95: M ./dft/solve.c -2 +4 cannam@95: M ./dft/vrank-geq1.c -3 +3 cannam@95: M ./kernel/align.c -6 +4 cannam@95: M ./kernel/ifftw.h -2 +9 cannam@95: M ./rdft/buffered.c -4 +4 cannam@95: M ./rdft/buffered2.c -3 +2 cannam@95: M ./rdft/solve.c -2 +2 cannam@95: M ./rdft/solve2.c -2 +2 cannam@95: M ./rdft/vrank-geq1-rdft2.c -3 +4 cannam@95: M ./rdft/vrank-geq1.c -3 +3 cannam@95: cannam@95: Fri Apr 4 13:12:58 EST 2003 athena cannam@95: * [project @ 2003-04-04 18:12:52 by athena] cannam@95: More conservative preservation of alignment cannam@95: cannam@95: M ./dft/buffered.c -1 +2 cannam@95: M ./dft/rader.c -4 +3 cannam@95: M ./dft/simd/n2b.c -1 cannam@95: M ./dft/simd/n2f.c -1 cannam@95: M ./dft/vrank-geq1.c -6 +4 cannam@95: M ./kernel/align.c -35 +8 cannam@95: M ./kernel/ifftw.h -3 +2 cannam@95: M ./rdft/buffered.c -1 +2 cannam@95: M ./rdft/buffered2.c -1 +4 cannam@95: M ./rdft/dht-rader.c -2 +2 cannam@95: M ./rdft/vrank-geq1-rdft2.c -5 +3 cannam@95: M ./rdft/vrank-geq1.c -5 +5 cannam@95: cannam@95: Thu Apr 3 23:16:27 EST 2003 stevenj cannam@95: * [project @ 2003-04-04 04:16:27 by stevenj] cannam@95: plan/execute with aligned stack cannam@95: cannam@95: M ./api/apiplan.c -19 +24 cannam@95: M ./api/execute-dft-c2r.c -2 +2 cannam@95: M ./api/execute-dft-r2c.c -2 +2 cannam@95: M ./api/execute-dft.c -2 +2 cannam@95: M ./api/execute-r2r.c -2 +2 cannam@95: M ./api/execute.c -2 +2 cannam@95: M ./api/f77funcs.h -10 +10 cannam@95: cannam@95: Thu Apr 3 15:40:01 EST 2003 stevenj cannam@95: * [project @ 2003-04-03 20:40:01 by stevenj] cannam@95: whoops, missed FFTW_MEASURE in fftw3.f cannam@95: cannam@95: M ./api/Makefile.am -1 +1 cannam@95: cannam@95: Thu Apr 3 13:44:46 EST 2003 stevenj cannam@95: * [project @ 2003-04-03 18:44:46 by stevenj] cannam@95: use WITH_ALIGNED_STACK for experimental semaphore stuff, too cannam@95: cannam@95: M ./threads/threads.c -8 +3 cannam@95: cannam@95: Thu Apr 3 09:04:23 EST 2003 athena cannam@95: * [project @ 2003-04-03 14:04:23 by athena] cannam@95: Removed old file cannam@95: cannam@95: R ./kernel/stack.c cannam@95: cannam@95: Thu Apr 3 07:50:43 EST 2003 athena cannam@95: * [project @ 2003-04-03 12:50:43 by athena] cannam@95: Improved stack-alignment hack cannam@95: cannam@95: M ./kernel/Makefile.am -3 +3 cannam@95: M ./kernel/ifftw.h -4 +52 cannam@95: M ./threads/ct-dit.c -7 +6 cannam@95: M ./threads/dft-vrank-geq1.c -7 +6 cannam@95: M ./threads/hc2hc-dif.c -7 +6 cannam@95: M ./threads/hc2hc-dit.c -7 +6 cannam@95: M ./threads/rdft-vrank-geq1.c -6 +4 cannam@95: M ./threads/threads.h -5 cannam@95: M ./threads/vrank-geq1-rdft2.c -7 +6 cannam@95: cannam@95: Thu Apr 3 02:37:57 EST 2003 stevenj cannam@95: * [project @ 2003-04-03 07:37:57 by stevenj] cannam@95: use aligned stack for experimental semaphores, too cannam@95: cannam@95: M ./threads/threads.c -2 +8 cannam@95: cannam@95: Thu Apr 3 02:17:58 EST 2003 stevenj cannam@95: * [project @ 2003-04-03 07:17:58 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./kernel/ifftw.h -2 +4 cannam@95: M ./kernel/stack.c -2 +2 cannam@95: M ./threads/ct-dit.c -2 +4 cannam@95: M ./threads/dft-vrank-geq1.c -2 +4 cannam@95: M ./threads/hc2hc-dif.c -2 +4 cannam@95: M ./threads/hc2hc-dit.c -2 +4 cannam@95: M ./threads/rdft-vrank-geq1.c -2 +4 cannam@95: M ./threads/threads.c -24 +3 cannam@95: M ./threads/threads.h +5 cannam@95: M ./threads/vrank-geq1-rdft2.c -2 +4 cannam@95: cannam@95: Thu Apr 3 01:58:32 EST 2003 stevenj cannam@95: * [project @ 2003-04-03 06:58:32 by stevenj] cannam@95: fix(?) for SIMD thread problems cannam@95: cannam@95: M ./kernel/ifftw.h -2 +2 cannam@95: M ./kernel/stack.c -3 +3 cannam@95: M ./threads/ct-dit.c -7 +10 cannam@95: M ./threads/hc2hc-dif.c -2 +2 cannam@95: M ./threads/hc2hc-dit.c -2 +2 cannam@95: M ./threads/threads.c -6 +27 cannam@95: cannam@95: Wed Apr 2 20:33:12 EST 2003 stevenj cannam@95: * [project @ 2003-04-03 01:33:12 by stevenj] cannam@95: noted n=1 REDFT01 case cannam@95: cannam@95: M ./doc/fftw3.texi -1 +11 cannam@95: cannam@95: Wed Apr 2 20:32:07 EST 2003 stevenj cannam@95: * [project @ 2003-04-03 01:32:07 by stevenj] cannam@95: note about n=2 REDFT00 formula cannam@95: cannam@95: M ./doc/fftw3.texi -2 +3 cannam@95: cannam@95: Wed Apr 2 20:30:10 EST 2003 stevenj cannam@95: * [project @ 2003-04-03 01:30:10 by stevenj] cannam@95: note about undefined REDFT00 cannam@95: cannam@95: M ./doc/fftw3.texi -1 +2 cannam@95: cannam@95: Wed Apr 2 20:18:03 EST 2003 stevenj cannam@95: * [project @ 2003-04-03 01:18:03 by stevenj] cannam@95: noted n=1 RODFT01 case cannam@95: cannam@95: M ./doc/fftw3.texi -2 +12 cannam@95: cannam@95: Wed Apr 2 20:14:07 EST 2003 stevenj cannam@95: * [project @ 2003-04-03 01:14:07 by stevenj] cannam@95: corrected definitions cannam@95: cannam@95: M ./doc/equation-redft11.png cannam@95: M ./doc/equation-rodft01.png cannam@95: M ./doc/equation-rodft11.png cannam@95: M ./doc/fftw3.texi -6 +6 cannam@95: cannam@95: Wed Apr 2 19:43:59 EST 2003 stevenj cannam@95: * [project @ 2003-04-03 00:43:59 by stevenj] cannam@95: added REODFT_KINDP, fixed nontrivial test for R2HC11 and HC2R11 (not that we support these yet anyway) cannam@95: cannam@95: M ./rdft/codelet-rdft.h +1 cannam@95: M ./rdft/problem.c -3 +3 cannam@95: M ./rdft/vrank-geq1.c -3 +3 cannam@95: cannam@95: Wed Apr 2 19:16:54 EST 2003 stevenj cannam@95: * [project @ 2003-04-03 00:16:53 by stevenj] cannam@95: size 2 hc2r and dht are equivalent to r2hc cannam@95: cannam@95: M ./rdft/codelets/hc2r/Makefile.am -3 +5 cannam@95: M ./rdft/problem.c -3 +5 cannam@95: cannam@95: Wed Apr 2 15:09:08 EST 2003 stevenj cannam@95: * [project @ 2003-04-02 20:09:08 by stevenj] cannam@95: noted overwriting in upgrading section cannam@95: cannam@95: M ./doc/fftw3.texi -1 +4 cannam@95: cannam@95: Wed Apr 2 05:25:56 EST 2003 athena cannam@95: * [project @ 2003-04-02 10:25:56 by athena] cannam@95: Moved with_aligned_stack to its own file cannam@95: cannam@95: A ./kernel/stack.c cannam@95: M ./kernel/Makefile.am -3 +3 cannam@95: M ./kernel/align.c -46 +1 cannam@95: M ./kernel/stack.c +67 cannam@95: cannam@95: Tue Apr 1 21:11:31 EST 2003 athena cannam@95: * [project @ 2003-04-02 02:11:31 by athena] cannam@95: Fixed comments cannam@95: cannam@95: M ./kernel/align.c -6 +2 cannam@95: M ./libbench2/aligned-main.c -6 +2 cannam@95: cannam@95: Tue Apr 1 20:57:39 EST 2003 athena cannam@95: * [project @ 2003-04-02 01:57:39 by athena] cannam@95: Alignment hacks cannam@95: cannam@95: M ./kernel/align.c -1 +50 cannam@95: M ./kernel/ifftw.h -2 +3 cannam@95: M ./libbench2/aligned-main.c -26 +14 cannam@95: M ./libbench2/bench-main.c -1 +9 cannam@95: cannam@95: Tue Apr 1 14:26:48 EST 2003 stevenj cannam@95: * [project @ 2003-04-01 19:26:48 by stevenj] cannam@95: phew, no, previous version was okay cannam@95: cannam@95: M ./threads/threads.c -1 +1 cannam@95: cannam@95: Tue Apr 1 14:26:15 EST 2003 stevenj cannam@95: * [project @ 2003-04-01 19:26:15 by stevenj] cannam@95: whoops, crap cannam@95: cannam@95: M ./threads/threads.c -2 +2 cannam@95: cannam@95: Tue Apr 1 08:01:06 EST 2003 athena cannam@95: * [project @ 2003-04-01 13:01:06 by athena] cannam@95: support sse2 in forthcoming gcc-3.3 cannam@95: cannam@95: M ./simd/simd-sse2.h -3 +11 cannam@95: cannam@95: Tue Apr 1 01:17:15 EST 2003 stevenj cannam@95: * [project @ 2003-04-01 06:17:15 by stevenj] cannam@95: comment cannam@95: cannam@95: M ./kernel/cycle.h -2 +2 cannam@95: cannam@95: Tue Apr 1 01:16:46 EST 2003 stevenj cannam@95: * [project @ 2003-04-01 06:16:46 by stevenj] cannam@95: noted ac_check_headers cannam@95: cannam@95: M ./kernel/cycle.h -1 +4 cannam@95: cannam@95: Tue Apr 1 01:11:31 EST 2003 stevenj cannam@95: * [project @ 2003-04-01 06:11:31 by stevenj] cannam@95: comment cannam@95: cannam@95: M ./kernel/cycle.h -1 +2 cannam@95: cannam@95: Tue Apr 1 01:06:53 EST 2003 stevenj cannam@95: * [project @ 2003-04-01 06:06:53 by stevenj] cannam@95: documented autoconf tests, so that cycle.h can be distributed separately cannam@95: cannam@95: M ./kernel/cycle.h -1 +27 cannam@95: cannam@95: Sat Aug 12 23:11:17 EDT 2006 Unknown tagger cannam@95: tagged fftw-3-0-beta2 cannam@95: cannam@95: cannam@95: Mon Mar 31 22:12:02 EST 2003 stevenj cannam@95: * [project @ 2003-04-01 03:12:02 by stevenj] cannam@95: IRIX is all-caps cannam@95: cannam@95: M ./NEWS -1 +1 cannam@95: cannam@95: Mon Mar 31 22:11:42 EST 2003 stevenj cannam@95: * [project @ 2003-04-01 03:11:42 by stevenj] cannam@95: noted Irix fix cannam@95: cannam@95: M ./NEWS +3 cannam@95: cannam@95: Mon Mar 31 22:10:33 EST 2003 stevenj cannam@95: * [project @ 2003-04-01 03:10:33 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./threads/api.c -1 +1 cannam@95: M ./threads/threads.h -1 +1 cannam@95: cannam@95: Mon Mar 31 22:04:35 EST 2003 stevenj cannam@95: * [project @ 2003-04-01 03:04:35 by stevenj] cannam@95: use ithreads_init so as not to confuse fftw 2 users cannam@95: cannam@95: M ./threads/threads.c -6 +8 cannam@95: cannam@95: Mon Mar 31 22:00:42 EST 2003 stevenj cannam@95: * [project @ 2003-04-01 03:00:42 by stevenj] cannam@95: IRIX lossage cannam@95: cannam@95: M ./threads/threads.c -1 +5 cannam@95: cannam@95: Mon Mar 31 21:19:20 EST 2003 stevenj cannam@95: * [project @ 2003-04-01 02:19:20 by stevenj] cannam@95: check for -openmp (icc) among the OpenMP flags (TODO: make this a cannam@95: separate macro, with a loop instead of repeated checks) cannam@95: cannam@95: M ./configure.ac +5 cannam@95: cannam@95: Mon Mar 31 17:12:19 EST 2003 stevenj cannam@95: * [project @ 2003-03-31 22:12:19 by stevenj] cannam@95: clarification cannam@95: cannam@95: M ./doc/fftw3.texi -1 +4 cannam@95: cannam@95: Mon Mar 31 17:01:16 EST 2003 athena cannam@95: * [project @ 2003-03-31 22:01:16 by athena] cannam@95: More liberal test for solaris CC cannam@95: cannam@95: M ./acinclude.m4 -1 +1 cannam@95: cannam@95: Mon Mar 31 15:13:33 EST 2003 athena cannam@95: * [project @ 2003-03-31 20:13:33 by athena] cannam@95: Allow x86-64 simd cannam@95: cannam@95: M ./simd/simd-sse.h -1 +1 cannam@95: M ./simd/simd-sse2.h -1 +1 cannam@95: cannam@95: Mon Mar 31 15:13:21 EST 2003 athena cannam@95: * [project @ 2003-03-31 20:13:21 by athena] cannam@95: Added x86-64 timer code cannam@95: cannam@95: M ./kernel/cycle.h -1 +22 cannam@95: cannam@95: Mon Mar 31 13:10:54 EST 2003 stevenj cannam@95: * [project @ 2003-03-31 18:10:54 by stevenj] cannam@95: updated cannam@95: cannam@95: M ./NEWS +4 cannam@95: cannam@95: Mon Mar 31 13:07:19 EST 2003 stevenj cannam@95: * [project @ 2003-03-31 18:07:19 by stevenj] cannam@95: updated cannam@95: cannam@95: M ./ChangeLog -1378 +3497 cannam@95: cannam@95: Mon Mar 31 13:05:27 EST 2003 stevenj cannam@95: * [project @ 2003-03-31 18:05:27 by stevenj] cannam@95: colon cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -1 +1 cannam@95: cannam@95: Mon Mar 31 07:20:20 EST 2003 athena cannam@95: * [project @ 2003-03-31 12:20:20 by athena] cannam@95: Reorganized compiler bugs section (which is growing out of control) cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -11 +18 cannam@95: cannam@95: Mon Mar 31 07:15:20 EST 2003 athena cannam@95: * [project @ 2003-03-31 12:15:20 by athena] cannam@95: solaris gcc bug appears to be also in 2.95.2 cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -1 +1 cannam@95: cannam@95: Mon Mar 31 07:13:45 EST 2003 athena cannam@95: * [project @ 2003-03-31 12:13:45 by athena] cannam@95: Workaround works---there is another gcc/sparc bug elsehwere cannam@95: cannam@95: M ./kernel/planner.c -3 +3 cannam@95: cannam@95: Mon Mar 31 07:08:56 EST 2003 athena cannam@95: * [project @ 2003-03-31 12:08:56 by athena] cannam@95: Grrr, workaround does not work. cannam@95: cannam@95: M ./kernel/planner.c -3 +3 cannam@95: cannam@95: Mon Mar 31 07:02:23 EST 2003 athena cannam@95: * [project @ 2003-03-31 12:02:23 by athena] cannam@95: ADDMOD is now function, which seems to avoid gcc bugs. cannam@95: cannam@95: M ./kernel/planner.c -10 +16 cannam@95: cannam@95: Sun Mar 30 16:40:26 EST 2003 athena cannam@95: * [project @ 2003-03-30 21:40:26 by athena] cannam@95: Workaround sparc gcc bug cannam@95: cannam@95: M ./kernel/planner.c -3 +5 cannam@95: cannam@95: Sun Mar 30 15:51:59 EST 2003 stevenj cannam@95: * [project @ 2003-03-30 20:50:59 by stevenj] cannam@95: note cannam@95: cannam@95: M ./doc/fftw3.texi -4 +5 cannam@95: cannam@95: Sun Mar 30 15:34:57 EST 2003 stevenj cannam@95: * [project @ 2003-03-30 20:34:57 by stevenj] cannam@95: make non-square UGLY, for now cannam@95: cannam@95: M ./dft/vrank2-transpose.c -4 +4 cannam@95: M ./dft/vrank3-transpose.c -1 +4 cannam@95: cannam@95: Sun Mar 30 15:33:57 EST 2003 stevenj cannam@95: * [project @ 2003-03-30 20:33:57 by stevenj] cannam@95: added -o amnesia to forget_wisdom before each plan cannam@95: cannam@95: M ./tests/bench.c +5 cannam@95: cannam@95: Sun Mar 30 09:41:27 EST 2003 athena cannam@95: * [project @ 2003-03-30 14:41:27 by athena] cannam@95: Report setup time in benchmark cannam@95: cannam@95: M ./libbench2/bench-user.h -1 +4 cannam@95: M ./libbench2/report.c -2 +2 cannam@95: M ./libbench2/speed.c -1 +4 cannam@95: cannam@95: Sat Mar 29 20:21:15 EST 2003 stevenj cannam@95: * [project @ 2003-03-30 01:20:52 by stevenj] cannam@95: comment cannam@95: cannam@95: M ./kernel/transpose.c -1 +2 cannam@95: cannam@95: Sat Mar 29 19:11:10 EST 2003 stevenj cannam@95: * [project @ 2003-03-30 00:11:10 by stevenj] cannam@95: slight change cannam@95: cannam@95: M ./doc/fftw3.texi -4 +4 cannam@95: cannam@95: Sat Mar 29 18:46:16 EST 2003 athena cannam@95: * [project @ 2003-03-29 23:46:16 by athena] cannam@95: More relaxed definition of UGLYness cannam@95: cannam@95: M ./kernel/ct.c -1 +1 cannam@95: cannam@95: Sat Mar 29 15:28:01 EST 2003 stevenj cannam@95: * [project @ 2003-03-29 20:28:01 by stevenj] cannam@95: no more cvs id strings in header files...I'm tired of having to rebuild everything after a commit cannam@95: cannam@95: M ./rdft/codelet-rdft.h -2 cannam@95: M ./rdft/hc2hc.h -2 cannam@95: M ./rdft/rdft.h -2 cannam@95: M ./reodft/reodft.h -2 cannam@95: M ./threads/threads.h -2 cannam@95: cannam@95: Sat Mar 29 15:22:28 EST 2003 stevenj cannam@95: * [project @ 2003-03-29 20:22:28 by stevenj] cannam@95: rdft2 stride unification cannam@95: cannam@95: A ./rdft/rdft2-strides.c cannam@95: M ./rdft/Makefile.am -1 +1 cannam@95: M ./rdft/buffered2.c -20 +11 cannam@95: M ./rdft/direct2.c -5 +5 cannam@95: M ./rdft/rdft.h -1 +2 cannam@95: M ./rdft/rdft2-inplace-strides.c -9 +2 cannam@95: M ./rdft/rdft2-strides.c +38 cannam@95: M ./rdft/rdft2-tensor-max-index.c -7 +4 cannam@95: M ./rdft/vrank-geq1-rdft2.c -26 +9 cannam@95: M ./threads/vrank-geq1-rdft2.c -9 +3 cannam@95: cannam@95: Sat Mar 29 14:38:23 EST 2003 stevenj cannam@95: * [project @ 2003-03-29 19:38:23 by stevenj] cannam@95: preserve in-place-ness cannam@95: cannam@95: M ./rdft/vrank-geq1-rdft2.c -3 +3 cannam@95: cannam@95: Sat Mar 29 14:23:31 EST 2003 stevenj cannam@95: * [project @ 2003-03-29 19:23:31 by stevenj] cannam@95: make nowisdom the default cannam@95: cannam@95: M ./tests/Makefile.am -7 +7 cannam@95: M ./tests/bench.c -2 +2 cannam@95: M ./tests/check.pl -3 +3 cannam@95: cannam@95: Sat Mar 29 14:13:18 EST 2003 athena cannam@95: * [project @ 2003-03-29 19:13:18 by athena] cannam@95: --verbose in paranoid-check produces too much output. Make it quiet. cannam@95: cannam@95: M ./tests/Makefile.am -1 +1 cannam@95: cannam@95: Sat Mar 29 13:45:13 EST 2003 stevenj cannam@95: * [project @ 2003-03-29 18:45:13 by stevenj] cannam@95: fixed transpose bugs...need to check ri-ii before deciding whether Ntuple fits cannam@95: cannam@95: M ./dft/vrank2-transpose.c -5 +9 cannam@95: M ./dft/vrank3-transpose.c -7 +13 cannam@95: M ./kernel/ifftw.h -3 +5 cannam@95: M ./kernel/transpose.c -11 +21 cannam@95: cannam@95: Sat Mar 29 08:10:40 EST 2003 athena cannam@95: * [project @ 2003-03-29 13:10:40 by athena] cannam@95: try more 2^k cannam@95: cannam@95: M ./tests/check.pl +2 cannam@95: cannam@95: Sat Mar 29 08:05:41 EST 2003 athena cannam@95: * [project @ 2003-03-29 13:05:41 by athena] cannam@95: MIN_ALIGNMENT was defined after being used, causing crash in sse2. cannam@95: cannam@95: M ./kernel/ifftw.h -4 +4 cannam@95: cannam@95: Sat Mar 29 03:07:34 EST 2003 stevenj cannam@95: * [project @ 2003-03-29 08:07:34 by stevenj] cannam@95: real transposes are currently unused, and are not needed for MPI code either cannam@95: cannam@95: M ./kernel/Makefile.am -2 +2 cannam@95: M ./kernel/ifftw.h -2 +1 cannam@95: R ./kernel/tensor10.c cannam@95: M ./kernel/transpose.c -180 +18 cannam@95: M ./rdft/Makefile.am -2 +3 cannam@95: M ./rdft/conf.c -3 +5 cannam@95: cannam@95: Sat Mar 29 02:58:39 EST 2003 stevenj cannam@95: * [project @ 2003-03-29 07:58:39 by stevenj] cannam@95: added general transpose cannam@95: cannam@95: A ./kernel/transpose.c cannam@95: M ./dft/vrank2-transpose.c -9 +43 cannam@95: M ./dft/vrank3-transpose.c -25 +57 cannam@95: M ./kernel/Makefile.am -2 +2 cannam@95: M ./kernel/ifftw.h -1 +13 cannam@95: M ./kernel/transpose.c +523 cannam@95: cannam@95: Fri Mar 28 22:49:04 EST 2003 stevenj cannam@95: * [project @ 2003-03-29 03:49:04 by stevenj] cannam@95: added transposition option cannam@95: cannam@95: M ./libbench2/problem.c -1 +18 cannam@95: cannam@95: Fri Mar 28 22:09:22 EST 2003 stevenj cannam@95: * [project @ 2003-03-29 03:09:22 by stevenj] cannam@95: yikes, fixed incorrect applicability of transpose plans cannam@95: cannam@95: A ./kernel/tensor10.c cannam@95: M ./dft/vrank2-transpose.c -3 +3 cannam@95: M ./dft/vrank3-transpose.c -2 +2 cannam@95: M ./kernel/Makefile.am -2 +2 cannam@95: M ./kernel/ifftw.h -1 +2 cannam@95: M ./kernel/tensor10.c +31 cannam@95: M ./rdft/vrank2-transpose.c -2 +2 cannam@95: M ./rdft/vrank3-transpose.c -2 +2 cannam@95: cannam@95: Fri Mar 28 22:06:14 EST 2003 stevenj cannam@95: * [project @ 2003-03-29 03:06:14 by stevenj] cannam@95: in the future, we might want to allow sz->rnk == 0, vecsz->rnk arbitrary to be converted to r2hc (the apply function already should work for this case)...disabled for now, though cannam@95: cannam@95: M ./rdft/dft-r2hc.c -7 +21 cannam@95: cannam@95: Fri Mar 28 19:12:08 EST 2003 stevenj cannam@95: * [project @ 2003-03-29 00:12:08 by stevenj] cannam@95: use most_unaligned in rdft2 cannam@95: cannam@95: M ./kernel/align.c -6 +1 cannam@95: M ./kernel/ifftw.h -2 +1 cannam@95: M ./rdft/vrank-geq1-rdft2.c -6 +11 cannam@95: cannam@95: Fri Mar 28 19:11:47 EST 2003 stevenj cannam@95: * [project @ 2003-03-29 00:11:47 by stevenj] cannam@95: slight change cannam@95: cannam@95: M ./tests/Makefile.am -2 +2 cannam@95: cannam@95: Fri Mar 28 19:00:21 EST 2003 stevenj cannam@95: * [project @ 2003-03-29 00:00:21 by stevenj] cannam@95: output message when checks pass cannam@95: cannam@95: M ./tests/Makefile.am +12 cannam@95: cannam@95: Fri Mar 28 17:21:47 EST 2003 stevenj cannam@95: * [project @ 2003-03-28 22:21:47 by stevenj] cannam@95: added ifndef alloca around alloca stuff cannam@95: cannam@95: M ./kernel/ifftw.h -1 +3 cannam@95: cannam@95: Fri Mar 28 13:45:50 EST 2003 athena cannam@95: * [project @ 2003-03-28 18:45:43 by athena] cannam@95: Proper alignment in rader cannam@95: cannam@95: M ./dft/rader.c -2 +4 cannam@95: M ./dft/vrank-geq1.c -4 +3 cannam@95: M ./kernel/align.c -3 +2 cannam@95: M ./kernel/ifftw.h -2 +2 cannam@95: M ./rdft/dht-rader.c -2 +5 cannam@95: cannam@95: Fri Mar 28 12:43:23 EST 2003 stevenj cannam@95: * [project @ 2003-03-28 17:43:23 by stevenj] cannam@95: whitespace cannam@95: cannam@95: M ./kernel/ifftw.h -1 +2 cannam@95: cannam@95: Fri Mar 28 12:41:39 EST 2003 stevenj cannam@95: * [project @ 2003-03-28 17:41:39 by stevenj] cannam@95: whoops, alloca stuff inside HAVE_ALLOCA cannam@95: cannam@95: M ./kernel/ifftw.h -3 +4 cannam@95: cannam@95: Fri Mar 28 12:35:21 EST 2003 stevenj cannam@95: * [project @ 2003-03-28 17:35:21 by stevenj] cannam@95: make check can afford to be a little bigger cannam@95: cannam@95: M ./tests/Makefile.am -2 +2 cannam@95: cannam@95: Fri Mar 28 12:31:32 EST 2003 stevenj cannam@95: * [project @ 2003-03-28 17:31:32 by stevenj] cannam@95: use same alloca macrology as configure script cannam@95: cannam@95: M ./kernel/ifftw.h -28 +45 cannam@95: cannam@95: Fri Mar 28 03:05:15 EST 2003 stevenj cannam@95: * [project @ 2003-03-28 08:05:15 by stevenj] cannam@95: fallback is no longer needed for mingw cannam@95: cannam@95: M ./kernel/ifftw.h -6 +1 cannam@95: cannam@95: Fri Mar 28 02:58:45 EST 2003 stevenj cannam@95: * [project @ 2003-03-28 07:58:45 by stevenj] cannam@95: alloca fallback for gcc cannam@95: cannam@95: M ./kernel/ifftw.h -3 +4 cannam@95: cannam@95: Fri Mar 28 02:49:59 EST 2003 stevenj cannam@95: * [project @ 2003-03-28 07:49:59 by stevenj] cannam@95: _alloca was added for MinGW, but it causes problems there cannam@95: cannam@95: M ./configure.ac -1 +1 cannam@95: cannam@95: Thu Mar 27 22:06:07 EST 2003 stevenj cannam@95: * [project @ 2003-03-28 03:06:07 by stevenj] cannam@95: fixed most_unaligned for split format cannam@95: cannam@95: M ./kernel/align.c -5 +8 cannam@95: cannam@95: Thu Mar 27 19:01:58 EST 2003 stevenj cannam@95: * [project @ 2003-03-28 00:01:58 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./Makefile.am -1 +1 cannam@95: cannam@95: Thu Mar 27 19:00:20 EST 2003 stevenj cannam@95: * [project @ 2003-03-28 00:00:20 by stevenj] cannam@95: added pkg-config cannam@95: cannam@95: A ./fftw.pc.in cannam@95: M ./Makefile.am -1 +6 cannam@95: M ./configure.ac +1 cannam@95: M ./fftw.pc.in +10 cannam@95: cannam@95: Thu Mar 27 15:59:01 EST 2003 stevenj cannam@95: * [project @ 2003-03-27 20:59:01 by stevenj] cannam@95: fixed asserts cannam@95: cannam@95: M ./dft/vrank-geq1.c -2 +2 cannam@95: M ./rdft/vrank-geq1-rdft2.c -2 +3 cannam@95: M ./rdft/vrank-geq1.c -2 +2 cannam@95: cannam@95: Thu Mar 27 15:49:53 EST 2003 athena cannam@95: * [project @ 2003-03-27 20:49:53 by athena] cannam@95: Do not adjust r/i pointers separately. cannam@95: cannam@95: M ./kernel/align.c -4 +5 cannam@95: cannam@95: Thu Mar 27 15:17:40 EST 2003 athena cannam@95: * [project @ 2003-03-27 20:17:40 by athena] cannam@95: iForgot to add files cannam@95: cannam@95: A ./dft/simd/n2b.h cannam@95: A ./dft/simd/n2f.h cannam@95: M ./dft/simd/n2b.h +25 cannam@95: M ./dft/simd/n2f.h +25 cannam@95: cannam@95: Thu Mar 27 15:10:41 EST 2003 athena cannam@95: * [project @ 2003-03-27 20:10:36 by athena] cannam@95: Specialized n simd codelets for unit vector stride. cannam@95: cannam@95: A ./dft/simd/n2b.c cannam@95: A ./dft/simd/n2f.c cannam@95: M ./configure.ac +4 cannam@95: M ./dft/simd/Makefile.am -2 +2 cannam@95: M ./dft/simd/codelets/Makefile.am -8 +54 cannam@95: M ./dft/simd/n1b.c -3 +3 cannam@95: M ./dft/simd/n1b.h +4 cannam@95: M ./dft/simd/n1f.c -3 +3 cannam@95: M ./dft/simd/n1f.h +4 cannam@95: M ./dft/simd/n2b.c +49 cannam@95: M ./dft/simd/n2f.c +49 cannam@95: cannam@95: Thu Mar 27 08:22:03 EST 2003 athena cannam@95: * [project @ 2003-03-27 13:22:03 by athena] cannam@95: Changed version number to beta2 cannam@95: cannam@95: M ./configure.ac -1 +1 cannam@95: cannam@95: Thu Mar 27 06:37:07 EST 2003 athena cannam@95: * [project @ 2003-03-27 11:37:07 by athena] cannam@95: Changed alignment requirements for n1 simd codelets. Changed cannam@95: mechanism for detecting lack of alignment. cannam@95: cannam@95: M ./api/mapflags.c -1 +1 cannam@95: M ./dft/simd/n1b.c -4 +4 cannam@95: M ./dft/simd/n1b.h -3 cannam@95: M ./dft/simd/n1f.c -4 +4 cannam@95: M ./dft/simd/n1f.h -3 cannam@95: M ./dft/simd/q1b.c -1 +1 cannam@95: M ./dft/simd/q1f.c -1 +1 cannam@95: M ./dft/simd/t1b.c -1 +1 cannam@95: M ./dft/simd/t1f.c -1 +1 cannam@95: M ./dft/vrank-geq1.c -7 +7 cannam@95: M ./kernel/align.c -5 +42 cannam@95: M ./kernel/ifftw.h -3 +5 cannam@95: M ./rdft/vrank-geq1-rdft2.c -4 +7 cannam@95: M ./rdft/vrank-geq1.c -5 +5 cannam@95: cannam@95: Thu Mar 27 04:25:06 EST 2003 athena cannam@95: * [project @ 2003-03-27 09:25:06 by athena] cannam@95: Oops, wrong place for hook cannam@95: cannam@95: M ./tests/bench.c -2 +1 cannam@95: cannam@95: Thu Mar 27 02:37:52 EST 2003 stevenj cannam@95: * [project @ 2003-03-27 07:37:52 by stevenj] cannam@95: added comments to codelet makefiles, to aid people wanting to generate their own code cannam@95: cannam@95: M ./dft/codelets/inplace/Makefile.am -1 +23 cannam@95: M ./dft/codelets/standard/Makefile.am +24 cannam@95: M ./dft/k7/codelets/Makefile.am +25 cannam@95: M ./dft/simd/codelets/Makefile.am -1 +32 cannam@95: M ./rdft/codelets/hc2r/Makefile.am +25 cannam@95: M ./rdft/codelets/r2hc/Makefile.am +25 cannam@95: M ./rdft/codelets/r2r/Makefile.am +18 cannam@95: cannam@95: Thu Mar 27 01:42:27 EST 2003 stevenj cannam@95: * [project @ 2003-03-27 06:42:27 by stevenj] cannam@95: Matteo is also a copyright holder cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -1 +1 cannam@95: cannam@95: Thu Mar 27 01:41:08 EST 2003 stevenj cannam@95: * [project @ 2003-03-27 06:41:08 by stevenj] cannam@95: FORTRAN is officially Fortran, these days cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -2 +2 cannam@95: cannam@95: Thu Mar 27 01:40:32 EST 2003 stevenj cannam@95: * [project @ 2003-03-27 06:40:32 by stevenj] cannam@95: punctuation cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -1 +1 cannam@95: cannam@95: Thu Mar 27 01:40:14 EST 2003 stevenj cannam@95: * [project @ 2003-03-27 06:40:14 by stevenj] cannam@95: don't use "wrapper" cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -1 +1 cannam@95: cannam@95: Thu Mar 27 01:37:53 EST 2003 stevenj cannam@95: * [project @ 2003-03-27 06:37:53 by stevenj] cannam@95: plural cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -1 +1 cannam@95: cannam@95: Thu Mar 27 01:35:32 EST 2003 stevenj cannam@95: * [project @ 2003-03-27 06:35:32 by stevenj] cannam@95: grammar cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -1 +1 cannam@95: cannam@95: Thu Mar 27 01:33:35 EST 2003 stevenj cannam@95: * [project @ 2003-03-27 06:33:35 by stevenj] cannam@95: better phrasing cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -1 +1 cannam@95: cannam@95: Wed Mar 26 22:47:58 EST 2003 stevenj cannam@95: * [project @ 2003-03-27 03:47:58 by stevenj] cannam@95: stddef.h should not be needed anymore for this file cannam@95: cannam@95: M ./kernel/align.c -2 +1 cannam@95: cannam@95: Wed Mar 26 22:13:48 EST 2003 stevenj cannam@95: * [project @ 2003-03-27 03:13:48 by stevenj] cannam@95: added comments for Franz mode cannam@95: cannam@95: M ./dft/codelets/standard/Makefile.am -5 +5 cannam@95: cannam@95: Wed Mar 26 22:11:58 EST 2003 stevenj cannam@95: * [project @ 2003-03-27 03:10:11 by stevenj] cannam@95: clarification cannam@95: cannam@95: M ./dft/simd/codelets/Makefile.am -3 +4 cannam@95: cannam@95: Wed Mar 26 22:08:51 EST 2003 stevenj cannam@95: * [project @ 2003-03-27 03:08:51 by stevenj] cannam@95: commented on FRANZ codelets cannam@95: cannam@95: M ./dft/simd/codelets/Makefile.am -5 +5 cannam@95: cannam@95: Wed Mar 26 22:06:45 EST 2003 stevenj cannam@95: * [project @ 2003-03-27 03:06:45 by stevenj] cannam@95: updated cannam@95: cannam@95: M ./NEWS +23 cannam@95: cannam@95: Wed Mar 26 21:51:15 EST 2003 stevenj cannam@95: * [project @ 2003-03-27 02:51:15 by stevenj] cannam@95: disable DIF codelets, since they are never used (apparently) except cannam@95: for some non-power-of-two sizes...improve support for the latter by cannam@95: adding size 3, 5, and 6 q^2 codelets. cannam@95: cannam@95: M ./dft/codelets/inplace/Makefile.am -4 +8 cannam@95: cannam@95: Wed Mar 26 20:07:11 EST 2003 stevenj cannam@95: * [project @ 2003-03-27 01:07:11 by stevenj] cannam@95: DHT has no forward/backward cannam@95: cannam@95: M ./doc/fftw3.texi -3 +3 cannam@95: cannam@95: Wed Mar 26 19:46:12 EST 2003 fftw cannam@95: * [project @ 2003-03-27 00:46:12 by fftw] cannam@95: added hacky way to use an arbitrary flag cannam@95: cannam@95: M ./tests/bench.c +1 cannam@95: cannam@95: Wed Mar 26 19:44:31 EST 2003 athena cannam@95: * [project @ 2003-03-27 00:44:31 by athena] cannam@95: Better place to install hook cannam@95: cannam@95: M ./tests/bench.c -2 +2 cannam@95: cannam@95: Wed Mar 26 19:40:28 EST 2003 stevenj cannam@95: * [project @ 2003-03-27 00:40:28 by stevenj] cannam@95: noted that the user should run make check if they think FFTW has a bug cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -6 +8 cannam@95: cannam@95: Wed Mar 26 17:31:16 EST 2003 athena cannam@95: * [project @ 2003-03-26 22:31:16 by athena] cannam@95: Oops, what am I thinking cannam@95: cannam@95: M ./kernel/planner.c -5 +8 cannam@95: cannam@95: Wed Mar 26 17:23:56 EST 2003 athena cannam@95: * [project @ 2003-03-26 22:23:56 by athena] cannam@95: Grrr.... fixed bug in estimator cannam@95: cannam@95: M ./kernel/planner.c -8 +5 cannam@95: cannam@95: Wed Mar 26 17:16:19 EST 2003 athena cannam@95: * [project @ 2003-03-26 22:16:19 by athena] cannam@95: Oops---the flop count was right. The estimator is broken elsewhere. cannam@95: cannam@95: M ./genfft/c.ml -5 +2 cannam@95: cannam@95: Wed Mar 26 14:28:41 EST 2003 athena cannam@95: * [project @ 2003-03-26 19:28:41 by athena] cannam@95: Fixed SIMD estimator cannam@95: cannam@95: M ./genfft/c.ml -2 +5 cannam@95: cannam@95: Wed Mar 26 07:45:03 EST 2003 athena cannam@95: * [project @ 2003-03-26 12:45:03 by athena] cannam@95: Added twidsq simd codelets cannam@95: cannam@95: A ./dft/simd/q1b.c cannam@95: A ./dft/simd/q1b.h cannam@95: A ./dft/simd/q1f.c cannam@95: A ./dft/simd/q1f.h cannam@95: A ./genfft/gen_twidsq_c.ml cannam@95: M ./dft/simd/Makefile.am -1 +2 cannam@95: M ./dft/simd/codelets/Makefile.am -1 +12 cannam@95: M ./dft/simd/q1b.c +44 cannam@95: M ./dft/simd/q1b.h +25 cannam@95: M ./dft/simd/q1f.c +44 cannam@95: M ./dft/simd/q1f.h +25 cannam@95: M ./dft/simd/t1b.c +1 cannam@95: M ./dft/simd/t1f.c +1 cannam@95: M ./genfft/Makefile.am -8 +14 cannam@95: M ./genfft/gen_twiddle_c.ml -3 +3 cannam@95: M ./genfft/gen_twidsq_c.ml +195 cannam@95: M ./support/Makefile.codelets -1 +2 cannam@95: cannam@95: Tue Mar 25 23:33:03 EST 2003 stevenj cannam@95: * [project @ 2003-03-26 04:33:03 by stevenj] cannam@95: gensrc -> genfft cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Tue Mar 25 23:32:16 EST 2003 stevenj cannam@95: * [project @ 2003-03-26 04:32:16 by stevenj] cannam@95: newline cannam@95: cannam@95: M ./TODO +1 cannam@95: cannam@95: Tue Mar 25 19:17:08 EST 2003 athena cannam@95: * [project @ 2003-03-26 00:17:08 by athena] cannam@95: Noted need to add dif simd codelets cannam@95: cannam@95: M ./TODO +1 cannam@95: cannam@95: Tue Mar 25 13:03:47 EST 2003 stevenj cannam@95: * [project @ 2003-03-25 18:03:47 by stevenj] cannam@95: noted shift cannam@95: cannam@95: M ./doc/fftw3.texi -3 +2 cannam@95: cannam@95: Tue Mar 25 13:02:47 EST 2003 stevenj cannam@95: * [project @ 2003-03-25 18:02:47 by stevenj] cannam@95: clarification cannam@95: cannam@95: M ./doc/fftw3.texi -4 +14 cannam@95: cannam@95: Tue Mar 25 12:46:44 EST 2003 stevenj cannam@95: * [project @ 2003-03-25 17:46:44 by stevenj] cannam@95: need make after bootstrap cannam@95: cannam@95: M ./doc/fftw3.texi -2 +3 cannam@95: cannam@95: Tue Mar 25 12:31:49 EST 2003 stevenj cannam@95: * [project @ 2003-03-25 17:31:49 by stevenj] cannam@95: slight change cannam@95: cannam@95: M ./doc/fftw3.texi -3 +3 cannam@95: cannam@95: Tue Mar 25 12:30:56 EST 2003 stevenj cannam@95: * [project @ 2003-03-25 17:30:56 by stevenj] cannam@95: libtool is also needed cannam@95: cannam@95: M ./doc/fftw3.texi -4 +4 cannam@95: cannam@95: Tue Mar 25 12:29:52 EST 2003 stevenj cannam@95: * [project @ 2003-03-25 17:29:52 by stevenj] cannam@95: added code generator introduction cannam@95: cannam@95: M ./doc/fftw3.texi -13 +63 cannam@95: cannam@95: Tue Mar 25 11:51:49 EST 2003 stevenj cannam@95: * [project @ 2003-03-25 16:51:49 by stevenj] cannam@95: added support for REDFT/RODFT/DHT direct codelets cannam@95: cannam@95: A ./genfft/gen_r2r.ml cannam@95: A ./rdft/kr2r.c cannam@95: A ./rdft/codelets/r2r/ cannam@95: A ./rdft/codelets/r2r.c cannam@95: A ./rdft/codelets/r2r.h cannam@95: A ./rdft/codelets/r2r/Makefile.am cannam@95: M ./Makefile.am +1 cannam@95: M ./configure.ac +1 cannam@95: M ./genfft/Makefile.am -6 +6 cannam@95: M ./genfft/complex.ml -1 +8 cannam@95: M ./genfft/complex.mli -1 +2 cannam@95: M ./genfft/gen_r2r.ml +240 cannam@95: R ./genfft/gen_trig.ml cannam@95: M ./genfft/trig.ml -60 +47 cannam@95: M ./rdft/Makefile.am -3 +3 cannam@95: M ./rdft/codelet-rdft.h -2 +2 cannam@95: M ./rdft/codelets/Makefile.am -2 +2 cannam@95: M ./rdft/codelets/r2r/Makefile.am +72 cannam@95: M ./rdft/codelets/r2r.c +38 cannam@95: M ./rdft/codelets/r2r.h +23 cannam@95: M ./rdft/conf.c -1 +2 cannam@95: M ./rdft/direct.c -16 +58 cannam@95: M ./rdft/kr2r.c +28 cannam@95: M ./rdft/rdft.h -1 +2 cannam@95: M ./support/Makefile.codelets -1 +2 cannam@95: cannam@95: Tue Mar 25 11:29:29 EST 2003 stevenj cannam@95: * [project @ 2003-03-25 16:29:29 by stevenj] cannam@95: noted ARM bug; thanks to Jay Treacy cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -1 +2 cannam@95: cannam@95: Tue Mar 25 07:55:54 EST 2003 athena cannam@95: * [project @ 2003-03-25 12:55:54 by athena] cannam@95: bugfix from Stefan cannam@95: cannam@95: M ./genfft-k7/vK7Optimization.ml -1 +2 cannam@95: cannam@95: Mon Mar 24 15:59:08 EST 2003 stevenj cannam@95: * [project @ 2003-03-24 20:59:08 by stevenj] cannam@95: slight change cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Mon Mar 24 15:58:44 EST 2003 stevenj cannam@95: * [project @ 2003-03-24 20:58:44 by stevenj] cannam@95: caveat cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Mon Mar 24 15:58:04 EST 2003 stevenj cannam@95: * [project @ 2003-03-24 20:58:04 by stevenj] cannam@95: warning about DHT cannam@95: cannam@95: M ./doc/fftw3.texi -1 +8 cannam@95: cannam@95: Mon Mar 24 08:34:14 EST 2003 athena cannam@95: * [project @ 2003-03-24 13:34:14 by athena] cannam@95: Oops cannam@95: cannam@95: M ./dft/k7/codelets/Makefile.am -2 +2 cannam@95: cannam@95: Mon Mar 24 08:13:15 EST 2003 athena cannam@95: * [project @ 2003-03-24 13:13:15 by athena] cannam@95: Regression test for p4fftwgel cannam@95: cannam@95: M ./dft/k7/codelets/Makefile.am -2 +2 cannam@95: M ./tests/Makefile.am +3 cannam@95: M ./tests/check.pl +4 cannam@95: cannam@95: Mon Mar 24 03:09:06 EST 2003 stevenj cannam@95: * [project @ 2003-03-24 08:09:06 by stevenj] cannam@95: make check is faster, old tests are in make bigcheck cannam@95: cannam@95: M ./tests/Makefile.am -1 +6 cannam@95: cannam@95: Sat Mar 22 00:41:21 EST 2003 stevenj cannam@95: * [project @ 2003-03-22 05:41:21 by stevenj] cannam@95: note cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -1 +1 cannam@95: cannam@95: Sat Mar 22 00:40:05 EST 2003 stevenj cannam@95: * [project @ 2003-03-22 05:40:05 by stevenj] cannam@95: whoops, line wrapping cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -2 +1 cannam@95: cannam@95: Fri Mar 21 15:10:00 EST 2003 athena cannam@95: * [project @ 2003-03-21 20:10:00 by athena] cannam@95: Franz-mode codelets even without SIMD. (disabled) cannam@95: cannam@95: M ./dft/codelets/standard/Makefile.am -2 +59 cannam@95: M ./genfft/gen_notw.ml -5 +6 cannam@95: M ./genfft/gen_twiddle.ml -4 +5 cannam@95: cannam@95: Fri Mar 21 09:09:30 EST 2003 athena cannam@95: * [project @ 2003-03-21 14:09:30 by athena] cannam@95: Bug is in netbsd-1.6, not 1.5 cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -1 +1 cannam@95: cannam@95: Fri Mar 21 07:45:48 EST 2003 athena cannam@95: * [project @ 2003-03-21 12:45:48 by athena] cannam@95: const cast, should placate c++ compilers. cannam@95: cannam@95: M ./simd/simd-altivec.h -1 +1 cannam@95: cannam@95: Thu Mar 20 18:49:49 EST 2003 stevenj cannam@95: * [project @ 2003-03-20 23:49:49 by stevenj] cannam@95: added FAQ on why plans are array-specific cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn +26 cannam@95: cannam@95: Thu Mar 20 16:12:56 EST 2003 stevenj cannam@95: * [project @ 2003-03-20 21:12:56 by stevenj] cannam@95: comment fix cannam@95: cannam@95: M ./reodft/reodft010e-r2hc.c -3 +3 cannam@95: cannam@95: Thu Mar 20 16:12:15 EST 2003 stevenj cannam@95: * [project @ 2003-03-20 21:12:15 by stevenj] cannam@95: noted comparison to NR cannam@95: cannam@95: M ./reodft/reodft010e-r2hc.c -1 +5 cannam@95: cannam@95: Wed Mar 19 20:13:16 EST 2003 stevenj cannam@95: * [project @ 2003-03-20 01:13:16 by stevenj] cannam@95: whoops, C99 complex didn't work if complex is a macro (as it is with glibc); thanks to Keh-Cheng Chu for the bug report cannam@95: cannam@95: M ./api/fftw3.h -5 +4 cannam@95: cannam@95: Wed Mar 19 16:52:54 EST 2003 stevenj cannam@95: * [project @ 2003-03-19 21:52:54 by stevenj] cannam@95: noted in help that --enable-k7 enables 3dnow, and that --enable-3dnow is only a fallback cannam@95: cannam@95: M ./configure.ac -2 +2 cannam@95: cannam@95: Wed Mar 19 15:09:52 EST 2003 athena cannam@95: * [project @ 2003-03-19 20:09:52 by athena] cannam@95: New gcc bug. html.refs was not in repository/distribution. cannam@95: cannam@95: A ./doc/FAQ/html.refs cannam@95: M ./doc/FAQ/Makefile.am -1 +1 cannam@95: M ./doc/FAQ/fftw-faq.bfnn -1 +13 cannam@95: M ./doc/FAQ/html.refs +6 cannam@95: cannam@95: Wed Mar 19 10:09:16 EST 2003 athena cannam@95: * [project @ 2003-03-19 15:09:16 by athena] cannam@95: Don't write wisdom if you don't have it. cannam@95: cannam@95: M ./tests/bench.c -1 +1 cannam@95: cannam@95: Tue Mar 18 15:44:41 EST 2003 athena cannam@95: * [project @ 2003-03-18 20:44:41 by athena] cannam@95: Added index entries for DHT. Similarly for DCT, DST cannam@95: cannam@95: M ./doc/fftw3.texi -1 +6 cannam@95: cannam@95: Tue Mar 18 14:50:04 EST 2003 stevenj cannam@95: * [project @ 2003-03-18 19:50:04 by stevenj] cannam@95: execute should not go through C api, for efficiency cannam@95: cannam@95: M ./api/f77api.c +2 cannam@95: M ./api/f77funcs.h -5 +10 cannam@95: cannam@95: Tue Mar 18 06:14:51 EST 2003 athena cannam@95: * [project @ 2003-03-18 11:14:51 by athena] cannam@95: Renamed FFTW_IODIM, FFTW_R2R_KIND cannam@95: cannam@95: M ./api/fftw3.h -6 +6 cannam@95: cannam@95: Tue Mar 18 00:30:17 EST 2003 stevenj cannam@95: * [project @ 2003-03-18 05:30:17 by stevenj] cannam@95: added rfftwnd.eps to dist, so that transfig is not required for people trying to build other formats (e.g. ps); thanks to Brian Gough for the bug report cannam@95: cannam@95: M ./doc/Makefile.am -1 +1 cannam@95: cannam@95: Mon Mar 17 15:17:59 EST 2003 stevenj cannam@95: * [project @ 2003-03-17 20:17:59 by stevenj] cannam@95: pointer to upgrading section from tutorial cannam@95: cannam@95: M ./doc/fftw3.texi -1 +4 cannam@95: cannam@95: Mon Mar 17 14:44:40 EST 2003 stevenj cannam@95: * [project @ 2003-03-17 19:44:40 by stevenj] cannam@95: make print_plan and fprint_plan, so that the former can be more easily called from other languages cannam@95: cannam@95: M ./api/f77funcs.h -2 +2 cannam@95: M ./api/fftw3.h -2 +3 cannam@95: M ./api/print-plan.c -1 +6 cannam@95: M ./doc/fftw3.texi -3 +5 cannam@95: M ./tests/bench.c -1 +1 cannam@95: cannam@95: Mon Mar 17 14:19:10 EST 2003 stevenj cannam@95: * [project @ 2003-03-17 19:19:10 by stevenj] cannam@95: whoops, forgot to change equation image links to .png cannam@95: cannam@95: M ./doc/fftw3.texi -14 +14 cannam@95: cannam@95: Mon Mar 17 04:15:50 EST 2003 athena cannam@95: * [project @ 2003-03-17 09:15:50 by athena] cannam@95: fixed c++ linkage problems cannam@95: cannam@95: M ./api/fftw3.h -4 +4 cannam@95: M ./api/version.c -8 +5 cannam@95: M ./support/Makefile.codelets -1 +2 cannam@95: cannam@95: Mon Mar 17 03:25:17 EST 2003 athena cannam@95: * [project @ 2003-03-17 08:25:17 by athena] cannam@95: Removed ``const'', otherwise c++ link fails cannam@95: cannam@95: M ./api/fftw3.h -4 +4 cannam@95: M ./api/version.c -4 +4 cannam@95: cannam@95: Sun Mar 16 20:24:31 EST 2003 stevenj cannam@95: * [project @ 2003-03-17 01:24:31 by stevenj] cannam@95: fixed C++ annoyances: void* casts, and global variables are static by default(?!?) cannam@95: cannam@95: M ./api/f77api.c -4 +5 cannam@95: M ./api/f77funcs.h -3 +5 cannam@95: M ./api/version.c -4 +4 cannam@95: M ./libbench2/allocate.c -9 +9 cannam@95: M ./libbench2/getopt-utils.c -2 +2 cannam@95: M ./libbench2/problem.c -3 +3 cannam@95: M ./libbench2/speed.c -2 +2 cannam@95: M ./libbench2/timer.c -2 +2 cannam@95: M ./libbench2/verify-r2r.c -3 +3 cannam@95: M ./libbench2/zero.c -9 +9 cannam@95: M ./support/Makefile.codelets -1 +1 cannam@95: M ./tests/bench.c -37 +66 cannam@95: M ./tests/hook.c -2 +2 cannam@95: M ./tools/fftw-wisdom.c -1 +1 cannam@95: cannam@95: Sat Aug 12 23:06:56 EDT 2006 Unknown tagger cannam@95: tagged fftw-3-0-beta1 cannam@95: cannam@95: cannam@95: Sun Mar 16 15:29:11 EST 2003 stevenj cannam@95: * [project @ 2003-03-16 20:29:11 by stevenj] cannam@95: ranlib bug is in binutils cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -4 +5 cannam@95: cannam@95: Sun Mar 16 15:26:42 EST 2003 stevenj cannam@95: * [project @ 2003-03-16 20:26:42 by stevenj] cannam@95: ranlib Irix bug cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn +5 cannam@95: cannam@95: Sun Mar 16 15:13:35 EST 2003 stevenj cannam@95: * [project @ 2003-03-16 20:13:35 by stevenj] cannam@95: start with random tests cannam@95: cannam@95: M ./tests/check.pl -1 +1 cannam@95: cannam@95: Sun Mar 16 15:00:04 EST 2003 stevenj cannam@95: * [project @ 2003-03-16 20:00:03 by stevenj] cannam@95: silenced some compiler warnings, eliminated unused variables, and fixed Makefile.am for f77funcs.h cannam@95: cannam@95: M ./api/Makefile.am -5 +1 cannam@95: M ./dft/direct.c -3 +2 cannam@95: M ./kernel/ifftw.h -4 +4 cannam@95: M ./libbench2/verify-r2r.c -3 cannam@95: M ./rdft/direct.c -3 +2 cannam@95: M ./rdft/direct2.c -3 +2 cannam@95: M ./threads/Makefile.am -3 +1 cannam@95: cannam@95: Sun Mar 16 14:55:13 EST 2003 stevenj cannam@95: * [project @ 2003-03-16 19:55:13 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -1 +1 cannam@95: cannam@95: Sun Mar 16 14:28:22 EST 2003 stevenj cannam@95: * [project @ 2003-03-16 19:28:22 by stevenj] cannam@95: 3dnow is float cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Sun Mar 16 14:27:45 EST 2003 stevenj cannam@95: * [project @ 2003-03-16 19:27:45 by stevenj] cannam@95: fixed k7 docs cannam@95: cannam@95: M ./doc/fftw3.texi -17 +10 cannam@95: cannam@95: Sun Mar 16 14:19:10 EST 2003 stevenj cannam@95: * [project @ 2003-03-16 19:19:10 by stevenj] cannam@95: SGI compilers now support inline cannam@95: cannam@95: M ./kernel/cycle.h -3 +3 cannam@95: cannam@95: Sun Mar 16 14:18:32 EST 2003 stevenj cannam@95: * [project @ 2003-03-16 19:18:32 by stevenj] cannam@95: cruft cannam@95: cannam@95: M ./kernel/cycle.h -2 +1 cannam@95: cannam@95: Sun Mar 16 14:15:47 EST 2003 stevenj cannam@95: * [project @ 2003-03-16 19:15:47 by stevenj] cannam@95: texinfo doesn't like commas in nodes cannam@95: cannam@95: M ./doc/fftw3.texi -16 +15 cannam@95: cannam@95: Sun Mar 16 13:52:04 EST 2003 stevenj cannam@95: * [project @ 2003-03-16 18:48:30 by stevenj] cannam@95: updated cannam@95: cannam@95: M ./ChangeLog -47 +1389 cannam@95: M ./README +28 cannam@95: cannam@95: Sun Mar 16 13:47:44 EST 2003 stevenj cannam@95: * [project @ 2003-03-16 18:47:44 by stevenj] cannam@95: f77funcs.c -> f77funcs.h so that people don't try to compile it cannam@95: cannam@95: A ./api/f77funcs.h cannam@95: A ./threads/f77funcs.h cannam@95: M ./api/f77api.c -4 +4 cannam@95: R ./api/f77funcs.c cannam@95: M ./api/f77funcs.h +366 cannam@95: M ./threads/f77api.c -4 +4 cannam@95: R ./threads/f77funcs.c cannam@95: M ./threads/f77funcs.h +39 cannam@95: cannam@95: Sun Mar 16 13:46:11 EST 2003 stevenj cannam@95: * [project @ 2003-03-16 18:46:11 by stevenj] cannam@95: minor changes cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -7 +4 cannam@95: cannam@95: Sun Mar 16 13:39:24 EST 2003 stevenj cannam@95: * [project @ 2003-03-16 18:39:24 by stevenj] cannam@95: updated compiler bug list cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -40 +20 cannam@95: cannam@95: Sun Mar 16 13:39:07 EST 2003 stevenj cannam@95: * [project @ 2003-03-16 18:39:07 by stevenj] cannam@95: noted how to set CC cannam@95: cannam@95: M ./doc/fftw3.texi -4 +10 cannam@95: cannam@95: Sun Mar 16 13:01:01 EST 2003 stevenj cannam@95: * [project @ 2003-03-16 18:01:01 by stevenj] cannam@95: TODONE cannam@95: cannam@95: M ./TODO -10 cannam@95: cannam@95: Sun Mar 16 13:00:42 EST 2003 stevenj cannam@95: * [project @ 2003-03-16 18:00:42 by stevenj] cannam@95: yikes, bugfix cannam@95: cannam@95: M ./threads/vrank-geq1-rdft2.c -19 +11 cannam@95: cannam@95: Sun Mar 16 10:26:28 EST 2003 stevenj cannam@95: * [project @ 2003-03-16 15:26:28 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./kernel/ifftw.h -2 +2 cannam@95: cannam@95: Sun Mar 16 09:24:19 EST 2003 athena cannam@95: * [project @ 2003-03-16 14:24:19 by athena] cannam@95: Report SIMD extensions in version string cannam@95: cannam@95: M ./api/version.c -2 +26 cannam@95: cannam@95: Sat Mar 15 18:56:11 EST 2003 stevenj cannam@95: * [project @ 2003-03-15 23:56:11 by stevenj] cannam@95: more verbose output cannam@95: cannam@95: M ./tests/bench.c +2 cannam@95: cannam@95: Sat Mar 15 17:41:25 EST 2003 stevenj cannam@95: * [project @ 2003-03-15 22:41:25 by stevenj] cannam@95: a couple of additional non-Unix instructions cannam@95: cannam@95: M ./doc/fftw3.texi -1 +9 cannam@95: cannam@95: Sat Mar 15 17:15:26 EST 2003 stevenj cannam@95: * [project @ 2003-03-15 22:15:26 by stevenj] cannam@95: hyphen cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -1 +1 cannam@95: cannam@95: Sat Mar 15 17:12:29 EST 2003 stevenj cannam@95: * [project @ 2003-03-15 22:12:29 by stevenj] cannam@95: softened cannam@95: cannam@95: M ./doc/FAQ/fftw-faq.bfnn -3 +3 cannam@95: cannam@95: Sat Mar 15 17:09:44 EST 2003 stevenj cannam@95: * [project @ 2003-03-15 22:09:44 by stevenj] cannam@95: added FAQ, used PNGs cannam@95: cannam@95: A ./doc/equation-dft.png cannam@95: A ./doc/equation-dht.png cannam@95: A ./doc/equation-idft.png cannam@95: A ./doc/equation-redft00.png cannam@95: A ./doc/equation-redft01.png cannam@95: A ./doc/equation-redft10.png cannam@95: A ./doc/equation-redft11.png cannam@95: A ./doc/equation-rodft00.png cannam@95: A ./doc/equation-rodft01.png cannam@95: A ./doc/equation-rodft10.png cannam@95: A ./doc/FAQ/ cannam@95: A ./doc/FAQ/Makefile.am cannam@95: A ./doc/FAQ/bfnnconv.pl cannam@95: A ./doc/FAQ/fftw-faq.bfnn cannam@95: A ./doc/FAQ/m-ascii.pl cannam@95: A ./doc/FAQ/m-html.pl cannam@95: A ./doc/FAQ/m-info.pl cannam@95: A ./doc/FAQ/m-lout.pl cannam@95: A ./doc/FAQ/m-post.pl cannam@95: A ./doc/equation-rodft11.png cannam@95: M ./configure.ac +1 cannam@95: M ./doc/FAQ/Makefile.am +14 cannam@95: M ./doc/FAQ/bfnnconv.pl +298 cannam@95: M ./doc/FAQ/fftw-faq.bfnn +492 cannam@95: M ./doc/FAQ/m-ascii.pl +189 cannam@95: M ./doc/FAQ/m-html.pl +337 cannam@95: M ./doc/FAQ/m-info.pl +226 cannam@95: M ./doc/FAQ/m-lout.pl +242 cannam@95: M ./doc/FAQ/m-post.pl +189 cannam@95: M ./doc/Makefile.am -4 +6 cannam@95: R ./doc/equation-dft.gif cannam@95: M ./doc/equation-dft.png cannam@95: R ./doc/equation-dht.gif cannam@95: M ./doc/equation-dht.png cannam@95: R ./doc/equation-idft.gif cannam@95: M ./doc/equation-idft.png cannam@95: R ./doc/equation-redft00.gif cannam@95: M ./doc/equation-redft00.png cannam@95: R ./doc/equation-redft01.gif cannam@95: M ./doc/equation-redft01.png cannam@95: R ./doc/equation-redft10.gif cannam@95: M ./doc/equation-redft10.png cannam@95: R ./doc/equation-redft11.gif cannam@95: M ./doc/equation-redft11.png cannam@95: R ./doc/equation-rodft00.gif cannam@95: M ./doc/equation-rodft00.png cannam@95: R ./doc/equation-rodft01.gif cannam@95: M ./doc/equation-rodft01.png cannam@95: R ./doc/equation-rodft10.gif cannam@95: M ./doc/equation-rodft10.png cannam@95: R ./doc/equation-rodft11.gif cannam@95: M ./doc/equation-rodft11.png cannam@95: cannam@95: Sat Mar 15 15:29:43 EST 2003 stevenj cannam@95: * [project @ 2003-03-15 20:29:42 by stevenj] cannam@95: great copyright update cannam@95: cannam@95: M ./COPYRIGHT -2 +2 cannam@95: M ./TODO -6 cannam@95: M ./api/api.h -2 +2 cannam@95: M ./api/apiplan.c -2 +2 cannam@95: M ./api/configure.c -2 +2 cannam@95: M ./api/execute-dft-c2r.c -2 +2 cannam@95: M ./api/execute-dft-r2c.c -2 +2 cannam@95: M ./api/execute-dft.c -2 +2 cannam@95: M ./api/execute-r2r.c -2 +2 cannam@95: M ./api/execute.c -2 +2 cannam@95: M ./api/export-wisdom-to-file.c -2 +2 cannam@95: M ./api/export-wisdom-to-string.c -2 +2 cannam@95: M ./api/export-wisdom.c -2 +2 cannam@95: M ./api/extract-reim.c -2 +2 cannam@95: M ./api/f77api.c -2 +2 cannam@95: M ./api/f77funcs.c -2 +2 cannam@95: M ./api/fftw3.h -3 +3 cannam@95: M ./api/flops.c -2 +2 cannam@95: M ./api/forget-wisdom.c -2 +2 cannam@95: M ./api/import-system-wisdom.c -2 +2 cannam@95: M ./api/import-wisdom-from-file.c -2 +2 cannam@95: M ./api/import-wisdom-from-string.c -2 +2 cannam@95: M ./api/import-wisdom.c -2 +2 cannam@95: M ./api/map-r2r-kind.c -2 +2 cannam@95: M ./api/mapflags.c -2 +2 cannam@95: M ./api/mkprinter-file.c -2 +2 cannam@95: M ./api/mktensor-iodims.c -2 +2 cannam@95: M ./api/mktensor-rowmajor.c -2 +2 cannam@95: M ./api/plan-dft-1d.c -2 +2 cannam@95: M ./api/plan-dft-2d.c -2 +2 cannam@95: M ./api/plan-dft-3d.c -2 +2 cannam@95: M ./api/plan-dft-c2r-1d.c -2 +2 cannam@95: M ./api/plan-dft-c2r-2d.c -2 +2 cannam@95: M ./api/plan-dft-c2r-3d.c -2 +2 cannam@95: M ./api/plan-dft-c2r.c -2 +2 cannam@95: M ./api/plan-dft-r2c-1d.c -2 +2 cannam@95: M ./api/plan-dft-r2c-2d.c -2 +2 cannam@95: M ./api/plan-dft-r2c-3d.c -2 +2 cannam@95: M ./api/plan-dft-r2c.c -2 +2 cannam@95: M ./api/plan-dft.c -2 +2 cannam@95: M ./api/plan-guru-dft-c2r.c -2 +2 cannam@95: M ./api/plan-guru-dft-r2c.c -2 +2 cannam@95: M ./api/plan-guru-dft.c -2 +2 cannam@95: M ./api/plan-guru-r2r.c -2 +2 cannam@95: M ./api/plan-many-dft-c2r.c -2 +2 cannam@95: M ./api/plan-many-dft-r2c.c -2 +2 cannam@95: M ./api/plan-many-dft.c -2 +2 cannam@95: M ./api/plan-many-r2r.c -2 +2 cannam@95: M ./api/plan-r2r-1d.c -2 +2 cannam@95: M ./api/plan-r2r-2d.c -2 +2 cannam@95: M ./api/plan-r2r-3d.c -2 +2 cannam@95: M ./api/plan-r2r.c -2 +2 cannam@95: M ./api/print-plan.c -2 +2 cannam@95: M ./api/rdft2-pad.c -2 +2 cannam@95: M ./api/the-planner.c -2 +2 cannam@95: M ./api/version.c -3 +3 cannam@95: M ./api/x77.h -2 +2 cannam@95: M ./dft/buffered.c -3 +3 cannam@95: M ./dft/codelet-dft.h -3 +3 cannam@95: M ./dft/codelets/n.c -2 +2 cannam@95: M ./dft/codelets/n.h -2 +2 cannam@95: M ./dft/codelets/t.c -2 +2 cannam@95: M ./dft/codelets/t.h -2 +2 cannam@95: M ./dft/conf.c -3 +3 cannam@95: M ./dft/ct-dif.c -3 +3 cannam@95: M ./dft/ct-dit.c -3 +3 cannam@95: M ./dft/ct-ditbuf.c -3 +3 cannam@95: M ./dft/ct-ditf.c -3 +3 cannam@95: M ./dft/ct.c -3 +3 cannam@95: M ./dft/ct.h -3 +3 cannam@95: M ./dft/dft.h -3 +3 cannam@95: M ./dft/direct.c -3 +3 cannam@95: M ./dft/generic.c -2 +2 cannam@95: M ./dft/indirect.c -3 +3 cannam@95: M ./dft/k7/k7.c -3 +3 cannam@95: M ./dft/kdft-dif.c -3 +3 cannam@95: M ./dft/kdft-difsq.c -3 +3 cannam@95: M ./dft/kdft-dit.c -3 +3 cannam@95: M ./dft/kdft.c -3 +3 cannam@95: M ./dft/nop.c -3 +3 cannam@95: M ./dft/plan.c -3 +3 cannam@95: M ./dft/problem.c -3 +3 cannam@95: M ./dft/rader-omega.c -2 +2 cannam@95: M ./dft/rader.c -2 +2 cannam@95: M ./dft/rank-geq2.c -3 +3 cannam@95: M ./dft/rank0.c -3 +3 cannam@95: M ./dft/simd/n1b.c -2 +2 cannam@95: M ./dft/simd/n1b.h -2 +2 cannam@95: M ./dft/simd/n1f.c -2 +2 cannam@95: M ./dft/simd/n1f.h -2 +2 cannam@95: M ./dft/simd/t1b.c -2 +2 cannam@95: M ./dft/simd/t1b.h -2 +2 cannam@95: M ./dft/simd/t1f.c -2 +2 cannam@95: M ./dft/simd/t1f.h -2 +2 cannam@95: M ./dft/solve.c -3 +3 cannam@95: M ./dft/vrank-geq1.c -3 +3 cannam@95: M ./dft/vrank2-transpose.c -3 +3 cannam@95: M ./dft/vrank3-transpose.c -3 +3 cannam@95: M ./dft/zero.c -3 +3 cannam@95: M ./doc/f77_wisdom.f -2 +2 cannam@95: M ./doc/fftw3.texi -4 +4 cannam@95: M ./genfft/algsimp.ml -4 +4 cannam@95: M ./genfft/algsimp.mli -3 +3 cannam@95: M ./genfft/annotate.ml -4 +4 cannam@95: M ./genfft/annotate.mli -3 +3 cannam@95: M ./genfft/assoctable.ml -3 +3 cannam@95: M ./genfft/assoctable.mli -3 +3 cannam@95: M ./genfft/c.ml -3 +3 cannam@95: M ./genfft/c.mli -3 +3 cannam@95: M ./genfft/complex.ml -3 +3 cannam@95: M ./genfft/complex.mli -3 +3 cannam@95: M ./genfft/conv.ml -2 +2 cannam@95: M ./genfft/conv.mli -3 +3 cannam@95: M ./genfft/dag.ml -3 +3 cannam@95: M ./genfft/dag.mli -3 +3 cannam@95: M ./genfft/expr.ml -3 +3 cannam@95: M ./genfft/expr.mli -3 +3 cannam@95: M ./genfft/fft.ml -4 +4 cannam@95: M ./genfft/fft.mli -3 +3 cannam@95: M ./genfft/gen_athnotw.ml -4 +4 cannam@95: M ./genfft/gen_athtw.ml -4 +4 cannam@95: M ./genfft/gen_conv.ml -4 +4 cannam@95: M ./genfft/gen_hc2hc.ml -4 +4 cannam@95: M ./genfft/gen_hc2r.ml -4 +4 cannam@95: M ./genfft/gen_notw.ml -4 +4 cannam@95: M ./genfft/gen_notw_c.ml -4 +4 cannam@95: M ./genfft/gen_r2hc.ml -4 +4 cannam@95: M ./genfft/gen_trig.ml -4 +4 cannam@95: M ./genfft/gen_twiddle.ml -4 +4 cannam@95: M ./genfft/gen_twiddle_c.ml -4 +4 cannam@95: M ./genfft/gen_twidsq.ml -4 +4 cannam@95: M ./genfft/genutil.ml -3 +3 cannam@95: M ./genfft/littlesimp.ml -3 +3 cannam@95: M ./genfft/littlesimp.mli -3 +3 cannam@95: M ./genfft/magic.ml -3 +3 cannam@95: M ./genfft/monads.ml -3 +3 cannam@95: M ./genfft/number.ml -3 +3 cannam@95: M ./genfft/number.mli -3 +3 cannam@95: M ./genfft/oracle.ml -3 +3 cannam@95: M ./genfft/oracle.mli -3 +3 cannam@95: M ./genfft/schedule.ml -3 +3 cannam@95: M ./genfft/schedule.mli -3 +3 cannam@95: M ./genfft/simd.ml -3 +3 cannam@95: M ./genfft/simd.mli -3 +3 cannam@95: M ./genfft/simdmagic.ml -3 +3 cannam@95: M ./genfft/to_alist.ml -3 +3 cannam@95: M ./genfft/to_alist.mli -3 +3 cannam@95: M ./genfft/trig.ml -3 +3 cannam@95: M ./genfft/trig.mli -3 +3 cannam@95: M ./genfft/twiddle.ml -3 +3 cannam@95: M ./genfft/twiddle.mli -3 +3 cannam@95: M ./genfft/unique.ml -3 +3 cannam@95: M ./genfft/unique.mli -3 +3 cannam@95: M ./genfft/util.ml -3 +3 cannam@95: M ./genfft/util.mli -3 +3 cannam@95: M ./genfft/variable.ml -3 +3 cannam@95: M ./genfft/variable.mli -3 +3 cannam@95: M ./genfft-k7/algsimp.ml -4 +4 cannam@95: M ./genfft-k7/algsimp.mli -3 +3 cannam@95: M ./genfft-k7/assoctable.ml -3 +3 cannam@95: M ./genfft-k7/assoctable.mli -3 +3 cannam@95: M ./genfft-k7/expr.ml -3 +3 cannam@95: M ./genfft-k7/expr.mli -3 +3 cannam@95: M ./genfft-k7/fft.ml -4 +4 cannam@95: M ./genfft-k7/littlesimp.ml -3 +3 cannam@95: M ./genfft-k7/littlesimp.mli -3 +3 cannam@95: M ./genfft-k7/monads.ml -3 +3 cannam@95: M ./genfft-k7/number.ml -3 +3 cannam@95: M ./genfft-k7/number.mli -3 +3 cannam@95: M ./genfft-k7/oracle.ml -3 +3 cannam@95: M ./genfft-k7/oracle.mli -3 +3 cannam@95: M ./genfft-k7/to_alist.ml -3 +3 cannam@95: M ./genfft-k7/to_alist.mli -3 +3 cannam@95: M ./genfft-k7/twiddle.ml -3 +3 cannam@95: M ./genfft-k7/twiddle.mli -3 +3 cannam@95: M ./kernel/align.c -3 +3 cannam@95: M ./kernel/alloc.c -3 +3 cannam@95: M ./kernel/assert.c -3 +3 cannam@95: M ./kernel/awake.c -3 +3 cannam@95: M ./kernel/buffered.c -2 +2 cannam@95: M ./kernel/ct.c -2 +2 cannam@95: M ./kernel/cycle.h -3 +3 cannam@95: M ./kernel/debug.c -3 +3 cannam@95: M ./kernel/hash.c -2 +2 cannam@95: M ./kernel/iabs.c -3 +3 cannam@95: M ./kernel/ifftw.h -3 +3 cannam@95: M ./kernel/md5-1.c -2 +2 cannam@95: M ./kernel/md5.c -2 +2 cannam@95: M ./kernel/minmax.c -3 +3 cannam@95: M ./kernel/ops.c -3 +3 cannam@95: M ./kernel/pickdim.c -3 +3 cannam@95: M ./kernel/plan.c -3 +3 cannam@95: M ./kernel/planner.c -2 +2 cannam@95: M ./kernel/primes.c -3 +3 cannam@95: M ./kernel/print.c -3 +3 cannam@95: M ./kernel/problem.c -3 +3 cannam@95: M ./kernel/rader.c -2 +2 cannam@95: M ./kernel/scan.c -3 +3 cannam@95: M ./kernel/solver.c -3 +3 cannam@95: M ./kernel/solvtab.c -3 +3 cannam@95: M ./kernel/square.c -3 +3 cannam@95: M ./kernel/stride.c -3 +3 cannam@95: M ./kernel/tensor.c -3 +3 cannam@95: M ./kernel/tensor1.c -3 +3 cannam@95: M ./kernel/tensor2.c -3 +3 cannam@95: M ./kernel/tensor4.c -3 +3 cannam@95: M ./kernel/tensor5.c -3 +3 cannam@95: M ./kernel/tensor7.c -3 +3 cannam@95: M ./kernel/tensor8.c -3 +3 cannam@95: M ./kernel/tensor9.c -3 +3 cannam@95: M ./kernel/timer.c -3 +3 cannam@95: M ./kernel/trig.c -3 +3 cannam@95: M ./kernel/trig1.c -3 +3 cannam@95: M ./kernel/twiddle.c -3 +3 cannam@95: M ./libbench/bench-main.c -2 +2 cannam@95: M ./libbench/bench-user.h -2 +2 cannam@95: M ./libbench/bench.h -2 +2 cannam@95: M ./libbench/can-do.c -2 +2 cannam@95: M ./libbench/getopt-utils.c -2 +2 cannam@95: M ./libbench/info.c -2 +2 cannam@95: M ./libbench/main.c -2 +2 cannam@95: M ./libbench/prime.c -2 +2 cannam@95: M ./libbench/problem.c -2 +2 cannam@95: M ./libbench/report.c -2 +2 cannam@95: M ./libbench/speed.c -2 +2 cannam@95: M ./libbench/timer.c -2 +2 cannam@95: M ./libbench/util.c -1 +1 cannam@95: M ./libbench/verify.c -2 +2 cannam@95: M ./libbench/zero.c -2 +2 cannam@95: M ./libbench2/aligned-main.c -2 +2 cannam@95: M ./libbench2/bench-main.c -2 +2 cannam@95: M ./libbench2/bench-user.h -2 +2 cannam@95: M ./libbench2/bench.h -2 +2 cannam@95: M ./libbench2/can-do.c -2 +2 cannam@95: M ./libbench2/dotens2.c -3 +3 cannam@95: M ./libbench2/getopt-utils.c -2 +2 cannam@95: M ./libbench2/info.c -2 +2 cannam@95: M ./libbench2/main.c -2 +2 cannam@95: M ./libbench2/problem.c -2 +2 cannam@95: M ./libbench2/report.c -2 +2 cannam@95: M ./libbench2/speed.c -2 +2 cannam@95: M ./libbench2/tensor.c -2 +2 cannam@95: M ./libbench2/timer.c -2 +2 cannam@95: M ./libbench2/useropt.c -2 +2 cannam@95: M ./libbench2/util.c -1 +1 cannam@95: M ./libbench2/verify-dft.c -3 +3 cannam@95: M ./libbench2/verify-lib.c -3 +3 cannam@95: M ./libbench2/verify-r2r.c -2 +2 cannam@95: M ./libbench2/verify-rdft2.c -3 +3 cannam@95: M ./libbench2/verify.c -2 +2 cannam@95: M ./libbench2/verify.h -2 +2 cannam@95: M ./libbench2/zero.c -2 +2 cannam@95: M ./rdft/buffered.c -3 +3 cannam@95: M ./rdft/buffered2.c -3 +3 cannam@95: M ./rdft/codelet-rdft.h -3 +3 cannam@95: M ./rdft/codelets/hb.h -2 +2 cannam@95: M ./rdft/codelets/hc2r.c -2 +2 cannam@95: M ./rdft/codelets/hc2r.h -2 +2 cannam@95: M ./rdft/codelets/hc2rIII.h -2 +2 cannam@95: M ./rdft/codelets/hf.h -2 +2 cannam@95: M ./rdft/codelets/hfb.c -2 +2 cannam@95: M ./rdft/codelets/r2hc.c -2 +2 cannam@95: M ./rdft/codelets/r2hc.h -2 +2 cannam@95: M ./rdft/codelets/r2hcII.h -2 +2 cannam@95: M ./rdft/conf.c -3 +3 cannam@95: M ./rdft/dft-r2hc.c -3 +3 cannam@95: M ./rdft/dht-r2hc.c -3 +3 cannam@95: M ./rdft/dht-rader.c -2 +2 cannam@95: M ./rdft/direct.c -3 +3 cannam@95: M ./rdft/direct2.c -3 +3 cannam@95: M ./rdft/generic.c -2 +2 cannam@95: M ./rdft/hc2hc-buf.c -3 +3 cannam@95: M ./rdft/hc2hc-dif.c -3 +3 cannam@95: M ./rdft/hc2hc-dit.c -3 +3 cannam@95: M ./rdft/hc2hc.c -3 +3 cannam@95: M ./rdft/hc2hc.h -3 +3 cannam@95: M ./rdft/indirect.c -3 +3 cannam@95: M ./rdft/khc2hc-dif.c -3 +3 cannam@95: M ./rdft/khc2hc-dit.c -3 +3 cannam@95: M ./rdft/khc2r.c -3 +3 cannam@95: M ./rdft/kr2hc.c -3 +3 cannam@95: M ./rdft/nop.c -3 +3 cannam@95: M ./rdft/nop2.c -3 +3 cannam@95: M ./rdft/plan.c -3 +3 cannam@95: M ./rdft/plan2.c -3 +3 cannam@95: M ./rdft/problem.c -3 +3 cannam@95: M ./rdft/problem2.c -3 +3 cannam@95: M ./rdft/rader-hc2hc.c -2 +2 cannam@95: M ./rdft/rank-geq2-rdft2.c -3 +3 cannam@95: M ./rdft/rank-geq2.c -3 +3 cannam@95: M ./rdft/rank0-rdft2.c -3 +3 cannam@95: M ./rdft/rank0.c -3 +3 cannam@95: M ./rdft/rdft-dht.c -3 +3 cannam@95: M ./rdft/rdft.h -3 +3 cannam@95: M ./rdft/rdft2-inplace-strides.c -3 +3 cannam@95: M ./rdft/rdft2-radix2.c -3 +3 cannam@95: M ./rdft/rdft2-tensor-max-index.c -3 +3 cannam@95: M ./rdft/solve.c -3 +3 cannam@95: M ./rdft/solve2.c -3 +3 cannam@95: M ./rdft/vrank-geq1-rdft2.c -3 +3 cannam@95: M ./rdft/vrank-geq1.c -3 +3 cannam@95: M ./rdft/vrank2-transpose.c -3 +3 cannam@95: M ./rdft/vrank3-transpose.c -3 +3 cannam@95: M ./reodft/conf.c -3 +3 cannam@95: M ./reodft/redft00e-r2hc.c -3 +3 cannam@95: M ./reodft/reodft.h -3 +3 cannam@95: M ./reodft/reodft010e-r2hc.c -3 +3 cannam@95: M ./reodft/reodft11e-r2hc-odd.c -3 +3 cannam@95: M ./reodft/reodft11e-r2hc.c -3 +3 cannam@95: M ./reodft/reodft11e-radix2.c -3 +3 cannam@95: M ./reodft/rodft00e-r2hc.c -3 +3 cannam@95: M ./simd/3dnow.c -3 +3 cannam@95: M ./simd/altivec.c -3 +3 cannam@95: M ./simd/simd-3dnow.h -2 +2 cannam@95: M ./simd/simd-altivec.h -2 +2 cannam@95: M ./simd/simd-sse.h -2 +2 cannam@95: M ./simd/simd-sse2.h -2 +2 cannam@95: M ./simd/simd.h -2 +2 cannam@95: M ./simd/sse-aux.c -3 +3 cannam@95: M ./simd/sse.c -3 +3 cannam@95: M ./simd/sse2-aux.c -3 +3 cannam@95: M ./simd/sse2.c -3 +3 cannam@95: M ./threads/api.c -2 +2 cannam@95: M ./threads/conf.c -3 +3 cannam@95: M ./threads/ct-dit.c -3 +3 cannam@95: M ./threads/dft-vrank-geq1.c -3 +3 cannam@95: M ./threads/f77api.c -2 +2 cannam@95: M ./threads/f77funcs.c -2 +2 cannam@95: M ./threads/hc2hc-dif.c -3 +3 cannam@95: M ./threads/hc2hc-dit.c -3 +3 cannam@95: M ./threads/rdft-vrank-geq1.c -3 +3 cannam@95: M ./threads/threads.c -2 +2 cannam@95: M ./threads/threads.h -3 +3 cannam@95: M ./threads/vrank-geq1-rdft2.c -3 +3 cannam@95: M ./tools/fftw-wisdom-to-conf.1 -4 +4 cannam@95: M ./tools/fftw-wisdom-to-conf.in -2 +2 cannam@95: M ./tools/fftw-wisdom.c -2 +2 cannam@95: M ./tools/fftw_wisdom.1.in -4 +4 cannam@95: cannam@95: Sat Mar 15 15:14:02 EST 2003 stevenj cannam@95: * [project @ 2003-03-15 20:14:02 by stevenj] cannam@95: threads in make check cannam@95: cannam@95: M ./TODO -4 cannam@95: M ./tests/Makefile.am -1 +5 cannam@95: M ./tests/check.pl +3 cannam@95: cannam@95: Sat Mar 15 15:11:24 EST 2003 stevenj cannam@95: * [project @ 2003-03-15 20:11:24 by stevenj] cannam@95: fixed const warnings cannam@95: cannam@95: M ./threads/ct-dit.c -2 +2 cannam@95: M ./threads/hc2hc-dif.c -2 +2 cannam@95: M ./threads/hc2hc-dit.c -2 +2 cannam@95: cannam@95: Sat Mar 15 15:08:25 EST 2003 stevenj cannam@95: * [project @ 2003-03-15 20:08:25 by stevenj] cannam@95: make sure spawn_loop size > 1 (it has to be at least > 0 lest we crash, but > 1 is an optimization) cannam@95: cannam@95: M ./threads/ct-dit.c -1 +2 cannam@95: M ./threads/hc2hc-dif.c -1 +2 cannam@95: M ./threads/hc2hc-dit.c -1 +2 cannam@95: cannam@95: Sat Mar 15 14:00:17 EST 2003 athena cannam@95: * [project @ 2003-03-15 19:00:17 by athena] cannam@95: hpux seems to want machine/sys/inline.h as opposed to cannam@95: machine/inline.h. cannam@95: cannam@95: M ./kernel/cycle.h -2 +2 cannam@95: cannam@95: Sat Mar 15 13:36:56 EST 2003 stevenj cannam@95: * [project @ 2003-03-15 18:36:56 by stevenj] cannam@95: Sourceforge is really SourceForge.net, and is run by VA cannam@95: cannam@95: M ./doc/fftw3.texi -3 +4 cannam@95: cannam@95: Sat Mar 15 13:34:05 EST 2003 stevenj cannam@95: * [project @ 2003-03-15 18:33:07 by stevenj] cannam@95: comma cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Sat Mar 15 13:31:42 EST 2003 stevenj cannam@95: * [project @ 2003-03-15 18:31:42 by stevenj] cannam@95: fixed AMD company name cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Sat Mar 15 13:29:41 EST 2003 stevenj cannam@95: * [project @ 2003-03-15 18:29:41 by stevenj] cannam@95: minor changes cannam@95: cannam@95: M ./doc/fftw3.texi -19 +19 cannam@95: cannam@95: Sat Mar 15 13:13:55 EST 2003 stevenj cannam@95: * [project @ 2003-03-15 18:13:55 by stevenj] cannam@95: more emitter->read_char renaming cannam@95: cannam@95: M ./api/f77api.c -5 +5 cannam@95: M ./api/f77funcs.c -4 +4 cannam@95: cannam@95: Sat Mar 15 13:08:45 EST 2003 stevenj cannam@95: * [project @ 2003-03-15 18:08:45 by stevenj] cannam@95: more wisdom docs, noted wisdom utilities cannam@95: cannam@95: M ./doc/fftw3.texi -35 +95 cannam@95: cannam@95: Sat Mar 15 11:41:32 EST 2003 stevenj cannam@95: * [project @ 2003-03-15 16:41:32 by stevenj] cannam@95: compound adjectives are hyphenated cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Sat Mar 15 11:40:30 EST 2003 stevenj cannam@95: * [project @ 2003-03-15 16:40:30 by stevenj] cannam@95: fftw does support another type of packed array via r2r cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Sat Mar 15 11:29:12 EST 2003 stevenj cannam@95: * [project @ 2003-03-15 16:29:12 by stevenj] cannam@95: write_char/read_char for export/import functions cannam@95: cannam@95: M ./api/export-wisdom.c -4 +4 cannam@95: M ./api/f77api.c -5 +5 cannam@95: M ./api/f77funcs.c -4 +4 cannam@95: M ./api/fftw3.h -3 +3 cannam@95: M ./api/import-wisdom.c -4 +4 cannam@95: M ./doc/f77_wisdom.f -6 +6 cannam@95: M ./doc/fftw3.texi -5 +5 cannam@95: cannam@95: Sat Mar 15 11:19:19 EST 2003 stevenj cannam@95: * [project @ 2003-03-15 16:19:19 by stevenj] cannam@95: comments cannam@95: cannam@95: M ./threads/threads.c -4 +8 cannam@95: cannam@95: Sat Mar 15 10:08:26 EST 2003 athena cannam@95: * [project @ 2003-03-15 15:08:26 by athena] cannam@95: Enabled randomized-cse cannam@95: cannam@95: M ./support/Makefile.codelets -1 +1 cannam@95: cannam@95: Sat Mar 15 09:47:49 EST 2003 athena cannam@95: * [project @ 2003-03-15 14:47:49 by athena] cannam@95: Changed to 3.0-beta1 cannam@95: cannam@95: M ./configure.ac -1 +1 cannam@95: cannam@95: Sat Mar 15 09:07:31 EST 2003 athena cannam@95: * [project @ 2003-03-15 14:07:31 by athena] cannam@95: First complete draft cannam@95: cannam@95: M ./doc/fftw3.texi -18 +131 cannam@95: cannam@95: Sat Mar 15 08:37:52 EST 2003 athena cannam@95: * [project @ 2003-03-15 13:37:52 by athena] cannam@95: EMITTER is a misnomer cannam@95: cannam@95: M ./api/fftw3.h -2 +2 cannam@95: M ./api/import-wisdom.c -8 +6 cannam@95: cannam@95: Sat Mar 15 05:50:50 EST 2003 athena cannam@95: * [project @ 2003-03-15 10:50:50 by athena] cannam@95: Revision, wisdom tutorial, acks. cannam@95: cannam@95: M ./doc/fftw3.texi -44 +219 cannam@95: cannam@95: Fri Mar 14 22:59:04 EST 2003 stevenj cannam@95: * [project @ 2003-03-15 03:59:04 by stevenj] cannam@95: noted OpenMP cannam@95: cannam@95: M ./NEWS +2 cannam@95: cannam@95: Fri Mar 14 22:38:49 EST 2003 stevenj cannam@95: * [project @ 2003-03-15 03:38:49 by stevenj] cannam@95: comment cannam@95: cannam@95: M ./threads/threads.c -1 +1 cannam@95: cannam@95: Fri Mar 14 22:38:30 EST 2003 stevenj cannam@95: * [project @ 2003-03-15 03:38:30 by stevenj] cannam@95: comments cannam@95: cannam@95: M ./threads/threads.c -2 +2 cannam@95: cannam@95: Fri Mar 14 22:38:05 EST 2003 stevenj cannam@95: * [project @ 2003-03-15 03:38:05 by stevenj] cannam@95: reformatting cannam@95: cannam@95: M ./threads/threads.c -3 +1 cannam@95: cannam@95: Fri Mar 14 22:26:28 EST 2003 stevenj cannam@95: * [project @ 2003-03-15 03:26:28 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./threads/threads.c -2 +2 cannam@95: cannam@95: Fri Mar 14 22:11:23 EST 2003 stevenj cannam@95: * [project @ 2003-03-15 03:11:23 by stevenj] cannam@95: some threads fixes, and added experimental semaphore (pre-thread-spawning) and Linux spinlock support cannam@95: cannam@95: M ./tests/bench.c -1 +10 cannam@95: M ./threads/api.c +1 cannam@95: M ./threads/threads.c -8 +159 cannam@95: M ./threads/threads.h -1 +2 cannam@95: cannam@95: Fri Mar 14 20:50:46 EST 2003 stevenj cannam@95: * [project @ 2003-03-15 01:50:46 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./threads/f77funcs.c -2 +2 cannam@95: cannam@95: Fri Mar 14 18:23:03 EST 2003 stevenj cannam@95: * [project @ 2003-03-14 23:23:03 by stevenj] cannam@95: added note that FFTW_PATIENT will disable threads if they are not beneficial cannam@95: cannam@95: M ./doc/fftw3.texi -2 +5 cannam@95: cannam@95: Fri Mar 14 18:20:44 EST 2003 stevenj cannam@95: * [project @ 2003-03-14 23:20:44 by stevenj] cannam@95: made fftw_cleanup* more restrictive, in that we don't want to cannam@95: guarantee that previously created plans will still work (they won't, cannam@95: in the case of threaded plans and fftw_cleanup_threads), and there is cannam@95: no reason to provide such a guarantee anyway. cannam@95: cannam@95: M ./doc/fftw3.texi -6 +14 cannam@95: cannam@95: Fri Mar 14 17:23:13 EST 2003 athena cannam@95: * [project @ 2003-03-14 22:23:13 by athena] cannam@95: Moved version.c from kernel/ into api/ cannam@95: cannam@95: A ./api/version.c cannam@95: M ./api/Makefile.am -1 +2 cannam@95: M ./api/version.c +28 cannam@95: M ./kernel/Makefile.am -1 +1 cannam@95: M ./kernel/ifftw.h -4 +1 cannam@95: R ./kernel/version.c cannam@95: cannam@95: Fri Mar 14 17:19:50 EST 2003 athena cannam@95: * [project @ 2003-03-14 22:19:50 by athena] cannam@95: icc-7.0 requires -openmp cannam@95: cannam@95: M ./configure.ac +5 cannam@95: cannam@95: Fri Mar 14 14:47:52 EST 2003 athena cannam@95: * [project @ 2003-03-14 19:47:52 by athena] cannam@95: Ensure that one can do make dist given the distribution cannam@95: cannam@95: M ./doc/Makefile.am -5 +5 cannam@95: cannam@95: Fri Mar 14 14:38:11 EST 2003 athena cannam@95: * [project @ 2003-03-14 19:38:11 by athena] cannam@95: Dist fftw3.pdf, not fftw.pdf cannam@95: cannam@95: M ./doc/Makefile.am -1 +2 cannam@95: cannam@95: Fri Mar 14 14:36:25 EST 2003 athena cannam@95: * [project @ 2003-03-14 19:36:25 by athena] cannam@95: Support -onthreads=%d cannam@95: cannam@95: M ./tests/bench.c -4 +8 cannam@95: cannam@95: Fri Mar 14 14:34:21 EST 2003 stevenj cannam@95: * [project @ 2003-03-14 19:34:21 by stevenj] cannam@95: comment cannam@95: cannam@95: M ./kernel/alloc.c -2 +2 cannam@95: cannam@95: Fri Mar 14 14:33:27 EST 2003 stevenj cannam@95: * [project @ 2003-03-14 19:33:27 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./threads/Makefile.am -1 +3 cannam@95: cannam@95: Fri Mar 14 12:32:18 EST 2003 stevenj cannam@95: * [project @ 2003-03-14 17:32:18 by stevenj] cannam@95: fftw_real is gone cannam@95: cannam@95: M ./doc/rfftwnd.fig -1 +1 cannam@95: cannam@95: Fri Mar 14 12:26:04 EST 2003 stevenj cannam@95: * [project @ 2003-03-14 17:26:04 by stevenj] cannam@95: typos cannam@95: cannam@95: M ./doc/fftw3.texi -3 +3 cannam@95: cannam@95: Fri Mar 14 06:21:43 EST 2003 athena cannam@95: * [project @ 2003-03-14 11:21:43 by athena] cannam@95: More BENCH_DOC strings cannam@95: cannam@95: M ./api/fftw3.h -3 +8 cannam@95: M ./tests/bench.c -4 +11 cannam@95: cannam@95: Fri Mar 14 05:58:53 EST 2003 athena cannam@95: * [project @ 2003-03-14 10:58:53 by athena] cannam@95: Fixed xref's cannam@95: cannam@95: M ./doc/fftw3.texi -15 +15 cannam@95: cannam@95: Fri Mar 14 05:38:26 EST 2003 athena cannam@95: * [project @ 2003-03-14 10:38:26 by athena] cannam@95: Revised manual (esp. intro and tutorial), fixed texinfo hackery cannam@95: for figures. cannam@95: cannam@95: M ./doc/Makefile.am -8 +13 cannam@95: M ./doc/fftw3.texi -165 +280 cannam@95: R ./doc/rfftwnd.gif cannam@95: cannam@95: Wed Mar 12 02:42:33 EST 2003 stevenj cannam@95: * [project @ 2003-03-12 07:42:33 by stevenj] cannam@95: redirect users from guru execute to advanced interface, if possible cannam@95: cannam@95: M ./doc/fftw3.texi -1 +5 cannam@95: cannam@95: Wed Mar 12 02:35:22 EST 2003 stevenj cannam@95: * [project @ 2003-03-12 07:35:22 by stevenj] cannam@95: punctuation cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Wed Mar 12 02:28:51 EST 2003 stevenj cannam@95: * [project @ 2003-03-12 07:28:51 by stevenj] cannam@95: use correct heading level cannam@95: cannam@95: M ./doc/fftw3.texi -7 +7 cannam@95: cannam@95: Wed Mar 12 02:24:37 EST 2003 stevenj cannam@95: * [project @ 2003-03-12 07:24:37 by stevenj] cannam@95: html generation cannam@95: cannam@95: M ./doc/Makefile.am +15 cannam@95: M ./doc/fftw3.texi -91 +90 cannam@95: cannam@95: Wed Mar 12 01:44:00 EST 2003 stevenj cannam@95: * [project @ 2003-03-12 06:44:00 by stevenj] cannam@95: added equation GIFs cannam@95: cannam@95: A ./doc/equation-dft.gif cannam@95: A ./doc/equation-dht.gif cannam@95: A ./doc/equation-idft.gif cannam@95: A ./doc/equation-redft00.gif cannam@95: A ./doc/equation-redft01.gif cannam@95: A ./doc/equation-redft10.gif cannam@95: A ./doc/equation-redft11.gif cannam@95: A ./doc/equation-rodft00.gif cannam@95: A ./doc/equation-rodft01.gif cannam@95: A ./doc/equation-rodft10.gif cannam@95: A ./doc/equation-rodft11.gif cannam@95: M ./doc/equation-dft.gif cannam@95: M ./doc/equation-dht.gif cannam@95: M ./doc/equation-idft.gif cannam@95: M ./doc/equation-redft00.gif cannam@95: M ./doc/equation-redft01.gif cannam@95: M ./doc/equation-redft10.gif cannam@95: M ./doc/equation-redft11.gif cannam@95: M ./doc/equation-rodft00.gif cannam@95: M ./doc/equation-rodft01.gif cannam@95: M ./doc/equation-rodft10.gif cannam@95: M ./doc/equation-rodft11.gif cannam@95: cannam@95: Wed Mar 12 01:43:27 EST 2003 stevenj cannam@95: * [project @ 2003-03-12 06:43:27 by stevenj] cannam@95: punctuation cannam@95: cannam@95: M ./doc/fftw3.texi -18 +18 cannam@95: cannam@95: Wed Mar 12 01:26:46 EST 2003 stevenj cannam@95: * [project @ 2003-03-12 06:26:46 by stevenj] cannam@95: punctuation cannam@95: cannam@95: M ./doc/fftw3.texi -14 +14 cannam@95: cannam@95: Wed Mar 12 01:25:12 EST 2003 stevenj cannam@95: * [project @ 2003-03-12 06:25:12 by stevenj] cannam@95: added multi-dimensional transform definitions cannam@95: cannam@95: M ./doc/fftw3.texi -3 +124 cannam@95: cannam@95: Wed Mar 12 00:14:03 EST 2003 stevenj cannam@95: * [project @ 2003-03-12 05:14:03 by stevenj] cannam@95: slight changes cannam@95: cannam@95: M ./doc/fftw3.texi -3 +5 cannam@95: cannam@95: Wed Mar 12 00:06:34 EST 2003 stevenj cannam@95: * [project @ 2003-03-12 05:06:34 by stevenj] cannam@95: typo cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Tue Mar 11 23:50:43 EST 2003 stevenj cannam@95: * [project @ 2003-03-12 04:50:43 by stevenj] cannam@95: added 1d version of What FFTW Really Computes cannam@95: cannam@95: M ./doc/fftw3.texi -1 +432 cannam@95: cannam@95: Tue Mar 11 21:17:54 EST 2003 stevenj cannam@95: * [project @ 2003-03-12 02:17:54 by stevenj] cannam@95: note in upgrading section about FFTW_PATIENT cannam@95: cannam@95: M ./doc/fftw3.texi -1 +7 cannam@95: cannam@95: Tue Mar 11 15:18:39 EST 2003 stevenj cannam@95: * [project @ 2003-03-11 20:18:39 by stevenj] cannam@95: added cycle-counter section cannam@95: cannam@95: M ./doc/fftw3.texi -1 +32 cannam@95: cannam@95: Tue Mar 11 14:53:44 EST 2003 stevenj cannam@95: * [project @ 2003-03-11 19:53:44 by stevenj] cannam@95: more ideas cannam@95: cannam@95: M ./TODO +7 cannam@95: cannam@95: Mon Mar 10 17:41:35 EST 2003 stevenj cannam@95: * [project @ 2003-03-10 22:41:35 by stevenj] cannam@95: noted that indirect should probably be merged with rank-geq2, to make a rank-split solver cannam@95: cannam@95: M ./dft/indirect.c -1 +4 cannam@95: M ./rdft/indirect.c -1 +4 cannam@95: cannam@95: Fri Mar 7 03:01:52 EST 2003 stevenj cannam@95: * [project @ 2003-03-07 08:01:52 by stevenj] cannam@95: added non-Unix installation instructions cannam@95: cannam@95: M ./doc/fftw3.texi -1 +38 cannam@95: cannam@95: Fri Mar 7 02:30:59 EST 2003 stevenj cannam@95: * [project @ 2003-03-07 07:30:59 by stevenj] cannam@95: also talk about stack alignment with SSE/SSE2 cannam@95: cannam@95: M ./doc/fftw3.texi -1 +7 cannam@95: cannam@95: Fri Mar 7 02:24:07 EST 2003 stevenj cannam@95: * [project @ 2003-03-07 07:24:07 by stevenj] cannam@95: made warning more dire cannam@95: cannam@95: M ./doc/fftw3.texi -3 +4 cannam@95: cannam@95: Fri Mar 7 02:13:25 EST 2003 stevenj cannam@95: * [project @ 2003-03-07 07:13:25 by stevenj] cannam@95: fix cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Fri Mar 7 02:09:55 EST 2003 stevenj cannam@95: * [project @ 2003-03-07 07:09:55 by stevenj] cannam@95: number cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Fri Mar 7 02:09:08 EST 2003 stevenj cannam@95: * [project @ 2003-03-07 07:09:08 by stevenj] cannam@95: fix cannam@95: cannam@95: M ./doc/fftw3.texi -3 +3 cannam@95: cannam@95: Fri Mar 7 02:08:01 EST 2003 stevenj cannam@95: * [project @ 2003-03-07 07:08:01 by stevenj] cannam@95: minor cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Fri Mar 7 02:04:45 EST 2003 stevenj cannam@95: * [project @ 2003-03-07 07:04:45 by stevenj] cannam@95: minor fix cannam@95: cannam@95: M ./doc/fftw3.texi -3 +3 cannam@95: cannam@95: Fri Mar 7 01:58:15 EST 2003 stevenj cannam@95: * [project @ 2003-03-07 06:58:15 by stevenj] cannam@95: cross-ref cannam@95: cannam@95: M ./doc/fftw3.texi -2 +3 cannam@95: cannam@95: Fri Mar 7 01:57:31 EST 2003 stevenj cannam@95: * [project @ 2003-03-07 06:57:31 by stevenj] cannam@95: minor cannam@95: cannam@95: M ./doc/fftw3.texi -7 +8 cannam@95: cannam@95: Fri Mar 7 01:53:28 EST 2003 stevenj cannam@95: * [project @ 2003-03-07 06:53:28 by stevenj] cannam@95: more installation manual cannam@95: cannam@95: M ./doc/fftw3.texi -7 +151 cannam@95: cannam@95: Fri Mar 7 00:43:40 EST 2003 stevenj cannam@95: * [project @ 2003-03-07 05:43:40 by stevenj] cannam@95: GNU-lly correct cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Fri Mar 7 00:38:48 EST 2003 stevenj cannam@95: * [project @ 2003-03-07 05:38:48 by stevenj] cannam@95: started installation section cannam@95: cannam@95: M ./doc/fftw3.texi -3 +33 cannam@95: cannam@95: Fri Mar 7 00:25:02 EST 2003 stevenj cannam@95: * [project @ 2003-03-07 05:25:02 by stevenj] cannam@95: added --without-cycle-counter option as a last resort cannam@95: cannam@95: M ./configure.ac +5 cannam@95: M ./kernel/timer.c -11 +24 cannam@95: cannam@95: Fri Mar 7 00:07:12 EST 2003 stevenj cannam@95: * [project @ 2003-03-07 05:07:12 by stevenj] cannam@95: macros with () arguments were only standardized in C99, and we don't need them anyway cannam@95: cannam@95: M ./kernel/cycle.h -3 +3 cannam@95: cannam@95: Thu Mar 6 23:10:41 EST 2003 stevenj cannam@95: * [project @ 2003-03-07 04:10:41 by stevenj] cannam@95: wording cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Thu Mar 6 23:03:03 EST 2003 stevenj cannam@95: * [project @ 2003-03-07 04:03:03 by stevenj] cannam@95: parallelism cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Thu Mar 6 23:01:47 EST 2003 stevenj cannam@95: * [project @ 2003-03-07 04:01:47 by stevenj] cannam@95: additions to upgrading chapter cannam@95: cannam@95: M ./doc/fftw3.texi -1 +31 cannam@95: cannam@95: Thu Mar 6 22:39:36 EST 2003 stevenj cannam@95: * [project @ 2003-03-07 03:39:36 by stevenj] cannam@95: noted additional humility of FFTW 3 wisdom cannam@95: cannam@95: M ./doc/fftw3.texi -1 +9 cannam@95: cannam@95: Thu Mar 6 22:32:44 EST 2003 stevenj cannam@95: * [project @ 2003-03-07 03:32:44 by stevenj] cannam@95: renaming cannam@95: cannam@95: M ./doc/fftw3.texi -6 +6 cannam@95: cannam@95: Thu Mar 6 22:31:00 EST 2003 stevenj cannam@95: * [project @ 2003-03-07 03:31:00 by stevenj] cannam@95: added placeholder for wisdom reference cannam@95: cannam@95: M ./doc/fftw3.texi -3 +10 cannam@95: cannam@95: Thu Mar 6 22:29:38 EST 2003 stevenj cannam@95: * [project @ 2003-03-07 03:29:38 by stevenj] cannam@95: wrote upgrading chapter cannam@95: cannam@95: M ./doc/fftw3.texi -1 +139 cannam@95: cannam@95: Thu Mar 6 18:01:10 EST 2003 stevenj cannam@95: * [project @ 2003-03-06 23:01:10 by stevenj] cannam@95: slight change cannam@95: cannam@95: M ./doc/fftw3.texi -6 +6 cannam@95: cannam@95: Thu Mar 6 18:00:43 EST 2003 stevenj cannam@95: * [project @ 2003-03-06 23:00:43 by stevenj] cannam@95: placeholder for upgrade chapter cannam@95: cannam@95: M ./doc/fftw3.texi -3 +18 cannam@95: cannam@95: Thu Mar 6 13:47:49 EST 2003 stevenj cannam@95: * [project @ 2003-03-06 18:47:49 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./tools/fftw-wisdom.c -2 +2 cannam@95: cannam@95: Thu Mar 6 13:36:38 EST 2003 stevenj cannam@95: * [project @ 2003-03-06 18:36:38 by stevenj] cannam@95: strengthed warning about time cannam@95: cannam@95: M ./tools/fftw_wisdom.1.in -1 +1 cannam@95: cannam@95: Thu Mar 6 13:35:42 EST 2003 stevenj cannam@95: * [project @ 2003-03-06 18:35:42 by stevenj] cannam@95: noted -t in example cannam@95: cannam@95: M ./tools/fftw_wisdom.1.in -1 +3 cannam@95: cannam@95: Thu Mar 6 13:21:03 EST 2003 stevenj cannam@95: * [project @ 2003-03-06 18:21:03 by stevenj] cannam@95: pay attention to WINDOWS_F77_MANGLING cannam@95: cannam@95: M ./threads/f77api.c -3 +26 cannam@95: cannam@95: Thu Mar 6 02:52:30 EST 2003 stevenj cannam@95: * [project @ 2003-03-06 07:52:30 by stevenj] cannam@95: punctuation cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Thu Mar 6 02:51:02 EST 2003 stevenj cannam@95: * [project @ 2003-03-06 07:51:02 by stevenj] cannam@95: index cannam@95: cannam@95: M ./doc/fftw3.texi -1 +3 cannam@95: cannam@95: Thu Mar 6 02:50:38 EST 2003 stevenj cannam@95: * [project @ 2003-03-06 07:50:38 by stevenj] cannam@95: documented C++ usage cannam@95: cannam@95: M ./doc/fftw3.texi -3 +18 cannam@95: cannam@95: Thu Mar 6 02:25:32 EST 2003 stevenj cannam@95: * [project @ 2003-03-06 07:25:32 by stevenj] cannam@95: got rid of overfull hbox TeX warnings cannam@95: cannam@95: M ./doc/fftw3.texi -6 +6 cannam@95: cannam@95: Thu Mar 6 02:20:38 EST 2003 stevenj cannam@95: * [project @ 2003-03-06 07:20:38 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Thu Mar 6 02:20:13 EST 2003 stevenj cannam@95: * [project @ 2003-03-06 07:20:13 by stevenj] cannam@95: noted fftw_iodim split for Fortran guru interface cannam@95: cannam@95: M ./doc/fftw3.texi -5 +16 cannam@95: cannam@95: Thu Mar 6 02:14:21 EST 2003 stevenj cannam@95: * [project @ 2003-03-06 07:14:21 by stevenj] cannam@95: added guru reference cannam@95: cannam@95: M ./doc/fftw3.texi -3 +187 cannam@95: cannam@95: Wed Mar 5 22:56:05 EST 2003 stevenj cannam@95: * [project @ 2003-03-06 03:56:05 by stevenj] cannam@95: minor cannam@95: cannam@95: M ./doc/fftw3.texi -3 +3 cannam@95: cannam@95: Wed Mar 5 22:45:31 EST 2003 stevenj cannam@95: * [project @ 2003-03-06 03:45:31 by stevenj] cannam@95: use @r{...} for comment text in code examples cannam@95: cannam@95: M ./doc/fftw3.texi -3 +3 cannam@95: cannam@95: Wed Mar 5 13:14:04 EST 2003 stevenj cannam@95: * [project @ 2003-03-05 18:14:04 by stevenj] cannam@95: eliminate warning cannam@95: cannam@95: M ./simd/sse.c -1 +2 cannam@95: cannam@95: Wed Mar 5 13:12:56 EST 2003 stevenj cannam@95: * [project @ 2003-03-05 18:12:56 by stevenj] cannam@95: SIMD_CFLAGS only for simd code cannam@95: cannam@95: M ./configure.ac -5 +8 cannam@95: M ./dft/simd/Makefile.am -1 +1 cannam@95: M ./dft/simd/codelets/Makefile.am -1 +1 cannam@95: M ./kernel/align.c -12 +9 cannam@95: M ./simd/Makefile.am +1 cannam@95: cannam@95: Wed Mar 5 11:06:41 EST 2003 athena cannam@95: * [project @ 2003-03-05 16:06:41 by athena] cannam@95: Minor changes. cannam@95: cannam@95: M ./doc/fftw3.texi -89 +96 cannam@95: cannam@95: Wed Mar 5 02:13:34 EST 2003 stevenj cannam@95: * [project @ 2003-03-05 07:13:34 by stevenj] cannam@95: cross-compiling with MinGW can't detect f77 mangling, so add an option to use what seems to be the most common styles cannam@95: cannam@95: M ./api/f77api.c -3 +34 cannam@95: M ./configure.ac +5 cannam@95: cannam@95: Tue Mar 4 20:00:31 EST 2003 stevenj cannam@95: * [project @ 2003-03-05 01:00:31 by stevenj] cannam@95: comment cannam@95: cannam@95: M ./libbench2/util.c -2 +3 cannam@95: cannam@95: Tue Mar 4 20:00:13 EST 2003 stevenj cannam@95: * [project @ 2003-03-05 01:00:13 by stevenj] cannam@95: we only use our-malloc-16 on machines where size_t == uintptr_t, so don't bother doing the right thing with the benchmark cannam@95: cannam@95: M ./libbench2/util.c -1 +1 cannam@95: cannam@95: Tue Mar 4 19:46:09 EST 2003 stevenj cannam@95: * [project @ 2003-03-05 00:46:09 by stevenj] cannam@95: support WITH_OUR_MALLOC16 cannam@95: cannam@95: M ./libbench2/util.c -2 +19 cannam@95: cannam@95: Tue Mar 4 18:50:53 EST 2003 fftw cannam@95: * [project @ 2003-03-04 23:50:53 by fftw] cannam@95: automatically add -msse etcetera for --enable-sse etcetera cannam@95: cannam@95: M ./configure.ac +22 cannam@95: cannam@95: Tue Mar 4 18:24:26 EST 2003 fftw cannam@95: * [project @ 2003-03-04 23:24:26 by fftw] cannam@95: got rid of const warning cannam@95: cannam@95: M ./tools/fftw-wisdom.c -2 +2 cannam@95: cannam@95: Tue Mar 4 18:22:48 EST 2003 fftw cannam@95: * [project @ 2003-03-04 23:22:48 by fftw] cannam@95: missing header cannam@95: cannam@95: M ./libbench2/problem.c -1 +2 cannam@95: cannam@95: Tue Mar 4 15:55:47 EST 2003 stevenj cannam@95: * [project @ 2003-03-04 20:55:47 by stevenj] cannam@95: fixes cannam@95: cannam@95: M ./doc/fftw3.texi -15 +49 cannam@95: cannam@95: Tue Mar 4 15:53:26 EST 2003 stevenj cannam@95: * [project @ 2003-03-04 20:53:26 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./api/import-system-wisdom.c -1 +1 cannam@95: cannam@95: Tue Mar 4 02:22:14 EST 2003 stevenj cannam@95: * [project @ 2003-03-04 07:22:14 by stevenj] cannam@95: started guru reference cannam@95: cannam@95: M ./doc/fftw3.texi -1 +110 cannam@95: cannam@95: Tue Mar 4 01:44:09 EST 2003 stevenj cannam@95: * [project @ 2003-03-04 06:44:09 by stevenj] cannam@95: use same FFTW_IODIM between precisions cannam@95: cannam@95: M ./api/fftw3.h -6 +8 cannam@95: cannam@95: Tue Mar 4 00:25:57 EST 2003 stevenj cannam@95: * [project @ 2003-03-04 05:25:57 by stevenj] cannam@95: renamed section cannam@95: cannam@95: M ./doc/fftw3.texi -5 +5 cannam@95: cannam@95: Tue Mar 4 00:21:49 EST 2003 stevenj cannam@95: * [project @ 2003-03-04 05:21:49 by stevenj] cannam@95: no need for "advanced" in subheadings cannam@95: cannam@95: M ./doc/fftw3.texi -4 +4 cannam@95: cannam@95: Tue Mar 4 00:20:05 EST 2003 stevenj cannam@95: * [project @ 2003-03-04 05:20:05 by stevenj] cannam@95: typo cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Tue Mar 4 00:17:23 EST 2003 stevenj cannam@95: * [project @ 2003-03-04 05:17:23 by stevenj] cannam@95: finished advanced interface cannam@95: cannam@95: M ./doc/fftw3.texi -1 +34 cannam@95: cannam@95: Mon Mar 3 23:26:12 EST 2003 stevenj cannam@95: * [project @ 2003-03-04 04:26:12 by stevenj] cannam@95: more advance interface docs cannam@95: cannam@95: M ./doc/fftw3.texi -1 +46 cannam@95: cannam@95: Mon Mar 3 23:12:09 EST 2003 stevenj cannam@95: * [project @ 2003-03-04 04:12:09 by stevenj] cannam@95: fail for win32 cannam@95: cannam@95: M ./api/import-system-wisdom.c -2 +2 cannam@95: cannam@95: Mon Mar 3 17:18:48 EST 2003 fftw cannam@95: * [project @ 2003-03-03 22:18:48 by fftw] cannam@95: shortened help string cannam@95: cannam@95: M ./configure.ac -1 +1 cannam@95: cannam@95: Mon Mar 3 17:16:17 EST 2003 fftw cannam@95: * [project @ 2003-03-03 22:16:17 by fftw] cannam@95: fixed cross-refs cannam@95: cannam@95: M ./doc/fftw3.texi -3 +3 cannam@95: cannam@95: Mon Mar 3 17:07:27 EST 2003 fftw cannam@95: * [project @ 2003-03-03 22:07:27 by fftw] cannam@95: FFTW_POSSIBLY_UNALIGNED -> simpler FFTW_UNALIGNED in API, added bench option cannam@95: cannam@95: M ./api/fftw3.h -2 +2 cannam@95: M ./api/mapflags.c -1 +1 cannam@95: M ./doc/fftw3.texi -3 +3 cannam@95: M ./tests/bench.c +1 cannam@95: cannam@95: Mon Mar 3 16:58:07 EST 2003 fftw cannam@95: * [project @ 2003-03-03 21:58:07 by fftw] cannam@95: whoops cannam@95: cannam@95: M ./kernel/alloc.c -3 +3 cannam@95: cannam@95: Mon Mar 3 16:52:58 EST 2003 fftw cannam@95: * [project @ 2003-03-03 21:52:58 by fftw] cannam@95: noted assumption cannam@95: cannam@95: M ./kernel/alloc.c -2 +2 cannam@95: cannam@95: Mon Mar 3 16:50:33 EST 2003 fftw cannam@95: * [project @ 2003-03-03 21:50:33 by fftw] cannam@95: provide our own malloc16 routine because of Windows lossage cannam@95: cannam@95: M ./configure.ac +5 cannam@95: M ./kernel/alloc.c -3 +35 cannam@95: cannam@95: Mon Mar 3 13:28:12 EST 2003 stevenj cannam@95: * [project @ 2003-03-03 18:28:12 by stevenj] cannam@95: capitalization cannam@95: cannam@95: M ./doc/fftw3.texi -6 +6 cannam@95: cannam@95: Mon Mar 3 13:26:32 EST 2003 stevenj cannam@95: * [project @ 2003-03-03 18:26:32 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Mon Mar 3 12:55:57 EST 2003 stevenj cannam@95: * [project @ 2003-03-03 17:55:57 by stevenj] cannam@95: vertical skip looks better than indenting for setting off short paragraphs cannam@95: cannam@95: M ./doc/fftw3.texi -1 +6 cannam@95: cannam@95: Mon Mar 3 06:34:09 EST 2003 athena cannam@95: * [project @ 2003-03-03 11:34:09 by athena] cannam@95: Removed franz-mode. Automake was distributing franz files cannam@95: whether franz mode was enabled or not. cannam@95: cannam@95: M ./configure.ac -3 cannam@95: M ./dft/simd/codelets/Makefile.am -7 +2 cannam@95: cannam@95: Mon Mar 3 01:44:00 EST 2003 stevenj cannam@95: * [project @ 2003-03-03 06:44:00 by stevenj] cannam@95: made output boundary conditions more prominent; they are important, cannam@95: because they make the different transform types inequivalent in cannam@95: parity cannam@95: cannam@95: M ./doc/fftw3.texi -4 +8 cannam@95: cannam@95: Mon Mar 3 01:17:28 EST 2003 stevenj cannam@95: * [project @ 2003-03-03 06:17:28 by stevenj] cannam@95: clarification cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Mon Mar 3 01:17:07 EST 2003 stevenj cannam@95: * [project @ 2003-03-03 06:17:07 by stevenj] cannam@95: typo cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Mon Mar 3 01:10:28 EST 2003 stevenj cannam@95: * [project @ 2003-03-03 06:10:28 by stevenj] cannam@95: started advanced reference cannam@95: cannam@95: M ./doc/fftw3.texi -6 +51 cannam@95: cannam@95: Mon Mar 3 00:52:02 EST 2003 stevenj cannam@95: * [project @ 2003-03-03 05:52:02 by stevenj] cannam@95: r2r reference cannam@95: cannam@95: M ./doc/fftw3.texi -2 +209 cannam@95: cannam@95: Sun Mar 2 23:51:21 EST 2003 stevenj cannam@95: * [project @ 2003-03-03 04:51:21 by stevenj] cannam@95: workaround for info formatting bug cannam@95: cannam@95: M ./doc/fftw3.texi -5 +4 cannam@95: cannam@95: Sun Mar 2 23:47:19 EST 2003 stevenj cannam@95: * [project @ 2003-03-03 04:47:19 by stevenj] cannam@95: noted lack of fftw_malloc in Fortran cannam@95: cannam@95: M ./doc/fftw3.texi -3 +9 cannam@95: cannam@95: Sun Mar 2 23:42:52 EST 2003 stevenj cannam@95: * [project @ 2003-03-03 04:42:32 by stevenj] cannam@95: parallelism cannam@95: cannam@95: M ./doc/fftw3.texi -4 +4 cannam@95: cannam@95: Sun Mar 2 23:39:54 EST 2003 stevenj cannam@95: * [project @ 2003-03-03 04:39:05 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./doc/fftw3.texi -7 +7 cannam@95: cannam@95: Sun Mar 2 23:33:02 EST 2003 stevenj cannam@95: * [project @ 2003-03-03 04:33:02 by stevenj] cannam@95: r2c/c2r reference cannam@95: cannam@95: M ./doc/fftw3.texi -40 +248 cannam@95: cannam@95: Sun Mar 2 22:44:10 EST 2003 stevenj cannam@95: * [project @ 2003-03-03 03:44:10 by stevenj] cannam@95: table of contents was being included twice cannam@95: cannam@95: M ./doc/fftw3.texi -2 +1 cannam@95: cannam@95: Sun Mar 2 22:42:29 EST 2003 stevenj cannam@95: * [project @ 2003-03-03 03:42:29 by stevenj] cannam@95: minor changes cannam@95: cannam@95: M ./doc/fftw3.texi -15 +16 cannam@95: cannam@95: Sun Mar 2 21:54:13 EST 2003 stevenj cannam@95: * [project @ 2003-03-03 02:54:13 by stevenj] cannam@95: started reference section cannam@95: cannam@95: M ./doc/fftw3.texi -2 +238 cannam@95: cannam@95: Sun Mar 2 19:10:02 EST 2003 stevenj cannam@95: * [project @ 2003-03-03 00:10:02 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./doc/Makefile.am -2 +2 cannam@95: cannam@95: Sun Mar 2 19:03:23 EST 2003 stevenj cannam@95: * [project @ 2003-03-03 00:03:23 by stevenj] cannam@95: started ref. section cannam@95: cannam@95: M ./doc/fftw3.texi -2 +146 cannam@95: cannam@95: Sun Mar 2 18:50:58 EST 2003 stevenj cannam@95: * [project @ 2003-03-02 23:50:58 by stevenj] cannam@95: fftw_flops takes const plan cannam@95: cannam@95: M ./api/fftw3.h -2 +2 cannam@95: M ./api/flops.c -1 +1 cannam@95: cannam@95: Sun Mar 2 15:54:14 EST 2003 stevenj cannam@95: * [project @ 2003-03-02 20:54:14 by stevenj] cannam@95: typo cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Sun Mar 2 15:52:41 EST 2003 stevenj cannam@95: * [project @ 2003-03-02 20:52:41 by stevenj] cannam@95: added "Wisdom of Fortran?" section cannam@95: cannam@95: M ./doc/fftw3.texi -1 +39 cannam@95: cannam@95: Sun Mar 2 15:50:37 EST 2003 stevenj cannam@95: * [project @ 2003-03-02 20:50:37 by stevenj] cannam@95: typo cannam@95: cannam@95: M ./doc/f77_wisdom.f -2 +2 cannam@95: cannam@95: Sun Mar 2 15:49:57 EST 2003 stevenj cannam@95: * [project @ 2003-03-02 20:49:57 by stevenj] cannam@95: wording cannam@95: cannam@95: M ./doc/f77_wisdom.f -1 +1 cannam@95: cannam@95: Sun Mar 2 15:46:13 EST 2003 stevenj cannam@95: * [project @ 2003-03-02 20:46:13 by stevenj] cannam@95: added comments cannam@95: cannam@95: M ./doc/f77_wisdom.f +25 cannam@95: cannam@95: Sun Mar 2 15:44:01 EST 2003 stevenj cannam@95: * [project @ 2003-03-02 20:44:01 by stevenj] cannam@95: added example file cannam@95: cannam@95: A ./doc/f77_wisdom.f cannam@95: cannam@95: Sun Mar 2 15:37:32 EST 2003 stevenj cannam@95: * [project @ 2003-03-02 20:37:32 by stevenj] cannam@95: don't print out READ WISDOM unless we have cannam@95: cannam@95: M ./tests/bench.c -6 +11 cannam@95: cannam@95: Sun Mar 2 15:36:28 EST 2003 stevenj cannam@95: * [project @ 2003-03-02 20:36:28 by stevenj] cannam@95: EOF is not a space cannam@95: cannam@95: M ./kernel/scan.c -2 +2 cannam@95: cannam@95: Sun Mar 2 09:14:37 EST 2003 athena cannam@95: * [project @ 2003-03-02 14:14:37 by athena] cannam@95: Turn on inline by default cannam@95: cannam@95: M ./kernel/ifftw.h -2 +3 cannam@95: cannam@95: Sun Mar 2 07:11:56 EST 2003 athena cannam@95: * [project @ 2003-03-02 12:11:56 by athena] cannam@95: Optionally inline loop in notw codelets cannam@95: cannam@95: M ./genfft/gen_hc2r.ml -3 +3 cannam@95: M ./genfft/gen_notw.ml -3 +3 cannam@95: M ./genfft/gen_notw_c.ml -3 +3 cannam@95: M ./genfft/gen_r2hc.ml -3 +3 cannam@95: M ./kernel/ifftw.h -1 +3 cannam@95: cannam@95: Sun Mar 2 01:37:41 EST 2003 stevenj cannam@95: * [project @ 2003-03-02 06:37:41 by stevenj] cannam@95: updated nodes cannam@95: cannam@95: M ./doc/fftw3.texi -5 +9 cannam@95: cannam@95: Sun Mar 2 01:37:19 EST 2003 stevenj cannam@95: * [project @ 2003-03-02 06:37:19 by stevenj] cannam@95: wrote most of Fortran chapter cannam@95: cannam@95: M ./doc/fftw3.texi -1 +208 cannam@95: cannam@95: Sun Mar 2 00:58:37 EST 2003 stevenj cannam@95: * [project @ 2003-03-02 05:58:37 by stevenj] cannam@95: citation cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Sun Mar 2 00:57:22 EST 2003 stevenj cannam@95: * [project @ 2003-03-02 05:57:22 by stevenj] cannam@95: added parallel FFTW chapter cannam@95: cannam@95: M ./doc/fftw3.texi -3 +209 cannam@95: cannam@95: Sat Mar 1 20:42:23 EST 2003 stevenj cannam@95: * [project @ 2003-03-02 01:42:23 by stevenj] cannam@95: typo cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Sat Mar 1 20:34:38 EST 2003 stevenj cannam@95: * [project @ 2003-03-02 01:34:38 by stevenj] cannam@95: added inlining to TODO cannam@95: cannam@95: M ./TODO +2 cannam@95: cannam@95: Sat Mar 1 19:36:26 EST 2003 stevenj cannam@95: * [project @ 2003-03-02 00:36:26 by stevenj] cannam@95: added K cannam@95: cannam@95: M ./CONVENTIONS -1 +2 cannam@95: cannam@95: Sat Mar 1 19:15:18 EST 2003 stevenj cannam@95: * [project @ 2003-03-02 00:15:18 by stevenj] cannam@95: use K for constants cannam@95: cannam@95: M ./dft/zero.c -3 +3 cannam@95: M ./kernel/trig1.c -5 +5 cannam@95: M ./rdft/generic.c -8 +8 cannam@95: M ./rdft/problem.c -3 +3 cannam@95: M ./rdft/rdft-dht.c -3 +3 cannam@95: M ./rdft/rdft2-radix2.c -11 +11 cannam@95: M ./reodft/redft00e-r2hc.c -3 +3 cannam@95: M ./reodft/reodft010e-r2hc.c -11 +11 cannam@95: M ./reodft/reodft11e-r2hc.c -7 +7 cannam@95: M ./reodft/reodft11e-radix2.c -11 +11 cannam@95: M ./reodft/rodft00e-r2hc.c -3 +3 cannam@95: cannam@95: Sat Mar 1 19:14:54 EST 2003 stevenj cannam@95: * [project @ 2003-03-02 00:14:54 by stevenj] cannam@95: fixed cross-ref cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Sat Mar 1 19:14:16 EST 2003 stevenj cannam@95: * [project @ 2003-03-02 00:14:16 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./doc/fftw3.texi -6 +6 cannam@95: cannam@95: Sat Mar 1 18:50:43 EST 2003 stevenj cannam@95: * [project @ 2003-03-01 23:50:43 by stevenj] cannam@95: cleanup cannam@95: cannam@95: M ./doc/fftw3.texi -8 +8 cannam@95: cannam@95: Sat Mar 1 18:46:38 EST 2003 stevenj cannam@95: * [project @ 2003-03-01 23:46:38 by stevenj] cannam@95: "words of wisdom" by itself is a little too obscure cannam@95: cannam@95: M ./doc/fftw3.texi -6 +6 cannam@95: cannam@95: Sat Mar 1 18:43:21 EST 2003 stevenj cannam@95: * [project @ 2003-03-01 23:43:21 by stevenj] cannam@95: re-added multi-dimensional array stuff cannam@95: cannam@95: M ./doc/fftw3.texi -1 +203 cannam@95: cannam@95: Sat Mar 1 18:15:22 EST 2003 stevenj cannam@95: * [project @ 2003-03-01 23:15:22 by stevenj] cannam@95: added alignment section cannam@95: cannam@95: M ./doc/fftw3.texi -4 +93 cannam@95: cannam@95: Sat Mar 1 16:34:21 EST 2003 stevenj cannam@95: * [project @ 2003-03-01 21:34:21 by stevenj] cannam@95: shrunk code cannam@95: cannam@95: M ./reodft/reodft11e-r2hc-odd.c -147 +57 cannam@95: cannam@95: Fri Feb 28 20:22:00 EST 2003 stevenj cannam@95: * [project @ 2003-03-01 01:22:00 by stevenj] cannam@95: slight compression cannam@95: cannam@95: M ./reodft/reodft11e-r2hc-odd.c -146 +122 cannam@95: cannam@95: Fri Feb 28 19:01:20 EST 2003 stevenj cannam@95: * [project @ 2003-03-01 00:01:20 by stevenj] cannam@95: style cannam@95: cannam@95: M ./doc/fftw3.texi -7 +7 cannam@95: M ./reodft/reodft11e-radix2.c -2 +2 cannam@95: cannam@95: Fri Feb 28 18:46:53 EST 2003 stevenj cannam@95: * [project @ 2003-02-28 23:46:53 by stevenj] cannam@95: noted not in API cannam@95: cannam@95: M ./CONVENTIONS -1 +1 cannam@95: cannam@95: Fri Feb 28 18:43:14 EST 2003 stevenj cannam@95: * [project @ 2003-02-28 23:43:14 by stevenj] cannam@95: more updates cannam@95: cannam@95: M ./CONVENTIONS -3 +4 cannam@95: cannam@95: Fri Feb 28 18:38:42 EST 2003 stevenj cannam@95: * [project @ 2003-02-28 23:38:42 by stevenj] cannam@95: slight updates cannam@95: cannam@95: M ./CONVENTIONS -3 +4 cannam@95: cannam@95: Fri Feb 28 18:28:58 EST 2003 stevenj cannam@95: * [project @ 2003-02-28 23:28:58 by stevenj] cannam@95: great const-ification of apply/solve and print cannam@95: cannam@95: M ./api/f77funcs.c -1 +1 cannam@95: M ./api/fftw3.h -2 +2 cannam@95: M ./api/print-plan.c -1 +1 cannam@95: M ./dft/buffered.c -5 +5 cannam@95: M ./dft/ct-dif.c -3 +3 cannam@95: M ./dft/ct-dit.c -3 +3 cannam@95: M ./dft/ct-ditbuf.c -3 +3 cannam@95: M ./dft/ct-ditf.c -3 +3 cannam@95: M ./dft/ct.c -3 +3 cannam@95: M ./dft/dft.h -3 +3 cannam@95: M ./dft/direct.c -5 +5 cannam@95: M ./dft/generic.c -4 +4 cannam@95: M ./dft/indirect.c -7 +7 cannam@95: M ./dft/nop.c -3 +3 cannam@95: M ./dft/rader.c -12 +12 cannam@95: M ./dft/rank-geq2.c -5 +5 cannam@95: M ./dft/rank0.c -12 +12 cannam@95: M ./dft/solve.c -3 +3 cannam@95: M ./dft/vrank-geq1.c -5 +5 cannam@95: M ./dft/vrank2-transpose.c -5 +5 cannam@95: M ./dft/vrank3-transpose.c -5 +5 cannam@95: M ./kernel/ifftw.h -3 +3 cannam@95: M ./rdft/buffered.c -5 +5 cannam@95: M ./rdft/buffered2.c -7 +7 cannam@95: M ./rdft/dft-r2hc.c -5 +5 cannam@95: M ./rdft/dht-r2hc.c -5 +5 cannam@95: M ./rdft/dht-rader.c -4 +4 cannam@95: M ./rdft/direct.c -7 +7 cannam@95: M ./rdft/direct2.c -7 +7 cannam@95: M ./rdft/generic.c -6 +6 cannam@95: M ./rdft/hc2hc-buf.c -5 +5 cannam@95: M ./rdft/hc2hc-dif.c -3 +3 cannam@95: M ./rdft/hc2hc-dit.c -3 +3 cannam@95: M ./rdft/hc2hc.c -3 +3 cannam@95: M ./rdft/indirect.c -7 +7 cannam@95: M ./rdft/nop.c -3 +3 cannam@95: M ./rdft/nop2.c -3 +3 cannam@95: M ./rdft/rader-hc2hc.c -6 +6 cannam@95: M ./rdft/rank-geq2-rdft2.c -7 +7 cannam@95: M ./rdft/rank-geq2.c -5 +5 cannam@95: M ./rdft/rank0-rdft2.c -9 +9 cannam@95: M ./rdft/rank0.c -8 +8 cannam@95: M ./rdft/rdft-dht.c -9 +9 cannam@95: M ./rdft/rdft.h -5 +5 cannam@95: M ./rdft/rdft2-radix2.c -12 +12 cannam@95: M ./rdft/solve.c -3 +3 cannam@95: M ./rdft/solve2.c -3 +3 cannam@95: M ./rdft/vrank-geq1-rdft2.c -7 +7 cannam@95: M ./rdft/vrank-geq1.c -5 +5 cannam@95: M ./rdft/vrank2-transpose.c -5 +5 cannam@95: M ./rdft/vrank3-transpose.c -5 +5 cannam@95: M ./reodft/redft00e-r2hc.c -5 +5 cannam@95: M ./reodft/reodft010e-r2hc.c -11 +11 cannam@95: M ./reodft/reodft11e-r2hc-odd.c -7 +7 cannam@95: M ./reodft/reodft11e-r2hc.c -7 +7 cannam@95: M ./reodft/reodft11e-radix2.c -9 +9 cannam@95: M ./reodft/rodft00e-r2hc.c -5 +5 cannam@95: M ./threads/ct-dit.c -3 +3 cannam@95: M ./threads/dft-vrank-geq1.c -5 +5 cannam@95: M ./threads/hc2hc-dif.c -3 +3 cannam@95: M ./threads/hc2hc-dit.c -3 +3 cannam@95: M ./threads/rdft-vrank-geq1.c -5 +5 cannam@95: M ./threads/vrank-geq1-rdft2.c -7 +7 cannam@95: cannam@95: Fri Feb 28 17:51:15 EST 2003 stevenj cannam@95: * [project @ 2003-02-28 22:51:15 by stevenj] cannam@95: make fftw_execute take a const plan, to remind the user that it is re-entrant (or should be)... cannam@95: cannam@95: M ./api/execute-dft-c2r.c -1 +1 cannam@95: M ./api/execute-dft-r2c.c -1 +1 cannam@95: M ./api/execute-dft.c -1 +1 cannam@95: M ./api/execute-r2r.c -1 +1 cannam@95: M ./api/execute.c -1 +1 cannam@95: M ./api/f77funcs.c -5 +6 cannam@95: M ./api/fftw3.h -6 +6 cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Fri Feb 28 17:29:40 EST 2003 stevenj cannam@95: * [project @ 2003-02-28 22:29:40 by stevenj] cannam@95: weakening cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Fri Feb 28 17:28:48 EST 2003 stevenj cannam@95: * [project @ 2003-02-28 22:28:48 by stevenj] cannam@95: note cannam@95: cannam@95: M ./doc/fftw3.texi -4 +4 cannam@95: cannam@95: Fri Feb 28 17:27:10 EST 2003 stevenj cannam@95: * [project @ 2003-02-28 22:27:10 by stevenj] cannam@95: footnote about why DHT is provided cannam@95: cannam@95: M ./doc/fftw3.texi -7 +12 cannam@95: cannam@95: Fri Feb 28 15:07:03 EST 2003 stevenj cannam@95: * [project @ 2003-02-28 20:07:03 by stevenj] cannam@95: index cannam@95: cannam@95: M ./doc/fftw3.texi -1 +2 cannam@95: cannam@95: Fri Feb 28 15:05:48 EST 2003 stevenj cannam@95: * [project @ 2003-02-28 20:05:48 by stevenj] cannam@95: added DHT tutorial cannam@95: cannam@95: M ./doc/fftw3.texi -2 +32 cannam@95: cannam@95: Fri Feb 28 14:36:45 EST 2003 stevenj cannam@95: * [project @ 2003-02-28 19:36:45 by stevenj] cannam@95: fixed O(n log n) cannam@95: cannam@95: M ./doc/fftw3.texi -4 +16 cannam@95: cannam@95: Fri Feb 28 14:12:15 EST 2003 stevenj cannam@95: * [project @ 2003-02-28 19:08:03 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./doc/fftw3.texi -4 +10 cannam@95: cannam@95: Fri Feb 28 14:06:22 EST 2003 stevenj cannam@95: * [project @ 2003-02-28 19:06:22 by stevenj] cannam@95: slight improvements cannam@95: cannam@95: M ./doc/fftw3.texi -2 +12 cannam@95: cannam@95: Fri Feb 28 00:55:50 EST 2003 stevenj cannam@95: * [project @ 2003-02-28 05:55:50 by stevenj] cannam@95: addition cannam@95: cannam@95: M ./doc/fftw3.texi -3 +5 cannam@95: cannam@95: Fri Feb 28 00:54:09 EST 2003 stevenj cannam@95: * [project @ 2003-02-28 05:54:09 by stevenj] cannam@95: clarification cannam@95: cannam@95: M ./doc/fftw3.texi -7 +8 cannam@95: cannam@95: Thu Feb 27 23:49:37 EST 2003 stevenj cannam@95: * [project @ 2003-02-28 04:49:37 by stevenj] cannam@95: fix cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Thu Feb 27 23:43:56 EST 2003 stevenj cannam@95: * [project @ 2003-02-28 04:43:56 by stevenj] cannam@95: slight changes cannam@95: cannam@95: M ./doc/fftw3.texi -15 +21 cannam@95: cannam@95: Thu Feb 27 23:27:48 EST 2003 stevenj cannam@95: * [project @ 2003-02-28 04:27:48 by stevenj] cannam@95: added R{E,O}DFTab tutorial cannam@95: cannam@95: M ./doc/fftw3.texi -9 +105 cannam@95: cannam@95: Thu Feb 27 17:24:20 EST 2003 stevenj cannam@95: * [project @ 2003-02-27 22:24:20 by stevenj] cannam@95: fixes cannam@95: cannam@95: M ./doc/fftw3.texi -12 +11 cannam@95: cannam@95: Thu Feb 27 17:20:42 EST 2003 stevenj cannam@95: * [project @ 2003-02-27 22:20:42 by stevenj] cannam@95: fixes cannam@95: cannam@95: M ./doc/fftw3.texi -24 +26 cannam@95: cannam@95: Thu Feb 27 17:11:54 EST 2003 stevenj cannam@95: * [project @ 2003-02-27 22:11:22 by stevenj] cannam@95: slight change cannam@95: cannam@95: M ./doc/fftw3.texi -4 +4 cannam@95: cannam@95: Thu Feb 27 17:07:45 EST 2003 stevenj cannam@95: * [project @ 2003-02-27 22:07:45 by stevenj] cannam@95: documented r2hc/hc2r cannam@95: cannam@95: M ./doc/fftw3.texi -6 +86 cannam@95: cannam@95: Thu Feb 27 16:19:16 EST 2003 stevenj cannam@95: * [project @ 2003-02-27 21:19:16 by stevenj] cannam@95: minor changes cannam@95: cannam@95: M ./doc/fftw3.texi -24 +26 cannam@95: cannam@95: Thu Feb 27 13:54:06 EST 2003 stevenj cannam@95: * [project @ 2003-02-27 18:54:06 by stevenj] cannam@95: timed planner and unifying radix-2 butterfly loops are not critical for release cannam@95: cannam@95: M ./TODO -8 +8 cannam@95: cannam@95: Thu Feb 27 13:51:20 EST 2003 stevenj cannam@95: * [project @ 2003-02-27 18:51:20 by stevenj] cannam@95: reodft/verify.c no longer exists cannam@95: cannam@95: M ./TODO -2 cannam@95: cannam@95: Thu Feb 27 13:44:19 EST 2003 stevenj cannam@95: * [project @ 2003-02-27 18:44:19 by stevenj] cannam@95: optimization: REDFT00 of size 2 is same as R2HC cannam@95: cannam@95: M ./rdft/problem.c -1 +5 cannam@95: cannam@95: Thu Feb 27 12:35:33 EST 2003 stevenj cannam@95: * [project @ 2003-02-27 17:35:33 by stevenj] cannam@95: R{E,O}DFT01 of size-1 is identity cannam@95: cannam@95: M ./rdft/problem.c -3 +4 cannam@95: cannam@95: Thu Feb 27 12:15:10 EST 2003 stevenj cannam@95: * [project @ 2003-02-27 17:15:10 by stevenj] cannam@95: minor simplification cannam@95: cannam@95: M ./reodft/reodft11e-r2hc-odd.c -31 +31 cannam@95: cannam@95: Thu Feb 27 02:46:31 EST 2003 stevenj cannam@95: * [project @ 2003-02-27 07:46:31 by stevenj] cannam@95: fixed add count cannam@95: cannam@95: M ./reodft/reodft11e-r2hc-odd.c -2 +2 cannam@95: cannam@95: Thu Feb 27 02:25:04 EST 2003 stevenj cannam@95: * [project @ 2003-02-27 07:25:04 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./reodft/reodft11e-r2hc-odd.c -5 +5 cannam@95: cannam@95: Thu Feb 27 02:22:03 EST 2003 stevenj cannam@95: * [project @ 2003-02-27 07:22:03 by stevenj] cannam@95: another optimization cannam@95: cannam@95: M ./reodft/reodft11e-r2hc-odd.c -53 +53 cannam@95: cannam@95: Thu Feb 27 01:43:00 EST 2003 stevenj cannam@95: * [project @ 2003-02-27 06:43:00 by stevenj] cannam@95: added op counts cannam@95: cannam@95: M ./reodft/reodft11e-r2hc-odd.c -2 +4 cannam@95: M ./reodft/reodft11e-radix2.c -2 +9 cannam@95: cannam@95: Thu Feb 27 01:29:32 EST 2003 stevenj cannam@95: * [project @ 2003-02-27 06:29:32 by stevenj] cannam@95: cleanup cannam@95: cannam@95: M ./reodft/reodft11e-r2hc-odd.c -53 +55 cannam@95: cannam@95: Thu Feb 27 01:17:23 EST 2003 stevenj cannam@95: * [project @ 2003-02-27 06:17:23 by stevenj] cannam@95: typo in comment cannam@95: cannam@95: M ./reodft/reodft11e-r2hc-odd.c -2 +2 cannam@95: cannam@95: Thu Feb 27 01:13:49 EST 2003 stevenj cannam@95: * [project @ 2003-02-27 06:13:49 by stevenj] cannam@95: fixed comment cannam@95: cannam@95: M ./reodft/reodft11e-r2hc-odd.c -2 +2 cannam@95: cannam@95: Thu Feb 27 01:12:05 EST 2003 stevenj cannam@95: * [project @ 2003-02-27 06:12:05 by stevenj] cannam@95: use E instead of R cannam@95: cannam@95: M ./reodft/reodft11e-r2hc-odd.c -9 +9 cannam@95: cannam@95: Thu Feb 27 01:05:39 EST 2003 stevenj cannam@95: * [project @ 2003-02-27 06:05:39 by stevenj] cannam@95: more unrolling to eliminate if statements in loops, for speedups of 25-40% cannam@95: cannam@95: M ./reodft/reodft11e-r2hc-odd.c -25 +28 cannam@95: cannam@95: Thu Feb 27 00:27:00 EST 2003 stevenj cannam@95: * [project @ 2003-02-27 05:27:00 by stevenj] cannam@95: some loop splitting to touch each element of output buf only once and eliminate some conditionals...speeds up by 30-40% cannam@95: cannam@95: M ./reodft/reodft11e-r2hc-odd.c -31 +167 cannam@95: cannam@95: Wed Feb 26 17:48:26 EST 2003 stevenj cannam@95: * [project @ 2003-02-26 22:48:26 by stevenj] cannam@95: comma cannam@95: cannam@95: M ./reodft/reodft11e-r2hc-odd.c -2 +2 cannam@95: cannam@95: Wed Feb 26 17:46:17 EST 2003 stevenj cannam@95: * [project @ 2003-02-26 22:46:17 by stevenj] cannam@95: pointer to odd case cannam@95: cannam@95: M ./reodft/reodft11e-radix2.c -1 +3 cannam@95: cannam@95: Wed Feb 26 17:40:54 EST 2003 stevenj cannam@95: * [project @ 2003-02-26 22:40:54 by stevenj] cannam@95: precision -> accuracy (c.f. Kahan) cannam@95: cannam@95: M ./reodft/reodft11e-r2hc.c -2 +2 cannam@95: cannam@95: Wed Feb 26 17:36:13 EST 2003 stevenj cannam@95: * [project @ 2003-02-26 22:36:13 by stevenj] cannam@95: added time limit for wisdom generation cannam@95: cannam@95: M ./Makefile.am -1 +4 cannam@95: M ./libbench2/bench-user.h -1 +2 cannam@95: M ./libbench2/problem.c -1 +4 cannam@95: M ./tools/fftw-wisdom.c -16 +59 cannam@95: M ./tools/fftw_wisdom.1.in +6 cannam@95: cannam@95: Wed Feb 26 13:24:36 EST 2003 stevenj cannam@95: * [project @ 2003-02-26 18:24:36 by stevenj] cannam@95: caps cannam@95: cannam@95: M ./reodft/reodft11e-r2hc-odd.c -2 +2 cannam@95: cannam@95: Tue Feb 25 20:56:01 EST 2003 stevenj cannam@95: * [project @ 2003-02-26 01:56:01 by stevenj] cannam@95: another note cannam@95: cannam@95: M ./reodft/reodft11e-r2hc-odd.c -2 +3 cannam@95: cannam@95: Tue Feb 25 20:54:57 EST 2003 stevenj cannam@95: * [project @ 2003-02-26 01:54:57 by stevenj] cannam@95: note cannam@95: cannam@95: M ./reodft/reodft11e-r2hc-odd.c -4 +5 cannam@95: cannam@95: Tue Feb 25 20:42:08 EST 2003 stevenj cannam@95: * [project @ 2003-02-26 01:42:08 by stevenj] cannam@95: added new, more accurate (hopefully) reodft11 algorithms; added --disable-debug-malloc; added --impulse-accuracy-rounds=rounds flags to libbench2 for impulse-response accuracy tests cannam@95: cannam@95: A ./reodft/reodft11e-r2hc-odd.c cannam@95: A ./reodft/reodft11e-radix2.c cannam@95: M ./configure.ac +8 cannam@95: M ./kernel/alloc.c -2 +2 cannam@95: M ./kernel/ifftw.h -3 +3 cannam@95: M ./libbench2/bench-main.c -2 +8 cannam@95: M ./libbench2/bench.h -2 +2 cannam@95: M ./libbench2/verify-dft.c -3 +4 cannam@95: M ./libbench2/verify-lib.c -18 +76 cannam@95: M ./libbench2/verify-r2r.c -7 +39 cannam@95: M ./libbench2/verify-rdft2.c -3 +4 cannam@95: M ./libbench2/verify.c -7 +10 cannam@95: M ./libbench2/verify.h -4 +7 cannam@95: M ./reodft/Makefile.am -1 +4 cannam@95: M ./reodft/conf.c -2 +4 cannam@95: M ./reodft/reodft.h -1 +3 cannam@95: M ./reodft/reodft11e-r2hc-odd.c +273 cannam@95: M ./reodft/reodft11e-r2hc.c -2 +8 cannam@95: M ./reodft/reodft11e-radix2.c +506 cannam@95: M ./tests/bench.c -1 +1 cannam@95: cannam@95: Sun Feb 23 14:07:48 EST 2003 athena cannam@95: * [project @ 2003-02-23 19:07:48 by athena] cannam@95: fftw_wisdom.1 is in $builddir, not $srcdir cannam@95: cannam@95: M ./tools/Makefile.am -1 +1 cannam@95: cannam@95: Mon Feb 17 03:42:19 EST 2003 stevenj cannam@95: * [project @ 2003-02-17 08:42:19 by stevenj] cannam@95: pde cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Mon Feb 17 03:40:19 EST 2003 stevenj cannam@95: * [project @ 2003-02-17 08:40:19 by stevenj] cannam@95: consistent number cannam@95: cannam@95: M ./doc/fftw3.texi -4 +4 cannam@95: cannam@95: Mon Feb 17 03:39:02 EST 2003 stevenj cannam@95: * [project @ 2003-02-17 08:39:02 by stevenj] cannam@95: started r2r doc cannam@95: cannam@95: M ./doc/fftw3.texi -1 +65 cannam@95: cannam@95: Mon Feb 17 02:31:51 EST 2003 stevenj cannam@95: * [project @ 2003-02-17 07:31:50 by stevenj] cannam@95: rfftwnd cannam@95: cannam@95: A ./doc/rfftwnd.fig cannam@95: A ./doc/rfftwnd.gif cannam@95: M ./doc/Makefile.am +11 cannam@95: M ./doc/fftw3.texi -26 +147 cannam@95: M ./doc/rfftwnd.fig +1148 cannam@95: M ./doc/rfftwnd.gif cannam@95: cannam@95: Sat Feb 15 17:02:07 EST 2003 stevenj cannam@95: * [project @ 2003-02-15 22:02:07 by stevenj] cannam@95: continued cannam@95: cannam@95: M ./doc/fftw3.texi -4 +32 cannam@95: cannam@95: Sat Feb 15 15:16:26 EST 2003 stevenj cannam@95: * [project @ 2003-02-15 20:16:26 by stevenj] cannam@95: started r2c/c2r docs cannam@95: cannam@95: M ./doc/fftw3.texi -8 +87 cannam@95: cannam@95: Sat Feb 15 01:12:52 EST 2003 stevenj cannam@95: * [project @ 2003-02-15 06:12:52 by stevenj] cannam@95: added r{e,o}dft11 accuracy test cannam@95: cannam@95: M ./libbench2/verify-r2r.c -23 +45 cannam@95: cannam@95: Sat Feb 15 00:42:48 EST 2003 stevenj cannam@95: * [project @ 2003-02-15 05:42:48 by stevenj] cannam@95: added more r2r accuracy checks cannam@95: cannam@95: M ./libbench2/verify-dft.c -2 +2 cannam@95: M ./libbench2/verify-lib.c -10 +4 cannam@95: M ./libbench2/verify-r2r.c -1 +127 cannam@95: M ./libbench2/verify-rdft2.c -2 +2 cannam@95: M ./libbench2/verify.h -2 +3 cannam@95: cannam@95: Fri Feb 14 19:19:54 EST 2003 athena cannam@95: * [project @ 2003-02-15 00:19:54 by athena] cannam@95: $< is a GNUism cannam@95: cannam@95: M ./tools/Makefile.am -1 +1 cannam@95: cannam@95: Wed Feb 12 21:02:16 EST 2003 stevenj cannam@95: * [project @ 2003-02-13 02:02:16 by stevenj] cannam@95: r2r test cases are in cannam@95: cannam@95: M ./TODO -1 +1 cannam@95: cannam@95: Wed Feb 12 21:01:28 EST 2003 stevenj cannam@95: * [project @ 2003-02-13 02:01:28 by stevenj] cannam@95: added vector radix to TODO cannam@95: cannam@95: M ./TODO +2 cannam@95: cannam@95: Wed Feb 12 17:21:33 EST 2003 stevenj cannam@95: * [project @ 2003-02-12 22:21:33 by stevenj] cannam@95: fixed cross-ref cannam@95: cannam@95: M ./tools/fftw_wisdom.1.in -1 +1 cannam@95: cannam@95: Wed Feb 12 17:19:56 EST 2003 stevenj cannam@95: * [project @ 2003-02-12 22:19:56 by stevenj] cannam@95: shorter synopsis cannam@95: cannam@95: M ./tools/fftw_wisdom.1.in -1 +1 cannam@95: cannam@95: Wed Feb 12 12:53:19 EST 2003 stevenj cannam@95: * [project @ 2003-02-12 17:53:19 by stevenj] cannam@95: obsolete cannam@95: cannam@95: R ./tests/debug.h cannam@95: cannam@95: Wed Feb 12 12:52:53 EST 2003 stevenj cannam@95: * [project @ 2003-02-12 17:52:53 by stevenj] cannam@95: removed old dotens cannam@95: cannam@95: R ./tests/dotens.c cannam@95: R ./tests/dotens2.c cannam@95: cannam@95: Wed Feb 12 12:52:16 EST 2003 stevenj cannam@95: * [project @ 2003-02-12 17:52:16 by stevenj] cannam@95: removed old verify files cannam@95: cannam@95: R ./tests/verify-dft.c cannam@95: R ./tests/verify-lib.c cannam@95: R ./tests/verify-rdft.c cannam@95: R ./tests/verify-reodft.c cannam@95: R ./tests/verify.h cannam@95: cannam@95: Wed Feb 12 12:37:17 EST 2003 stevenj cannam@95: * [project @ 2003-02-12 17:37:17 by stevenj] cannam@95: disable threads support by default cannam@95: cannam@95: M ./tools/fftw-wisdom.c -4 +11 cannam@95: M ./tools/fftw_wisdom.1.in -7 cannam@95: cannam@95: Wed Feb 12 11:03:28 EST 2003 athena cannam@95: * [project @ 2003-02-12 16:03:28 by athena] cannam@95: Removed old test program cannam@95: cannam@95: M ./tests/bench.c -390 cannam@95: cannam@95: Tue Feb 11 22:30:55 EST 2003 stevenj cannam@95: * [project @ 2003-02-12 03:30:55 by stevenj] cannam@95: joke cannam@95: cannam@95: M ./tools/fftw-wisdom-to-conf.in -1 +1 cannam@95: cannam@95: Tue Feb 11 22:27:44 EST 2003 stevenj cannam@95: * [project @ 2003-02-12 03:27:44 by stevenj] cannam@95: add --help and --version, to be GNU-lly correct cannam@95: cannam@95: M ./tools/fftw-wisdom-to-conf.1 +7 cannam@95: M ./tools/fftw-wisdom-to-conf.in +36 cannam@95: cannam@95: Tue Feb 11 22:27:18 EST 2003 stevenj cannam@95: * [project @ 2003-02-12 03:27:18 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./tools/fftw_wisdom.1.in -1 cannam@95: cannam@95: Tue Feb 11 22:17:35 EST 2003 stevenj cannam@95: * [project @ 2003-02-12 03:15:03 by stevenj] cannam@95: better help cannam@95: cannam@95: M ./tools/fftw-wisdom.c -10 +43 cannam@95: cannam@95: Tue Feb 11 21:47:35 EST 2003 stevenj cannam@95: * [project @ 2003-02-12 02:47:35 by stevenj] cannam@95: comma cannam@95: cannam@95: M ./tools/fftw-wisdom-to-conf.1 -2 +1 cannam@95: cannam@95: Tue Feb 11 21:46:12 EST 2003 stevenj cannam@95: * [project @ 2003-02-12 02:46:12 by stevenj] cannam@95: formatting cannam@95: cannam@95: M ./tools/fftw-wisdom-to-conf.1 -1 +1 cannam@95: cannam@95: Tue Feb 11 21:45:23 EST 2003 stevenj cannam@95: * [project @ 2003-02-12 02:45:23 by stevenj] cannam@95: man pages for tools cannam@95: cannam@95: A ./tools/fftw-wisdom-to-conf.1 cannam@95: A ./tools/fftw_wisdom.1.in cannam@95: M ./configure.ac +1 cannam@95: M ./tools/Makefile.am -1 +7 cannam@95: M ./tools/fftw-wisdom-to-conf.1 +85 cannam@95: M ./tools/fftw_wisdom.1.in +190 cannam@95: cannam@95: Tue Feb 11 19:07:12 EST 2003 stevenj cannam@95: * [project @ 2003-02-12 00:07:12 by stevenj] cannam@95: added -V cannam@95: cannam@95: M ./tools/fftw-wisdom.c +25 cannam@95: cannam@95: Tue Feb 11 18:42:17 EST 2003 stevenj cannam@95: * [project @ 2003-02-11 23:42:17 by stevenj] cannam@95: added install-wisdom target cannam@95: cannam@95: M ./Makefile.am +10 cannam@95: cannam@95: Tue Feb 11 18:23:02 EST 2003 stevenj cannam@95: * [project @ 2003-02-11 23:23:02 by stevenj] cannam@95: another note cannam@95: cannam@95: M ./NEWS -1 +2 cannam@95: cannam@95: Tue Feb 11 17:32:56 EST 2003 stevenj cannam@95: * [project @ 2003-02-11 22:32:56 by stevenj] cannam@95: started r2r accuracy tests (only three kinds covered so far) cannam@95: cannam@95: M ./libbench2/verify-dft.c -2 +2 cannam@95: M ./libbench2/verify-lib.c -10 +13 cannam@95: M ./libbench2/verify-r2r.c +103 cannam@95: M ./libbench2/verify-rdft2.c -2 +2 cannam@95: M ./libbench2/verify.c -2 +2 cannam@95: M ./libbench2/verify.h -1 +4 cannam@95: cannam@95: Mon Feb 10 22:04:18 EST 2003 stevenj cannam@95: * [project @ 2003-02-11 03:04:18 by stevenj] cannam@95: silence warning cannam@95: cannam@95: M ./kernel/ifftw.h -3 +3 cannam@95: cannam@95: Mon Feb 10 20:55:20 EST 2003 athena cannam@95: * [project @ 2003-02-11 01:55:20 by athena] cannam@95: gcc bug is now avoided. cannam@95: cannam@95: M ./TODO -5 cannam@95: cannam@95: Mon Feb 10 20:37:54 EST 2003 athena cannam@95: * [project @ 2003-02-11 01:37:54 by athena] cannam@95: Accuracy test cannam@95: cannam@95: A ./libbench2/mp.c cannam@95: M ./libbench2/Makefile.am -5 +6 cannam@95: M ./libbench2/bench-user.h -7 +1 cannam@95: M ./libbench2/mp.c +592 cannam@95: M ./libbench2/verify-dft.c -1 +22 cannam@95: M ./libbench2/verify-lib.c -16 +39 cannam@95: M ./libbench2/verify-r2r.c -17 +2 cannam@95: M ./libbench2/verify-rdft2.c -1 +23 cannam@95: M ./libbench2/verify.c -2 +21 cannam@95: M ./libbench2/verify.h +27 cannam@95: cannam@95: Mon Feb 10 07:59:57 EST 2003 athena cannam@95: * [project @ 2003-02-10 12:59:57 by athena] cannam@95: There is no point in precomputing strides for the long-double code, as cannam@95: multiplication by sizeof(long double) cannot be folded into the cannam@95: addressing mode. This change also fixes the gcc-2.95 bug that causes cannam@95: miscompilation of certain codelets. cannam@95: cannam@95: M ./kernel/ifftw.h -2 +2 cannam@95: cannam@95: Mon Feb 10 02:54:35 EST 2003 stevenj cannam@95: * [project @ 2003-02-10 07:54:35 by stevenj] cannam@95: added random r2r tests cannam@95: cannam@95: M ./tests/check.pl -3 +19 cannam@95: cannam@95: Mon Feb 10 02:44:58 EST 2003 stevenj cannam@95: * [project @ 2003-02-10 07:44:58 by stevenj] cannam@95: whoops, bugfix: missing stride for ro10 cannam@95: cannam@95: M ./reodft/reodft010e-r2hc.c -2 +2 cannam@95: cannam@95: Mon Feb 10 02:21:50 EST 2003 stevenj cannam@95: * [project @ 2003-02-10 07:21:50 by stevenj] cannam@95: formatting cannam@95: cannam@95: M ./api/mapflags.c -1 +2 cannam@95: cannam@95: Sun Feb 9 23:24:52 EST 2003 stevenj cannam@95: * [project @ 2003-02-10 04:24:52 by stevenj] cannam@95: flop counts for reodft cannam@95: cannam@95: M ./reodft/redft00e-r2hc.c -5 +12 cannam@95: M ./reodft/reodft010e-r2hc.c -3 +16 cannam@95: M ./reodft/reodft11e-r2hc.c -3 +10 cannam@95: M ./reodft/rodft00e-r2hc.c -3 +12 cannam@95: cannam@95: Sun Feb 9 23:22:15 EST 2003 stevenj cannam@95: * [project @ 2003-02-10 04:22:15 by stevenj] cannam@95: declare aligned_main cannam@95: cannam@95: M ./libbench2/bench.h -1 +2 cannam@95: cannam@95: Sun Feb 9 20:56:06 EST 2003 stevenj cannam@95: * [project @ 2003-02-10 01:56:06 by stevenj] cannam@95: corrected rader op counts cannam@95: cannam@95: M ./rdft/dht-rader.c -6 +5 cannam@95: M ./rdft/rader-hc2hc.c -2 +2 cannam@95: cannam@95: Sun Feb 9 20:25:32 EST 2003 stevenj cannam@95: * [project @ 2003-02-10 01:25:32 by stevenj] cannam@95: punctuation cannam@95: cannam@95: M ./TODO -1 +1 cannam@95: cannam@95: Sun Feb 9 20:25:17 EST 2003 stevenj cannam@95: * [project @ 2003-02-10 01:25:17 by stevenj] cannam@95: noted need for better estimator cannam@95: cannam@95: M ./TODO +4 cannam@95: cannam@95: Sun Feb 9 19:58:59 EST 2003 stevenj cannam@95: * [project @ 2003-02-10 00:58:59 by stevenj] cannam@95: noted F77 api fix for g77 mangling incompatibility cannam@95: cannam@95: M ./NEWS +3 cannam@95: cannam@95: Sun Feb 9 19:30:55 EST 2003 stevenj cannam@95: * [project @ 2003-02-10 00:30:55 by stevenj] cannam@95: build f77 header file of constants from fftw3.h cannam@95: cannam@95: M ./api/Makefile.am -1 +10 cannam@95: cannam@95: Sun Feb 9 19:04:53 EST 2003 stevenj cannam@95: * [project @ 2003-02-10 00:04:53 by stevenj] cannam@95: updates cannam@95: cannam@95: M ./TODO -7 +7 cannam@95: cannam@95: Sun Feb 9 19:03:34 EST 2003 stevenj cannam@95: * [project @ 2003-02-10 00:03:34 by stevenj] cannam@95: threads f77 api cannam@95: cannam@95: A ./api/x77.h cannam@95: A ./threads/f77api.c cannam@95: A ./threads/f77funcs.c cannam@95: M ./api/Makefile.am -1 +1 cannam@95: M ./api/f77api.c -12 +1 cannam@95: M ./api/x77.h +32 cannam@95: M ./threads/Makefile.am -1 +1 cannam@95: M ./threads/f77api.c +39 cannam@95: M ./threads/f77funcs.c +39 cannam@95: cannam@95: Sun Feb 9 18:54:00 EST 2003 stevenj cannam@95: * [project @ 2003-02-09 23:54:00 by stevenj] cannam@95: finished f77 serial api cannam@95: cannam@95: M ./api/f77api.c +15 cannam@95: M ./api/f77funcs.c +220 cannam@95: cannam@95: Sun Feb 9 18:32:26 EST 2003 stevenj cannam@95: * [project @ 2003-02-09 23:32:26 by stevenj] cannam@95: added flops, slight cleanups cannam@95: cannam@95: M ./api/f77api.c -8 +5 cannam@95: M ./api/f77funcs.c -21 +27 cannam@95: cannam@95: Sun Feb 9 18:11:48 EST 2003 athena cannam@95: * [project @ 2003-02-09 23:11:48 by athena] cannam@95: Oops, forgot #include cannam@95: cannam@95: M ./libbench2/aligned-main.c -1 +3 cannam@95: cannam@95: Sun Feb 9 18:08:26 EST 2003 athena cannam@95: * [project @ 2003-02-09 23:08:26 by athena] cannam@95: Removed duplication of stack-alignment code cannam@95: cannam@95: A ./libbench2/aligned-main.c cannam@95: M ./libbench2/Makefile.am -5 +5 cannam@95: M ./libbench2/aligned-main.c +81 cannam@95: M ./libbench2/bench-main.c -62 +2 cannam@95: M ./libbench2/main.c -2 +2 cannam@95: M ./tools/fftw-wisdom.c -61 +1 cannam@95: cannam@95: Sun Feb 9 15:48:15 EST 2003 stevenj cannam@95: * [project @ 2003-02-09 20:48:15 by stevenj] cannam@95: allow - to read problems from stdin cannam@95: cannam@95: M ./tools/fftw-wisdom.c -2 +14 cannam@95: cannam@95: Sun Feb 9 15:22:23 EST 2003 stevenj cannam@95: * [project @ 2003-02-09 20:22:23 by stevenj] cannam@95: added fftw-wisdom tool cannam@95: cannam@95: A ./tools/fftw-wisdom.c cannam@95: M ./tools/Makefile.am +15 cannam@95: M ./tools/fftw-wisdom.c +298 cannam@95: cannam@95: Sun Feb 9 15:06:38 EST 2003 stevenj cannam@95: * [project @ 2003-02-09 20:06:38 by stevenj] cannam@95: elim. warning cannam@95: cannam@95: M ./tests/bench.c -1 +1 cannam@95: cannam@95: Sun Feb 9 14:24:19 EST 2003 stevenj cannam@95: * [project @ 2003-02-09 19:24:19 by stevenj] cannam@95: destroy_input should not contaminate flags of other problems cannam@95: cannam@95: M ./tests/bench.c -1 +5 cannam@95: cannam@95: Sun Feb 9 13:06:11 EST 2003 stevenj cannam@95: * [project @ 2003-02-09 18:06:11 by stevenj] cannam@95: updated cannam@95: cannam@95: M ./ChangeLog -53 +1740 cannam@95: cannam@95: Sun Feb 9 13:01:45 EST 2003 stevenj cannam@95: * [project @ 2003-02-09 18:01:44 by stevenj] cannam@95: removed overzealous inplace check, which caused problems for rdft2 cannam@95: cannam@95: M ./dft/rank-geq2.c -12 +1 cannam@95: M ./rdft/rank-geq2-rdft2.c -3 +2 cannam@95: M ./rdft/rank-geq2.c -12 +1 cannam@95: cannam@95: Sun Feb 9 08:14:03 EST 2003 athena cannam@95: * [project @ 2003-02-09 13:14:03 by athena] cannam@95: Consistent syntax for RNK_MINFTY tensors cannam@95: cannam@95: M ./kernel/tensor.c -4 +4 cannam@95: cannam@95: Sun Feb 9 07:31:13 EST 2003 athena cannam@95: * [project @ 2003-02-09 12:31:13 by athena] cannam@95: lisply-correct tensor print. We no longer need to parse tensors. cannam@95: cannam@95: M ./kernel/tensor.c -3 +9 cannam@95: cannam@95: Sun Feb 9 03:35:56 EST 2003 stevenj cannam@95: * [project @ 2003-02-09 08:35:56 by stevenj] cannam@95: removed completed items cannam@95: cannam@95: M ./TODO -4 cannam@95: cannam@95: Sun Feb 9 03:27:56 EST 2003 stevenj cannam@95: * [project @ 2003-02-09 08:27:56 by stevenj] cannam@95: slight renaming cannam@95: cannam@95: M ./libbench2/verify-r2r.c -4 +4 cannam@95: cannam@95: Sun Feb 9 03:15:28 EST 2003 stevenj cannam@95: * [project @ 2003-02-09 08:15:28 by stevenj] cannam@95: multi-dimensional r2r verifier cannam@95: cannam@95: M ./libbench2/problem.c -2 +2 cannam@95: M ./libbench2/verify-r2r.c -81 +118 cannam@95: cannam@95: Sun Feb 9 02:40:22 EST 2003 stevenj cannam@95: * [project @ 2003-02-09 07:40:22 by stevenj] cannam@95: comments cannam@95: cannam@95: M ./libbench2/verify-r2r.c -1 +3 cannam@95: cannam@95: Sun Feb 9 02:38:26 EST 2003 stevenj cannam@95: * [project @ 2003-02-09 07:38:26 by stevenj] cannam@95: slight simplification cannam@95: cannam@95: M ./libbench2/verify-r2r.c -7 +2 cannam@95: cannam@95: Sun Feb 9 02:36:25 EST 2003 stevenj cannam@95: * [project @ 2003-02-09 07:36:25 by stevenj] cannam@95: added 1d r2r verifier (triple ugh) cannam@95: cannam@95: A ./libbench2/verify-r2r.c cannam@95: M ./libbench2/Makefile.am -2 +2 cannam@95: M ./libbench2/allocate.c -1 +17 cannam@95: M ./libbench2/bench-user.h -2 +10 cannam@95: M ./libbench2/mflops.c +1 cannam@95: M ./libbench2/problem.c -7 +68 cannam@95: M ./libbench2/verify-r2r.c +616 cannam@95: M ./libbench2/verify.c -2 +2 cannam@95: M ./libbench2/zero.c -1 +4 cannam@95: M ./tests/bench.c +101 cannam@95: cannam@95: Sat Feb 8 22:23:00 EST 2003 stevenj cannam@95: * [project @ 2003-02-09 03:23:00 by stevenj] cannam@95: added vector transforms to random tests cannam@95: cannam@95: M ./tests/check.pl -3 +11 cannam@95: cannam@95: Sat Feb 8 20:59:07 EST 2003 stevenj cannam@95: * [project @ 2003-02-09 01:59:07 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./rdft/direct2.c -2 +1 cannam@95: cannam@95: Sat Feb 8 19:52:58 EST 2003 stevenj cannam@95: * [project @ 2003-02-09 00:52:58 by stevenj] cannam@95: fixed interaction between dwims for sz/vecsz with rdft2 transforms cannam@95: cannam@95: M ./libbench2/problem.c -15 +17 cannam@95: cannam@95: Sat Feb 8 19:35:56 EST 2003 stevenj cannam@95: * [project @ 2003-02-09 00:35:56 by stevenj] cannam@95: added destroy_input flag/check cannam@95: cannam@95: M ./libbench2/bench-user.h -1 +2 cannam@95: M ./libbench2/problem.c -1 +3 cannam@95: M ./libbench2/verify-dft.c -1 +4 cannam@95: M ./libbench2/verify-lib.c -1 +35 cannam@95: M ./libbench2/verify-rdft2.c -1 +5 cannam@95: M ./libbench2/verify.h +2 cannam@95: M ./tests/bench.c +7 cannam@95: cannam@95: Sat Feb 8 19:11:58 EST 2003 stevenj cannam@95: * [project @ 2003-02-09 00:11:57 by stevenj] cannam@95: added rdft2 verifier cannam@95: cannam@95: A ./api/extract-reim.c cannam@95: A ./libbench2/aset.c cannam@95: A ./libbench2/verify-rdft2.c cannam@95: M ./api/Makefile.am -3 +3 cannam@95: R ./api/dfthelp.c cannam@95: M ./api/extract-reim.c +36 cannam@95: M ./api/plan-guru-dft-c2r.c -1 +2 cannam@95: M ./api/plan-many-dft-c2r.c +2 cannam@95: M ./libbench2/Makefile.am -5 +5 cannam@95: M ./libbench2/allocate.c -7 +44 cannam@95: M ./libbench2/aset.c +10 cannam@95: M ./libbench2/bench-user.h -1 +4 cannam@95: M ./libbench2/bench.h -1 +2 cannam@95: M ./libbench2/problem.c -10 +42 cannam@95: M ./libbench2/tensor.c -1 +47 cannam@95: M ./libbench2/verify-dft.c -1 +3 cannam@95: M ./libbench2/verify-lib.c -16 +13 cannam@95: M ./libbench2/verify-rdft2.c +255 cannam@95: M ./libbench2/verify.c -2 +2 cannam@95: M ./libbench2/verify.h -1 +1 cannam@95: M ./libbench2/zero.c -2 +8 cannam@95: M ./tests/bench.c -8 +189 cannam@95: M ./tests/check.pl -1 +7 cannam@95: cannam@95: Sat Feb 8 13:31:14 EST 2003 stevenj cannam@95: * [project @ 2003-02-08 18:31:14 by stevenj] cannam@95: an additional check for in-place case cannam@95: cannam@95: M ./rdft/rdft2-radix2.c -3 +7 cannam@95: cannam@95: Fri Feb 7 17:36:56 EST 2003 stevenj cannam@95: * [project @ 2003-02-07 22:36:56 by stevenj] cannam@95: slight fix: hc2r constraints are mostly determined by sub-plan cannam@95: cannam@95: M ./rdft/rank0-rdft2.c -6 +7 cannam@95: cannam@95: Fri Feb 7 16:28:55 EST 2003 stevenj cannam@95: * [project @ 2003-02-07 21:28:55 by stevenj] cannam@95: make radix2-dft inapplicable to in-place/split case (r == rio, iio >= rio + n/2+1 != r + 1) cannam@95: cannam@95: M ./rdft/rdft2-radix2.c -3 +22 cannam@95: cannam@95: Tue Feb 4 06:36:29 EST 2003 athena cannam@95: * [project @ 2003-02-04 11:36:29 by athena] cannam@95: Allow plnr->hook to be 0 cannam@95: cannam@95: M ./kernel/planner.c -12 +12 cannam@95: M ./tests/hook.c -7 +2 cannam@95: cannam@95: Tue Feb 4 03:25:36 EST 2003 stevenj cannam@95: * [project @ 2003-02-04 08:25:36 by stevenj] cannam@95: moved dft stuff into verify-dft cannam@95: cannam@95: M ./libbench2/bench-user.h -4 +2 cannam@95: M ./libbench2/verify-dft.c -9 +84 cannam@95: M ./libbench2/verify.c -84 +2 cannam@95: cannam@95: Tue Feb 4 03:25:00 EST 2003 stevenj cannam@95: * [project @ 2003-02-04 08:25:00 by stevenj] cannam@95: cruft cannam@95: cannam@95: M ./tests/hook.c -1 cannam@95: cannam@95: Tue Feb 4 03:18:28 EST 2003 stevenj cannam@95: * [project @ 2003-02-04 08:18:28 by stevenj] cannam@95: further unify libbench2 and paranoid verifiers cannam@95: cannam@95: M ./libbench2/bench-user.h -1 +6 cannam@95: M ./libbench2/problem.c -1 +2 cannam@95: M ./libbench2/verify.c -16 +21 cannam@95: M ./tests/bench.c -3 +5 cannam@95: M ./tests/hook.c -125 +58 cannam@95: cannam@95: Sun Feb 2 01:45:37 EST 2003 stevenj cannam@95: * [project @ 2003-02-02 06:45:37 by stevenj] cannam@95: typo in comment cannam@95: cannam@95: M ./api/import-wisdom-from-file.c -1 +1 cannam@95: cannam@95: Sat Feb 1 09:30:03 EST 2003 athena cannam@95: * [project @ 2003-02-01 14:30:03 by athena] cannam@95: Fixed p==2 case cannam@95: cannam@95: M ./kernel/primes.c -1 +4 cannam@95: cannam@95: Sat Feb 1 09:23:43 EST 2003 athena cannam@95: * [project @ 2003-02-01 14:23:43 by athena] cannam@95: Incorporated new find_generator by Greg Dionne. cannam@95: cannam@95: M ./kernel/primes.c -21 +35 cannam@95: cannam@95: Fri Jan 31 20:46:24 EST 2003 athena cannam@95: * [project @ 2003-02-01 01:46:24 by athena] cannam@95: Removed nonportable call to gettext() cannam@95: cannam@95: M ./libbench2/getopt.c +7 cannam@95: cannam@95: Wed Jan 29 19:03:43 EST 2003 athena cannam@95: * [project @ 2003-01-30 00:03:43 by athena] cannam@95: uintptr_t is in in openbsd cannam@95: cannam@95: M ./kernel/ifftw.h -1 +5 cannam@95: cannam@95: Wed Jan 29 15:41:56 EST 2003 athena cannam@95: * [project @ 2003-01-29 20:41:56 by athena] cannam@95: Huge speedups in wisdom I/O. cannam@95: cannam@95: M ./api/export-wisdom-to-string.c -2 +2 cannam@95: M ./api/export-wisdom.c -1 +1 cannam@95: M ./api/import-wisdom-from-file.c -4 +27 cannam@95: M ./api/mkprinter-file.c -5 +25 cannam@95: M ./kernel/debug.c -2 +2 cannam@95: M ./kernel/ifftw.h -2 +5 cannam@95: M ./kernel/planner.c -5 +9 cannam@95: M ./kernel/print.c -2 +7 cannam@95: M ./kernel/scan.c -4 +25 cannam@95: M ./tests/bench.c -1 +5 cannam@95: cannam@95: Tue Jan 28 19:36:51 EST 2003 athena cannam@95: * [project @ 2003-01-29 00:36:51 by athena] cannam@95: Added appropriate warning against likely future bug. cannam@95: cannam@95: M ./kernel/planner.c -1 +4 cannam@95: cannam@95: Tue Jan 28 19:00:24 EST 2003 athena cannam@95: * [project @ 2003-01-29 00:00:24 by athena] cannam@95: Don't attempt to remove bogus wisdom entries. cannam@95: cannam@95: M ./kernel/planner.c -10 +3 cannam@95: cannam@95: Tue Jan 28 18:16:24 EST 2003 athena cannam@95: * [project @ 2003-01-28 23:16:24 by athena] cannam@95: Fixed a couple of very very very nasty bugs---pointers became cannam@95: invalid after the hash table was relocated. cannam@95: cannam@95: M ./kernel/planner.c -19 +29 cannam@95: cannam@95: Tue Jan 28 07:34:10 EST 2003 athena cannam@95: * [project @ 2003-01-28 12:34:10 by athena] cannam@95: Read wisdom at can_do() time, otherwise wisdom is destroyed. cannam@95: cannam@95: M ./tests/bench.c -2 +5 cannam@95: cannam@95: Tue Jan 28 06:54:38 EST 2003 athena cannam@95: * [project @ 2003-01-28 11:54:38 by athena] cannam@95: More conservative inheritance of blessings cannam@95: cannam@95: M ./kernel/planner.c -19 +24 cannam@95: cannam@95: Tue Jan 28 06:50:20 EST 2003 athena cannam@95: * [project @ 2003-01-28 11:50:20 by athena] cannam@95: Print the same info as it is hashed cannam@95: cannam@95: M ./dft/problem.c -3 +4 cannam@95: cannam@95: Tue Jan 28 06:49:48 EST 2003 athena cannam@95: * [project @ 2003-01-28 11:49:48 by athena] cannam@95: Print name of executable when FAILURE cannam@95: cannam@95: M ./tests/check.pl -2 +2 cannam@95: cannam@95: Mon Jan 27 06:59:40 EST 2003 athena cannam@95: * [project @ 2003-01-27 11:59:40 by athena] cannam@95: New NO_SEARCH planner flag, which avoids searching altogether. cannam@95: A wisdom entry must lead to a NO_SEARCH-grade plan, or else the cannam@95: wisdom entry is bogus. cannam@95: cannam@95: M ./kernel/ifftw.h -5 +10 cannam@95: M ./kernel/planner.c -13 +30 cannam@95: cannam@95: Sun Jan 26 20:45:21 EST 2003 athena cannam@95: * [project @ 2003-01-27 01:45:21 by athena] cannam@95: Use cosl()/sinl() when appropriate cannam@95: cannam@95: M ./libbench2/verify-lib.c -3 +20 cannam@95: cannam@95: Sun Jan 26 16:29:18 EST 2003 athena cannam@95: * [project @ 2003-01-26 21:29:18 by athena] cannam@95: Use null pointers when estimating. The estimator should never cannam@95: time anything. cannam@95: cannam@95: M ./kernel/planner.c -1 +6 cannam@95: M ./libbench2/problem.c -2 +1 cannam@95: M ./libbench2/speed.c -1 +2 cannam@95: M ./libbench2/verify.c -1 +3 cannam@95: cannam@95: Sun Jan 26 15:19:01 EST 2003 stevenj cannam@95: * [project @ 2003-01-26 20:19:01 by stevenj] cannam@95: note cannam@95: cannam@95: M ./api/f77api.c -1 +1 cannam@95: cannam@95: Sun Jan 26 15:16:22 EST 2003 stevenj cannam@95: * [project @ 2003-01-26 20:16:22 by stevenj] cannam@95: support multiple mangling schemes with g77 cannam@95: cannam@95: A ./api/f77funcs.c cannam@95: M ./api/Makefile.am +4 cannam@95: M ./api/f77api.c -130 +41 cannam@95: M ./api/f77funcs.c +139 cannam@95: M ./configure.ac +6 cannam@95: cannam@95: Sun Jan 26 12:58:57 EST 2003 stevenj cannam@95: * [project @ 2003-01-26 17:58:57 by stevenj] cannam@95: fixed verbose, made random tests only use selected rank, use rank <= 4, fixed final flush_problems call cannam@95: cannam@95: M ./tests/check.pl -4 +9 cannam@95: cannam@95: Sun Jan 26 12:42:49 EST 2003 stevenj cannam@95: * [project @ 2003-01-26 17:42:49 by stevenj] cannam@95: fixed typo (count instead of maxcount) cannam@95: cannam@95: M ./tests/check.pl -2 +2 cannam@95: cannam@95: Sun Jan 26 12:12:07 EST 2003 stevenj cannam@95: * [project @ 2003-01-26 17:12:07 by stevenj] cannam@95: hypot is no longer used cannam@95: cannam@95: M ./configure.ac -1 +1 cannam@95: cannam@95: Sun Jan 26 12:07:43 EST 2003 stevenj cannam@95: * [project @ 2003-01-26 17:07:43 by stevenj] cannam@95: check for _alloca (MSVC) cannam@95: cannam@95: M ./configure.ac -1 +1 cannam@95: M ./kernel/ifftw.h -1 +5 cannam@95: cannam@95: Sun Jan 26 11:56:53 EST 2003 stevenj cannam@95: * [project @ 2003-01-26 16:56:53 by stevenj] cannam@95: slight fix in assert cannam@95: cannam@95: M ./kernel/alloc.c -2 +2 cannam@95: cannam@95: Sun Jan 26 11:55:39 EST 2003 athena cannam@95: * [project @ 2003-01-26 16:55:39 by athena] cannam@95: Allocate problem in all cases--- can_do may need correct pointers. cannam@95: cannam@95: M ./libbench2/problem.c -1 +2 cannam@95: M ./libbench2/speed.c -2 +1 cannam@95: M ./libbench2/verify.c -3 +1 cannam@95: M ./tests/bench.c -6 cannam@95: cannam@95: Sun Jan 26 11:51:27 EST 2003 athena cannam@95: * [project @ 2003-01-26 16:51:27 by athena] cannam@95: Nastier checks cannam@95: cannam@95: M ./tests/bench.c -1 +8 cannam@95: M ./tests/check.pl -9 +23 cannam@95: cannam@95: Sun Jan 26 11:51:16 EST 2003 athena cannam@95: * [project @ 2003-01-26 16:51:16 by athena] cannam@95: X(use_plan) is a relic. cannam@95: cannam@95: M ./kernel/ifftw.h -3 +1 cannam@95: M ./kernel/plan.c -11 +3 cannam@95: M ./kernel/planner.c -5 +2 cannam@95: cannam@95: Sun Jan 26 09:23:16 EST 2003 athena cannam@95: * [project @ 2003-01-26 14:23:16 by athena] cannam@95: Print full pathname of the bench executable, so that I don't get cannam@95: confused when running multiple tests for different configurations. cannam@95: cannam@95: M ./tests/Makefile.am -1 +1 cannam@95: cannam@95: Sun Jan 26 07:35:46 EST 2003 athena cannam@95: * [project @ 2003-01-26 12:35:46 by athena] cannam@95: Split done() into done() and cleanup(), in order to test cannam@95: multiple problems with the same planner from the command line. cannam@95: cannam@95: M ./libbench2/bench-main.c -1 +2 cannam@95: M ./libbench2/bench-user.h -1 +2 cannam@95: M ./tests/bench.c -1 +8 cannam@95: cannam@95: Sat Jan 25 20:44:49 EST 2003 athena cannam@95: * [project @ 2003-01-26 01:44:49 by athena] cannam@95: Improved readability cannam@95: cannam@95: M ./kernel/alloc.c -6 +16 cannam@95: cannam@95: Sat Jan 25 19:17:26 EST 2003 stevenj cannam@95: * [project @ 2003-01-26 00:17:26 by stevenj] cannam@95: comment cannam@95: cannam@95: M ./kernel/alloc.c -3 +3 cannam@95: cannam@95: Sat Jan 25 19:16:53 EST 2003 stevenj cannam@95: * [project @ 2003-01-26 00:16:53 by stevenj] cannam@95: added macos9 mpallocatealigned function cannam@95: cannam@95: M ./kernel/alloc.c -1 +19 cannam@95: cannam@95: Sat Jan 25 18:59:55 EST 2003 stevenj cannam@95: * [project @ 2003-01-25 23:59:55 by stevenj] cannam@95: sometimes __APPLE__ is defined instead of __MACOSX__ cannam@95: cannam@95: M ./kernel/alloc.c -2 +3 cannam@95: cannam@95: Sat Jan 25 18:54:39 EST 2003 stevenj cannam@95: * [project @ 2003-01-25 23:54:39 by stevenj] cannam@95: macos x malloc is already 16-byte aligned cannam@95: cannam@95: M ./kernel/alloc.c -2 +3 cannam@95: cannam@95: Sat Jan 25 13:38:32 EST 2003 athena cannam@95: * [project @ 2003-01-25 18:38:32 by athena] cannam@95: Include because uintptr_t is defined there cannam@95: on solaris. cannam@95: cannam@95: M ./kernel/ifftw.h -2 +6 cannam@95: cannam@95: Sat Jan 25 13:22:59 EST 2003 athena cannam@95: * [project @ 2003-01-25 18:22:59 by athena] cannam@95: Oops---forgot getopt_long cannam@95: cannam@95: A ./libbench2/getopt1.c cannam@95: M ./libbench2/Makefile.am -2 +2 cannam@95: M ./libbench2/getopt1.c +188 cannam@95: cannam@95: Sat Jan 25 13:17:29 EST 2003 athena cannam@95: * [project @ 2003-01-25 18:17:29 by athena] cannam@95: Include default includes when checking for uintptr_t. cannam@95: (Otherwise solaris breaks.) cannam@95: cannam@95: M ./configure.ac -1 +1 cannam@95: cannam@95: Sat Jan 25 12:39:52 EST 2003 athena cannam@95: * [project @ 2003-01-25 17:39:52 by athena] cannam@95: distribute check.pl cannam@95: cannam@95: M ./tests/Makefile.am +1 cannam@95: cannam@95: Sat Jan 25 12:38:34 EST 2003 athena cannam@95: * [project @ 2003-01-25 17:38:34 by athena] cannam@95: Check split format, too. cannam@95: cannam@95: M ./tests/check.pl +4 cannam@95: cannam@95: Sat Jan 25 11:48:19 EST 2003 athena cannam@95: * [project @ 2003-01-25 16:48:19 by athena] cannam@95: New tests, added make check cannam@95: cannam@95: M ./tests/Makefile.am -1 +2 cannam@95: M ./tests/check.pl -3 +66 cannam@95: cannam@95: Thu Jan 23 08:34:24 EST 2003 athena cannam@95: * [project @ 2003-01-23 13:34:24 by athena] cannam@95: More tests cannam@95: cannam@95: M ./tests/check.pl -17 +83 cannam@95: cannam@95: Tue Jan 21 20:32:12 EST 2003 athena cannam@95: * [project @ 2003-01-22 01:32:09 by athena] cannam@95: Deal with rnk(sz)=-infinity cannam@95: cannam@95: M ./api/mktensor-iodims.c -4 +9 cannam@95: M ./api/mktensor-rowmajor.c -4 +6 cannam@95: M ./libbench2/problem.c -5 +2 cannam@95: cannam@95: Tue Jan 21 10:07:16 EST 2003 athena cannam@95: * [project @ 2003-01-21 15:07:16 by athena] cannam@95: Crazy idea cannam@95: cannam@95: M ./TODO +3 cannam@95: cannam@95: Tue Jan 21 07:14:22 EST 2003 athena cannam@95: * [project @ 2003-01-21 12:14:22 by athena] cannam@95: Test program, still barely worthy of the name. cannam@95: cannam@95: A ./tests/check.pl cannam@95: cannam@95: Mon Jan 20 08:29:21 EST 2003 athena cannam@95: * [project @ 2003-01-20 13:29:21 by athena] cannam@95: Stylistic changes cannam@95: cannam@95: M ./libbench2/problem.c -15 +10 cannam@95: cannam@95: Mon Jan 20 07:03:38 EST 2003 athena cannam@95: * [project @ 2003-01-20 12:03:38 by athena] cannam@95: Implemented flops api cannam@95: cannam@95: A ./api/flops.c cannam@95: M ./api/Makefile.am -10 +10 cannam@95: M ./api/fftw3.h -2 +4 cannam@95: M ./api/flops.c +27 cannam@95: M ./tests/bench.c +3 cannam@95: cannam@95: Sun Jan 19 14:27:21 EST 2003 stevenj cannam@95: * [project @ 2003-01-19 19:27:21 by stevenj] cannam@95: cleanup cannam@95: cannam@95: M ./libbench2/problem.c -31 +24 cannam@95: cannam@95: Sun Jan 19 14:14:49 EST 2003 stevenj cannam@95: * [project @ 2003-01-19 19:14:49 by stevenj] cannam@95: 'v' syntax now defaults to an 'internal' (stride 1) vector, which is a more interesting case and corresponds more closely to the intuitive notion of a 'vector' transform, while '*' does the old 'external' (stride n) vector cannam@95: cannam@95: M ./libbench2/problem.c -4 +24 cannam@95: cannam@95: Sun Jan 19 13:55:35 EST 2003 stevenj cannam@95: * [project @ 2003-01-19 18:55:35 by stevenj] cannam@95: removed '/' overloading cannam@95: cannam@95: M ./libbench2/problem.c -2 +2 cannam@95: cannam@95: Sun Jan 19 13:52:09 EST 2003 stevenj cannam@95: * [project @ 2003-01-19 18:52:09 by stevenj] cannam@95: get rid of '*' and ',' synonyms for 'x' in problem parser; there's no need to clutter the namespace with syntax we never use cannam@95: cannam@95: M ./libbench2/problem.c -2 +2 cannam@95: cannam@95: Sun Jan 19 07:28:27 EST 2003 athena cannam@95: * [project @ 2003-01-19 12:28:27 by athena] cannam@95: Signed/unsigned fixes. cannam@95: cannam@95: M ./kernel/planner.c -4 +4 cannam@95: cannam@95: Sun Jan 19 07:09:54 EST 2003 athena cannam@95: * [project @ 2003-01-19 12:09:54 by athena] cannam@95: Test split arrays. cannam@95: cannam@95: M ./libbench2/bench-user.h -3 +9 cannam@95: M ./libbench2/verify-dft.c -14 +11 cannam@95: M ./libbench2/verify.c -13 +43 cannam@95: M ./libbench2/verify.h +1 cannam@95: M ./tests/bench.c -13 +50 cannam@95: M ./tests/hook.c -1 +4 cannam@95: cannam@95: Sat Jan 18 23:46:57 EST 2003 stevenj cannam@95: * [project @ 2003-01-19 04:46:57 by stevenj] cannam@95: clarification cannam@95: cannam@95: M ./doc/fftw3.texi -17 +16 cannam@95: cannam@95: Sat Jan 18 21:53:18 EST 2003 stevenj cannam@95: * [project @ 2003-01-19 02:53:18 by stevenj] cannam@95: caps cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Sat Jan 18 21:52:51 EST 2003 stevenj cannam@95: * [project @ 2003-01-19 02:52:51 by stevenj] cannam@95: brackets cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Sat Jan 18 21:52:32 EST 2003 stevenj cannam@95: * [project @ 2003-01-19 02:52:32 by stevenj] cannam@95: quote cannam@95: cannam@95: M ./doc/fftw3.texi -1 +7 cannam@95: cannam@95: Sat Jan 18 20:53:11 EST 2003 stevenj cannam@95: * [project @ 2003-01-19 01:53:11 by stevenj] cannam@95: referencing cannam@95: cannam@95: M ./doc/fftw3.texi -4 +7 cannam@95: cannam@95: Sat Jan 18 20:33:28 EST 2003 stevenj cannam@95: * [project @ 2003-01-19 01:33:28 by stevenj] cannam@95: fix cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Sat Jan 18 20:31:41 EST 2003 stevenj cannam@95: * [project @ 2003-01-19 01:31:41 by stevenj] cannam@95: slight change cannam@95: cannam@95: M ./doc/fftw3.texi -8 +8 cannam@95: cannam@95: Sat Jan 18 20:31:22 EST 2003 athena cannam@95: * [project @ 2003-01-19 01:31:22 by athena] cannam@95: Print errors when --verify. cannam@95: cannam@95: M ./libbench2/verify-dft.c -7 +12 cannam@95: M ./libbench2/verify-lib.c -30 +38 cannam@95: M ./libbench2/verify.h -12 +14 cannam@95: M ./tests/bench.c -5 +5 cannam@95: cannam@95: Sat Jan 18 20:30:27 EST 2003 stevenj cannam@95: * [project @ 2003-01-19 01:30:27 by stevenj] cannam@95: improved description, noted that FFTW_ESTIMATE does not destroy arrays cannam@95: cannam@95: M ./doc/fftw3.texi -7 +8 cannam@95: cannam@95: Sat Jan 18 20:23:12 EST 2003 stevenj cannam@95: * [project @ 2003-01-19 01:23:12 by stevenj] cannam@95: FFTW_DEFAULTS isn't really needed cannam@95: cannam@95: M ./api/fftw3.h -2 +1 cannam@95: cannam@95: Sat Jan 18 20:21:09 EST 2003 stevenj cannam@95: * [project @ 2003-01-19 01:21:09 by stevenj] cannam@95: added FFTW_MEASURE synonym for FFTW_DEFAULTS cannam@95: cannam@95: M ./api/fftw3.h -1 +2 cannam@95: M ./doc/fftw3.texi -6 +6 cannam@95: cannam@95: Sat Jan 18 20:18:29 EST 2003 stevenj cannam@95: * [project @ 2003-01-19 01:18:29 by stevenj] cannam@95: slight change cannam@95: cannam@95: M ./kernel/alloc.c -2 +2 cannam@95: cannam@95: Sat Jan 18 20:16:08 EST 2003 athena cannam@95: * [project @ 2003-01-19 01:16:08 by athena] cannam@95: Clearer name cannam@95: cannam@95: M ./tests/bench.c -2 +2 cannam@95: cannam@95: Sat Jan 18 20:13:14 EST 2003 athena cannam@95: * [project @ 2003-01-19 01:13:14 by athena] cannam@95: Completed dft api test cannam@95: cannam@95: M ./api/fftw3.h -12 +12 cannam@95: M ./libbench2/tensor.c -3 +3 cannam@95: M ./tests/bench.c -8 +65 cannam@95: cannam@95: Sat Jan 18 20:07:33 EST 2003 stevenj cannam@95: * [project @ 2003-01-19 01:07:33 by stevenj] cannam@95: index cannam@95: cannam@95: M ./doc/fftw3.texi -1 +2 cannam@95: cannam@95: Sat Jan 18 20:05:50 EST 2003 stevenj cannam@95: * [project @ 2003-01-19 01:05:50 by stevenj] cannam@95: fix cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Sat Jan 18 20:04:11 EST 2003 stevenj cannam@95: * [project @ 2003-01-19 01:04:11 by stevenj] cannam@95: parallel structure cannam@95: cannam@95: M ./doc/fftw3.texi -2 +2 cannam@95: cannam@95: Sat Jan 18 20:03:18 EST 2003 stevenj cannam@95: * [project @ 2003-01-19 01:03:18 by stevenj] cannam@95: fix cannam@95: cannam@95: M ./doc/fftw3.texi -4 +4 cannam@95: cannam@95: Sat Jan 18 20:00:24 EST 2003 stevenj cannam@95: * [project @ 2003-01-19 01:00:24 by stevenj] cannam@95: joke cannam@95: cannam@95: M ./doc/fftw3.texi -5 +6 cannam@95: cannam@95: Sat Jan 18 19:59:28 EST 2003 stevenj cannam@95: * [project @ 2003-01-19 00:59:28 by stevenj] cannam@95: recommendation to read tutorial in-order cannam@95: cannam@95: M ./doc/fftw3.texi -1 +6 cannam@95: cannam@95: Sat Jan 18 19:54:55 EST 2003 stevenj cannam@95: * [project @ 2003-01-19 00:54:55 by stevenj] cannam@95: expanded outline cannam@95: cannam@95: M ./doc/fftw3.texi -7 +42 cannam@95: cannam@95: Sat Jan 18 19:35:52 EST 2003 stevenj cannam@95: * [project @ 2003-01-19 00:35:52 by stevenj] cannam@95: clarification cannam@95: cannam@95: M ./doc/fftw3.texi -3 +5 cannam@95: cannam@95: Sat Jan 18 19:17:27 EST 2003 stevenj cannam@95: * [project @ 2003-01-19 00:17:27 by stevenj] cannam@95: draft complex-dft tutorial cannam@95: cannam@95: M ./doc/fftw3.texi -25 +183 cannam@95: cannam@95: Sat Jan 18 17:27:15 EST 2003 athena cannam@95: * [project @ 2003-01-18 22:27:15 by athena] cannam@95: Paranoid mode is back. Fixed dwim to do what I mean. cannam@95: cannam@95: A ./tests/hook.c cannam@95: M ./libbench2/allocate.c -2 +1 cannam@95: M ./libbench2/bench-main.c -2 +1 cannam@95: M ./libbench2/bench-user.h -8 +1 cannam@95: M ./libbench2/bench.h -2 +1 cannam@95: M ./libbench2/can-do.c -2 +1 cannam@95: M ./libbench2/dotens2.c -2 +2 cannam@95: M ./libbench2/info.c -2 +1 cannam@95: M ./libbench2/problem.c -18 +21 cannam@95: M ./libbench2/report.c -2 +1 cannam@95: M ./libbench2/speed.c -2 +1 cannam@95: M ./libbench2/timer.c -2 +1 cannam@95: M ./libbench2/util.c -1 cannam@95: M ./libbench2/verify.c -4 +3 cannam@95: M ./libbench2/verify.h +7 cannam@95: M ./libbench2/zero.c -2 +1 cannam@95: M ./tests/Makefile.am -1 +1 cannam@95: M ./tests/bench.c +4 cannam@95: M ./tests/hook.c +191 cannam@95: cannam@95: Sat Jan 18 17:13:51 EST 2003 stevenj cannam@95: * [project @ 2003-01-18 22:13:51 by stevenj] cannam@95: started tut. cannam@95: cannam@95: M ./doc/fftw3.texi -14 +46 cannam@95: cannam@95: Sat Jan 18 16:13:15 EST 2003 athena cannam@95: * [project @ 2003-01-18 21:13:15 by athena] cannam@95: Great renaming, so that we can include both bench-user.h and cannam@95: ifftw.h to implement the paranoid-mode hook. cannam@95: cannam@95: M ./libbench2/allocate.c -5 +5 cannam@95: M ./libbench2/bench-user.h -30 +28 cannam@95: M ./libbench2/bench.h -12 +12 cannam@95: M ./libbench2/can-do.c -2 +2 cannam@95: M ./libbench2/dotens2.c -3 +3 cannam@95: M ./libbench2/mflops.c -1 +1 cannam@95: M ./libbench2/problem.c -12 +12 cannam@95: M ./libbench2/report.c -10 +10 cannam@95: M ./libbench2/speed.c -2 +2 cannam@95: M ./libbench2/tensor.c -22 +22 cannam@95: M ./libbench2/verify-dft.c -2 +2 cannam@95: M ./libbench2/verify-lib.c -6 +6 cannam@95: M ./libbench2/verify.c -9 +9 cannam@95: M ./libbench2/verify.h -2 +2 cannam@95: M ./libbench2/zero.c -2 +2 cannam@95: M ./tests/bench.c -29 +29 cannam@95: cannam@95: Sat Jan 18 15:41:18 EST 2003 athena cannam@95: * [project @ 2003-01-18 20:41:18 by athena] cannam@95: Trying to tweak the verifier so that I can use it in cannam@95: bench.c for paranoid mode cannam@95: cannam@95: M ./libbench2/bench-user.h -1 +23 cannam@95: M ./libbench2/problem.c -3 +3 cannam@95: M ./libbench2/verify-dft.c -66 +9 cannam@95: M ./libbench2/verify-lib.c -21 +18 cannam@95: M ./libbench2/verify.c -5 +62 cannam@95: M ./libbench2/verify.h -18 +3 cannam@95: cannam@95: Sat Jan 18 10:24:05 EST 2003 athena cannam@95: * [project @ 2003-01-18 15:24:05 by athena] cannam@95: Added stride_factor for complex arrays. cannam@95: cannam@95: M ./tests/bench.c -5 +5 cannam@95: cannam@95: Sat Jan 18 10:02:11 EST 2003 athena cannam@95: * [project @ 2003-01-18 15:02:11 by athena] cannam@95: can_do now calls the planner. cannam@95: cannam@95: M ./tests/bench.c -1 +6 cannam@95: cannam@95: Sat Jan 18 09:59:24 EST 2003 athena cannam@95: * [project @ 2003-01-18 14:59:24 by athena] cannam@95: Call guru api in bench.c cannam@95: cannam@95: M ./api/plan-guru-dft.c -2 +1 cannam@95: M ./tests/bench.c -3 +47 cannam@95: cannam@95: Sat Jan 18 08:17:23 EST 2003 athena cannam@95: * [project @ 2003-01-18 13:17:23 by athena] cannam@95: Fixed prototype. cannam@95: cannam@95: M ./libbench2/bench.h -1 +3 cannam@95: M ./libbench2/zero.c -2 +2 cannam@95: cannam@95: Sat Jan 18 08:14:48 EST 2003 athena cannam@95: * [project @ 2003-01-18 13:14:48 by athena] cannam@95: Attempt to make the signed/unsigned use of flags consistent. cannam@95: cannam@95: M ./api/api.h -2 +2 cannam@95: M ./api/apiplan.c -1 +1 cannam@95: M ./api/fftw3.h -25 +25 cannam@95: M ./api/mapflags.c -8 +10 cannam@95: M ./api/plan-dft-1d.c -1 +1 cannam@95: M ./api/plan-dft-2d.c -1 +1 cannam@95: M ./api/plan-dft-3d.c -1 +1 cannam@95: M ./api/plan-dft-c2r-1d.c -1 +1 cannam@95: M ./api/plan-dft-c2r-2d.c -1 +1 cannam@95: M ./api/plan-dft-c2r-3d.c -1 +1 cannam@95: M ./api/plan-dft-c2r.c -1 +1 cannam@95: M ./api/plan-dft-r2c-1d.c -1 +1 cannam@95: M ./api/plan-dft-r2c-2d.c -1 +1 cannam@95: M ./api/plan-dft-r2c-3d.c -1 +1 cannam@95: M ./api/plan-dft-r2c.c -1 +1 cannam@95: M ./api/plan-dft.c -1 +1 cannam@95: M ./api/plan-guru-dft-c2r.c -1 +1 cannam@95: M ./api/plan-guru-dft-r2c.c -1 +1 cannam@95: M ./api/plan-guru-dft.c -1 +1 cannam@95: M ./api/plan-guru-r2r.c -1 +1 cannam@95: M ./api/plan-many-dft-c2r.c -1 +1 cannam@95: M ./api/plan-many-dft-r2c.c -1 +1 cannam@95: M ./api/plan-many-dft.c -1 +1 cannam@95: M ./api/plan-many-r2r.c -1 +1 cannam@95: M ./api/plan-r2r-1d.c -1 +1 cannam@95: M ./api/plan-r2r-2d.c -1 +1 cannam@95: M ./api/plan-r2r-3d.c -1 +1 cannam@95: M ./api/plan-r2r.c -1 +1 cannam@95: M ./kernel/ifftw.h -2 +2 cannam@95: cannam@95: Sat Jan 18 08:03:07 EST 2003 athena cannam@95: * [project @ 2003-01-18 13:03:07 by athena] cannam@95: Implemented useropt. cannam@95: cannam@95: A ./libbench2/useropt.c cannam@95: M ./libbench2/Makefile.am -2 +3 cannam@95: M ./libbench2/bench-main.c -1 +5 cannam@95: M ./libbench2/bench-user.h -1 +2 cannam@95: M ./libbench2/useropt.c +30 cannam@95: M ./tests/bench.c -7 +24 cannam@95: cannam@95: Sat Jan 18 08:02:05 EST 2003 athena cannam@95: * [project @ 2003-01-18 13:02:05 by athena] cannam@95: The first map_flags pass must be transitive, i.e., always use the cannam@95: latest flags value as opposed to the original value. (I think.) cannam@95: cannam@95: M ./api/mapflags.c -11 +12 cannam@95: cannam@95: Sat Jan 18 07:20:19 EST 2003 athena cannam@95: * [project @ 2003-01-18 12:20:18 by athena] cannam@95: Started working on verifier cannam@95: cannam@95: A ./libbench2/dotens2.c cannam@95: A ./libbench2/verify-dft.c cannam@95: A ./libbench2/verify-lib.c cannam@95: A ./libbench2/verify.h cannam@95: M ./libbench2/Makefile.am -3 +3 cannam@95: M ./libbench2/bench-user.h -1 +4 cannam@95: M ./libbench2/dotens2.c +55 cannam@95: M ./libbench2/tensor.c -1 +8 cannam@95: M ./libbench2/verify-dft.c +119 cannam@95: M ./libbench2/verify-lib.c +399 cannam@95: M ./libbench2/verify.c -2 +9 cannam@95: M ./libbench2/verify.h +65 cannam@95: M ./tests/Makefile.am -4 +1 cannam@95: cannam@95: Fri Jan 17 14:53:28 EST 2003 stevenj cannam@95: * [project @ 2003-01-17 19:53:28 by stevenj] cannam@95: added X(threads_cleanup) cannam@95: cannam@95: M ./api/fftw3.h -1 +2 cannam@95: M ./threads/api.c +9 cannam@95: M ./threads/threads.c -6 +19 cannam@95: M ./threads/threads.h -1 +2 cannam@95: cannam@95: Fri Jan 17 10:35:56 EST 2003 athena cannam@95: * [project @ 2003-01-17 15:35:56 by athena] cannam@95: Use C style for upper and lower array bounds. Free tensors properly. cannam@95: cannam@95: M ./libbench2/allocate.c -3 +5 cannam@95: M ./libbench2/tensor.c -2 +2 cannam@95: cannam@95: Fri Jan 17 08:50:42 EST 2003 athena cannam@95: * [project @ 2003-01-17 13:50:42 by athena] cannam@95: Fixed ambiguous syntax cannam@95: cannam@95: M ./libbench2/problem.c -2 +2 cannam@95: cannam@95: Fri Jan 17 08:20:57 EST 2003 athena cannam@95: * [project @ 2003-01-17 13:20:57 by athena] cannam@95: Parse minus sign, bugfixes cannam@95: cannam@95: M ./libbench2/problem.c -4 +17 cannam@95: cannam@95: Fri Jan 17 08:11:56 EST 2003 athena cannam@95: * [project @ 2003-01-17 13:11:56 by athena] cannam@95: Skeleton libbench2 implemented (probably still buggy) cannam@95: cannam@95: A ./libbench2/ cannam@95: A ./libbench2/Makefile.am cannam@95: A ./libbench2/allocate.c cannam@95: A ./libbench2/bench-main.c cannam@95: A ./libbench2/bench-user.h cannam@95: A ./libbench2/bench.h cannam@95: A ./libbench2/can-do.c cannam@95: A ./libbench2/caset.c cannam@95: A ./libbench2/getopt-utils.c cannam@95: A ./libbench2/getopt.c cannam@95: A ./libbench2/getopt.h cannam@95: A ./libbench2/info.c cannam@95: A ./libbench2/main.c cannam@95: A ./libbench2/mflops.c cannam@95: A ./libbench2/ovtpvt.c cannam@95: A ./libbench2/pow2.c cannam@95: A ./libbench2/problem.c cannam@95: A ./libbench2/report.c cannam@95: A ./libbench2/speed.c cannam@95: A ./libbench2/tensor.c cannam@95: A ./libbench2/timer.c cannam@95: A ./libbench2/util.c cannam@95: A ./libbench2/verify.c cannam@95: A ./libbench2/zero.c cannam@95: M ./Makefile.am -1 +1 cannam@95: M ./configure.ac -1 +1 cannam@95: M ./libbench2/Makefile.am +13 cannam@95: M ./libbench2/allocate.c +55 cannam@95: M ./libbench2/bench-main.c +250 cannam@95: M ./libbench2/bench-user.h +206 cannam@95: M ./libbench2/bench.h +67 cannam@95: M ./libbench2/can-do.c +33 cannam@95: M ./libbench2/caset.c +12 cannam@95: M ./libbench2/getopt-utils.c +104 cannam@95: M ./libbench2/getopt.c +1062 cannam@95: M ./libbench2/getopt.h +180 cannam@95: M ./libbench2/info.c +59 cannam@95: M ./libbench2/main.c +40 cannam@95: M ./libbench2/mflops.c +25 cannam@95: M ./libbench2/ovtpvt.c +13 cannam@95: M ./libbench2/pow2.c +6 cannam@95: M ./libbench2/problem.c +172 cannam@95: M ./libbench2/report.c +159 cannam@95: M ./libbench2/speed.c +73 cannam@95: M ./libbench2/tensor.c +173 cannam@95: M ./libbench2/timer.c +241 cannam@95: M ./libbench2/util.c +216 cannam@95: M ./libbench2/verify.c +53 cannam@95: M ./libbench2/zero.c +36 cannam@95: M ./tests/Makefile.am -2 +2 cannam@95: M ./tests/bench.c -56 +93 cannam@95: cannam@95: Fri Jan 17 04:23:37 EST 2003 athena cannam@95: * [project @ 2003-01-17 09:23:37 by athena] cannam@95: Formatting cannam@95: cannam@95: M ./kernel/tensor4.c -10 +10 cannam@95: cannam@95: Fri Jan 17 03:15:24 EST 2003 fftw cannam@95: * [project @ 2003-01-17 08:15:24 by fftw] cannam@95: slight updates cannam@95: cannam@95: M ./doc/fftw3.texi -27 +34 cannam@95: cannam@95: Fri Jan 17 01:44:44 EST 2003 stevenj cannam@95: * [project @ 2003-01-17 06:44:44 by stevenj] cannam@95: eliminated obsolete uimin/uimax cannam@95: cannam@95: M ./dft/vrank-geq1.c -2 +2 cannam@95: M ./kernel/buffered.c -2 +2 cannam@95: M ./kernel/ifftw.h -3 +1 cannam@95: M ./kernel/minmax.c -11 +1 cannam@95: M ./kernel/tensor4.c -4 +4 cannam@95: M ./rdft/buffered2.c -4 +4 cannam@95: M ./rdft/rdft2-inplace-strides.c -2 +2 cannam@95: M ./rdft/rdft2-tensor-max-index.c -4 +4 cannam@95: M ./rdft/vrank-geq1-rdft2.c -2 +2 cannam@95: M ./rdft/vrank-geq1.c -2 +2 cannam@95: cannam@95: Fri Jan 17 01:40:10 EST 2003 stevenj cannam@95: * [project @ 2003-01-17 06:40:10 by stevenj] cannam@95: threads needs to have its own library, lest all programs linking to libfftw3.so need -lpthread cannam@95: cannam@95: A ./threads/api.c cannam@95: M ./Makefile.am -1 cannam@95: M ./api/Makefile.am -4 +3 cannam@95: M ./api/configure.c -2 cannam@95: M ./api/fftw3.h -2 +3 cannam@95: R ./api/plan-with-nthreads.c cannam@95: M ./tests/Makefile.am -1 +8 cannam@95: M ./tests/bench.c +5 cannam@95: M ./threads/Makefile.am -9 +6 cannam@95: M ./threads/api.c +56 cannam@95: M ./threads/threads.c -1 +1 cannam@95: cannam@95: Thu Jan 16 19:53:46 EST 2003 stevenj cannam@95: * [project @ 2003-01-17 00:53:46 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./api/f77api.c -2 +2 cannam@95: cannam@95: Thu Jan 16 19:53:30 EST 2003 stevenj cannam@95: * [project @ 2003-01-17 00:53:30 by stevenj] cannam@95: better name cannam@95: cannam@95: M ./api/f77api.c -2 +2 cannam@95: cannam@95: Thu Jan 16 19:52:36 EST 2003 stevenj cannam@95: * [project @ 2003-01-17 00:52:36 by stevenj] cannam@95: added more functions cannam@95: cannam@95: M ./api/f77api.c +68 cannam@95: cannam@95: Thu Jan 16 16:57:06 EST 2003 stevenj cannam@95: * [project @ 2003-01-16 21:57:06 by stevenj] cannam@95: if 'long' is big enough, use it for mulmod in preference to 'long long' cannam@95: cannam@95: M ./kernel/ifftw.h -1 +3 cannam@95: cannam@95: Thu Jan 16 14:53:41 EST 2003 stevenj cannam@95: * [project @ 2003-01-16 19:53:41 by stevenj] cannam@95: use uintptr_t for pointer alignment arithmetic cannam@95: cannam@95: M ./configure.ac -2 +9 cannam@95: M ./kernel/align.c -2 +2 cannam@95: M ./kernel/ifftw.h -5 +22 cannam@95: cannam@95: Thu Jan 16 07:58:28 EST 2003 athena cannam@95: * [project @ 2003-01-16 12:58:28 by athena] cannam@95: More signed/unsigned cleanup cannam@95: cannam@95: M ./kernel/planner.c -10 +10 cannam@95: M ./kernel/print.c -3 +3 cannam@95: M ./kernel/tensor.c -2 +2 cannam@95: M ./kernel/twiddle.c -2 +2 cannam@95: M ./rdft/problem.c -2 +2 cannam@95: cannam@95: Thu Jan 16 07:57:40 EST 2003 athena cannam@95: * [project @ 2003-01-16 12:57:40 by athena] cannam@95: null function pointers are technically nonportable cannam@95: cannam@95: M ./kernel/solvtab.c -2 +2 cannam@95: cannam@95: Thu Jan 16 07:17:45 EST 2003 athena cannam@95: * [project @ 2003-01-16 12:17:45 by athena] cannam@95: Free short_options cannam@95: cannam@95: M ./libbench/bench-main.c -2 +3 cannam@95: cannam@95: Thu Jan 16 05:48:30 EST 2003 athena cannam@95: * [project @ 2003-01-16 10:48:30 by athena] cannam@95: Oops, forgot STACK_FREE cannam@95: cannam@95: M ./kernel/alloc.c -15 +17 cannam@95: M ./kernel/ifftw.h -2 +3 cannam@95: M ./tests/bench.c -1 +1 cannam@95: cannam@95: Thu Jan 16 05:40:39 EST 2003 athena cannam@95: * [project @ 2003-01-16 10:40:39 by athena] cannam@95: Do not require memalign() unless HAVE_SIMD cannam@95: cannam@95: M ./kernel/alloc.c -6 +15 cannam@95: M ./kernel/ifftw.h -13 +18 cannam@95: cannam@95: Thu Jan 16 01:03:31 EST 2003 stevenj cannam@95: * [project @ 2003-01-16 06:03:31 by stevenj] cannam@95: MS VC++ _aligned_malloc cannam@95: cannam@95: M ./kernel/alloc.c -1 +7 cannam@95: cannam@95: Thu Jan 16 00:44:45 EST 2003 stevenj cannam@95: * [project @ 2003-01-16 05:44:45 by stevenj] cannam@95: added api fftw_malloc/free cannam@95: cannam@95: M ./api/fftw3.h -2 +5 cannam@95: M ./kernel/alloc.c -2 +10 cannam@95: cannam@95: Thu Jan 16 00:43:48 EST 2003 stevenj cannam@95: * [project @ 2003-01-16 05:43:48 by stevenj] cannam@95: silence warning cannam@95: cannam@95: M ./api/map-r2r-kind.c -1 +1 cannam@95: cannam@95: Wed Jan 15 22:39:04 EST 2003 stevenj cannam@95: * [project @ 2003-01-16 03:39:04 by stevenj] cannam@95: send error output to stderr cannam@95: cannam@95: M ./tools/fftw-wisdom-to-conf.in -1 +1 cannam@95: cannam@95: Wed Jan 15 13:20:35 EST 2003 athena cannam@95: * [project @ 2003-01-15 18:20:35 by athena] cannam@95: Pure paranoia. cannam@95: cannam@95: M ./kernel/tensor7.c -3 +6 cannam@95: cannam@95: Wed Jan 15 06:51:34 EST 2003 athena cannam@95: * [project @ 2003-01-15 11:51:34 by athena] cannam@95: Fixed formatting that was messed up by the conversion uint->int. cannam@95: Ensure that iodims etc are kosher. cannam@95: cannam@95: A ./kernel/tensor9.c cannam@95: M ./api/api.h -13 +20 cannam@95: M ./api/apiplan.c -9 +9 cannam@95: M ./api/configure.c -5 +5 cannam@95: M ./api/dfthelp.c -1 +1 cannam@95: M ./api/execute-dft-c2r.c -1 +1 cannam@95: M ./api/execute-dft-r2c.c -1 +1 cannam@95: M ./api/execute-dft.c -1 +1 cannam@95: M ./api/execute-r2r.c -1 +1 cannam@95: M ./api/execute.c -1 +1 cannam@95: M ./api/export-wisdom-to-file.c -4 +4 cannam@95: M ./api/export-wisdom-to-string.c -6 +6 cannam@95: M ./api/export-wisdom.c -6 +6 cannam@95: M ./api/f77api.c -58 +46 cannam@95: M ./api/fftw3.h -14 +11 cannam@95: M ./api/forget-wisdom.c -2 +2 cannam@95: M ./api/import-system-wisdom.c -2 +2 cannam@95: M ./api/import-wisdom-from-file.c -5 +5 cannam@95: M ./api/import-wisdom-from-string.c -4 +4 cannam@95: M ./api/import-wisdom.c -6 +6 cannam@95: M ./api/map-r2r-kind.c -37 +17 cannam@95: M ./api/mapflags.c -4 +4 cannam@95: M ./api/mkprinter-file.c -2 +2 cannam@95: M ./api/mktensor-iodims.c -2 +23 cannam@95: M ./api/mktensor-rowmajor.c -4 +20 cannam@95: M ./api/plan-dft-1d.c -2 +2 cannam@95: M ./api/plan-dft-2d.c -2 +2 cannam@95: M ./api/plan-dft-3d.c -3 +3 cannam@95: M ./api/plan-dft-c2r-1d.c -2 +2 cannam@95: M ./api/plan-dft-c2r-2d.c -2 +2 cannam@95: M ./api/plan-dft-c2r-3d.c -3 +3 cannam@95: M ./api/plan-dft-c2r.c -3 +3 cannam@95: M ./api/plan-dft-r2c-1d.c -2 +2 cannam@95: M ./api/plan-dft-r2c-2d.c -2 +2 cannam@95: M ./api/plan-dft-r2c-3d.c -3 +3 cannam@95: M ./api/plan-dft-r2c.c -3 +5 cannam@95: M ./api/plan-dft.c -4 +6 cannam@95: M ./api/plan-guru-dft-c2r.c -10 +10 cannam@95: M ./api/plan-guru-dft-r2c.c -10 +11 cannam@95: M ./api/plan-guru-dft.c -10 +11 cannam@95: M ./api/plan-guru-r2r.c -14 +16 cannam@95: M ./api/plan-many-dft-c2r.c -34 +21 cannam@95: M ./api/plan-many-dft-r2c.c -34 +21 cannam@95: M ./api/plan-many-dft.c -18 +19 cannam@95: M ./api/plan-many-r2r.c -22 +21 cannam@95: M ./api/plan-r2r-1d.c -2 +2 cannam@95: M ./api/plan-r2r-2d.c -3 +3 cannam@95: M ./api/plan-r2r-3d.c -4 +4 cannam@95: M ./api/plan-r2r.c -4 +4 cannam@95: M ./api/plan-with-nthreads.c -3 +3 cannam@95: M ./api/print-plan.c -3 +3 cannam@95: M ./api/rdft2-pad.c -2 +2 cannam@95: M ./api/the-planner.c -5 +5 cannam@95: M ./dft/buffered.c -2 +2 cannam@95: M ./dft/ct.c -2 +2 cannam@95: M ./dft/direct.c -2 +2 cannam@95: M ./dft/generic.c -1 +1 cannam@95: M ./dft/problem.c -4 +5 cannam@95: M ./dft/rader.c -1 +1 cannam@95: M ./dft/vrank-geq1.c -2 +2 cannam@95: M ./dft/vrank2-transpose.c -2 +2 cannam@95: M ./dft/vrank3-transpose.c -2 +2 cannam@95: M ./kernel/Makefile.am -2 +2 cannam@95: M ./kernel/alloc.c -2 +2 cannam@95: M ./kernel/ifftw.h -1 +2 cannam@95: M ./kernel/planner.c -10 +10 cannam@95: M ./kernel/print.c -3 +3 cannam@95: M ./kernel/tensor.c -2 +4 cannam@95: M ./kernel/tensor9.c +37 cannam@95: M ./rdft/buffered.c -2 +2 cannam@95: M ./rdft/buffered2.c -2 +2 cannam@95: M ./rdft/dft-r2hc.c -2 +2 cannam@95: M ./rdft/dht-r2hc.c -2 +2 cannam@95: M ./rdft/dht-rader.c -1 +1 cannam@95: M ./rdft/direct.c -2 +2 cannam@95: M ./rdft/direct2.c -2 +2 cannam@95: M ./rdft/generic.c -1 +1 cannam@95: M ./rdft/hc2hc.c -2 +2 cannam@95: M ./rdft/problem.c -2 +6 cannam@95: M ./rdft/problem2.c -2 +5 cannam@95: M ./rdft/rader-hc2hc.c -1 +1 cannam@95: M ./rdft/rdft-dht.c -2 +2 cannam@95: M ./rdft/rdft2-radix2.c -2 +2 cannam@95: M ./rdft/vrank-geq1-rdft2.c -2 +2 cannam@95: M ./rdft/vrank-geq1.c -2 +2 cannam@95: M ./rdft/vrank2-transpose.c -2 +2 cannam@95: M ./rdft/vrank3-transpose.c -2 +2 cannam@95: M ./reodft/redft00e-r2hc.c -2 +2 cannam@95: M ./reodft/reodft010e-r2hc.c -2 +2 cannam@95: M ./reodft/reodft11e-r2hc.c -2 +2 cannam@95: M ./reodft/rodft00e-r2hc.c -2 +2 cannam@95: M ./threads/dft-vrank-geq1.c -2 +2 cannam@95: M ./threads/rdft-vrank-geq1.c -2 +2 cannam@95: M ./threads/vrank-geq1-rdft2.c -2 +2 cannam@95: cannam@95: Wed Jan 15 01:32:18 EST 2003 stevenj cannam@95: * [project @ 2003-01-15 06:32:18 by stevenj] cannam@95: added version stamp cannam@95: cannam@95: M ./tools/fftw-wisdom-to-conf.in -1 +2 cannam@95: cannam@95: Wed Jan 15 01:28:20 EST 2003 stevenj cannam@95: * [project @ 2003-01-15 06:28:20 by stevenj] cannam@95: added warning cannam@95: cannam@95: M ./tools/fftw-wisdom-to-conf.in +1 cannam@95: cannam@95: Wed Jan 15 01:23:25 EST 2003 stevenj cannam@95: * [project @ 2003-01-15 06:23:25 by stevenj] cannam@95: add fftw-wisdom-to-conf to BUILT_SOURCES cannam@95: cannam@95: M ./tools/Makefile.am -1 +1 cannam@95: cannam@95: Wed Jan 15 01:09:29 EST 2003 stevenj cannam@95: * [project @ 2003-01-15 06:09:29 by stevenj] cannam@95: added const cannam@95: cannam@95: M ./tools/fftw-wisdom-to-conf.in -2 +2 cannam@95: cannam@95: Wed Jan 15 01:04:10 EST 2003 stevenj cannam@95: * [project @ 2003-01-15 06:04:10 by stevenj] cannam@95: added wisdom-to-conf cannam@95: cannam@95: A ./tools/ cannam@95: A ./tools/Makefile.am cannam@95: A ./tools/fftw-wisdom-to-conf.in cannam@95: M ./Makefile.am -1 +1 cannam@95: M ./configure.ac +3 cannam@95: M ./tools/Makefile.am +3 cannam@95: M ./tools/fftw-wisdom-to-conf.in +37 cannam@95: cannam@95: Wed Jan 15 00:23:36 EST 2003 stevenj cannam@95: * [project @ 2003-01-15 05:23:36 by stevenj] cannam@95: include type prefix in wisdom preamble cannam@95: cannam@95: M ./kernel/planner.c -2 +2 cannam@95: cannam@95: Wed Jan 15 00:02:31 EST 2003 stevenj cannam@95: * [project @ 2003-01-15 05:02:31 by stevenj] cannam@95: updates cannam@95: cannam@95: M ./TODO -4 +2 cannam@95: cannam@95: Tue Jan 14 23:59:26 EST 2003 stevenj cannam@95: * [project @ 2003-01-15 04:59:26 by stevenj] cannam@95: check the_plan before printing cannam@95: cannam@95: M ./tests/bench.c -2 +3 cannam@95: cannam@95: Tue Jan 14 21:10:25 EST 2003 athena cannam@95: * [project @ 2003-01-15 02:10:25 by athena] cannam@95: Eliminated those unsigned values that would break LP64 machines. cannam@95: cannam@95: M ./ChangeLog -1 +1 cannam@95: M ./api/api.h -16 +16 cannam@95: M ./api/apiplan.c -19 +19 cannam@95: M ./api/configure.c -5 +5 cannam@95: M ./api/dfthelp.c -5 +5 cannam@95: M ./api/execute-dft-c2r.c -1 +1 cannam@95: M ./api/execute-dft-r2c.c -1 +1 cannam@95: M ./api/execute-dft.c -1 +1 cannam@95: M ./api/execute-r2r.c -1 +1 cannam@95: M ./api/execute.c -1 +1 cannam@95: M ./api/export-wisdom-to-file.c -4 +4 cannam@95: M ./api/export-wisdom-to-string.c -13 +13 cannam@95: M ./api/export-wisdom.c -7 +7 cannam@95: M ./api/f77api.c -71 +69 cannam@95: M ./api/fftw3.h -134 +136 cannam@95: M ./api/forget-wisdom.c -2 +2 cannam@95: M ./api/import-system-wisdom.c -8 +8 cannam@95: M ./api/import-wisdom-from-file.c -6 +6 cannam@95: M ./api/import-wisdom-from-string.c -6 +6 cannam@95: M ./api/import-wisdom.c -9 +11 cannam@95: M ./api/map-r2r-kind.c -17 +40 cannam@95: M ./api/mapflags.c -18 +18 cannam@95: M ./api/mkprinter-file.c -3 +3 cannam@95: M ./api/mktensor-iodims.c -9 +9 cannam@95: M ./api/mktensor-rowmajor.c -6 +5 cannam@95: M ./api/plan-dft-1d.c -3 +2 cannam@95: M ./api/plan-dft-2d.c -5 +5 cannam@95: M ./api/plan-dft-3d.c -5 +7 cannam@95: M ./api/plan-dft-c2r-1d.c -2 +2 cannam@95: M ./api/plan-dft-c2r-2d.c -5 +5 cannam@95: M ./api/plan-dft-c2r-3d.c -5 +7 cannam@95: M ./api/plan-dft-c2r.c -6 +3 cannam@95: M ./api/plan-dft-r2c-1d.c -2 +2 cannam@95: M ./api/plan-dft-r2c-2d.c -5 +5 cannam@95: M ./api/plan-dft-r2c-3d.c -5 +7 cannam@95: M ./api/plan-dft-r2c.c -6 +3 cannam@95: M ./api/plan-dft.c -6 +4 cannam@95: M ./api/plan-guru-dft-c2r.c -10 +10 cannam@95: M ./api/plan-guru-dft-r2c.c -10 +10 cannam@95: M ./api/plan-guru-dft.c -10 +10 cannam@95: M ./api/plan-guru-r2r.c -13 +14 cannam@95: M ./api/plan-many-dft-c2r.c -23 +36 cannam@95: M ./api/plan-many-dft-r2c.c -23 +36 cannam@95: M ./api/plan-many-dft.c -17 +18 cannam@95: M ./api/plan-many-r2r.c -18 +22 cannam@95: M ./api/plan-r2r-1d.c -3 +2 cannam@95: M ./api/plan-r2r-2d.c -7 +8 cannam@95: M ./api/plan-r2r-3d.c -8 +11 cannam@95: M ./api/plan-r2r.c -3 +4 cannam@95: M ./api/plan-with-nthreads.c -7 +7 cannam@95: M ./api/print-plan.c -3 +3 cannam@95: M ./api/rdft2-pad.c -10 +9 cannam@95: M ./api/the-planner.c -7 +7 cannam@95: M ./configure.ac -9 +3 cannam@95: M ./dft/buffered.c -11 +11 cannam@95: M ./dft/codelet-dft.h -11 +11 cannam@95: M ./dft/codelets/n.c -1 +1 cannam@95: M ./dft/codelets/t.c -1 +1 cannam@95: M ./dft/ct-dif.c -4 +4 cannam@95: M ./dft/ct-dit.c -4 +4 cannam@95: M ./dft/ct-ditbuf.c -6 +6 cannam@95: M ./dft/ct-ditf.c -2 +2 cannam@95: M ./dft/ct.c -5 +5 cannam@95: M ./dft/ct.h -2 +2 cannam@95: M ./dft/dft.h -2 +2 cannam@95: M ./dft/direct.c -3 +3 cannam@95: M ./dft/generic.c -5 +5 cannam@95: M ./dft/indirect.c -2 +2 cannam@95: M ./dft/k7/k7.c -9 +9 cannam@95: M ./dft/problem.c -3 +3 cannam@95: M ./dft/rader-omega.c -2 +2 cannam@95: M ./dft/rader.c -13 +13 cannam@95: M ./dft/rank-geq2.c -9 +9 cannam@95: M ./dft/rank0.c -7 +7 cannam@95: M ./dft/simd/n1b.c -1 +1 cannam@95: M ./dft/simd/n1f.c -1 +1 cannam@95: M ./dft/simd/t1b.c -1 +1 cannam@95: M ./dft/simd/t1f.c -1 +1 cannam@95: M ./dft/vrank-geq1.c -11 +11 cannam@95: M ./dft/vrank2-transpose.c -4 +4 cannam@95: M ./dft/vrank3-transpose.c -10 +10 cannam@95: M ./dft/zero.c -3 +3 cannam@95: M ./genfft/gen_hc2hc.ml -4 +4 cannam@95: M ./genfft/gen_hc2r.ml -4 +4 cannam@95: M ./genfft/gen_notw.ml -4 +4 cannam@95: M ./genfft/gen_notw_c.ml -4 +4 cannam@95: M ./genfft/gen_r2hc.ml -4 +4 cannam@95: M ./genfft/gen_twiddle.ml -4 +4 cannam@95: M ./genfft/gen_twiddle_c.ml -4 +4 cannam@95: M ./genfft/gen_twidsq.ml -4 +4 cannam@95: M ./kernel/align.c -5 +5 cannam@95: M ./kernel/buffered.c -3 +3 cannam@95: M ./kernel/ct.c -1 +1 cannam@95: M ./kernel/hash.c -3 +3 cannam@95: M ./kernel/iabs.c -3 +3 cannam@95: M ./kernel/ifftw.h -71 +70 cannam@95: M ./kernel/md5-1.c -3 +3 cannam@95: M ./kernel/md5.c -4 +4 cannam@95: M ./kernel/minmax.c -3 +3 cannam@95: M ./kernel/ops.c -4 +4 cannam@95: M ./kernel/pickdim.c -6 +6 cannam@95: M ./kernel/planner.c -32 +32 cannam@95: M ./kernel/primes.c -14 +14 cannam@95: M ./kernel/print.c -5 +5 cannam@95: M ./kernel/rader.c -3 +3 cannam@95: M ./kernel/scan.c -3 +3 cannam@95: M ./kernel/tensor.c -9 +9 cannam@95: M ./kernel/tensor1.c -2 +2 cannam@95: M ./kernel/tensor2.c -3 +3 cannam@95: M ./kernel/tensor4.c -10 +10 cannam@95: M ./kernel/tensor5.c -6 +6 cannam@95: M ./kernel/tensor7.c -4 +4 cannam@95: M ./kernel/trig.c -4 +4 cannam@95: M ./kernel/twiddle.c -14 +14 cannam@95: M ./libbench/acopy.c -2 +2 cannam@95: M ./libbench/allocate.c -2 +2 cannam@95: M ./libbench/ascale.c -2 +2 cannam@95: M ./libbench/aset.c -2 +2 cannam@95: M ./libbench/bench-user.h -23 +23 cannam@95: M ./libbench/bench.h -2 +2 cannam@95: M ./libbench/caadd.c -2 +2 cannam@95: M ./libbench/cacopy.c -2 +2 cannam@95: M ./libbench/cascale.c -2 +2 cannam@95: M ./libbench/caset.c -2 +2 cannam@95: M ./libbench/casub.c -2 +2 cannam@95: M ./libbench/copy-c2h-1d-fftpack.c -1 +1 cannam@95: M ./libbench/copy-c2h-1d-halfcomplex.c -1 +1 cannam@95: M ./libbench/copy-c2h-1d-packed.c -2 +2 cannam@95: M ./libbench/copy-c2h-1d-unpacked-ri.c -1 +1 cannam@95: M ./libbench/copy-c2h-unpacked.c -1 +1 cannam@95: M ./libbench/copy-c2r-packed.c -2 +2 cannam@95: M ./libbench/copy-c2r-unpacked.c -1 +1 cannam@95: M ./libbench/copy-c2ri.c -2 +2 cannam@95: M ./libbench/copy-h2c-1d-fftpack.c -1 +1 cannam@95: M ./libbench/copy-h2c-1d-halfcomplex.c -1 +1 cannam@95: M ./libbench/copy-h2c-1d-packed.c -2 +2 cannam@95: M ./libbench/copy-h2c-1d-unpacked-ri.c -1 +1 cannam@95: M ./libbench/copy-h2c-unpacked.c -9 +9 cannam@95: M ./libbench/copy-r2c-packed.c -2 +2 cannam@95: M ./libbench/copy-r2c-unpacked.c -1 +1 cannam@95: M ./libbench/copy-ri2c.c -2 +2 cannam@95: M ./libbench/getopt-utils.c -2 +2 cannam@95: M ./libbench/getopt.c -3 +3 cannam@95: M ./libbench/log2.c -2 +2 cannam@95: M ./libbench/mp.c -20 +20 cannam@95: M ./libbench/pow2.c -1 +1 cannam@95: M ./libbench/prime.c -4 +4 cannam@95: M ./libbench/problem.c -2 +2 cannam@95: M ./libbench/timer.c -2 +2 cannam@95: M ./libbench/verify.c -36 +36 cannam@95: M ./rdft/buffered.c -10 +10 cannam@95: M ./rdft/buffered2.c -22 +22 cannam@95: M ./rdft/codelet-rdft.h -17 +17 cannam@95: M ./rdft/codelets/hc2r.c -1 +1 cannam@95: M ./rdft/codelets/hfb.c -1 +1 cannam@95: M ./rdft/codelets/r2hc.c -1 +1 cannam@95: M ./rdft/dft-r2hc.c -4 +4 cannam@95: M ./rdft/dht-r2hc.c -3 +3 cannam@95: M ./rdft/dht-rader.c -6 +6 cannam@95: M ./rdft/direct.c -6 +6 cannam@95: M ./rdft/direct2.c -5 +5 cannam@95: M ./rdft/generic.c -8 +8 cannam@95: M ./rdft/hc2hc-buf.c -7 +7 cannam@95: M ./rdft/hc2hc-dif.c -4 +4 cannam@95: M ./rdft/hc2hc-dit.c -4 +4 cannam@95: M ./rdft/hc2hc.c -5 +5 cannam@95: M ./rdft/hc2hc.h -2 +2 cannam@95: M ./rdft/indirect.c -2 +2 cannam@95: M ./rdft/problem.c -11 +11 cannam@95: M ./rdft/problem2.c -3 +3 cannam@95: M ./rdft/rader-hc2hc.c -12 +12 cannam@95: M ./rdft/rank-geq2-rdft2.c -9 +9 cannam@95: M ./rdft/rank-geq2.c -9 +9 cannam@95: M ./rdft/rank0-rdft2.c -4 +4 cannam@95: M ./rdft/rank0.c -5 +5 cannam@95: M ./rdft/rdft-dht.c -5 +5 cannam@95: M ./rdft/rdft.h -4 +4 cannam@95: M ./rdft/rdft2-inplace-strides.c -4 +4 cannam@95: M ./rdft/rdft2-radix2.c -15 +15 cannam@95: M ./rdft/rdft2-tensor-max-index.c -4 +4 cannam@95: M ./rdft/vrank-geq1-rdft2.c -12 +12 cannam@95: M ./rdft/vrank-geq1.c -11 +11 cannam@95: M ./rdft/vrank2-transpose.c -4 +4 cannam@95: M ./rdft/vrank3-transpose.c -10 +10 cannam@95: M ./reodft/redft00e-r2hc.c -7 +7 cannam@95: M ./reodft/reodft010e-r2hc.c -16 +16 cannam@95: M ./reodft/reodft11e-r2hc.c -10 +10 cannam@95: M ./reodft/rodft00e-r2hc.c -7 +7 cannam@95: M ./simd/3dnow.c -5 +5 cannam@95: M ./simd/sse.c -3 +3 cannam@95: M ./simd/sse2.c -3 +3 cannam@95: M ./tests/bench.c -13 +7 cannam@95: M ./tests/dotens.c -3 +3 cannam@95: M ./tests/dotens2.c -3 +3 cannam@95: M ./tests/trigtest.c -7 +7 cannam@95: M ./tests/verify-dft.c -4 +4 cannam@95: M ./tests/verify-lib.c -41 +41 cannam@95: M ./tests/verify-rdft.c -13 +13 cannam@95: M ./tests/verify-reodft.c -48 +48 cannam@95: M ./tests/verify.h -15 +15 cannam@95: M ./threads/ct-dit.c -5 +5 cannam@95: M ./threads/dft-vrank-geq1.c -15 +15 cannam@95: M ./threads/hc2hc-dif.c -6 +6 cannam@95: M ./threads/hc2hc-dit.c -6 +6 cannam@95: M ./threads/rdft-vrank-geq1.c -15 +15 cannam@95: M ./threads/threads.c -4 +4 cannam@95: M ./threads/threads.h -3 +3 cannam@95: M ./threads/vrank-geq1-rdft2.c -15 +15 cannam@95: cannam@95: Tue Jan 14 15:14:29 EST 2003 stevenj cannam@95: * [project @ 2003-01-14 20:14:29 by stevenj] cannam@95: comments cannam@95: cannam@95: M ./kernel/primes.c -4 +4 cannam@95: cannam@95: Tue Jan 14 08:00:08 EST 2003 athena cannam@95: * [project @ 2003-01-14 13:00:08 by athena] cannam@95: Oops cannam@95: cannam@95: M ./dft/generic.c -1 +1 cannam@95: M ./rdft/generic.c -1 +1 cannam@95: cannam@95: Tue Jan 14 07:59:14 EST 2003 athena cannam@95: * [project @ 2003-01-14 12:59:14 by athena] cannam@95: int/uint confusion cannam@95: cannam@95: M ./dft/generic.c -1 +2 cannam@95: M ./rdft/generic.c -1 +2 cannam@95: cannam@95: Tue Jan 14 02:25:33 EST 2003 stevenj cannam@95: * [project @ 2003-01-14 07:25:33 by stevenj] cannam@95: updated introduction and some organization cannam@95: cannam@95: M ./doc/fftw3.texi -79 +117 cannam@95: cannam@95: Tue Jan 14 01:34:46 EST 2003 stevenj cannam@95: * [project @ 2003-01-14 06:34:46 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./api/f77api.c -18 +18 cannam@95: cannam@95: Tue Jan 14 01:33:04 EST 2003 stevenj cannam@95: * [project @ 2003-01-14 06:33:04 by stevenj] cannam@95: newline cannam@95: cannam@95: M ./Makefile.am +1 cannam@95: cannam@95: Tue Jan 14 00:23:04 EST 2003 stevenj cannam@95: * [project @ 2003-01-14 05:23:04 by stevenj] cannam@95: added win32 timer cannam@95: cannam@95: M ./libbench/timer.c -1 +23 cannam@95: cannam@95: Tue Jan 14 00:12:21 EST 2003 stevenj cannam@95: * [project @ 2003-01-14 05:12:21 by stevenj] cannam@95: sync with kernel/alloc.c cannam@95: cannam@95: M ./libbench/util.c -2 +5 cannam@95: cannam@95: Tue Jan 14 00:03:20 EST 2003 stevenj cannam@95: * [project @ 2003-01-14 05:03:20 by stevenj] cannam@95: handle missing F77_FUNC_ cannam@95: cannam@95: M ./api/f77api.c +4 cannam@95: cannam@95: Mon Jan 13 17:42:50 EST 2003 stevenj cannam@95: * [project @ 2003-01-13 22:42:50 by stevenj] cannam@95: used fint instead of int to make Fortran integer type easier to change cannam@95: cannam@95: M ./api/f77api.c -24 +31 cannam@95: cannam@95: Mon Jan 13 17:38:56 EST 2003 stevenj cannam@95: * [project @ 2003-01-13 22:38:56 by stevenj] cannam@95: slight abbreviation cannam@95: cannam@95: M ./api/f77api.c -4 +4 cannam@95: cannam@95: Mon Jan 13 17:35:20 EST 2003 stevenj cannam@95: * [project @ 2003-01-13 22:35:20 by stevenj] cannam@95: the great lengthening, part I: int -> long in api; mv mktensor-rowmajor to api cannam@95: cannam@95: A ./api/mktensor-rowmajor.c cannam@95: M ./api/Makefile.am -2 +2 cannam@95: M ./api/api.h -2 +5 cannam@95: M ./api/f77api.c -7 +7 cannam@95: M ./api/fftw3.h -44 +46 cannam@95: M ./api/mktensor-rowmajor.c +44 cannam@95: M ./api/plan-dft-1d.c -1 +1 cannam@95: M ./api/plan-dft-2d.c -2 +2 cannam@95: M ./api/plan-dft-3d.c -2 +2 cannam@95: M ./api/plan-dft-c2r-1d.c -1 +1 cannam@95: M ./api/plan-dft-c2r-2d.c -2 +2 cannam@95: M ./api/plan-dft-c2r-3d.c -2 +2 cannam@95: M ./api/plan-dft-c2r.c -1 +1 cannam@95: M ./api/plan-dft-r2c-1d.c -1 +1 cannam@95: M ./api/plan-dft-r2c-2d.c -2 +2 cannam@95: M ./api/plan-dft-r2c-3d.c -2 +2 cannam@95: M ./api/plan-dft-r2c.c -1 +1 cannam@95: M ./api/plan-dft.c -1 +1 cannam@95: M ./api/plan-many-dft-c2r.c -7 +7 cannam@95: M ./api/plan-many-dft-r2c.c -7 +7 cannam@95: M ./api/plan-many-dft.c -6 +6 cannam@95: M ./api/plan-many-r2r.c -6 +6 cannam@95: M ./api/plan-r2r-1d.c -1 +1 cannam@95: M ./api/plan-r2r-2d.c -2 +2 cannam@95: M ./api/plan-r2r-3d.c -2 +2 cannam@95: M ./api/plan-r2r.c -1 +1 cannam@95: M ./api/rdft2-pad.c -4 +4 cannam@95: M ./kernel/Makefile.am -3 +3 cannam@95: M ./kernel/ifftw.h -4 +1 cannam@95: R ./kernel/tensor3.c cannam@95: M ./tests/bench.c -1 +9 cannam@95: cannam@95: Mon Jan 13 15:23:22 EST 2003 stevenj cannam@95: * [project @ 2003-01-13 20:23:22 by stevenj] cannam@95: long types cannam@95: cannam@95: M ./configure.ac +4 cannam@95: cannam@95: Mon Jan 13 04:20:37 EST 2003 athena cannam@95: * [project @ 2003-01-13 09:20:36 by athena] cannam@95: Renamed fftw_malloc -> MALLOC, X(free) -> X(ifree), X(free0) -> cannam@95: X(ifree0), non_fftw_malloc -> NATIVE_MALLOC cannam@95: cannam@95: M ./api/apiplan.c -2 +2 cannam@95: M ./api/export-wisdom-to-string.c -1 +1 cannam@95: M ./api/f77api.c -8 +8 cannam@95: M ./api/map-r2r-kind.c -1 +1 cannam@95: M ./api/plan-guru-r2r.c -1 +1 cannam@95: M ./api/plan-many-dft-c2r.c -2 +2 cannam@95: M ./api/plan-many-dft-r2c.c -2 +2 cannam@95: M ./api/plan-many-r2r.c -1 +1 cannam@95: M ./api/rdft2-pad.c -1 +1 cannam@95: M ./dft/buffered.c -6 +6 cannam@95: M ./dft/generic.c -1 +1 cannam@95: M ./dft/problem.c -2 +2 cannam@95: M ./dft/rader-omega.c -1 +1 cannam@95: M ./dft/rader.c -10 +10 cannam@95: M ./kernel/alloc.c -6 +6 cannam@95: M ./kernel/ifftw.h -11 +11 cannam@95: M ./kernel/plan.c -3 +3 cannam@95: M ./kernel/planner.c -12 +12 cannam@95: M ./kernel/print.c -3 +3 cannam@95: M ./kernel/problem.c -2 +2 cannam@95: M ./kernel/rader.c -3 +3 cannam@95: M ./kernel/scan.c -3 +3 cannam@95: M ./kernel/solver.c -3 +3 cannam@95: M ./kernel/stride.c -3 +3 cannam@95: M ./kernel/tensor.c -9 +9 cannam@95: M ./kernel/twiddle.c -5 +5 cannam@95: M ./rdft/buffered.c -6 +6 cannam@95: M ./rdft/buffered2.c -8 +8 cannam@95: M ./rdft/dht-rader.c -6 +6 cannam@95: M ./rdft/generic.c -1 +1 cannam@95: M ./rdft/problem.c -4 +4 cannam@95: M ./rdft/problem2.c -2 +2 cannam@95: M ./rdft/rader-hc2hc.c -10 +10 cannam@95: M ./reodft/redft00e-r2hc.c -5 +5 cannam@95: M ./reodft/reodft010e-r2hc.c -11 +11 cannam@95: M ./reodft/reodft11e-r2hc.c -7 +7 cannam@95: M ./reodft/rodft00e-r2hc.c -5 +5 cannam@95: M ./tests/verify-dft.c -15 +15 cannam@95: M ./tests/verify-rdft.c -29 +29 cannam@95: M ./tests/verify-reodft.c -15 +15 cannam@95: M ./threads/dft-vrank-geq1.c -4 +4 cannam@95: M ./threads/rdft-vrank-geq1.c -4 +4 cannam@95: M ./threads/threads.c -1 +1 cannam@95: M ./threads/vrank-geq1-rdft2.c -4 +4 cannam@95: cannam@95: Mon Jan 13 02:37:22 EST 2003 stevenj cannam@95: * [project @ 2003-01-13 07:37:22 by stevenj] cannam@95: added beginning of Fortran interface cannam@95: cannam@95: A ./api/f77api.c cannam@95: M ./api/Makefile.am -1 +1 cannam@95: M ./api/f77api.c +145 cannam@95: cannam@95: Mon Jan 13 01:05:29 EST 2003 stevenj cannam@95: * [project @ 2003-01-13 06:05:29 by stevenj] cannam@95: add fortran mangling check cannam@95: cannam@95: M ./configure.ac +18 cannam@95: cannam@95: Mon Jan 13 00:33:28 EST 2003 stevenj cannam@95: * [project @ 2003-01-13 05:33:28 by stevenj] cannam@95: added guru r2r interface cannam@95: cannam@95: A ./api/execute-r2r.c cannam@95: A ./api/plan-guru-r2r.c cannam@95: M ./api/Makefile.am -3 +4 cannam@95: M ./api/execute-r2r.c +29 cannam@95: M ./api/fftw3.h -1 +8 cannam@95: M ./api/plan-guru-r2r.c +42 cannam@95: cannam@95: Mon Jan 13 00:23:26 EST 2003 stevenj cannam@95: * [project @ 2003-01-13 05:23:26 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./api/fftw3.h -5 +7 cannam@95: M ./api/plan-r2r-1d.c -2 +2 cannam@95: M ./api/plan-r2r-2d.c -1 +4 cannam@95: M ./api/plan-r2r-3d.c -2 +5 cannam@95: cannam@95: Mon Jan 13 00:16:20 EST 2003 stevenj cannam@95: * [project @ 2003-01-13 05:16:20 by stevenj] cannam@95: added r2r planner cannam@95: cannam@95: A ./api/map-r2r-kind.c cannam@95: A ./api/plan-many-r2r.c cannam@95: A ./api/plan-r2r-1d.c cannam@95: A ./api/plan-r2r-2d.c cannam@95: A ./api/plan-r2r-3d.c cannam@95: A ./api/plan-r2r.c cannam@95: M ./api/Makefile.am -2 +3 cannam@95: M ./api/fftw3.h -1 +29 cannam@95: M ./api/map-r2r-kind.c +47 cannam@95: M ./api/plan-many-r2r.c +48 cannam@95: M ./api/plan-r2r-1d.c +27 cannam@95: M ./api/plan-r2r-2d.c +29 cannam@95: M ./api/plan-r2r-3d.c +30 cannam@95: M ./api/plan-r2r.c +27 cannam@95: cannam@95: Sun Jan 12 22:58:18 EST 2003 stevenj cannam@95: * [project @ 2003-01-13 03:58:18 by stevenj] cannam@95: more long-double checks cannam@95: cannam@95: M ./configure.ac -2 +13 cannam@95: cannam@95: Sun Jan 12 20:01:51 EST 2003 stevenj cannam@95: * [project @ 2003-01-13 01:01:51 by stevenj] cannam@95: slight regrouping cannam@95: cannam@95: M ./kernel/planner.c -5 +7 cannam@95: cannam@95: Sun Jan 12 19:58:46 EST 2003 stevenj cannam@95: * [project @ 2003-01-13 00:58:46 by stevenj] cannam@95: added joke cannam@95: cannam@95: M ./kernel/planner.c -2 +2 cannam@95: cannam@95: Sun Jan 12 19:53:58 EST 2003 stevenj cannam@95: * [project @ 2003-01-13 00:53:58 by stevenj] cannam@95: simplified rdft2 padding cannam@95: cannam@95: A ./api/rdft2-pad.c cannam@95: M ./api/Makefile.am -7 +7 cannam@95: M ./api/api.h -4 +2 cannam@95: R ./api/mktensor-rowmajor-pad.c cannam@95: M ./api/plan-many-dft-c2r.c -19 +16 cannam@95: M ./api/plan-many-dft-r2c.c -6 +18 cannam@95: M ./api/rdft2-pad.c +40 cannam@95: cannam@95: Sun Jan 12 19:02:09 EST 2003 stevenj cannam@95: * [project @ 2003-01-13 00:02:09 by stevenj] cannam@95: added comment cannam@95: cannam@95: M ./api/fftw3.h -2 +2 cannam@95: cannam@95: Sun Jan 12 18:54:49 EST 2003 stevenj cannam@95: * [project @ 2003-01-12 23:54:49 by stevenj] cannam@95: use latest api cannam@95: cannam@95: M ./tests/bench.c -2 +1 cannam@95: cannam@95: Sun Jan 12 18:49:58 EST 2003 stevenj cannam@95: * [project @ 2003-01-12 23:49:58 by stevenj] cannam@95: nembed should only be in advanced (many) interface, not basic interface...only a handful of people over the years have ever requested that functionality. cannam@95: cannam@95: M ./api/fftw3.h -10 +4 cannam@95: M ./api/plan-dft-1d.c -1 +1 cannam@95: M ./api/plan-dft-2d.c -1 +1 cannam@95: M ./api/plan-dft-3d.c -1 +1 cannam@95: M ./api/plan-dft-c2r-1d.c -1 +1 cannam@95: M ./api/plan-dft-c2r-2d.c -1 +1 cannam@95: M ./api/plan-dft-c2r-3d.c -1 +1 cannam@95: M ./api/plan-dft-c2r.c -5 +3 cannam@95: M ./api/plan-dft-r2c-1d.c -1 +1 cannam@95: M ./api/plan-dft-r2c-2d.c -1 +1 cannam@95: M ./api/plan-dft-r2c-3d.c -1 +1 cannam@95: M ./api/plan-dft-r2c.c -5 +3 cannam@95: M ./api/plan-dft.c -5 +3 cannam@95: cannam@95: Sun Jan 12 18:41:57 EST 2003 stevenj cannam@95: * [project @ 2003-01-12 23:41:57 by stevenj] cannam@95: impatient is default; generalize mapping functions using xor trick cannam@95: cannam@95: M ./api/fftw3.h -2 +2 cannam@95: M ./api/mapflags.c -55 +61 cannam@95: cannam@95: Sun Jan 12 14:39:42 EST 2003 stevenj cannam@95: * [project @ 2003-01-12 19:39:41 by stevenj] cannam@95: use NULL nembed to signal padding cannam@95: cannam@95: M ./api/mktensor-rowmajor-pad.c -5 +9 cannam@95: M ./api/plan-dft-c2r-1d.c -1 +1 cannam@95: M ./api/plan-dft-c2r-2d.c -1 +1 cannam@95: M ./api/plan-dft-c2r-3d.c -1 +1 cannam@95: M ./api/plan-dft-r2c-1d.c -1 +1 cannam@95: M ./api/plan-dft-r2c-2d.c -1 +1 cannam@95: M ./api/plan-dft-r2c-3d.c -1 +1 cannam@95: cannam@95: Sun Jan 12 14:23:00 EST 2003 stevenj cannam@95: * [project @ 2003-01-12 19:23:00 by stevenj] cannam@95: accept NULL nembed cannam@95: cannam@95: M ./api/plan-many-dft.c -1 +4 cannam@95: cannam@95: Sun Jan 12 13:57:13 EST 2003 stevenj cannam@95: * [project @ 2003-01-12 18:57:13 by stevenj] cannam@95: added execute-dft-r2c/c2r cannam@95: cannam@95: A ./api/execute-dft-c2r.c cannam@95: A ./api/execute-dft-r2c.c cannam@95: M ./api/Makefile.am -5 +6 cannam@95: M ./api/execute-dft-c2r.c +29 cannam@95: M ./api/execute-dft-r2c.c +29 cannam@95: M ./api/fftw3.h -1 +4 cannam@95: cannam@95: Sun Jan 12 13:43:20 EST 2003 stevenj cannam@95: * [project @ 2003-01-12 18:43:20 by stevenj] cannam@95: don't need dft.h cannam@95: cannam@95: M ./api/plan-dft.c -1 cannam@95: cannam@95: Sun Jan 12 13:22:14 EST 2003 stevenj cannam@95: * [project @ 2003-01-12 18:22:14 by stevenj] cannam@95: tensors are compressed in the problem, duh cannam@95: cannam@95: M ./api/plan-many-dft-c2r.c -2 +1 cannam@95: M ./api/plan-many-dft-r2c.c -2 +1 cannam@95: cannam@95: Sun Jan 12 12:45:26 EST 2003 stevenj cannam@95: * [project @ 2003-01-12 17:45:26 by stevenj] cannam@95: noted that posix_memalign bug is now fixed, thanks to bug report by yours truly cannam@95: cannam@95: M ./kernel/alloc.c -2 +3 cannam@95: cannam@95: Sun Jan 12 12:44:43 EST 2003 athena cannam@95: * [project @ 2003-01-12 17:44:43 by athena] cannam@95: Bug: n[3] instead of n[2]. Bug was propagated by copy-and-paste. cannam@95: Grrr... cannam@95: cannam@95: M ./api/plan-dft-3d.c -1 +1 cannam@95: M ./api/plan-dft-c2r-3d.c -1 +1 cannam@95: M ./api/plan-dft-r2c-3d.c -1 +1 cannam@95: cannam@95: Sun Jan 12 12:41:43 EST 2003 athena cannam@95: * [project @ 2003-01-12 17:41:43 by athena] cannam@95: Express plan_dft() in terms of plan_many_dft() cannam@95: cannam@95: M ./api/plan-dft.c -11 +4 cannam@95: cannam@95: Sun Jan 12 12:19:53 EST 2003 stevenj cannam@95: * [project @ 2003-01-12 17:19:38 by stevenj] cannam@95: whoops cannam@95: cannam@95: A ./api/plan-guru-dft-c2r.c cannam@95: A ./api/plan-guru-dft-r2c.c cannam@95: A ./api/plan-guru-dft.c cannam@95: M ./api/plan-guru-dft-c2r.c +36 cannam@95: M ./api/plan-guru-dft-r2c.c +35 cannam@95: M ./api/plan-guru-dft.c +35 cannam@95: cannam@95: Sun Jan 12 06:00:46 EST 2003 athena cannam@95: * [project @ 2003-01-12 11:00:46 by athena] cannam@95: Manual skeleton. cannam@95: cannam@95: A ./doc/ cannam@95: A ./doc/Makefile.am cannam@95: A ./doc/fftw3.texi cannam@95: M ./Makefile.am -1 +1 cannam@95: M ./configure.ac +1 cannam@95: M ./doc/Makefile.am -1 +2 cannam@95: M ./doc/fftw3.texi +318 cannam@95: M ./genfft-k7/vK7Optimization.ml -1 +1 cannam@95: cannam@95: Sat Jan 11 23:46:34 EST 2003 stevenj cannam@95: * [project @ 2003-01-12 04:46:34 by stevenj] cannam@95: added r2c/c2r guru api cannam@95: cannam@95: M ./api/Makefile.am -1 +2 cannam@95: M ./api/fftw3.h -1 +12 cannam@95: cannam@95: Sat Jan 11 23:42:10 EST 2003 stevenj cannam@95: * [project @ 2003-01-12 04:42:10 by stevenj] cannam@95: FFTW_DESTROY_INPUT is default for c2r transforms cannam@95: cannam@95: M ./api/plan-many-dft-c2r.c -1 +1 cannam@95: cannam@95: Sat Jan 11 23:36:26 EST 2003 stevenj cannam@95: * [project @ 2003-01-12 04:36:26 by stevenj] cannam@95: added more of r2c/c2r api cannam@95: cannam@95: A ./api/plan-dft-c2r-1d.c cannam@95: A ./api/plan-dft-c2r-2d.c cannam@95: A ./api/plan-dft-c2r-3d.c cannam@95: A ./api/plan-dft-c2r.c cannam@95: A ./api/plan-dft-r2c-1d.c cannam@95: A ./api/plan-dft-r2c-2d.c cannam@95: A ./api/plan-dft-r2c-3d.c cannam@95: A ./api/plan-dft-r2c.c cannam@95: M ./api/Makefile.am -1 +3 cannam@95: M ./api/fftw3.h -69 +90 cannam@95: M ./api/plan-dft-c2r-1d.c +26 cannam@95: M ./api/plan-dft-c2r-2d.c +29 cannam@95: M ./api/plan-dft-c2r-3d.c +29 cannam@95: M ./api/plan-dft-c2r.c +32 cannam@95: M ./api/plan-dft-r2c-1d.c +26 cannam@95: M ./api/plan-dft-r2c-2d.c +29 cannam@95: M ./api/plan-dft-r2c-3d.c +29 cannam@95: M ./api/plan-dft-r2c.c +32 cannam@95: cannam@95: Sat Jan 11 21:09:41 EST 2003 stevenj cannam@95: * [project @ 2003-01-12 02:09:41 by stevenj] cannam@95: r2c doesn't have adjustible sign cannam@95: cannam@95: M ./api/fftw3.h -3 +3 cannam@95: M ./api/plan-many-dft-c2r.c -2 +2 cannam@95: M ./api/plan-many-dft-r2c.c -2 +2 cannam@95: cannam@95: Sat Jan 11 21:07:55 EST 2003 stevenj cannam@95: * [project @ 2003-01-12 02:07:55 by stevenj] cannam@95: note that copyright year is out of date cannam@95: cannam@95: M ./TODO -1 +1 cannam@95: cannam@95: Sat Jan 11 21:04:23 EST 2003 stevenj cannam@95: * [project @ 2003-01-12 02:04:23 by stevenj] cannam@95: updated api for r2c cannam@95: cannam@95: M ./api/fftw3.h -1 +17 cannam@95: cannam@95: Sat Jan 11 21:00:07 EST 2003 stevenj cannam@95: * [project @ 2003-01-12 02:00:07 by stevenj] cannam@95: removed annoying nophys == niphys case cannam@95: cannam@95: M ./api/mktensor-rowmajor-pad.c -11 cannam@95: cannam@95: Sat Jan 11 20:58:13 EST 2003 stevenj cannam@95: * [project @ 2003-01-12 01:58:13 by stevenj] cannam@95: added basic r2c/c2r planner cannam@95: cannam@95: A ./api/mktensor-rowmajor-pad.c cannam@95: A ./api/plan-many-dft-c2r.c cannam@95: A ./api/plan-many-dft-r2c.c cannam@95: M ./api/Makefile.am -2 +3 cannam@95: M ./api/api.h +5 cannam@95: M ./api/mktensor-rowmajor-pad.c +88 cannam@95: M ./api/plan-many-dft-c2r.c +59 cannam@95: M ./api/plan-many-dft-r2c.c +44 cannam@95: cannam@95: Sat Jan 11 19:34:14 EST 2003 stevenj cannam@95: * [project @ 2003-01-12 00:34:14 by stevenj] cannam@95: dist should be in terms of complex values cannam@95: cannam@95: M ./api/plan-many-dft.c -1 +1 cannam@95: cannam@95: Sat Jan 11 19:14:24 EST 2003 stevenj cannam@95: * [project @ 2003-01-12 00:14:24 by stevenj] cannam@95: added plan-with-nthreads cannam@95: cannam@95: M ./api/fftw3.h -2 +5 cannam@95: cannam@95: Sat Jan 11 19:12:51 EST 2003 stevenj cannam@95: * [project @ 2003-01-12 00:12:51 by stevenj] cannam@95: added function to set nthr cannam@95: cannam@95: A ./api/plan-with-nthreads.c cannam@95: M ./api/Makefile.am -1 +1 cannam@95: M ./api/plan-with-nthreads.c +39 cannam@95: cannam@95: Sat Jan 11 18:04:57 EST 2003 stevenj cannam@95: * [project @ 2003-01-11 23:04:57 by stevenj] cannam@95: slight cleanup cannam@95: cannam@95: M ./api/fftw3.h -4 +5 cannam@95: cannam@95: Sat Jan 11 17:57:29 EST 2003 stevenj cannam@95: * [project @ 2003-01-11 22:57:29 by stevenj] cannam@95: whoops cannam@95: cannam@95: A ./api/mktensor-iodims.c cannam@95: cannam@95: Sat Jan 11 17:55:39 EST 2003 stevenj cannam@95: * [project @ 2003-01-11 22:55:39 by stevenj] cannam@95: maxlen is maximum string length, not including null termination cannam@95: cannam@95: M ./kernel/scan.c -2 +2 cannam@95: cannam@95: Sat Jan 11 17:50:49 EST 2003 stevenj cannam@95: * [project @ 2003-01-11 22:50:49 by stevenj] cannam@95: imprt reverts hashtable on failure cannam@95: cannam@95: M ./kernel/planner.c -3 +14 cannam@95: cannam@95: Sat Jan 11 16:43:54 EST 2003 stevenj cannam@95: * [project @ 2003-01-11 21:43:54 by stevenj] cannam@95: slight move cannam@95: cannam@95: M ./api/fftw3.h -14 +14 cannam@95: cannam@95: Sat Jan 11 16:34:56 EST 2003 stevenj cannam@95: * [project @ 2003-01-11 21:34:56 by stevenj] cannam@95: stdio.h should be inlcuded outside of extern "C" cannam@95: cannam@95: M ./api/fftw3.h -3 +3 cannam@95: cannam@95: Sat Jan 11 16:26:35 EST 2003 stevenj cannam@95: * [project @ 2003-01-11 21:26:35 by stevenj] cannam@95: added guru planner API cannam@95: cannam@95: M ./api/Makefile.am -2 +3 cannam@95: M ./api/api.h +1 cannam@95: M ./api/fftw3.h -1 +15 cannam@95: cannam@95: Sat Jan 11 15:54:57 EST 2003 stevenj cannam@95: * [project @ 2003-01-11 20:54:57 by stevenj] cannam@95: added FFTW_FORWARD/BACKWARD cannam@95: cannam@95: M ./api/fftw3.h -1 +4 cannam@95: cannam@95: Sat Jan 11 15:52:17 EST 2003 stevenj cannam@95: * [project @ 2003-01-11 20:52:17 by stevenj] cannam@95: added plan_many_dft cannam@95: cannam@95: A ./api/plan-many-dft.c cannam@95: M ./api/Makefile.am -1 +1 cannam@95: M ./api/fftw3.h -1 +9 cannam@95: M ./api/plan-many-dft.c +43 cannam@95: cannam@95: Sat Jan 11 15:44:37 EST 2003 stevenj cannam@95: * [project @ 2003-01-11 20:44:37 by stevenj] cannam@95: indenting cannam@95: cannam@95: M ./kernel/tensor3.c -3 +3 cannam@95: cannam@95: Sat Jan 11 14:49:08 EST 2003 athena cannam@95: * [project @ 2003-01-11 19:49:08 by athena] cannam@95: Final \n cannam@95: cannam@95: M ./tests/bench.c -2 +4 cannam@95: cannam@95: Sat Jan 11 14:47:31 EST 2003 athena cannam@95: * [project @ 2003-01-11 19:47:31 by athena] cannam@95: Do not compile if not defined(FFTW_DEBUG), in order to avoid cannam@95: unused code in the shared library. cannam@95: cannam@95: M ./kernel/debug.c -1 +3 cannam@95: cannam@95: Sat Jan 11 14:45:56 EST 2003 athena cannam@95: * [project @ 2003-01-11 19:45:56 by athena] cannam@95: Implemented print_plan() cannam@95: cannam@95: A ./api/mkprinter-file.c cannam@95: A ./api/print-plan.c cannam@95: M ./api/Makefile.am -4 +4 cannam@95: M ./api/api.h +1 cannam@95: M ./api/export-wisdom-to-file.c -19 +1 cannam@95: M ./api/fftw3.h -2 +3 cannam@95: M ./api/mkprinter-file.c +39 cannam@95: M ./api/print-plan.c +29 cannam@95: M ./tests/bench.c -2 +3 cannam@95: cannam@95: Sat Jan 11 13:12:01 EST 2003 stevenj cannam@95: * [project @ 2003-01-11 18:12:01 by stevenj] cannam@95: changed the OOP-like plan_destroy to the more-grammatical destroy_plan cannam@95: cannam@95: M ./api/apiplan.c -1 +1 cannam@95: M ./api/fftw3.h -2 +2 cannam@95: M ./tests/bench.c -1 +1 cannam@95: cannam@95: Sat Jan 11 12:58:04 EST 2003 stevenj cannam@95: * [project @ 2003-01-11 17:55:49 by stevenj] cannam@95: added guru execute_dft cannam@95: cannam@95: A ./api/execute-dft.c cannam@95: M ./api/Makefile.am -4 +5 cannam@95: M ./api/execute-dft.c +29 cannam@95: M ./api/fftw3.h -2 +3 cannam@95: cannam@95: Sat Jan 11 12:38:40 EST 2003 stevenj cannam@95: * [project @ 2003-01-11 17:38:40 by stevenj] cannam@95: allow for malloc errors in wisdom string, since non-fftw-malloc cannam@95: cannam@95: M ./api/export-wisdom-to-string.c -4 +5 cannam@95: cannam@95: Sat Jan 11 12:16:05 EST 2003 stevenj cannam@95: * [project @ 2003-01-11 17:16:05 by stevenj] cannam@95: cleanup should reset plnr to zero so that fftw can be restarted cannam@95: cannam@95: M ./api/the-planner.c -1 +3 cannam@95: cannam@95: Sat Jan 11 12:13:18 EST 2003 stevenj cannam@95: * [project @ 2003-01-11 17:13:18 by stevenj] cannam@95: NO_UGLY is an internal planner flag cannam@95: cannam@95: M ./api/fftw3.h -4 +3 cannam@95: M ./api/mapflags.c -3 cannam@95: cannam@95: Sat Jan 11 11:23:13 EST 2003 athena cannam@95: * [project @ 2003-01-11 16:23:13 by athena] cannam@95: Written 1d api in terms of generic n-d api. The code is less compact cannam@95: but easier to test cannam@95: cannam@95: M ./api/plan-dft-1d.c -9 +1 cannam@95: M ./tests/bench.c -27 +36 cannam@95: cannam@95: Sat Jan 11 11:07:25 EST 2003 athena cannam@95: * [project @ 2003-01-11 16:07:24 by athena] cannam@95: Added wisdom to header file, made scanners/printer static. stdio.h cannam@95: no longer needed in fftw.h, removed. Probably the printer_file cannam@95: should be reintroduced in a separate file if we ever want to cannam@95: print plans... cannam@95: cannam@95: M ./api/export-wisdom-to-file.c -2 +2 cannam@95: M ./api/export-wisdom-to-string.c -4 +4 cannam@95: M ./api/fftw3.h -2 +13 cannam@95: M ./api/import-wisdom-from-file.c -2 +2 cannam@95: M ./api/import-wisdom-from-string.c -2 +2 cannam@95: M ./kernel/alloc.c -1 +3 cannam@95: M ./kernel/assert.c -2 +2 cannam@95: M ./kernel/debug.c -2 +22 cannam@95: M ./kernel/ifftw.h -15 +1 cannam@95: M ./kernel/print.c -1 +2 cannam@95: M ./kernel/scan.c -1 +2 cannam@95: M ./tests/bench.c -13 +49 cannam@95: M ./tests/verify-lib.c -1 +2 cannam@95: M ./tests/verify-reodft.c -1 +2 cannam@95: cannam@95: Sat Jan 11 09:49:30 EST 2003 athena cannam@95: * [project @ 2003-01-11 14:49:30 by athena] cannam@95: Implemented more APIs cannam@95: cannam@95: A ./api/plan-dft-2d.c cannam@95: A ./api/plan-dft-3d.c cannam@95: A ./api/plan-dft.c cannam@95: M ./api/Makefile.am -6 +5 cannam@95: M ./api/apiplan.c -1 +1 cannam@95: M ./api/fftw3.h -13 +25 cannam@95: M ./api/plan-dft-2d.c +30 cannam@95: M ./api/plan-dft-3d.c +30 cannam@95: M ./api/plan-dft.c +40 cannam@95: M ./tests/bench.c +18 cannam@95: cannam@95: Sat Jan 11 09:21:53 EST 2003 athena cannam@95: * [project @ 2003-01-11 14:21:53 by athena] cannam@95: Added cleanup() to API cannam@95: cannam@95: M ./api/fftw3.h -3 +3 cannam@95: M ./api/the-planner.c +6 cannam@95: M ./tests/bench.c +1 cannam@95: cannam@95: Sat Jan 11 09:17:34 EST 2003 athena cannam@95: * [project @ 2003-01-11 14:17:34 by athena] cannam@95: Started new bench.c. I had to rename plan_destroy -> cannam@95: plan_destroy_internal to avoid conflicts with API cannam@95: cannam@95: M ./api/api.h -1 cannam@95: M ./api/apiplan.c -3 +3 cannam@95: M ./api/fftw3.h -1 +3 cannam@95: M ./dft/buffered.c -7 +7 cannam@95: M ./dft/ct.c -2 +2 cannam@95: M ./dft/generic.c -2 +2 cannam@95: M ./dft/indirect.c -5 +5 cannam@95: M ./dft/rader.c -8 +8 cannam@95: M ./dft/rank-geq2.c -5 +5 cannam@95: M ./dft/vrank-geq1.c -2 +2 cannam@95: M ./kernel/ifftw.h -2 +2 cannam@95: M ./kernel/plan.c -2 +2 cannam@95: M ./kernel/planner.c -3 +3 cannam@95: M ./libbench/bench-user.h -1 +4 cannam@95: M ./libbench/bench.h -3 +1 cannam@95: M ./rdft/buffered.c -7 +7 cannam@95: M ./rdft/buffered2.c -5 +5 cannam@95: M ./rdft/dft-r2hc.c -2 +2 cannam@95: M ./rdft/dht-r2hc.c -2 +2 cannam@95: M ./rdft/dht-rader.c -6 +6 cannam@95: M ./rdft/generic.c -2 +2 cannam@95: M ./rdft/hc2hc.c -7 +7 cannam@95: M ./rdft/indirect.c -5 +5 cannam@95: M ./rdft/rader-hc2hc.c -7 +7 cannam@95: M ./rdft/rank-geq2-rdft2.c -5 +5 cannam@95: M ./rdft/rank-geq2.c -5 +5 cannam@95: M ./rdft/rank0-rdft2.c -2 +2 cannam@95: M ./rdft/rdft-dht.c -2 +2 cannam@95: M ./rdft/rdft2-radix2.c -2 +2 cannam@95: M ./rdft/vrank-geq1-rdft2.c -2 +2 cannam@95: M ./rdft/vrank-geq1.c -2 +2 cannam@95: M ./reodft/redft00e-r2hc.c -2 +2 cannam@95: M ./reodft/reodft010e-r2hc.c -2 +2 cannam@95: M ./reodft/reodft11e-r2hc.c -2 +2 cannam@95: M ./reodft/rodft00e-r2hc.c -2 +2 cannam@95: M ./tests/Makefile.am -1 +1 cannam@95: M ./tests/bench.c -2 +103 cannam@95: M ./threads/dft-vrank-geq1.c -3 +3 cannam@95: M ./threads/rdft-vrank-geq1.c -3 +3 cannam@95: M ./threads/vrank-geq1-rdft2.c -3 +3 cannam@95: cannam@95: Sat Jan 11 02:45:39 EST 2003 stevenj cannam@95: * [project @ 2003-01-11 07:45:39 by stevenj] cannam@95: fix types cannam@95: cannam@95: A ./api/import-system-wisdom.c cannam@95: M ./api/Makefile.am -3 +3 cannam@95: M ./api/export-wisdom.c -2 +2 cannam@95: M ./api/import-system-wisdom.c +46 cannam@95: M ./api/import-wisdom-from-file.c -2 +3 cannam@95: M ./api/import-wisdom-from-string.c -3 +4 cannam@95: M ./api/import-wisdom.c -3 +5 cannam@95: cannam@95: Sat Jan 11 02:13:25 EST 2003 stevenj cannam@95: * [project @ 2003-01-11 07:13:25 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./api/export-wisdom-to-string.c -1 +2 cannam@95: cannam@95: Sat Jan 11 02:10:50 EST 2003 stevenj cannam@95: * [project @ 2003-01-11 07:10:50 by stevenj] cannam@95: added wisdom api cannam@95: cannam@95: A ./api/export-wisdom-to-file.c cannam@95: A ./api/export-wisdom-to-string.c cannam@95: A ./api/export-wisdom.c cannam@95: A ./api/forget-wisdom.c cannam@95: A ./api/import-wisdom-from-file.c cannam@95: A ./api/import-wisdom-from-string.c cannam@95: A ./api/import-wisdom.c cannam@95: M ./api/Makefile.am -2 +6 cannam@95: M ./api/export-wisdom-to-file.c +47 cannam@95: M ./api/export-wisdom-to-string.c +80 cannam@95: M ./api/export-wisdom.c +44 cannam@95: M ./api/forget-wisdom.c +27 cannam@95: M ./api/import-wisdom-from-file.c +47 cannam@95: M ./api/import-wisdom-from-string.c +49 cannam@95: M ./api/import-wisdom.c +44 cannam@95: M ./kernel/Makefile.am -4 +4 cannam@95: M ./kernel/ifftw.h -1 +3 cannam@95: R ./kernel/printers.c cannam@95: R ./kernel/scanners.c cannam@95: cannam@95: Sat Jan 11 01:01:17 EST 2003 stevenj cannam@95: * [project @ 2003-01-11 06:01:17 by stevenj] cannam@95: grammar cannam@95: cannam@95: M ./api/mapflags.c -1 +1 cannam@95: cannam@95: Sat Jan 11 00:54:54 EST 2003 stevenj cannam@95: * [project @ 2003-01-11 05:54:54 by stevenj] cannam@95: slight change cannam@95: cannam@95: M ./api/mapflags.c -1 +1 cannam@95: cannam@95: Sat Jan 11 00:52:04 EST 2003 stevenj cannam@95: * [project @ 2003-01-11 05:52:04 by stevenj] cannam@95: implemented api/mapflags cannam@95: cannam@95: M ./api/fftw3.h -1 +23 cannam@95: M ./api/mapflags.c -1 +83 cannam@95: cannam@95: Sat Jan 11 00:48:27 EST 2003 stevenj cannam@95: * [project @ 2003-01-11 05:48:27 by stevenj] cannam@95: IMPATIENT is an api issue cannam@95: cannam@95: M ./kernel/ifftw.h -11 +1 cannam@95: cannam@95: Fri Jan 10 01:57:41 EST 2003 stevenj cannam@95: * [project @ 2003-01-10 06:57:41 by stevenj] cannam@95: removed un-needed headers cannam@95: cannam@95: M ./api/the-planner.c -4 cannam@95: cannam@95: Fri Jan 10 01:56:59 EST 2003 stevenj cannam@95: * [project @ 2003-01-10 06:56:59 by stevenj] cannam@95: mkplanner initializes nthr to 1 already cannam@95: cannam@95: M ./api/the-planner.c -1 cannam@95: cannam@95: Thu Jan 9 18:53:09 EST 2003 stevenj cannam@95: * [project @ 2003-01-09 23:53:09 by stevenj] cannam@95: boilerplate cannam@95: cannam@95: M ./api/fftw3.h -1 +14 cannam@95: cannam@95: Thu Jan 9 18:16:39 EST 2003 stevenj cannam@95: * [project @ 2003-01-09 23:16:39 by stevenj] cannam@95: fold vecloop into r{e,o}dft apply function to share buffer, etcetera cannam@95: cannam@95: M ./rdft/vrank-geq1.c -1 +6 cannam@95: M ./reodft/redft00e-r2hc.c -42 +45 cannam@95: M ./reodft/reodft010e-r2hc.c -138 +153 cannam@95: M ./reodft/reodft11e-r2hc.c -98 +107 cannam@95: M ./reodft/rodft00e-r2hc.c -39 +42 cannam@95: cannam@95: Thu Jan 9 18:10:19 EST 2003 stevenj cannam@95: * [project @ 2003-01-09 23:10:19 by stevenj] cannam@95: whoops, bugfix in impulse test for vecn > 1 cannam@95: cannam@95: M ./tests/verify-reodft.c -2 +2 cannam@95: cannam@95: Thu Jan 9 14:23:51 EST 2003 stevenj cannam@95: * [project @ 2003-01-09 19:23:51 by stevenj] cannam@95: bugfix, grr cannam@95: cannam@95: M ./rdft/hc2hc-buf.c -2 +2 cannam@95: cannam@95: Thu Jan 9 14:21:16 EST 2003 stevenj cannam@95: * [project @ 2003-01-09 19:21:16 by stevenj] cannam@95: fixed signed-ness enum problem cannam@95: cannam@95: M ./rdft/codelet-rdft.h -2 +5 cannam@95: cannam@95: Thu Jan 9 14:12:42 EST 2003 athena cannam@95: * [project @ 2003-01-09 19:12:42 by athena] cannam@95: Explicit cast cannam@95: cannam@95: M ./kernel/md5-1.c -1 +1 cannam@95: cannam@95: Thu Jan 9 13:41:51 EST 2003 athena cannam@95: * [project @ 2003-01-09 18:41:51 by athena] cannam@95: Added configure_planner(). mkplan() behaves properly when plan is null. cannam@95: cannam@95: A ./api/configure.c cannam@95: M ./api/Makefile.am -2 +2 cannam@95: M ./api/api.h +1 cannam@95: M ./api/apiplan.c -14 +21 cannam@95: M ./api/configure.c +33 cannam@95: M ./api/fftw3.h -3 +3 cannam@95: M ./api/the-planner.c -4 +1 cannam@95: cannam@95: Thu Jan 9 06:48:53 EST 2003 athena cannam@95: * [project @ 2003-01-09 11:48:53 by athena] cannam@95: More API work cannam@95: cannam@95: A ./api/apiplan.c cannam@95: A ./api/execute.c cannam@95: A ./api/mapflags.c cannam@95: M ./api/Makefile.am -4 +2 cannam@95: M ./api/api.h -2 +3 cannam@95: M ./api/apiplan.c +60 cannam@95: M ./api/execute.c +27 cannam@95: M ./api/fftw3.h -7 +3 cannam@95: M ./api/mapflags.c +26 cannam@95: M ./api/plan-dft-1d.c -11 +4 cannam@95: M ./tests/bench.c -1 +1 cannam@95: cannam@95: Thu Jan 9 05:40:34 EST 2003 athena cannam@95: * [project @ 2003-01-09 10:40:34 by athena] cannam@95: First skeleton of API infrastructure cannam@95: cannam@95: A ./api/ cannam@95: A ./api/Makefile.am cannam@95: A ./api/api.h cannam@95: A ./api/dfthelp.c cannam@95: A ./api/fftw3.h cannam@95: A ./api/plan-dft-1d.c cannam@95: A ./api/the-planner.c cannam@95: M ./Makefile.am -5 +7 cannam@95: M ./api/Makefile.am +10 cannam@95: M ./api/api.h +51 cannam@95: M ./api/dfthelp.c +37 cannam@95: M ./api/fftw3.h +67 cannam@95: M ./api/plan-dft-1d.c +43 cannam@95: M ./api/the-planner.c +42 cannam@95: M ./configure.ac +2 cannam@95: M ./kernel/ifftw.h -1 +2 cannam@95: M ./kernel/trig.c -3 +1 cannam@95: M ./tests/Makefile.am -1 +1 cannam@95: cannam@95: Thu Jan 9 03:19:35 EST 2003 stevenj cannam@95: * [project @ 2003-01-09 08:19:35 by stevenj] cannam@95: unsigned strikes again cannam@95: cannam@95: M ./rdft/rdft2-tensor-max-index.c -3 +3 cannam@95: cannam@95: Thu Jan 9 01:51:45 EST 2003 stevenj cannam@95: * [project @ 2003-01-09 06:51:45 by stevenj] cannam@95: put rdft2_inplace_strides and rdft2_tensor_max_index in their own files for tighter linking cannam@95: cannam@95: A ./rdft/rdft2-inplace-strides.c cannam@95: A ./rdft/rdft2-tensor-max-index.c cannam@95: M ./rdft/Makefile.am -1 +2 cannam@95: M ./rdft/problem2.c -46 +1 cannam@95: M ./rdft/rdft2-inplace-strides.c +68 cannam@95: M ./rdft/rdft2-tensor-max-index.c +47 cannam@95: M ./rdft/vrank-geq1-rdft2.c -25 +1 cannam@95: cannam@95: Thu Jan 9 01:43:13 EST 2003 stevenj cannam@95: * [project @ 2003-01-09 06:43:13 by stevenj] cannam@95: added rdft2_tensor_max_index...incorrect use of tensor_max_index was preventing proper loop ordering for rnk > 2 rdft2 cannam@95: cannam@95: M ./rdft/rank-geq2-rdft2.c -2 +3 cannam@95: M ./rdft/rdft.h -1 +2 cannam@95: M ./rdft/vrank-geq1-rdft2.c -4 +27 cannam@95: cannam@95: Thu Jan 9 00:44:45 EST 2003 stevenj cannam@95: * [project @ 2003-01-09 05:44:45 by stevenj] cannam@95: arbitrary spltrnk in rdft2 rank-geq2 cannam@95: cannam@95: M ./rdft/rank-geq2-rdft2.c -11 +53 cannam@95: cannam@95: Thu Jan 9 00:40:17 EST 2003 stevenj cannam@95: * [project @ 2003-01-09 05:40:17 by stevenj] cannam@95: don't mention wisdom when non-verbose cannam@95: cannam@95: M ./tests/bench.c -2 +4 cannam@95: cannam@95: Thu Jan 9 00:02:35 EST 2003 stevenj cannam@95: * [project @ 2003-01-09 05:02:35 by stevenj] cannam@95: bug fix: printing %T should pass tensor *, not tensor ** cannam@95: cannam@95: M ./dft/problem.c -3 +3 cannam@95: M ./rdft/problem.c -3 +3 cannam@95: M ./rdft/problem2.c -3 +3 cannam@95: cannam@95: Wed Jan 8 23:40:48 EST 2003 stevenj cannam@95: * [project @ 2003-01-09 04:40:48 by stevenj] cannam@95: correct(?) normalization for rodft00 ... all of the even/odd transforms should be normalized according to the expanded'' DFT of ~twice the length cannam@95: cannam@95: M ./reodft/rodft00e-r2hc.c -4 +4 cannam@95: M ./tests/verify-reodft.c -2 +2 cannam@95: cannam@95: Wed Jan 8 23:18:23 EST 2003 stevenj cannam@95: * [project @ 2003-01-09 04:18:23 by stevenj] cannam@95: fixed tests for n=1 cannam@95: cannam@95: M ./tests/verify-reodft.c -3 +6 cannam@95: cannam@95: Wed Jan 8 22:10:08 EST 2003 stevenj cannam@95: * [project @ 2003-01-09 03:10:08 by stevenj] cannam@95: fixed bug in vector tests for rdft(2) cannam@95: cannam@95: M ./tests/bench.c -3 +5 cannam@95: cannam@95: Wed Jan 8 20:12:00 EST 2003 stevenj cannam@95: * [project @ 2003-01-09 01:12:00 by stevenj] cannam@95: fixed handling when first rnk-1 dimensions compress to nothing (ugh) cannam@95: cannam@95: M ./rdft/problem2.c -3 +6 cannam@95: cannam@95: Wed Jan 8 20:02:35 EST 2003 stevenj cannam@95: * [project @ 2003-01-09 01:02:35 by stevenj] cannam@95: fixed incorrect/missing rdft2 rank-0 handling cannam@95: cannam@95: A ./rdft/rank0-rdft2.c cannam@95: M ./rdft/Makefile.am -2 +2 cannam@95: M ./rdft/conf.c -1 +2 cannam@95: M ./rdft/nop2.c -3 +6 cannam@95: M ./rdft/rank0-rdft2.c +194 cannam@95: M ./rdft/rdft.h -1 +2 cannam@95: cannam@95: Wed Jan 8 19:49:05 EST 2003 stevenj cannam@95: * [project @ 2003-01-09 00:49:05 by stevenj] cannam@95: bug fix: for rnk > 1, must compress rnk-1 dims separately (ugh) cannam@95: cannam@95: M ./rdft/problem2.c -2 +11 cannam@95: cannam@95: Wed Jan 8 17:39:14 EST 2003 stevenj cannam@95: * [project @ 2003-01-08 22:39:14 by stevenj] cannam@95: added trailing newline cannam@95: cannam@95: M ./configure.ac +1 cannam@95: cannam@95: Wed Jan 8 17:38:02 EST 2003 stevenj cannam@95: * [project @ 2003-01-08 22:38:02 by stevenj] cannam@95: updated cannam@95: cannam@95: M ./ChangeLog +66 cannam@95: cannam@95: Wed Jan 8 16:53:16 EST 2003 stevenj cannam@95: * [project @ 2003-01-08 21:53:16 by stevenj] cannam@95: got rid of compiler warning cannam@95: cannam@95: M ./rdft/problem.c -2 +2 cannam@95: cannam@95: Wed Jan 8 16:49:48 EST 2003 stevenj cannam@95: * [project @ 2003-01-08 21:49:48 by stevenj] cannam@95: whoops, test r2hc and not rodft00 by default cannam@95: cannam@95: M ./tests/bench.c -1 +1 cannam@95: cannam@95: Wed Jan 8 16:46:24 EST 2003 stevenj cannam@95: * [project @ 2003-01-08 21:46:24 by stevenj] cannam@95: got rid of real_n...use physical n everywhere in rdft; fixed rdft sz compression; fixed rodft00 verify bug cannam@95: cannam@95: M ./rdft/buffered.c -3 +2 cannam@95: M ./rdft/indirect.c -10 +6 cannam@95: M ./rdft/problem.c -44 +43 cannam@95: M ./rdft/rank-geq2.c -10 +6 cannam@95: M ./rdft/rdft.h -3 +1 cannam@95: M ./reodft/redft00e-r2hc.c -4 +5 cannam@95: M ./reodft/reodft010e-r2hc.c -4 +3 cannam@95: M ./reodft/reodft11e-r2hc.c -4 +3 cannam@95: M ./reodft/rodft00e-r2hc.c -4 +3 cannam@95: M ./tests/bench.c -2 +2 cannam@95: M ./tests/verify-reodft.c -33 +35 cannam@95: cannam@95: Wed Jan 8 07:20:47 EST 2003 athena cannam@95: * [project @ 2003-01-08 12:20:47 by athena] cannam@95: icc-6.0 bug workaround cannam@95: cannam@95: A ./simd/sse-aux.c cannam@95: A ./simd/sse2-aux.c cannam@95: M ./simd/Makefile.am -2 +2 cannam@95: M ./simd/sse-aux.c +33 cannam@95: M ./simd/sse.c -3 +1 cannam@95: M ./simd/sse2-aux.c +34 cannam@95: M ./simd/sse2.c -3 +1 cannam@95: cannam@95: Wed Jan 8 04:21:40 EST 2003 athena cannam@95: * [project @ 2003-01-08 09:21:40 by athena] cannam@95: Reclaimed the fftw_real identifier, because I need it for the API cannam@95: cannam@95: M ./kernel/ifftw.h -12 +9 cannam@95: M ./rdft/buffered2.c -5 +5 cannam@95: M ./rdft/rader-hc2hc.c -1 +1 cannam@95: M ./tests/bench.c +1 cannam@95: cannam@95: Wed Jan 8 04:14:55 EST 2003 athena cannam@95: * [project @ 2003-01-08 09:14:55 by athena] cannam@95: Use recommended AC_OUTPUT syntax cannam@95: cannam@95: M ./configure.ac -1 +2 cannam@95: cannam@95: Wed Jan 8 04:00:22 EST 2003 athena cannam@95: * [project @ 2003-01-08 09:00:22 by athena] cannam@95: Removed FFTW(foo) as a synonym for X(foo). This is an API issue. cannam@95: cannam@95: M ./kernel/ifftw.h -8 +7 cannam@95: M ./tests/bench.c +1 cannam@95: cannam@95: Tue Jan 7 17:45:52 EST 2003 stevenj cannam@95: * [project @ 2003-01-07 22:45:52 by stevenj] cannam@95: get rid of warning cannam@95: cannam@95: M ./simd/sse2.c -1 +2 cannam@95: cannam@95: Tue Jan 7 16:22:39 EST 2003 athena cannam@95: * [project @ 2003-01-07 21:22:39 by athena] cannam@95: Renamed conflicting files */codelet.h into dft/codelet-dft.h and cannam@95: rdft/codelet-rdft.h cannam@95: cannam@95: A ./dft/codelet-dft.h cannam@95: A ./rdft/codelet-rdft.h cannam@95: A ./support/codelet_prelude.dft cannam@95: A ./support/codelet_prelude.rdft cannam@95: M ./dft/Makefile.am -1 +1 cannam@95: M ./dft/codelet-dft.h +113 cannam@95: R ./dft/codelet.h cannam@95: M ./dft/codelets/inplace/Makefile.am -4 +4 cannam@95: M ./dft/codelets/n.c -1 +1 cannam@95: M ./dft/codelets/standard/Makefile.am -3 +3 cannam@95: M ./dft/codelets/t.c -1 +1 cannam@95: M ./dft/dft.h -2 +2 cannam@95: M ./dft/simd/codelets/Makefile.am -6 +6 cannam@95: M ./dft/simd/n1b.c -1 +1 cannam@95: M ./dft/simd/n1f.c -1 +1 cannam@95: M ./dft/simd/t1b.c -1 +1 cannam@95: M ./dft/simd/t1f.c -1 +1 cannam@95: M ./rdft/Makefile.am -5 +5 cannam@95: M ./rdft/codelet-rdft.h +190 cannam@95: R ./rdft/codelet.h cannam@95: M ./rdft/codelets/hc2r/Makefile.am -4 +4 cannam@95: M ./rdft/codelets/hc2r.c -1 +1 cannam@95: M ./rdft/codelets/hfb.c -1 +1 cannam@95: M ./rdft/codelets/r2hc/Makefile.am -4 +4 cannam@95: M ./rdft/codelets/r2hc.c -1 +1 cannam@95: M ./rdft/rdft.h -2 +2 cannam@95: M ./support/Makefile.am -1 +2 cannam@95: M ./support/Makefile.codelets -2 +4 cannam@95: R ./support/codelet_prelude cannam@95: M ./support/codelet_prelude.dft +8 cannam@95: M ./support/codelet_prelude.rdft +8 cannam@95: cannam@95: Tue Jan 7 16:21:16 EST 2003 stevenj cannam@95: * [project @ 2003-01-07 21:21:16 by stevenj] cannam@95: updated cannam@95: cannam@95: M ./ChangeLog +737 cannam@95: cannam@95: Tue Jan 7 15:47:24 EST 2003 athena cannam@95: * [project @ 2003-01-07 20:47:24 by athena] cannam@95: Silence warnings cannam@95: cannam@95: M ./simd/simd-3dnow.h -1 +1 cannam@95: M ./simd/simd-sse.h -1 cannam@95: M ./simd/simd-sse2.h -1 +1 cannam@95: M ./simd/sse2.c -2 +1 cannam@95: cannam@95: Tue Jan 7 15:00:14 EST 2003 stevenj cannam@95: * [project @ 2003-01-07 20:00:14 by stevenj] cannam@95: fftw2 used spltrnk=1 cannam@95: cannam@95: M ./dft/rank-geq2.c -2 +4 cannam@95: M ./rdft/rank-geq2.c -2 +4 cannam@95: cannam@95: Tue Jan 7 14:32:06 EST 2003 athena cannam@95: * [project @ 2003-01-07 19:32:06 by athena] cannam@95: Silence warning cannam@95: cannam@95: M ./dft/codelet.h -2 +1 cannam@95: M ./rdft/codelet.h -2 +1 cannam@95: M ./simd/simd-sse.h -1 +2 cannam@95: M ./simd/sse.c -2 +1 cannam@95: cannam@95: Tue Jan 7 12:13:50 EST 2003 stevenj cannam@95: * [project @ 2003-01-07 17:13:50 by stevenj] cannam@95: noted deficiency cannam@95: cannam@95: M ./TODO +2 cannam@95: cannam@95: Tue Jan 7 07:18:51 EST 2003 athena cannam@95: * [project @ 2003-01-07 12:18:51 by athena] cannam@95: Strengthened conditions for a problem to be POSSIBLY_UNALIGNED cannam@95: cannam@95: M ./rdft/vrank-geq1-rdft2.c -4 +2 cannam@95: M ./rdft/vrank-geq1.c -2 +2 cannam@95: cannam@95: Tue Jan 7 05:09:42 EST 2003 athena cannam@95: * [project @ 2003-01-07 10:09:42 by athena] cannam@95: Strengthened conditions for a plan to be POSSIBLY_UNALIGNED cannam@95: cannam@95: M ./dft/vrank-geq1.c -5 +2 cannam@95: M ./kernel/align.c -1 +6 cannam@95: M ./kernel/ifftw.h -1 +2 cannam@95: cannam@95: Sun Jan 5 02:43:45 EST 2003 stevenj cannam@95: * [project @ 2003-01-05 07:43:45 by stevenj] cannam@95: added copyright todo cannam@95: cannam@95: M ./TODO +2 cannam@95: cannam@95: Sun Jan 5 02:37:31 EST 2003 stevenj cannam@95: * [project @ 2003-01-05 07:37:31 by stevenj] cannam@95: modified comment cannam@95: cannam@95: M ./kernel/planner.c -2 +3 cannam@95: cannam@95: Sun Jan 5 02:34:36 EST 2003 stevenj cannam@95: * [project @ 2003-01-05 07:33:41 by stevenj] cannam@95: fixed comment cannam@95: cannam@95: M ./tests/verify-rdft.c -3 +3 cannam@95: cannam@95: Sun Jan 5 02:31:56 EST 2003 stevenj cannam@95: * [project @ 2003-01-05 07:31:56 by stevenj] cannam@95: implemented rdft2 verify cannam@95: cannam@95: M ./TODO -4 +2 cannam@95: M ./tests/verify-rdft.c -5 +15 cannam@95: cannam@95: Sat Jan 4 16:20:42 EST 2003 stevenj cannam@95: * [project @ 2003-01-04 21:20:42 by stevenj] cannam@95: fix --enable-single cannam@95: cannam@95: M ./configure.ac -1 +1 cannam@95: cannam@95: Wed Oct 23 12:59:12 EDT 2002 stevenj cannam@95: * [project @ 2002-10-23 16:59:12 by stevenj] cannam@95: slight fixes cannam@95: cannam@95: M ./threads/threads.c -20 +14 cannam@95: cannam@95: Wed Oct 23 12:42:39 EDT 2002 stevenj cannam@95: * [project @ 2002-10-23 16:42:39 by stevenj] cannam@95: typo cannam@95: cannam@95: M ./threads/threads.c -1 +1 cannam@95: cannam@95: Tue Oct 1 09:32:56 EDT 2002 athena cannam@95: * [project @ 2002-10-01 13:32:56 by athena] cannam@95: Experimental stuff cannam@95: cannam@95: M ./genfft/annotate.ml -2 +9 cannam@95: M ./genfft/annotate.mli -3 +2 cannam@95: M ./genfft/c.ml -4 +16 cannam@95: M ./genfft/genutil.ml -1 +11 cannam@95: M ./genfft/magic.ml -1 +5 cannam@95: cannam@95: Sat Sep 28 13:03:53 EDT 2002 athena cannam@95: * [project @ 2002-09-28 17:03:53 by athena] cannam@95: Experimental Franz mode cannam@95: cannam@95: M ./configure.ac +3 cannam@95: M ./dft/simd/codelets/Makefile.am -3 +68 cannam@95: M ./genfft/gen_notw_c.ml -7 +8 cannam@95: M ./genfft/gen_twiddle_c.ml -6 +7 cannam@95: M ./genfft/genutil.ml -2 +10 cannam@95: cannam@95: Thu Sep 26 15:14:38 EDT 2002 athena cannam@95: * [project @ 2002-09-26 19:14:38 by athena] cannam@95: const-correct cannam@95: cannam@95: M ./kernel/tensor.c -2 +2 cannam@95: cannam@95: Thu Sep 26 15:06:38 EDT 2002 athena cannam@95: * [project @ 2002-09-26 19:06:38 by athena] cannam@95: Reuse dimcmp routine for other purposes cannam@95: cannam@95: M ./dft/vrank2-transpose.c -4 +3 cannam@95: M ./dft/vrank3-transpose.c -4 +2 cannam@95: M ./kernel/ifftw.h -1 +2 cannam@95: M ./kernel/tensor7.c -5 +5 cannam@95: M ./rdft/vrank2-transpose.c -4 +2 cannam@95: M ./rdft/vrank3-transpose.c -4 +2 cannam@95: cannam@95: Wed Sep 25 07:37:38 EDT 2002 athena cannam@95: * [project @ 2002-09-25 11:37:38 by athena] cannam@95: Use tornk1 correctly. cannam@95: cannam@95: M ./dft/direct.c -3 +3 cannam@95: M ./kernel/ifftw.h -2 +2 cannam@95: M ./kernel/tensor.c -2 +3 cannam@95: M ./rdft/direct.c -3 +3 cannam@95: M ./rdft/direct2.c -3 +3 cannam@95: cannam@95: Wed Sep 25 07:36:38 EDT 2002 athena cannam@95: * [project @ 2002-09-25 11:36:38 by athena] cannam@95: Hmm... I thought I had fixed this before... cannam@95: cannam@95: M ./rdft/rdft2-radix2.c -2 +2 cannam@95: cannam@95: Tue Sep 24 21:27:49 EDT 2002 athena cannam@95: * [project @ 2002-09-25 01:27:49 by athena] cannam@95: Collect more common idioms cannam@95: cannam@95: M ./dft/buffered.c -5 +3 cannam@95: M ./dft/rank0.c -16 +3 cannam@95: M ./kernel/tensor.c -2 +2 cannam@95: M ./rdft/buffered.c -5 +2 cannam@95: M ./rdft/buffered2.c -5 +2 cannam@95: M ./rdft/rank0.c -16 +3 cannam@95: cannam@95: Tue Sep 24 21:15:57 EDT 2002 athena cannam@95: * [project @ 2002-09-25 01:15:57 by athena] cannam@95: Still collecting common idioms... cannam@95: cannam@95: M ./dft/direct.c -10 +3 cannam@95: M ./rdft/direct.c -10 +2 cannam@95: M ./rdft/direct2.c -9 +2 cannam@95: cannam@95: Tue Sep 24 21:13:00 EDT 2002 athena cannam@95: * [project @ 2002-09-25 01:13:00 by athena] cannam@95: More garbage collection. cannam@95: cannam@95: M ./dft/direct.c -6 +5 cannam@95: M ./rdft/direct.c -6 +5 cannam@95: M ./rdft/direct2.c -6 +5 cannam@95: cannam@95: Tue Sep 24 21:08:19 EDT 2002 athena cannam@95: * [project @ 2002-09-25 01:08:19 by athena] cannam@95: More compact code cannam@95: cannam@95: M ./dft/buffered.c -6 +3 cannam@95: cannam@95: Tue Sep 24 20:54:43 EDT 2002 athena cannam@95: * [project @ 2002-09-25 00:54:43 by athena] cannam@95: Collect common pattern if (foo) free(foo) ==> free0(foo) cannam@95: cannam@95: M ./dft/buffered.c -7 +6 cannam@95: M ./dft/generic.c -2 +1 cannam@95: M ./dft/rader.c -6 +3 cannam@95: M ./kernel/alloc.c -1 +7 cannam@95: M ./kernel/ifftw.h -1 +2 cannam@95: M ./kernel/planner.c -8 +4 cannam@95: M ./kernel/stride.c -3 +2 cannam@95: M ./kernel/tensor.c -5 +2 cannam@95: M ./rdft/buffered.c -3 +2 cannam@95: M ./rdft/buffered2.c -3 +2 cannam@95: M ./rdft/dht-rader.c -2 +1 cannam@95: M ./rdft/generic.c -2 +1 cannam@95: M ./rdft/problem.c -3 +2 cannam@95: M ./rdft/rader-hc2hc.c -6 +3 cannam@95: cannam@95: Tue Sep 24 20:08:44 EDT 2002 athena cannam@95: * [project @ 2002-09-25 00:08:44 by athena] cannam@95: Collect some common code in */buffered*.c cannam@95: cannam@95: A ./kernel/buffered.c cannam@95: M ./dft/buffered.c -18 +2 cannam@95: M ./kernel/Makefile.am -7 +6 cannam@95: M ./kernel/buffered.c +44 cannam@95: M ./kernel/ifftw.h -1 +2 cannam@95: M ./rdft/buffered.c -18 +3 cannam@95: M ./rdft/buffered2.c -18 +2 cannam@95: cannam@95: Tue Sep 24 19:39:22 EDT 2002 stevenj cannam@95: * [project @ 2002-09-24 23:39:22 by stevenj] cannam@95: use STRUCT_HACK #define to determing rdft kind[] allocation cannam@95: cannam@95: M ./rdft/problem.c -8 +23 cannam@95: M ./rdft/rdft.h -2 +8 cannam@95: cannam@95: Tue Sep 24 17:21:09 EDT 2002 stevenj cannam@95: * [project @ 2002-09-24 21:21:09 by stevenj] cannam@95: report total pcost of measured/estimated plans...epcost is especially useful to estimate the effects of various impatience flags on planning time for large transforms cannam@95: cannam@95: M ./kernel/ifftw.h -1 +2 cannam@95: M ./kernel/planner.c -1 +6 cannam@95: cannam@95: Mon Sep 23 18:49:10 EDT 2002 athena cannam@95: * [project @ 2002-09-23 22:49:10 by athena] cannam@95: Prevent unwanted inlining cannam@95: cannam@95: A ./kernel/trig1.c cannam@95: M ./kernel/Makefile.am -1 +1 cannam@95: M ./kernel/trig.c -54 +7 cannam@95: M ./kernel/trig1.c +70 cannam@95: cannam@95: Mon Sep 23 18:37:59 EDT 2002 athena cannam@95: * [project @ 2002-09-23 22:37:59 by athena] cannam@95: Space compaction cannam@95: cannam@95: M ./kernel/ifftw.h -2 +1 cannam@95: M ./kernel/trig.c -21 +27 cannam@95: cannam@95: Mon Sep 23 11:49:32 EDT 2002 athena cannam@95: * [project @ 2002-09-23 15:49:32 by athena] cannam@95: Still reducing size cannam@95: cannam@95: A ./kernel/hash.c cannam@95: M ./kernel/Makefile.am -1 +1 cannam@95: M ./kernel/hash.c +31 cannam@95: M ./kernel/ifftw.h -1 +2 cannam@95: M ./kernel/md5-1.c +1 cannam@95: M ./kernel/planner.c -12 +3 cannam@95: M ./kernel/scan.c -22 +1 cannam@95: cannam@95: Sun Sep 22 16:03:30 EDT 2002 athena cannam@95: * [project @ 2002-09-22 20:03:30 by athena] cannam@95: Saved another 5KB by redesigning opcnt protocol. (gasp!) cannam@95: cannam@95: M ./dft/buffered.c -5 +6 cannam@95: M ./dft/ct-dif.c -4 +3 cannam@95: M ./dft/ct-dit.c -4 +3 cannam@95: M ./dft/ct-ditbuf.c -6 +6 cannam@95: M ./dft/ct-ditf.c -4 +3 cannam@95: M ./dft/direct.c -2 +3 cannam@95: M ./dft/generic.c -4 +2 cannam@95: M ./dft/indirect.c -2 +2 cannam@95: M ./dft/nop.c -2 +2 cannam@95: M ./dft/rader.c -4 +3 cannam@95: M ./dft/rank-geq2.c -2 +2 cannam@95: M ./dft/rank0.c -2 +2 cannam@95: M ./dft/vrank-geq1.c -2 +3 cannam@95: M ./dft/vrank2-transpose.c -2 +2 cannam@95: M ./dft/vrank3-transpose.c -2 +2 cannam@95: M ./kernel/ifftw.h -6 +14 cannam@95: M ./kernel/ops.c -23 +28 cannam@95: M ./kernel/plan.c -2 +2 cannam@95: M ./rdft/buffered.c -5 +6 cannam@95: M ./rdft/buffered2.c -3 +3 cannam@95: M ./rdft/dht-rader.c -1 +1 cannam@95: M ./rdft/direct.c -3 +8 cannam@95: M ./rdft/direct2.c -3 +8 cannam@95: M ./rdft/generic.c -4 +3 cannam@95: M ./rdft/hc2hc-buf.c -10 +9 cannam@95: M ./rdft/hc2hc-dif.c -8 +8 cannam@95: M ./rdft/hc2hc-dit.c -8 +8 cannam@95: M ./rdft/indirect.c -2 +2 cannam@95: M ./rdft/nop.c -2 +2 cannam@95: M ./rdft/nop2.c -2 +2 cannam@95: M ./rdft/rader-hc2hc.c -8 +5 cannam@95: M ./rdft/rank-geq2-rdft2.c -2 +2 cannam@95: M ./rdft/rank-geq2.c -2 +2 cannam@95: M ./rdft/rank0.c -2 +2 cannam@95: M ./rdft/rdft2-radix2.c -5 +4 cannam@95: M ./rdft/vrank-geq1-rdft2.c -2 +3 cannam@95: M ./rdft/vrank-geq1.c -2 +3 cannam@95: M ./rdft/vrank2-transpose.c -2 +2 cannam@95: M ./rdft/vrank3-transpose.c -2 +2 cannam@95: M ./threads/ct-dit.c -4 +3 cannam@95: M ./threads/dft-vrank-geq1.c -4 +3 cannam@95: M ./threads/hc2hc-dif.c -8 +8 cannam@95: M ./threads/hc2hc-dit.c -8 +8 cannam@95: M ./threads/rdft-vrank-geq1.c -4 +3 cannam@95: M ./threads/vrank-geq1-rdft2.c -4 +3 cannam@95: cannam@95: Sun Sep 22 15:00:59 EDT 2002 athena cannam@95: * [project @ 2002-09-22 19:00:59 by athena] cannam@95: More code compression cannam@95: cannam@95: A ./kernel/tensor8.c cannam@95: M ./dft/buffered.c -4 +3 cannam@95: M ./dft/direct.c -3 +2 cannam@95: M ./dft/indirect.c -4 +3 cannam@95: M ./dft/problem.c -5 +3 cannam@95: M ./dft/rank-geq2.c -9 +3 cannam@95: M ./kernel/Makefile.am -2 +2 cannam@95: M ./kernel/ifftw.h -1 +5 cannam@95: M ./kernel/tensor1.c -1 +6 cannam@95: M ./kernel/tensor4.c -1 +6 cannam@95: M ./kernel/tensor8.c +35 cannam@95: M ./rdft/buffered.c -4 +3 cannam@95: M ./rdft/dft-r2hc.c -2 +2 cannam@95: M ./rdft/direct.c -3 +2 cannam@95: M ./rdft/hc2hc.c -7 +6 cannam@95: M ./rdft/indirect.c -4 +3 cannam@95: M ./rdft/problem.c -7 +4 cannam@95: M ./rdft/problem2.c -5 +3 cannam@95: M ./rdft/rank-geq2-rdft2.c -9 +3 cannam@95: M ./rdft/rank-geq2.c -13 +5 cannam@95: cannam@95: Sun Sep 22 13:27:46 EDT 2002 athena cannam@95: * [project @ 2002-09-22 17:27:46 by athena] cannam@95: Smaller code size. cannam@95: cannam@95: M ./kernel/ifftw.h -2 +4 cannam@95: M ./kernel/solver.c -1 +6 cannam@95: cannam@95: Sun Sep 22 12:50:36 EDT 2002 athena cannam@95: * [project @ 2002-09-22 16:50:36 by athena] cannam@95: Started unification of rader cannam@95: cannam@95: A ./dft/rader-omega.c cannam@95: M ./dft/Makefile.am -3 +3 cannam@95: M ./dft/dft.h -1 +5 cannam@95: M ./dft/rader-omega.c +57 cannam@95: M ./dft/rader.c -38 +3 cannam@95: M ./rdft/rader-hc2hc.c -37 +3 cannam@95: cannam@95: Sun Sep 22 12:35:30 EDT 2002 athena cannam@95: * [project @ 2002-09-22 16:35:30 by athena] cannam@95: Typo cannam@95: cannam@95: M ./rdft/rdft2-radix2.c -2 +2 cannam@95: cannam@95: Sun Sep 22 12:25:20 EDT 2002 athena cannam@95: * [project @ 2002-09-22 16:25:20 by athena] cannam@95: Changed protocol for destroy_plan so as to save space. cannam@95: cannam@95: M ./dft/buffered.c -8 +4 cannam@95: M ./dft/ct.c -2 +1 cannam@95: M ./dft/direct.c -2 +1 cannam@95: M ./dft/generic.c -3 +1 cannam@95: M ./dft/indirect.c -6 +3 cannam@95: M ./dft/nop.c -7 +2 cannam@95: M ./dft/rader.c -7 +3 cannam@95: M ./dft/rank-geq2.c -6 +3 cannam@95: M ./dft/rank0.c -7 +2 cannam@95: M ./dft/vrank-geq1.c -2 +1 cannam@95: M ./dft/vrank2-transpose.c -7 +2 cannam@95: M ./dft/vrank3-transpose.c -7 +2 cannam@95: M ./kernel/ifftw.h -2 +2 cannam@95: M ./kernel/plan.c -3 +11 cannam@95: M ./kernel/problem.c -2 +3 cannam@95: M ./rdft/buffered.c -8 +4 cannam@95: M ./rdft/buffered2.c -6 +3 cannam@95: M ./rdft/dft-r2hc.c -2 +1 cannam@95: M ./rdft/dht-r2hc.c -2 +1 cannam@95: M ./rdft/dht-rader.c -7 +3 cannam@95: M ./rdft/direct.c -2 +1 cannam@95: M ./rdft/direct2.c -2 +1 cannam@95: M ./rdft/generic.c -3 +1 cannam@95: M ./rdft/hc2hc.c -14 +7 cannam@95: M ./rdft/indirect.c -6 +3 cannam@95: M ./rdft/nop.c -7 +2 cannam@95: M ./rdft/nop2.c -7 +2 cannam@95: M ./rdft/rader-hc2hc.c -9 +4 cannam@95: M ./rdft/rank-geq2-rdft2.c -6 +3 cannam@95: M ./rdft/rank-geq2.c -6 +3 cannam@95: M ./rdft/rank0.c -7 +2 cannam@95: M ./rdft/rdft-dht.c -2 +1 cannam@95: M ./rdft/rdft2-radix2.c -2 +1 cannam@95: M ./rdft/vrank-geq1-rdft2.c -2 +1 cannam@95: M ./rdft/vrank-geq1.c -2 +1 cannam@95: M ./rdft/vrank2-transpose.c -7 +2 cannam@95: M ./rdft/vrank3-transpose.c -7 +2 cannam@95: M ./reodft/redft00e-r2hc.c -2 +1 cannam@95: M ./reodft/reodft010e-r2hc.c -2 +1 cannam@95: M ./reodft/reodft11e-r2hc.c -2 +1 cannam@95: M ./reodft/rodft00e-r2hc.c -2 +1 cannam@95: M ./threads/dft-vrank-geq1.c -4 +2 cannam@95: M ./threads/rdft-vrank-geq1.c -4 +2 cannam@95: M ./threads/vrank-geq1-rdft2.c -4 +2 cannam@95: cannam@95: Sun Sep 22 11:08:57 EDT 2002 athena cannam@95: * [project @ 2002-09-22 15:08:57 by athena] cannam@95: Introduced convenient function X(mkplan_d) cannam@95: cannam@95: M ./dft/buffered.c -31 +18 cannam@95: M ./dft/ct.c -5 +2 cannam@95: M ./dft/generic.c -10 +5 cannam@95: M ./dft/indirect.c -14 +8 cannam@95: M ./dft/rader.c -38 +20 cannam@95: M ./dft/rank-geq2.c -16 +12 cannam@95: M ./dft/vrank-geq1.c -9 +7 cannam@95: M ./kernel/ifftw.h -2 +4 cannam@95: M ./kernel/planner.c -1 +8 cannam@95: M ./rdft/buffered.c -30 +20 cannam@95: M ./rdft/buffered2.c -9 +3 cannam@95: M ./rdft/dft-r2hc.c -7 +5 cannam@95: M ./rdft/dht-r2hc.c -7 +4 cannam@95: M ./rdft/dht-rader.c -23 +12 cannam@95: M ./rdft/generic.c -4 +1 cannam@95: M ./rdft/hc2hc.c -16 +10 cannam@95: M ./rdft/indirect.c -14 +9 cannam@95: M ./rdft/rader-hc2hc.c -41 +21 cannam@95: M ./rdft/rank-geq2-rdft2.c -12 +8 cannam@95: M ./rdft/rank-geq2.c -16 +12 cannam@95: M ./rdft/rdft-dht.c -5 +3 cannam@95: M ./rdft/rdft2-radix2.c -7 +3 cannam@95: M ./rdft/vrank-geq1-rdft2.c -9 +7 cannam@95: M ./rdft/vrank-geq1.c -10 +7 cannam@95: M ./reodft/redft00e-r2hc.c -3 +2 cannam@95: M ./reodft/reodft010e-r2hc.c -3 +2 cannam@95: M ./reodft/reodft11e-r2hc.c -3 +2 cannam@95: M ./reodft/rodft00e-r2hc.c -3 +2 cannam@95: M ./threads/ct-dit.c -4 +4 cannam@95: M ./threads/dft-vrank-geq1.c -12 +10 cannam@95: M ./threads/hc2hc-dif.c -4 +4 cannam@95: M ./threads/hc2hc-dit.c -4 +4 cannam@95: M ./threads/rdft-vrank-geq1.c -11 +9 cannam@95: M ./threads/vrank-geq1-rdft2.c -11 +9 cannam@95: cannam@95: Sun Sep 22 10:21:36 EDT 2002 athena cannam@95: * [project @ 2002-09-22 14:21:36 by athena] cannam@95: Split tensor/md5 into separate files to allow independent linking cannam@95: and/or prevent undesidred inlining cannam@95: cannam@95: A ./kernel/md5-1.c cannam@95: A ./kernel/tensor1.c cannam@95: A ./kernel/tensor2.c cannam@95: A ./kernel/tensor3.c cannam@95: A ./kernel/tensor4.c cannam@95: A ./kernel/tensor5.c cannam@95: A ./kernel/tensor7.c cannam@95: M ./kernel/Makefile.am -3 +4 cannam@95: M ./kernel/md5-1.c +53 cannam@95: M ./kernel/md5.c -31 cannam@95: M ./kernel/tensor.c -270 +2 cannam@95: M ./kernel/tensor1.c +32 cannam@95: M ./kernel/tensor2.c +37 cannam@95: M ./kernel/tensor3.c +46 cannam@95: M ./kernel/tensor4.c +68 cannam@95: M ./kernel/tensor5.c +93 cannam@95: M ./kernel/tensor7.c +127 cannam@95: cannam@95: Sun Sep 22 09:49:09 EDT 2002 athena cannam@95: * [project @ 2002-09-22 13:49:08 by athena] cannam@95: Treat all tensors as dynamically allocated objects. They were cannam@95: dynamically allocated in part anyway, so there is no point in cannam@95: complicating the object code with the clumsy calling conventions cannam@95: for by-value structs. cannam@95: cannam@95: M ./dft/buffered.c -19 +19 cannam@95: M ./dft/ct-dif.c -5 +5 cannam@95: M ./dft/ct-dit.c -5 +5 cannam@95: M ./dft/ct-ditbuf.c -4 +5 cannam@95: M ./dft/ct-ditf.c -5 +5 cannam@95: M ./dft/ct.c -14 +14 cannam@95: M ./dft/dft.h -4 +4 cannam@95: M ./dft/direct.c -13 +13 cannam@95: M ./dft/generic.c -7 +7 cannam@95: M ./dft/indirect.c -18 +16 cannam@95: M ./dft/nop.c -5 +5 cannam@95: M ./dft/problem.c -12 +12 cannam@95: M ./dft/rader.c -12 +12 cannam@95: M ./dft/rank-geq2.c -22 +22 cannam@95: M ./dft/rank0.c -14 +14 cannam@95: M ./dft/vrank-geq1.c -11 +11 cannam@95: M ./dft/vrank2-transpose.c -9 +9 cannam@95: M ./dft/vrank3-transpose.c -12 +12 cannam@95: M ./dft/zero.c -18 +20 cannam@95: M ./kernel/ifftw.h -13 +22 cannam@95: M ./kernel/tensor.c -92 +102 cannam@95: M ./rdft/buffered.c -23 +21 cannam@95: M ./rdft/buffered2.c -24 +24 cannam@95: M ./rdft/dft-r2hc.c -11 +11 cannam@95: M ./rdft/dht-r2hc.c -6 +6 cannam@95: M ./rdft/dht-rader.c -7 +7 cannam@95: M ./rdft/direct.c -18 +18 cannam@95: M ./rdft/direct2.c -16 +16 cannam@95: M ./rdft/generic.c -8 +8 cannam@95: M ./rdft/hc2hc-buf.c -4 +5 cannam@95: M ./rdft/hc2hc-dif.c -5 +5 cannam@95: M ./rdft/hc2hc-dit.c -5 +5 cannam@95: M ./rdft/hc2hc.c -23 +23 cannam@95: M ./rdft/indirect.c -20 +18 cannam@95: M ./rdft/nop.c -5 +5 cannam@95: M ./rdft/nop2.c -4 +4 cannam@95: M ./rdft/problem.c -39 +40 cannam@95: M ./rdft/problem2.c -34 +34 cannam@95: M ./rdft/rader-hc2hc.c -11 +11 cannam@95: M ./rdft/rank-geq2-rdft2.c -23 +23 cannam@95: M ./rdft/rank-geq2.c -28 +28 cannam@95: M ./rdft/rank0.c -10 +10 cannam@95: M ./rdft/rdft-dht.c -10 +10 cannam@95: M ./rdft/rdft.h -9 +9 cannam@95: M ./rdft/rdft2-radix2.c -21 +21 cannam@95: M ./rdft/vrank-geq1-rdft2.c -11 +11 cannam@95: M ./rdft/vrank-geq1.c -11 +11 cannam@95: M ./rdft/vrank2-transpose.c -9 +9 cannam@95: M ./rdft/vrank3-transpose.c -12 +12 cannam@95: M ./reodft/redft00e-r2hc.c -10 +10 cannam@95: M ./reodft/reodft010e-r2hc.c -10 +10 cannam@95: M ./reodft/reodft11e-r2hc.c -10 +10 cannam@95: M ./reodft/rodft00e-r2hc.c -10 +10 cannam@95: M ./tests/debug.h -2 +2 cannam@95: M ./tests/dotens.c -4 +4 cannam@95: M ./tests/dotens2.c -5 +5 cannam@95: M ./tests/verify-dft.c -10 +11 cannam@95: M ./tests/verify-lib.c -15 +16 cannam@95: M ./tests/verify-rdft.c -43 +47 cannam@95: M ./tests/verify-reodft.c -28 +28 cannam@95: M ./tests/verify.h -3 +3 cannam@95: cannam@95: Sat Sep 21 18:24:55 EDT 2002 stevenj cannam@95: * [project @ 2002-09-21 22:24:55 by stevenj] cannam@95: typo cannam@95: cannam@95: M ./kernel/ifftw.h -2 +2 cannam@95: cannam@95: Sat Sep 21 18:10:07 EDT 2002 athena cannam@95: * [project @ 2002-09-21 22:10:07 by athena] cannam@95: Avoid generating NaN when n = 0. cannam@95: cannam@95: M ./tests/verify-lib.c -13 +16 cannam@95: cannam@95: Sat Sep 21 18:04:05 EDT 2002 athena cannam@95: * [project @ 2002-09-21 22:04:05 by athena] cannam@95: Saved more. cannam@95: cannam@95: M ./dft/dft.h -2 +2 cannam@95: M ./dft/problem.c -5 +5 cannam@95: M ./dft/rank-geq2.c -2 +2 cannam@95: M ./rdft/dft-r2hc.c -2 +2 cannam@95: M ./rdft/dht-r2hc.c -2 +2 cannam@95: M ./rdft/hc2hc.c -5 +7 cannam@95: M ./rdft/problem.c -11 +11 cannam@95: M ./rdft/problem2.c -5 +5 cannam@95: M ./rdft/rank-geq2.c -2 +2 cannam@95: M ./rdft/rdft-dht.c -3 +3 cannam@95: M ./rdft/rdft.h -4 +4 cannam@95: M ./reodft/redft00e-r2hc.c -2 +2 cannam@95: M ./reodft/reodft010e-r2hc.c -2 +2 cannam@95: M ./reodft/reodft11e-r2hc.c -2 +2 cannam@95: M ./reodft/rodft00e-r2hc.c -2 +2 cannam@95: M ./threads/dft-vrank-geq1.c -2 +2 cannam@95: cannam@95: Sat Sep 21 17:47:36 EDT 2002 athena cannam@95: * [project @ 2002-09-21 21:47:35 by athena] cannam@95: Save 1200 bytes of object code. Do not pass structs by value whenever cannam@95: practical, because the calling protocol generates clumsy code. cannam@95: cannam@95: M ./dft/buffered.c -6 +6 cannam@95: M ./dft/ct.c -5 +5 cannam@95: M ./dft/direct.c -3 +3 cannam@95: M ./dft/indirect.c -12 +12 cannam@95: M ./dft/nop.c -2 +2 cannam@95: M ./dft/problem.c -11 +11 cannam@95: M ./dft/rank-geq2.c -21 +21 cannam@95: M ./dft/vrank-geq1.c -6 +6 cannam@95: M ./dft/vrank3-transpose.c -8 +8 cannam@95: M ./kernel/ifftw.h -20 +22 cannam@95: M ./kernel/pickdim.c -9 +9 cannam@95: M ./kernel/print.c -2 +2 cannam@95: M ./kernel/tensor.c -69 +71 cannam@95: M ./rdft/buffered.c -6 +6 cannam@95: M ./rdft/buffered2.c -2 +2 cannam@95: M ./rdft/dft-r2hc.c -3 +3 cannam@95: M ./rdft/direct.c -3 +3 cannam@95: M ./rdft/hc2hc.c -9 +9 cannam@95: M ./rdft/indirect.c -14 +14 cannam@95: M ./rdft/nop.c -2 +2 cannam@95: M ./rdft/problem.c -15 +15 cannam@95: M ./rdft/problem2.c -15 +15 cannam@95: M ./rdft/rank-geq2-rdft2.c -19 +19 cannam@95: M ./rdft/rank-geq2.c -27 +27 cannam@95: M ./rdft/rdft-dht.c -3 +3 cannam@95: M ./rdft/rdft.h -2 +2 cannam@95: M ./rdft/rdft2-radix2.c -7 +7 cannam@95: M ./rdft/vrank-geq1-rdft2.c -6 +7 cannam@95: M ./rdft/vrank-geq1.c -6 +6 cannam@95: M ./rdft/vrank3-transpose.c -8 +8 cannam@95: M ./reodft/redft00e-r2hc.c -2 +2 cannam@95: M ./reodft/reodft010e-r2hc.c -2 +2 cannam@95: M ./reodft/reodft11e-r2hc.c -2 +2 cannam@95: M ./reodft/rodft00e-r2hc.c -2 +2 cannam@95: M ./tests/verify-dft.c -6 +6 cannam@95: M ./tests/verify-lib.c -2 +2 cannam@95: M ./tests/verify-rdft.c -21 +21 cannam@95: M ./tests/verify-reodft.c -10 +10 cannam@95: M ./threads/dft-vrank-geq1.c -3 +3 cannam@95: M ./threads/rdft-vrank-geq1.c -3 +3 cannam@95: M ./threads/vrank-geq1-rdft2.c -3 +4 cannam@95: cannam@95: Sat Sep 21 12:10:21 EDT 2002 athena cannam@95: * [project @ 2002-09-21 16:10:21 by athena] cannam@95: Do not allocate buffers for rader omegas. Let the planner do it cannam@95: if necessary. cannam@95: cannam@95: M ./rdft/dht-rader.c -15 +6 cannam@95: cannam@95: Sat Sep 21 12:03:46 EDT 2002 athena cannam@95: * [project @ 2002-09-21 16:03:46 by athena] cannam@95: Check rank *before* reading kind[0], which may be undefined if rnk < 1 cannam@95: cannam@95: M ./tests/verify-rdft.c -2 +2 cannam@95: M ./tests/verify-reodft.c -2 +2 cannam@95: cannam@95: Sat Sep 21 11:48:50 EDT 2002 athena cannam@95: * [project @ 2002-09-21 15:48:50 by athena] cannam@95: Second step towards rader unification. cannam@95: cannam@95: M ./dft/rader.c -17 +7 cannam@95: M ./rdft/rader-hc2hc.c -1 +1 cannam@95: cannam@95: Sat Sep 21 11:37:06 EDT 2002 athena cannam@95: * [project @ 2002-09-21 15:37:06 by athena] cannam@95: First step towards unification of Rader code cannam@95: cannam@95: A ./kernel/rader.c cannam@95: M ./dft/rader.c -57 +12 cannam@95: M ./kernel/Makefile.am -3 +3 cannam@95: M ./kernel/ifftw.h -1 +9 cannam@95: M ./kernel/rader.c +68 cannam@95: M ./rdft/dht-rader.c -50 +6 cannam@95: M ./rdft/rader-hc2hc.c -57 +11 cannam@95: cannam@95: Sat Sep 21 07:58:11 EDT 2002 athena cannam@95: * [project @ 2002-09-21 11:58:11 by athena] cannam@95: Fix ugliness condition for cooley-tukey. cannam@95: cannam@95: A ./kernel/ct.c cannam@95: M ./dft/ct-dif.c -5 +3 cannam@95: M ./dft/ct-dit.c -4 +2 cannam@95: M ./dft/ct-ditbuf.c -5 +3 cannam@95: M ./kernel/Makefile.am -4 +5 cannam@95: M ./kernel/ct.c +31 cannam@95: M ./kernel/ifftw.h -1 +2 cannam@95: M ./kernel/planner.c -3 +2 cannam@95: M ./rdft/dht-r2hc.c -3 +3 cannam@95: M ./rdft/dht-rader.c -3 +3 cannam@95: M ./rdft/hc2hc-buf.c -9 +3 cannam@95: M ./rdft/hc2hc-dif.c -4 +2 cannam@95: M ./rdft/hc2hc-dit.c -4 +2 cannam@95: M ./rdft/rdft-dht.c -4 +2 cannam@95: M ./reodft/redft00e-r2hc.c -4 +2 cannam@95: M ./reodft/reodft010e-r2hc.c -4 +2 cannam@95: M ./reodft/reodft11e-r2hc.c -4 +2 cannam@95: M ./reodft/rodft00e-r2hc.c -4 +2 cannam@95: M ./threads/ct-dit.c -5 +3 cannam@95: M ./threads/hc2hc-dif.c -5 +3 cannam@95: M ./threads/hc2hc-dit.c -5 +3 cannam@95: cannam@95: Fri Sep 20 16:53:45 EDT 2002 athena cannam@95: * [project @ 2002-09-20 20:53:45 by athena] cannam@95: Removed RADER_MIN_GOOD and associated machinery cannam@95: cannam@95: M ./dft/rader.c -29 +6 cannam@95: M ./kernel/ifftw.h -2 +1 cannam@95: M ./rdft/dht-rader.c -14 +3 cannam@95: M ./rdft/rader-hc2hc.c -17 +5 cannam@95: cannam@95: Fri Sep 20 14:49:12 EDT 2002 athena cannam@95: * [project @ 2002-09-20 18:49:12 by athena] cannam@95: Proper cast cannam@95: cannam@95: M ./rdft/dht-r2hc.c -4 +3 cannam@95: cannam@95: Fri Sep 20 14:45:54 EDT 2002 athena cannam@95: * [project @ 2002-09-20 18:45:54 by athena] cannam@95: Typo cannam@95: cannam@95: M ./kernel/planner.c -2 +2 cannam@95: cannam@95: Fri Sep 20 14:38:13 EDT 2002 athena cannam@95: * [project @ 2002-09-20 18:38:13 by athena] cannam@95: Implemented NO_LARGE_GENERIC cannam@95: cannam@95: M ./dft/generic.c -1 +8 cannam@95: M ./kernel/ifftw.h -20 +23 cannam@95: M ./rdft/dht-rader.c -1 +1 cannam@95: M ./rdft/generic.c -2 +8 cannam@95: M ./rdft/rdft-dht.c -10 +2 cannam@95: M ./tests/bench.c +1 cannam@95: cannam@95: Thu Sep 19 07:48:25 EDT 2002 athena cannam@95: * [project @ 2002-09-19 11:48:24 by athena] cannam@95: Consistent macroization of NO_DHT_R2HC cannam@95: cannam@95: M ./kernel/ifftw.h -1 +2 cannam@95: M ./rdft/dht-r2hc.c -4 +4 cannam@95: cannam@95: Wed Sep 18 21:47:17 EDT 2002 athena cannam@95: * [project @ 2002-09-19 01:47:17 by athena] cannam@95: NO_DHT_R2HC is a planner flag, otherwise the EXHAUSTIVE planner loops. cannam@95: cannam@95: M ./kernel/ifftw.h -7 +6 cannam@95: M ./kernel/planner.c -10 +6 cannam@95: M ./rdft/dht-r2hc.c -3 +3 cannam@95: M ./tests/bench.c +1 cannam@95: cannam@95: Wed Sep 18 20:47:31 EDT 2002 athena cannam@95: * [project @ 2002-09-19 00:47:31 by athena] cannam@95: Resurrected NO_EXHAUSTIVE cannam@95: cannam@95: M ./kernel/ifftw.h -1 +3 cannam@95: M ./kernel/planner.c -8 +24 cannam@95: cannam@95: Wed Sep 18 19:31:57 EDT 2002 stevenj cannam@95: * [project @ 2002-09-18 23:31:57 by stevenj] cannam@95: au revoir, score() cannam@95: cannam@95: M ./threads/ct-dit.c -25 +22 cannam@95: M ./threads/dft-vrank-geq1.c -14 +10 cannam@95: M ./threads/hc2hc-dif.c -25 +21 cannam@95: M ./threads/hc2hc-dit.c -25 +21 cannam@95: M ./threads/rdft-vrank-geq1.c -14 +10 cannam@95: M ./threads/vrank-geq1-rdft2.c -14 +10 cannam@95: cannam@95: Wed Sep 18 19:31:05 EDT 2002 stevenj cannam@95: * [project @ 2002-09-18 23:31:05 by stevenj] cannam@95: eliminated unused cannam@95: cannam@95: M ./tests/bench.c +2 cannam@95: M ./tests/verify-reodft.c -7 +7 cannam@95: cannam@95: Wed Sep 18 18:28:44 EDT 2002 stevenj cannam@95: * [project @ 2002-09-18 22:28:44 by stevenj] cannam@95: capitalize and parenthesize SUBSUMES cannam@95: cannam@95: M ./kernel/planner.c -9 +8 cannam@95: cannam@95: Wed Sep 18 18:26:58 EDT 2002 stevenj cannam@95: * [project @ 2002-09-18 22:26:58 by stevenj] cannam@95: comment cannam@95: cannam@95: M ./kernel/ifftw.h -2 +2 cannam@95: cannam@95: Wed Sep 18 18:03:18 EDT 2002 athena cannam@95: * [project @ 2002-09-18 22:03:18 by athena] cannam@95: Use flags from wisdom if wisdom is applicable. cannam@95: cannam@95: M ./kernel/ifftw.h -2 +3 cannam@95: M ./kernel/planner.c -70 +32 cannam@95: cannam@95: Wed Sep 18 17:16:17 EDT 2002 athena cannam@95: * [project @ 2002-09-18 21:16:16 by athena] cannam@95: Removed score() machinery cannam@95: cannam@95: M ./dft/buffered.c -16 +11 cannam@95: M ./dft/ct-dif.c -23 +21 cannam@95: M ./dft/ct-dit.c -29 +25 cannam@95: M ./dft/ct-ditbuf.c -30 +24 cannam@95: M ./dft/ct-ditf.c -10 +2 cannam@95: M ./dft/direct.c -8 +2 cannam@95: M ./dft/generic.c -9 +7 cannam@95: M ./dft/indirect.c -9 +12 cannam@95: M ./dft/nop.c -8 +2 cannam@95: M ./dft/rader.c -23 +25 cannam@95: M ./dft/rank-geq2.c -14 +13 cannam@95: M ./dft/rank0.c -8 +2 cannam@95: M ./dft/vrank-geq1.c -30 +29 cannam@95: M ./dft/vrank2-transpose.c -8 +2 cannam@95: M ./dft/vrank3-transpose.c -13 +15 cannam@95: M ./kernel/ifftw.h -12 +2 cannam@95: M ./kernel/planner.c -31 +19 cannam@95: M ./rdft/buffered.c -18 +11 cannam@95: M ./rdft/buffered2.c -18 +11 cannam@95: M ./rdft/dft-r2hc.c -12 +14 cannam@95: M ./rdft/dht-r2hc.c -8 +8 cannam@95: M ./rdft/dht-rader.c -12 +12 cannam@95: M ./rdft/direct.c -9 +3 cannam@95: M ./rdft/direct2.c -9 +3 cannam@95: M ./rdft/generic.c -8 +7 cannam@95: M ./rdft/hc2hc-buf.c -30 +29 cannam@95: M ./rdft/hc2hc-dif.c -33 +25 cannam@95: M ./rdft/hc2hc-dit.c -32 +27 cannam@95: M ./rdft/indirect.c -9 +14 cannam@95: M ./rdft/nop.c -8 +2 cannam@95: M ./rdft/nop2.c -8 +2 cannam@95: M ./rdft/rader-hc2hc.c -13 +12 cannam@95: M ./rdft/rank-geq2-rdft2.c -15 +17 cannam@95: M ./rdft/rank-geq2.c -17 +19 cannam@95: M ./rdft/rank0.c -8 +2 cannam@95: M ./rdft/rdft-dht.c -11 +13 cannam@95: M ./rdft/rdft2-radix2.c -12 +2 cannam@95: M ./rdft/vrank-geq1-rdft2.c -32 +29 cannam@95: M ./rdft/vrank-geq1.c -31 +31 cannam@95: M ./rdft/vrank2-transpose.c -8 +2 cannam@95: M ./rdft/vrank3-transpose.c -18 +14 cannam@95: M ./reodft/redft00e-r2hc.c -6 +7 cannam@95: M ./reodft/reodft010e-r2hc.c -6 +7 cannam@95: M ./reodft/reodft11e-r2hc.c -6 +7 cannam@95: M ./reodft/rodft00e-r2hc.c -6 +7 cannam@95: M ./tests/bench.c -2 cannam@95: cannam@95: Wed Sep 18 14:12:21 EDT 2002 athena cannam@95: * [project @ 2002-09-18 18:12:21 by athena] cannam@95: Revised planner hack cannam@95: cannam@95: M ./kernel/planner.c -4 +7 cannam@95: cannam@95: Wed Sep 18 10:14:41 EDT 2002 athena cannam@95: * [project @ 2002-09-18 14:14:41 by athena] cannam@95: Fix warning cannam@95: cannam@95: M ./simd/simd-altivec.h +2 cannam@95: cannam@95: Tue Sep 17 17:54:07 EDT 2002 athena cannam@95: * [project @ 2002-09-17 21:54:07 by athena] cannam@95: Type qualifiers. cannam@95: cannam@95: M ./dft/indirect.c -2 +2 cannam@95: M ./rdft/indirect.c -2 +2 cannam@95: cannam@95: Tue Sep 17 16:17:55 EDT 2002 athena cannam@95: * [project @ 2002-09-17 20:17:55 by athena] cannam@95: ESTIMATE is no longer subsumed by everything else. cannam@95: cannam@95: M ./kernel/planner.c -2 +1 cannam@95: cannam@95: Tue Sep 17 10:55:15 EDT 2002 athena cannam@95: * [project @ 2002-09-17 14:55:15 by athena] cannam@95: NO_BUFFERING is a planner flag, not a problem flag cannam@95: cannam@95: M ./dft/indirect.c -2 +2 cannam@95: M ./rdft/indirect.c -2 +2 cannam@95: cannam@95: Tue Sep 17 09:36:16 EDT 2002 athena cannam@95: * [project @ 2002-09-17 13:36:16 by athena] cannam@95: Maintain flags in canonical form. cannam@95: cannam@95: M ./kernel/ifftw.h -4 +6 cannam@95: M ./kernel/planner.c -13 +15 cannam@95: cannam@95: Tue Sep 17 09:09:57 EDT 2002 athena cannam@95: * [project @ 2002-09-17 13:09:56 by athena] cannam@95: In dramatic break with tradition, SUBSUME is now a partial order. I cannam@95: swear. cannam@95: cannam@95: M ./kernel/ifftw.h -4 +1 cannam@95: M ./kernel/planner.c -15 +54 cannam@95: cannam@95: Tue Sep 17 07:29:00 EDT 2002 athena cannam@95: * [project @ 2002-09-17 11:29:00 by athena] cannam@95: Added comment cannam@95: cannam@95: M ./kernel/planner.c -1 +3 cannam@95: cannam@95: Tue Sep 17 07:27:17 EDT 2002 athena cannam@95: * [project @ 2002-09-17 11:27:17 by athena] cannam@95: Inverted ESTIMATE flag, renamed USE_SCORE for consistency with the cannam@95: convention that 0 subsumes 1. cannam@95: cannam@95: M ./kernel/ifftw.h -9 +6 cannam@95: M ./kernel/planner.c -10 +7 cannam@95: M ./tests/bench.c +2 cannam@95: cannam@95: Tue Sep 17 02:50:15 EDT 2002 stevenj cannam@95: * [project @ 2002-09-17 06:50:15 by stevenj] cannam@95: NO_INDIRECT -> NO_INDIRECT_OP (out-of-place only) cannam@95: cannam@95: M ./dft/indirect.c -4 +5 cannam@95: M ./kernel/ifftw.h -3 +3 cannam@95: M ./rdft/indirect.c -4 +5 cannam@95: M ./tests/bench.c -1 +1 cannam@95: cannam@95: Tue Sep 17 00:40:04 EDT 2002 stevenj cannam@95: * [project @ 2002-09-17 04:40:04 by stevenj] cannam@95: hpux needs -D_REENTRANT (thanks to Clinton Roy for the bug report) cannam@95: cannam@95: M ./acx_pthread.m4 -2 +2 cannam@95: cannam@95: Mon Sep 16 23:54:34 EDT 2002 athena cannam@95: * [project @ 2002-09-17 03:54:34 by athena] cannam@95: Oops. cannam@95: cannam@95: M ./kernel/planner.c -2 +2 cannam@95: cannam@95: Mon Sep 16 23:44:47 EDT 2002 athena cannam@95: * [project @ 2002-09-17 03:44:47 by athena] cannam@95: Yet another attempt at getting the planner right. cannam@95: cannam@95: M ./kernel/ifftw.h -3 +3 cannam@95: M ./kernel/planner.c -28 +27 cannam@95: cannam@95: Mon Sep 16 21:56:14 EDT 2002 athena cannam@95: * [project @ 2002-09-17 01:56:14 by athena] cannam@95: Better coding. cannam@95: cannam@95: M ./kernel/planner.c -21 +11 cannam@95: cannam@95: Mon Sep 16 21:51:06 EDT 2002 athena cannam@95: * [project @ 2002-09-17 01:51:06 by athena] cannam@95: NO_UGLY is no longer a flag, but a separate planner field that does not cannam@95: interfere with wisdom. cannam@95: cannam@95: M ./kernel/ifftw.h -3 +2 cannam@95: M ./kernel/planner.c -14 +15 cannam@95: cannam@95: Mon Sep 16 19:04:41 EDT 2002 athena cannam@95: * [project @ 2002-09-16 23:04:41 by athena] cannam@95: Did not compile without FFTW_DEBUG cannam@95: cannam@95: M ./tests/verify-reodft.c -3 +1 cannam@95: cannam@95: Mon Sep 16 18:37:06 EDT 2002 athena cannam@95: * [project @ 2002-09-16 22:37:06 by athena] cannam@95: Changed scoring mechanism. cannam@95: cannam@95: M ./kernel/ifftw.h -5 +4 cannam@95: M ./kernel/plan.c -5 +1 cannam@95: M ./kernel/planner.c -51 +28 cannam@95: M ./tests/bench.c -5 +2 cannam@95: cannam@95: Mon Sep 16 17:13:45 EDT 2002 athena cannam@95: * [project @ 2002-09-16 21:13:45 by athena] cannam@95: Count infeasible plans cannam@95: cannam@95: M ./kernel/planner.c -4 +12 cannam@95: cannam@95: Mon Sep 16 16:36:12 EDT 2002 athena cannam@95: * [project @ 2002-09-16 20:36:12 by athena] cannam@95: curse subsumed plans before export cannam@95: cannam@95: M ./kernel/planner.c -27 +35 cannam@95: cannam@95: Mon Sep 16 15:40:46 EDT 2002 stevenj cannam@95: * [project @ 2002-09-16 19:40:46 by stevenj] cannam@95: removed ESTIMATE_BIT vs. ESTIMATE... ESTIMATE | IMPATIENT is a UI issue cannam@95: cannam@95: M ./kernel/ifftw.h -6 +4 cannam@95: M ./kernel/planner.c -2 +2 cannam@95: cannam@95: Mon Sep 16 15:31:39 EDT 2002 stevenj cannam@95: * [project @ 2002-09-16 19:31:39 by stevenj] cannam@95: cleanup cannam@95: cannam@95: M ./rdft/buffered2.c -6 +3 cannam@95: cannam@95: Mon Sep 16 15:28:47 EDT 2002 stevenj cannam@95: * [project @ 2002-09-16 19:28:47 by stevenj] cannam@95: use CONSERVE_MEMORY flag to prevent buffered for large sizes cannam@95: cannam@95: M ./dft/buffered.c -4 +7 cannam@95: M ./rdft/buffered.c -4 +7 cannam@95: M ./rdft/buffered2.c -5 +9 cannam@95: cannam@95: Mon Sep 16 15:16:16 EDT 2002 stevenj cannam@95: * [project @ 2002-09-16 19:16:16 by stevenj] cannam@95: moved NO_DHT_R2HC back into planner flags: there's no reason we would want this flag to block plan reuse cannam@95: cannam@95: M ./kernel/ifftw.h -4 +4 cannam@95: cannam@95: Mon Sep 16 14:59:14 EDT 2002 stevenj cannam@95: * [project @ 2002-09-16 18:59:14 by stevenj] cannam@95: whoops, commas cannam@95: cannam@95: M ./kernel/ifftw.h -3 +3 cannam@95: cannam@95: Mon Sep 16 14:58:26 EDT 2002 stevenj cannam@95: * [project @ 2002-09-16 18:58:26 by stevenj] cannam@95: problem_flags == checked in applicable, planner_flags == checked in score cannam@95: cannam@95: M ./kernel/ifftw.h -5 +7 cannam@95: cannam@95: Mon Sep 16 14:53:16 EDT 2002 stevenj cannam@95: * [project @ 2002-09-16 18:53:16 by stevenj] cannam@95: ESTIMATE should not *include* all impatience flags, even if it subsumes them; some impatience flags, like NO_INDIRECT, might make a problem unsolvable cannam@95: cannam@95: M ./kernel/ifftw.h -6 +6 cannam@95: M ./kernel/planner.c -3 +4 cannam@95: cannam@95: Mon Sep 16 00:56:29 EDT 2002 stevenj cannam@95: * [project @ 2002-09-16 04:56:29 by stevenj] cannam@95: quotatio marks cannam@95: cannam@95: M ./kernel/planner.c -3 +3 cannam@95: cannam@95: Sun Sep 15 23:55:44 EDT 2002 stevenj cannam@95: * [project @ 2002-09-16 03:55:44 by stevenj] cannam@95: delete blank line cannam@95: cannam@95: M ./kernel/planner.c -2 +1 cannam@95: cannam@95: Sun Sep 15 23:51:14 EDT 2002 stevenj cannam@95: * [project @ 2002-09-16 03:51:14 by stevenj] cannam@95: substitution cannam@95: cannam@95: M ./kernel/planner.c -2 +2 cannam@95: cannam@95: Sun Sep 15 23:49:50 EDT 2002 stevenj cannam@95: * [project @ 2002-09-16 03:49:50 by stevenj] cannam@95: note that we are not GNUlly correct cannam@95: cannam@95: M ./kernel/planner.c -1 +10 cannam@95: cannam@95: Sun Sep 15 23:41:01 EDT 2002 stevenj cannam@95: * [project @ 2002-09-16 03:41:01 by stevenj] cannam@95: indenting cannam@95: cannam@95: M ./kernel/planner.c -3 +3 cannam@95: cannam@95: Sun Sep 15 23:37:46 EDT 2002 stevenj cannam@95: * [project @ 2002-09-16 03:37:46 by stevenj] cannam@95: more jokes cannam@95: cannam@95: M ./kernel/planner.c -2 +5 cannam@95: cannam@95: Sun Sep 15 23:20:14 EDT 2002 stevenj cannam@95: * [project @ 2002-09-16 03:20:14 by stevenj] cannam@95: NONTHREADED_ICKYP includes nthr > 1 check cannam@95: cannam@95: M ./dft/ct-dit.c -2 +2 cannam@95: M ./dft/vrank-geq1.c -2 +2 cannam@95: M ./kernel/ifftw.h -2 +3 cannam@95: M ./rdft/hc2hc-dif.c -2 +2 cannam@95: M ./rdft/hc2hc-dit.c -2 +2 cannam@95: M ./rdft/vrank-geq1-rdft2.c -2 +2 cannam@95: M ./rdft/vrank-geq1.c -2 +2 cannam@95: cannam@95: Sun Sep 15 22:56:44 EDT 2002 stevenj cannam@95: * [project @ 2002-09-16 02:56:44 by stevenj] cannam@95: use md5sig cannam@95: cannam@95: M ./kernel/md5.c -1 +1 cannam@95: cannam@95: Sun Sep 15 22:55:41 EDT 2002 stevenj cannam@95: * [project @ 2002-09-16 02:55:41 by stevenj] cannam@95: md5sig typedef cannam@95: cannam@95: M ./kernel/ifftw.h -2 +4 cannam@95: M ./kernel/planner.c -9 +9 cannam@95: cannam@95: Sun Sep 15 22:35:13 EDT 2002 stevenj cannam@95: * [project @ 2002-09-16 02:35:13 by stevenj] cannam@95: updated cannam@95: cannam@95: M ./ChangeLog +324 cannam@95: cannam@95: Sun Sep 15 22:30:26 EDT 2002 stevenj cannam@95: * [project @ 2002-09-16 02:30:26 by stevenj] cannam@95: partially-ordered impatience cannam@95: cannam@95: M ./dft/buffered.c -2 +2 cannam@95: M ./dft/ct-dif.c -3 +3 cannam@95: M ./dft/ct-dit.c -3 +3 cannam@95: M ./dft/ct-ditbuf.c -2 +2 cannam@95: M ./dft/ct.c -5 +1 cannam@95: M ./dft/indirect.c -4 +4 cannam@95: M ./dft/rank-geq2.c -3 +2 cannam@95: M ./dft/vrank-geq1.c -13 +3 cannam@95: M ./kernel/ifftw.h -18 +51 cannam@95: M ./kernel/planner.c -11 +49 cannam@95: M ./rdft/buffered.c -2 +2 cannam@95: M ./rdft/buffered2.c -2 +2 cannam@95: M ./rdft/dft-r2hc.c -2 +2 cannam@95: M ./rdft/dht-r2hc.c -3 +3 cannam@95: M ./rdft/hc2hc-buf.c -4 +3 cannam@95: M ./rdft/hc2hc-dif.c -4 +4 cannam@95: M ./rdft/hc2hc-dit.c -3 +3 cannam@95: M ./rdft/hc2hc.c -5 +1 cannam@95: M ./rdft/indirect.c -4 +4 cannam@95: M ./rdft/rank-geq2-rdft2.c -3 +2 cannam@95: M ./rdft/rank-geq2.c -3 +2 cannam@95: M ./rdft/rdft-dht.c -3 +3 cannam@95: M ./rdft/rdft2-radix2.c -2 +2 cannam@95: M ./rdft/vrank-geq1-rdft2.c -13 +3 cannam@95: M ./rdft/vrank-geq1.c -13 +3 cannam@95: M ./tests/bench.c -3 cannam@95: M ./threads/dft-vrank-geq1.c -13 +2 cannam@95: M ./threads/hc2hc-dif.c -2 +2 cannam@95: M ./threads/rdft-vrank-geq1.c -12 +2 cannam@95: M ./threads/vrank-geq1-rdft2.c -12 +2 cannam@95: cannam@95: Sat Sep 14 19:47:56 EDT 2002 athena cannam@95: * [project @ 2002-09-14 23:47:56 by athena] cannam@95: Removed all that planner inheritance crap. cannam@95: cannam@95: M ./kernel/Makefile.am -4 +3 cannam@95: M ./kernel/ifftw.h -15 +5 cannam@95: R ./kernel/planner-naive.c cannam@95: R ./kernel/planner-score.c cannam@95: M ./kernel/planner.c -35 +128 cannam@95: M ./tests/bench.c -2 +2 cannam@95: cannam@95: Sat Sep 14 16:35:28 EDT 2002 stevenj cannam@95: * [project @ 2002-09-14 20:35:28 by stevenj] cannam@95: string.h is used for more than strlen cannam@95: cannam@95: M ./kernel/planner.c -2 +2 cannam@95: cannam@95: Sat Sep 14 12:19:13 EDT 2002 athena cannam@95: * [project @ 2002-09-14 16:19:13 by athena] cannam@95: Reduced hashtable size by 1/6 (on 32-bit machines) at the expense cannam@95: of messier planner. cannam@95: cannam@95: M ./kernel/ifftw.h -13 +17 cannam@95: M ./kernel/planner.c -80 +87 cannam@95: cannam@95: Sat Sep 14 08:31:29 EDT 2002 athena cannam@95: * [project @ 2002-09-14 12:31:29 by athena] cannam@95: Only print wisdom if verbose > 3 cannam@95: cannam@95: M ./tests/bench.c -2 +4 cannam@95: cannam@95: Sat Sep 14 07:56:56 EDT 2002 athena cannam@95: * [project @ 2002-09-14 11:56:56 by athena] cannam@95: Changed syntax of temporaries to avoid shadowing library functions cannam@95: (which is harmless but I hate the warning) cannam@95: cannam@95: M ./genfft/variable.ml -2 +2 cannam@95: M ./genfft-k7/variable.ml -1 +1 cannam@95: cannam@95: Fri Sep 13 23:07:39 EDT 2002 stevenj cannam@95: * [project @ 2002-09-14 03:07:39 by stevenj] cannam@95: only add warnings in debug/maintainer mode, and add a few more warning flags; eliminate more warnings; add support for posix_memalign (broken in glibc, grrr) cannam@95: cannam@95: M ./acinclude.m4 -1 +1 cannam@95: M ./configure.ac -1 +9 cannam@95: M ./dft/rader.c -1 +1 cannam@95: M ./kernel/alloc.c -3 +16 cannam@95: M ./kernel/assert.c -2 +2 cannam@95: M ./kernel/ifftw.h -5 +5 cannam@95: M ./kernel/md5.c -1 +1 cannam@95: M ./kernel/planner-score.c -11 +13 cannam@95: M ./kernel/primes.c -4 +4 cannam@95: M ./kernel/scan.c -2 +2 cannam@95: M ./libbench/bench-user.h -2 +2 cannam@95: M ./libbench/bench.h -3 +4 cannam@95: M ./libbench/report.c -5 +5 cannam@95: M ./libbench/timer.c -1 +2 cannam@95: M ./libbench/util.c -14 +26 cannam@95: M ./libbench/verify.c -4 +1 cannam@95: M ./rdft/rader-hc2hc.c -6 +6 cannam@95: M ./tests/bench.c -3 cannam@95: M ./tests/verify-lib.c -1 +4 cannam@95: cannam@95: Fri Sep 13 21:57:50 EDT 2002 athena cannam@95: * [project @ 2002-09-14 01:57:50 by athena] cannam@95: Explicit cast cannam@95: cannam@95: M ./kernel/twiddle.c -2 +2 cannam@95: cannam@95: Fri Sep 13 21:54:50 EDT 2002 athena cannam@95: * [project @ 2002-09-14 01:54:50 by athena] cannam@95: Use double-hashing. This allows a slightly higher load factor cannam@95: at the expense of a messier computation of the hashtable size. cannam@95: cannam@95: M ./kernel/ifftw.h -1 +2 cannam@95: M ./kernel/planner.c -57 +63 cannam@95: M ./kernel/primes.c -1 +7 cannam@95: cannam@95: Fri Sep 13 17:53:13 EDT 2002 stevenj cannam@95: * [project @ 2002-09-13 21:53:13 by stevenj] cannam@95: typo cannam@95: cannam@95: M ./genfft/magic.ml -2 +2 cannam@95: cannam@95: Fri Sep 13 15:36:07 EDT 2002 athena cannam@95: * [project @ 2002-09-13 19:36:07 by athena] cannam@95: Slight change in hash table growth functions. cannam@95: cannam@95: M ./kernel/planner.c -3 +12 cannam@95: cannam@95: Fri Sep 13 14:58:22 EDT 2002 athena cannam@95: * [project @ 2002-09-13 18:58:22 by athena] cannam@95: More statistics. cannam@95: cannam@95: M ./kernel/ifftw.h -1 +2 cannam@95: M ./kernel/planner.c -1 +8 cannam@95: cannam@95: Fri Sep 13 10:13:02 EDT 2002 athena cannam@95: * [project @ 2002-09-13 14:13:02 by athena] cannam@95: Clearer logic. cannam@95: cannam@95: M ./kernel/planner.c -2 +3 cannam@95: cannam@95: Fri Sep 13 10:11:10 EDT 2002 athena cannam@95: * [project @ 2002-09-13 14:11:10 by athena] cannam@95: Oops. cannam@95: cannam@95: M ./kernel/planner.c -2 +1 cannam@95: cannam@95: Fri Sep 13 09:31:46 EDT 2002 athena cannam@95: * [project @ 2002-09-13 13:31:46 by athena] cannam@95: Cleaned up cannam@95: cannam@95: M ./kernel/planner.c -5 +6 cannam@95: cannam@95: Fri Sep 13 09:16:07 EDT 2002 athena cannam@95: * [project @ 2002-09-13 13:16:07 by athena] cannam@95: Deal properly with infeasible problems. cannam@95: cannam@95: M ./kernel/planner.c -4 +7 cannam@95: cannam@95: Fri Sep 13 07:15:06 EDT 2002 athena cannam@95: * [project @ 2002-09-13 11:15:06 by athena] cannam@95: Redundantly initialize hash table to prevent valgrind warnings. cannam@95: cannam@95: M ./kernel/planner.c -3 +7 cannam@95: cannam@95: Thu Sep 12 19:00:22 EDT 2002 athena cannam@95: * [project @ 2002-09-12 23:00:22 by athena] cannam@95: Removed relics from past. cannam@95: cannam@95: M ./kernel/md5.c -5 +3 cannam@95: cannam@95: Thu Sep 12 18:53:44 EDT 2002 athena cannam@95: * [project @ 2002-09-12 22:53:44 by athena] cannam@95: md5hash a problem only once. cannam@95: cannam@95: M ./kernel/ifftw.h -6 +11 cannam@95: M ./kernel/planner.c -39 +32 cannam@95: cannam@95: Thu Sep 12 16:33:49 EDT 2002 athena cannam@95: * [project @ 2002-09-12 20:33:49 by athena] cannam@95: Renamed k7 codelets cannam@95: cannam@95: M ./genfft-k7/genUtil.ml -1 +1 cannam@95: M ./genfft-k7/gen_notw.ml -3 +3 cannam@95: M ./genfft-k7/gen_twiddle.ml -4 +4 cannam@95: cannam@95: Thu Sep 12 16:32:03 EDT 2002 stevenj cannam@95: * [project @ 2002-09-12 20:32:03 by stevenj] cannam@95: FORBID_DHT_R2HC -> DHT_R2HC_VERBOTEN for consistency cannam@95: cannam@95: M ./kernel/ifftw.h -2 +2 cannam@95: M ./rdft/dht-r2hc.c -3 +3 cannam@95: cannam@95: Thu Sep 12 16:28:43 EDT 2002 stevenj cannam@95: * [project @ 2002-09-12 20:28:43 by stevenj] cannam@95: removed obsolete macro cannam@95: cannam@95: M ./kernel/ifftw.h -3 +1 cannam@95: cannam@95: Thu Sep 12 16:20:39 EDT 2002 athena cannam@95: * [project @ 2002-09-12 20:20:39 by athena] cannam@95: Split flags in SIMD code. cannam@95: cannam@95: M ./dft/simd/n1b.c -1 +1 cannam@95: M ./dft/simd/n1f.c -1 +1 cannam@95: M ./dft/simd/t1b.c -1 +1 cannam@95: M ./dft/simd/t1f.c -1 +1 cannam@95: cannam@95: Thu Sep 12 16:18:51 EDT 2002 athena cannam@95: * [project @ 2002-09-12 20:18:51 by athena] cannam@95: Forgot to fix threads cannam@95: cannam@95: M ./threads/dft-vrank-geq1.c -5 +6 cannam@95: M ./threads/hc2hc-dif.c -2 +2 cannam@95: M ./threads/rdft-vrank-geq1.c -5 +6 cannam@95: M ./threads/vrank-geq1-rdft2.c -5 +6 cannam@95: cannam@95: Thu Sep 12 16:10:05 EDT 2002 athena cannam@95: * [project @ 2002-09-12 20:10:05 by athena] cannam@95: Split flags into planner_flags and problem_flags cannam@95: cannam@95: M ./dft/buffered.c -2 +2 cannam@95: M ./dft/ct-dif.c -3 +3 cannam@95: M ./dft/ct-dit.c -3 +3 cannam@95: M ./dft/ct-ditbuf.c -2 +2 cannam@95: M ./dft/ct.c -3 +3 cannam@95: M ./dft/indirect.c -4 +4 cannam@95: M ./dft/rader.c -1 +1 cannam@95: M ./dft/rank-geq2.c -2 +3 cannam@95: M ./dft/rank0.c -2 +2 cannam@95: M ./dft/vrank-geq1.c -7 +8 cannam@95: M ./kernel/ifftw.h -22 +21 cannam@95: M ./kernel/planner-naive.c -3 +3 cannam@95: M ./kernel/planner-score.c -3 +3 cannam@95: M ./kernel/planner.c -33 +28 cannam@95: M ./rdft/buffered.c -2 +2 cannam@95: M ./rdft/buffered2.c -3 +3 cannam@95: M ./rdft/dft-r2hc.c -2 +2 cannam@95: M ./rdft/dht-r2hc.c -3 +5 cannam@95: M ./rdft/dht-rader.c -1 +1 cannam@95: M ./rdft/hc2hc-buf.c -3 +3 cannam@95: M ./rdft/hc2hc-dif.c -4 +4 cannam@95: M ./rdft/hc2hc-dit.c -3 +3 cannam@95: M ./rdft/hc2hc.c -3 +3 cannam@95: M ./rdft/indirect.c -4 +4 cannam@95: M ./rdft/rank-geq2-rdft2.c -2 +3 cannam@95: M ./rdft/rank-geq2.c -2 +3 cannam@95: M ./rdft/rdft-dht.c -3 +3 cannam@95: M ./rdft/rdft2-radix2.c -2 +2 cannam@95: M ./rdft/vrank-geq1-rdft2.c -7 +8 cannam@95: M ./rdft/vrank-geq1.c -7 +8 cannam@95: M ./tests/bench.c -8 +8 cannam@95: cannam@95: Thu Sep 12 15:46:56 EDT 2002 stevenj cannam@95: * [project @ 2002-09-12 19:46:56 by stevenj] cannam@95: tetrameter cannam@95: cannam@95: M ./kernel/planner.c -2 +3 cannam@95: cannam@95: Thu Sep 12 15:11:21 EDT 2002 athena cannam@95: * [project @ 2002-09-12 19:11:21 by athena] cannam@95: Overwrite less impatient solutions properly. cannam@95: cannam@95: M ./kernel/planner.c -23 +15 cannam@95: cannam@95: Thu Sep 12 11:29:16 EDT 2002 athena cannam@95: * [project @ 2002-09-12 15:29:16 by athena] cannam@95: Oops. cannam@95: cannam@95: M ./kernel/planner.c -3 +3 cannam@95: cannam@95: Thu Sep 12 10:58:56 EDT 2002 athena cannam@95: * [project @ 2002-09-12 14:58:56 by athena] cannam@95: Keep less impatient solution in case of conflict. Paranoid cannam@95: cast to uint in certain places. cannam@95: cannam@95: M ./kernel/planner.c -4 +5 cannam@95: cannam@95: Thu Sep 12 10:02:51 EDT 2002 athena cannam@95: * [project @ 2002-09-12 14:02:51 by athena] cannam@95: Complete reimplementation of planner hash table. cannam@95: cannam@95: M ./kernel/ifftw.h -6 +7 cannam@95: M ./kernel/planner.c -143 +146 cannam@95: M ./tests/bench.c -2 +4 cannam@95: cannam@95: Thu Sep 12 07:58:45 EDT 2002 athena cannam@95: * [project @ 2002-09-12 11:58:45 by athena] cannam@95: planner->cnt was not properly decremented. cannam@95: cannam@95: M ./kernel/planner.c -13 +14 cannam@95: cannam@95: Wed Sep 11 17:52:39 EDT 2002 stevenj cannam@95: * [project @ 2002-09-11 21:52:39 by stevenj] cannam@95: typo cannam@95: cannam@95: M ./NEWS -1 +1 cannam@95: cannam@95: Mon Sep 9 17:10:45 EDT 2002 athena cannam@95: * [project @ 2002-09-09 21:10:45 by athena] cannam@95: Simplified cannam@95: cannam@95: M ./kernel/planner.c -15 +5 cannam@95: cannam@95: Mon Sep 9 17:03:32 EDT 2002 athena cannam@95: * [project @ 2002-09-09 21:03:32 by athena] cannam@95: Always overwrite old wisdom with new, in case the old is cannam@95: corrupt/conclicting. cannam@95: cannam@95: M ./kernel/planner.c -16 +8 cannam@95: cannam@95: Mon Sep 9 16:56:03 EDT 2002 stevenj cannam@95: * [project @ 2002-09-09 20:56:03 by stevenj] cannam@95: added quote/joke cannam@95: cannam@95: M ./kernel/plan.c -1 +5 cannam@95: cannam@95: Mon Sep 9 15:04:47 EDT 2002 athena cannam@95: * [project @ 2002-09-09 19:04:47 by athena] cannam@95: Completed wisdom import cannam@95: cannam@95: M ./kernel/ifftw.h -10 +11 cannam@95: M ./kernel/md5.c -1 +1 cannam@95: M ./kernel/planner.c -37 +54 cannam@95: M ./kernel/print.c -2 +3 cannam@95: M ./kernel/scan.c -73 +25 cannam@95: M ./tests/bench.c -3 +3 cannam@95: cannam@95: Mon Sep 9 10:14:22 EDT 2002 athena cannam@95: * [project @ 2002-09-09 14:14:22 by athena] cannam@95: Slight cleanup of md5 interface. cannam@95: cannam@95: M ./dft/problem.c -2 +2 cannam@95: M ./kernel/ifftw.h -2 +4 cannam@95: M ./kernel/md5.c -12 +24 cannam@95: M ./rdft/problem.c -2 +2 cannam@95: M ./rdft/problem2.c -2 +2 cannam@95: cannam@95: Tue Sep 3 22:32:43 EDT 2002 athena cannam@95: * [project @ 2002-09-04 02:32:43 by athena] cannam@95: More consistent protocol between planner and inferior. cannam@95: cannam@95: M ./kernel/planner-naive.c -2 +5 cannam@95: M ./kernel/planner-score.c -1 +2 cannam@95: M ./kernel/planner.c -3 +1 cannam@95: cannam@95: Tue Sep 3 21:08:30 EDT 2002 athena cannam@95: * [project @ 2002-09-04 01:08:30 by athena] cannam@95: I can't think of any situation where saving infeasible problems would cannam@95: be desirable. Removed relevant code. cannam@95: cannam@95: M ./kernel/planner.c -9 +3 cannam@95: cannam@95: Tue Sep 3 20:57:03 EDT 2002 athena cannam@95: * [project @ 2002-09-04 00:57:03 by athena] cannam@95: Encoder registrar's names in wisdom. Remove export_conf, since cannam@95: a separate program can now generate it. cannam@95: cannam@95: M ./kernel/ifftw.h -5 +5 cannam@95: M ./kernel/planner.c -108 +32 cannam@95: M ./kernel/solvtab.c -1 +2 cannam@95: M ./tests/bench.c -3 +1 cannam@95: cannam@95: Tue Sep 3 15:11:06 EDT 2002 athena cannam@95: * [project @ 2002-09-03 19:11:06 by athena] cannam@95: Fixed typo cannam@95: cannam@95: M ./kernel/planner.c -2 +2 cannam@95: cannam@95: Tue Sep 3 14:52:45 EDT 2002 athena cannam@95: * [project @ 2002-09-03 18:52:45 by athena] cannam@95: Fixed broken trochaic meter. cannam@95: cannam@95: M ./kernel/planner.c -2 +2 cannam@95: cannam@95: Tue Sep 3 09:49:50 EDT 2002 athena cannam@95: * [project @ 2002-09-03 13:49:50 by athena] cannam@95: Initialize planner->score. It is correct to leave it uninitialized, cannam@95: but I don't want people to send reports about purify complaining. cannam@95: cannam@95: M ./kernel/planner.c -1 +2 cannam@95: cannam@95: Tue Sep 3 09:03:46 EDT 2002 athena cannam@95: * [project @ 2002-09-03 13:03:46 by athena] cannam@95: More latin silliness cannam@95: cannam@95: M ./kernel/planner.c -3 +7 cannam@95: cannam@95: Mon Sep 2 17:57:32 EDT 2002 stevenj cannam@95: * [project @ 2002-09-02 21:57:32 by stevenj] cannam@95: updated cannam@95: cannam@95: M ./ChangeLog +193 cannam@95: cannam@95: Mon Sep 2 17:33:49 EDT 2002 stevenj cannam@95: * [project @ 2002-09-02 21:33:49 by stevenj] cannam@95: added clock() getseconds timer cannam@95: cannam@95: M ./kernel/timer.c -1 +16 cannam@95: cannam@95: Mon Sep 2 16:16:58 EDT 2002 athena cannam@95: * [project @ 2002-09-02 20:16:58 by athena] cannam@95: Oops cannam@95: cannam@95: M ./rdft/indirect.c -2 +1 cannam@95: cannam@95: Mon Sep 2 15:58:19 EDT 2002 athena cannam@95: * [project @ 2002-09-02 19:58:19 by athena] cannam@95: Experimental INDIRECT_VERBOTEN flag (not used) cannam@95: cannam@95: M ./dft/indirect.c -1 +3 cannam@95: M ./kernel/ifftw.h -5 +6 cannam@95: M ./rdft/indirect.c -3 +4 cannam@95: cannam@95: Mon Sep 2 15:36:21 EDT 2002 athena cannam@95: * [project @ 2002-09-02 19:36:21 by athena] cannam@95: Do not allow buffering in children of indirect solvers. cannam@95: cannam@95: M ./dft/buffered.c -1 +4 cannam@95: M ./dft/indirect.c -1 +3 cannam@95: M ./kernel/ifftw.h -1 +2 cannam@95: M ./rdft/buffered.c -1 +4 cannam@95: M ./rdft/buffered2.c -1 +4 cannam@95: M ./rdft/indirect.c -1 +3 cannam@95: cannam@95: Mon Sep 2 15:02:11 EDT 2002 athena cannam@95: * [project @ 2002-09-02 19:02:11 by athena] cannam@95: Oops cannam@95: cannam@95: M ./kernel/planner.c -2 +2 cannam@95: cannam@95: Mon Sep 2 14:32:28 EDT 2002 athena cannam@95: * [project @ 2002-09-02 18:32:28 by athena] cannam@95: Hash sizeof(R) as part of wisdom. cannam@95: cannam@95: M ./kernel/planner.c -1 +2 cannam@95: cannam@95: Mon Sep 2 13:47:57 EDT 2002 stevenj cannam@95: * [project @ 2002-09-02 17:47:57 by stevenj] cannam@95: added --enable-float synonym for --enable-single (since with have --enable-long-double) cannam@95: cannam@95: M ./configure.ac +1 cannam@95: cannam@95: Mon Sep 2 13:46:08 EDT 2002 athena cannam@95: * [project @ 2002-09-02 17:46:08 by athena] cannam@95: zerotens is now in its own file, so it does not cause dft to be linked cannam@95: in if only rdft is used. cannam@95: cannam@95: A ./dft/zero.c cannam@95: M ./dft/Makefile.am -1 +1 cannam@95: M ./dft/problem.c -25 +1 cannam@95: M ./dft/zero.c +49 cannam@95: cannam@95: Mon Sep 2 11:56:37 EDT 2002 athena cannam@95: * [project @ 2002-09-02 15:56:37 by athena] cannam@95: Removed unused var. cannam@95: cannam@95: M ./kernel/planner.c -2 +1 cannam@95: cannam@95: Mon Sep 2 11:55:33 EDT 2002 athena cannam@95: * [project @ 2002-09-02 15:55:33 by athena] cannam@95: Split insert() in preparation for wisdom import cannam@95: cannam@95: M ./kernel/planner.c -5 +11 cannam@95: cannam@95: Mon Sep 2 11:46:57 EDT 2002 athena cannam@95: * [project @ 2002-09-02 15:46:57 by athena] cannam@95: Moved debugging infrastructure to test directory so that it is not cannam@95: linked into the shared library. cannam@95: cannam@95: A ./tests/debug.h cannam@95: A ./tests/dotens.c cannam@95: A ./tests/dotens2.c cannam@95: A ./tests/verify-dft.c cannam@95: A ./tests/verify-lib.c cannam@95: A ./tests/verify-rdft.c cannam@95: A ./tests/verify-reodft.c cannam@95: A ./tests/verify.h cannam@95: M ./dft/Makefile.am -1 +1 cannam@95: M ./dft/dft.h -4 +1 cannam@95: R ./dft/verify.c cannam@95: M ./kernel/Makefile.am -6 +5 cannam@95: R ./kernel/dotens.c cannam@95: R ./kernel/dotens2.c cannam@95: M ./kernel/ifftw.h -17 +1 cannam@95: R ./kernel/verify-lib.c cannam@95: R ./kernel/verify.h cannam@95: M ./rdft/Makefile.am -2 +2 cannam@95: M ./rdft/rdft.h -4 +1 cannam@95: R ./rdft/verify.c cannam@95: M ./reodft/Makefile.am -1 +1 cannam@95: M ./reodft/reodft.h -4 +1 cannam@95: R ./reodft/verify.c cannam@95: M ./tests/Makefile.am -2 +7 cannam@95: M ./tests/bench.c +3 cannam@95: M ./tests/debug.h +18 cannam@95: M ./tests/dotens.c +48 cannam@95: M ./tests/dotens2.c +56 cannam@95: M ./tests/verify-dft.c +131 cannam@95: M ./tests/verify-lib.c +386 cannam@95: M ./tests/verify-rdft.c +420 cannam@95: M ./tests/verify-reodft.c +536 cannam@95: M ./tests/verify.h +83 cannam@95: cannam@95: Mon Sep 2 11:04:54 EDT 2002 athena cannam@95: * [project @ 2002-09-02 15:04:53 by athena] cannam@95: Reactivated wisdom export cannam@95: cannam@95: M ./kernel/planner.c -42 +9 cannam@95: M ./kernel/print.c -1 +12 cannam@95: cannam@95: Sun Sep 1 21:30:58 EDT 2002 athena cannam@95: * [project @ 2002-09-02 01:30:58 by athena] cannam@95: Dump errors to stderr, not stdout. cannam@95: cannam@95: M ./kernel/verify-lib.c -2 +3 cannam@95: cannam@95: Sun Sep 1 21:26:38 EDT 2002 athena cannam@95: * [project @ 2002-09-02 01:26:38 by athena] cannam@95: Removed traverse.c. cannam@95: traverse.c is no longer need for plan blessing. I figured out cannam@95: a way to avoid using it in planner-score.c, so the file is cannam@95: now redundant. cannam@95: cannam@95: M ./kernel/Makefile.am -2 +2 cannam@95: M ./kernel/ifftw.h -9 +2 cannam@95: M ./kernel/planner-score.c -30 +37 cannam@95: R ./kernel/traverse.c cannam@95: M ./tests/bench.c -20 cannam@95: cannam@95: Sun Sep 1 19:51:50 EDT 2002 athena cannam@95: * [project @ 2002-09-01 23:51:50 by athena] cannam@95: Removed code made obsolete by new MD5 scheme: problem equality cannam@95: tests, scanners, and associated list of problem kinds. cannam@95: cannam@95: M ./dft/conf.c -3 +1 cannam@95: M ./dft/dft.h -3 +1 cannam@95: M ./dft/problem.c -58 +3 cannam@95: M ./kernel/align.c -7 +1 cannam@95: M ./kernel/ifftw.h -24 +4 cannam@95: M ./kernel/planner.c -25 +10 cannam@95: M ./kernel/problem.c -11 +2 cannam@95: M ./kernel/scan.c -44 +2 cannam@95: M ./kernel/scanners.c -5 +5 cannam@95: M ./kernel/tensor.c -46 +1 cannam@95: M ./rdft/conf.c -4 +1 cannam@95: M ./rdft/problem.c -78 +3 cannam@95: M ./rdft/problem2.c -58 +3 cannam@95: M ./rdft/rdft.h -4 +1 cannam@95: cannam@95: Sun Sep 1 19:22:54 EDT 2002 athena cannam@95: * [project @ 2002-09-01 23:22:53 by athena] cannam@95: Started md5 implementation cannam@95: cannam@95: A ./kernel/md5.c cannam@95: M ./dft/problem.c -9 +9 cannam@95: M ./kernel/Makefile.am -1 +1 cannam@95: M ./kernel/ifftw.h -3 +29 cannam@95: M ./kernel/md5.c +164 cannam@95: M ./kernel/planner-naive.c -3 +3 cannam@95: M ./kernel/planner-score.c -7 +11 cannam@95: M ./kernel/planner.c -27 +47 cannam@95: M ./kernel/tensor.c -9 +7 cannam@95: M ./rdft/problem.c -12 +11 cannam@95: M ./rdft/problem2.c -10 +11 cannam@95: cannam@95: Sat Aug 31 14:00:04 EDT 2002 athena cannam@95: * [project @ 2002-08-31 18:00:04 by athena] cannam@95: Keep track of hit rate cannam@95: cannam@95: M ./kernel/ifftw.h -1 +2 cannam@95: M ./kernel/planner.c -3 +9 cannam@95: cannam@95: Sat Aug 31 12:44:04 EDT 2002 athena cannam@95: * [project @ 2002-08-31 16:44:04 by athena] cannam@95: Only dump when verbose > 4 cannam@95: cannam@95: M ./kernel/planner.c -2 +2 cannam@95: cannam@95: Sat Aug 31 09:55:57 EDT 2002 athena cannam@95: * [project @ 2002-08-31 13:55:57 by athena] cannam@95: Debugging infrastructure cannam@95: cannam@95: M ./dft/indirect.c -2 +1 cannam@95: M ./kernel/ifftw.h -5 +6 cannam@95: M ./kernel/plan.c -1 +6 cannam@95: M ./kernel/planner.c -7 +12 cannam@95: M ./tests/bench.c -3 +7 cannam@95: cannam@95: Sat Aug 31 09:21:48 EDT 2002 athena cannam@95: * [project @ 2002-08-31 13:21:48 by athena] cannam@95: Use debug infrastructure to dump planner. cannam@95: cannam@95: M ./kernel/planner.c -19 +10 cannam@95: M ./kernel/print.c -2 +5 cannam@95: cannam@95: Fri Aug 30 21:29:10 EDT 2002 athena cannam@95: * [project @ 2002-08-31 01:29:10 by athena] cannam@95: Do not store plans in planner, plus general planner cleanup. cannam@95: cannam@95: M ./kernel/alloc.c -2 +2 cannam@95: M ./kernel/ifftw.h -34 +37 cannam@95: M ./kernel/plan.c -15 +1 cannam@95: M ./kernel/planner-naive.c -7 +7 cannam@95: M ./kernel/planner-score.c -8 +7 cannam@95: M ./kernel/planner.c -103 +80 cannam@95: M ./kernel/scan.c -3 +3 cannam@95: M ./kernel/scanners.c -3 +3 cannam@95: M ./tests/bench.c -3 +16 cannam@95: cannam@95: Fri Aug 30 18:07:52 EDT 2002 stevenj cannam@95: * [project @ 2002-08-30 22:07:52 by stevenj] cannam@95: renamed IN_DHT_R2HC to the more general FORBID_DHT_R2HC cannam@95: cannam@95: M ./kernel/ifftw.h -2 +2 cannam@95: M ./rdft/dht-r2hc.c -3 +3 cannam@95: cannam@95: Fri Aug 30 18:07:21 EDT 2002 stevenj cannam@95: * [project @ 2002-08-30 22:07:21 by stevenj] cannam@95: eliminated unused var cannam@95: cannam@95: M ./kernel/planner.c -2 +1 cannam@95: cannam@95: Fri Aug 30 12:09:48 EDT 2002 athena cannam@95: * [project @ 2002-08-30 16:09:48 by athena] cannam@95: Score planner was not working correctly when using wisdom. Fixed. cannam@95: cannam@95: M ./kernel/planner-naive.c -2 +8 cannam@95: M ./kernel/planner-score.c -2 +13 cannam@95: M ./kernel/planner.c -13 +18 cannam@95: cannam@95: Fri Aug 30 08:20:48 EDT 2002 athena cannam@95: * [project @ 2002-08-30 12:20:48 by athena] cannam@95: Use hash table in debug malloc cannam@95: cannam@95: M ./kernel/alloc.c -12 +25 cannam@95: M ./kernel/ifftw.h -1 +2 cannam@95: M ./kernel/stride.c -2 +2 cannam@95: cannam@95: Fri Aug 30 02:45:15 EDT 2002 stevenj cannam@95: * [project @ 2002-08-30 06:45:15 by stevenj] cannam@95: listed some good stuff cannam@95: cannam@95: M ./NEWS +47 cannam@95: cannam@95: Fri Aug 30 02:17:30 EDT 2002 stevenj cannam@95: * [project @ 2002-08-30 06:17:30 by stevenj] cannam@95: timed planner cannam@95: cannam@95: M ./TODO +4 cannam@95: cannam@95: Fri Aug 30 02:07:00 EDT 2002 stevenj cannam@95: * [project @ 2002-08-30 06:07:00 by stevenj] cannam@95: fma? cannam@95: cannam@95: M ./TODO +2 cannam@95: cannam@95: Fri Aug 30 02:05:55 EDT 2002 stevenj cannam@95: * [project @ 2002-08-30 06:05:55 by stevenj] cannam@95: update cannam@95: cannam@95: M ./TODO -3 +3 cannam@95: cannam@95: Fri Aug 30 01:31:47 EDT 2002 stevenj cannam@95: * [project @ 2002-08-30 05:31:47 by stevenj] cannam@95: rader-dht -> dht-rader cannam@95: cannam@95: A ./rdft/dht-rader.c cannam@95: M ./rdft/Makefile.am -4 +4 cannam@95: M ./rdft/conf.c -2 +2 cannam@95: M ./rdft/dht-rader.c +422 cannam@95: R ./rdft/rader-dht.c cannam@95: M ./rdft/rdft.h -2 +2 cannam@95: cannam@95: Fri Aug 30 01:21:37 EDT 2002 stevenj cannam@95: * [project @ 2002-08-30 05:21:37 by stevenj] cannam@95: add DHT solver, and break up rader-dht and r2hc-hc2r cannam@95: cannam@95: A ./rdft/dht-r2hc.c cannam@95: A ./rdft/rdft-dht.c cannam@95: M ./kernel/ifftw.h -2 +3 cannam@95: M ./rdft/Makefile.am -6 +6 cannam@95: M ./rdft/buffered2.c -2 +4 cannam@95: M ./rdft/conf.c -4 +5 cannam@95: M ./rdft/dht-r2hc.c +151 cannam@95: R ./rdft/r2hc-hc2r.c cannam@95: M ./rdft/rader-dht.c -73 +21 cannam@95: M ./rdft/rank-geq2.c -1 +12 cannam@95: M ./rdft/rdft-dht.c +229 cannam@95: M ./rdft/rdft.h -3 +4 cannam@95: cannam@95: Thu Aug 29 23:20:35 EDT 2002 stevenj cannam@95: * [project @ 2002-08-30 03:20:35 by stevenj] cannam@95: another option cannam@95: cannam@95: M ./tests/bench.c +1 cannam@95: cannam@95: Thu Aug 29 22:55:29 EDT 2002 stevenj cannam@95: * [project @ 2002-08-30 02:55:29 by stevenj] cannam@95: generalized indirect solvers for fftw2-like buffering and more cannam@95: cannam@95: M ./dft/indirect.c -22 +35 cannam@95: M ./kernel/ifftw.h -1 +3 cannam@95: M ./kernel/tensor.c -13 +17 cannam@95: M ./rdft/indirect.c -18 +35 cannam@95: cannam@95: Thu Aug 29 18:08:16 EDT 2002 stevenj cannam@95: * [project @ 2002-08-29 22:08:16 by stevenj] cannam@95: tensor_max_index and tensor_min_stride are now both unsigned cannam@95: cannam@95: M ./dft/vrank-geq1.c -2 +3 cannam@95: M ./kernel/ifftw.h -3 +3 cannam@95: M ./kernel/tensor.c -11 +7 cannam@95: M ./rdft/vrank-geq1-rdft2.c -2 +3 cannam@95: M ./rdft/vrank-geq1.c -2 +3 cannam@95: cannam@95: Thu Aug 29 17:58:35 EDT 2002 stevenj cannam@95: * [project @ 2002-08-29 21:58:35 by stevenj] cannam@95: added iabs.c, and tensor_min_stride returns min absolute value cannam@95: cannam@95: A ./kernel/iabs.c cannam@95: M ./kernel/Makefile.am -5 +5 cannam@95: M ./kernel/iabs.c +28 cannam@95: M ./kernel/ifftw.h -1 +5 cannam@95: M ./kernel/tensor.c -12 +6 cannam@95: M ./rdft/buffered2.c -9 +4 cannam@95: M ./rdft/problem2.c -8 +3 cannam@95: cannam@95: Thu Aug 29 17:31:39 EDT 2002 stevenj cannam@95: * [project @ 2002-08-29 21:31:39 by stevenj] cannam@95: bug fix in cldrest hc2c/c2hc copy loops cannam@95: cannam@95: M ./rdft/buffered2.c -9 +14 cannam@95: cannam@95: Thu Aug 29 13:45:08 EDT 2002 athena cannam@95: * [project @ 2002-08-29 17:45:08 by athena] cannam@95: Added things to do. cannam@95: cannam@95: M ./TODO -1 +1 cannam@95: cannam@95: Thu Aug 29 13:10:04 EDT 2002 stevenj cannam@95: * [project @ 2002-08-29 17:10:04 by stevenj] cannam@95: added automake prereq cannam@95: cannam@95: M ./configure.ac -1 +1 cannam@95: cannam@95: Thu Aug 29 08:36:36 EDT 2002 athena cannam@95: * [project @ 2002-08-29 12:36:36 by athena] cannam@95: Use indexed addressing cannam@95: cannam@95: M ./rdft/rdft2-radix2.c -61 +61 cannam@95: cannam@95: Thu Aug 29 08:20:55 EDT 2002 athena cannam@95: * [project @ 2002-08-29 12:20:55 by athena] cannam@95: Ooops cannam@95: cannam@95: M ./libbench/verify.c -3 +6 cannam@95: M ./rdft/rdft2-radix2.c -4 +7 cannam@95: cannam@95: Thu Aug 29 07:45:37 EDT 2002 athena cannam@95: * [project @ 2002-08-29 11:45:37 by athena] cannam@95: Oops cannam@95: cannam@95: M ./kernel/ifftw.h -10 +10 cannam@95: cannam@95: Thu Aug 29 02:32:13 EDT 2002 stevenj cannam@95: * [project @ 2002-08-29 06:32:13 by stevenj] cannam@95: updates to win32 threads code (ick) cannam@95: cannam@95: M ./threads/threads.c -4 +22 cannam@95: cannam@95: Thu Aug 29 01:44:33 EDT 2002 stevenj cannam@95: * [project @ 2002-08-29 05:44:33 by stevenj] cannam@95: added threaded version cannam@95: cannam@95: A ./acx_pthread.m4 cannam@95: A ./threads/ cannam@95: A ./threads/Makefile.am cannam@95: A ./threads/conf.c cannam@95: A ./threads/ct-dit.c cannam@95: A ./threads/dft-vrank-geq1.c cannam@95: A ./threads/hc2hc-dif.c cannam@95: A ./threads/hc2hc-dit.c cannam@95: A ./threads/rdft-vrank-geq1.c cannam@95: A ./threads/threads.c cannam@95: A ./threads/threads.h cannam@95: A ./threads/vrank-geq1-rdft2.c cannam@95: M ./Makefile.am -2 +3 cannam@95: M ./acx_pthread.m4 +226 cannam@95: M ./configure.ac +56 cannam@95: M ./dft/ct-dif.c -2 +2 cannam@95: M ./dft/ct-dit.c -2 +5 cannam@95: M ./dft/ct-ditbuf.c -2 +2 cannam@95: M ./dft/ct-ditf.c -2 +2 cannam@95: M ./dft/ct.c -2 +3 cannam@95: M ./dft/ct.h -1 +2 cannam@95: M ./dft/dft.h -1 +4 cannam@95: M ./dft/kdft-dif.c -1 +5 cannam@95: M ./dft/kdft-dit.c -1 +5 cannam@95: M ./dft/vrank-geq1.c -1 +4 cannam@95: M ./kernel/alloc.c -42 +56 cannam@95: M ./kernel/ifftw.h -7 +18 cannam@95: M ./kernel/planner.c -15 +24 cannam@95: M ./rdft/hc2hc-buf.c -1 +3 cannam@95: M ./rdft/hc2hc-dif.c -1 +5 cannam@95: M ./rdft/hc2hc-dit.c -1 +5 cannam@95: M ./rdft/hc2hc.c -2 +3 cannam@95: M ./rdft/hc2hc.h -1 +2 cannam@95: M ./rdft/khc2hc-dif.c -1 +5 cannam@95: M ./rdft/khc2hc-dit.c -1 +5 cannam@95: M ./rdft/rdft.h -1 +4 cannam@95: M ./rdft/vrank-geq1-rdft2.c -1 +4 cannam@95: M ./rdft/vrank-geq1.c -1 +4 cannam@95: M ./tests/Makefile.am -2 +4 cannam@95: M ./tests/bench.c -1 +5 cannam@95: M ./threads/Makefile.am +15 cannam@95: M ./threads/conf.c +41 cannam@95: M ./threads/ct-dit.c +151 cannam@95: M ./threads/dft-vrank-geq1.c +249 cannam@95: M ./threads/hc2hc-dif.c +168 cannam@95: M ./threads/hc2hc-dit.c +166 cannam@95: M ./threads/rdft-vrank-geq1.c +246 cannam@95: M ./threads/threads.c +467 cannam@95: M ./threads/threads.h +53 cannam@95: M ./threads/vrank-geq1-rdft2.c +267 cannam@95: cannam@95: Wed Aug 28 19:47:21 EDT 2002 stevenj cannam@95: * [project @ 2002-08-28 23:47:21 by stevenj] cannam@95: fix make dist cannam@95: cannam@95: M ./kernel/Makefile.am -2 +3 cannam@95: cannam@95: Wed Aug 28 15:09:03 EDT 2002 stevenj cannam@95: * [project @ 2002-08-28 19:09:03 by stevenj] cannam@95: whoops, bugfix for inverse cannam@95: cannam@95: M ./rdft/rank-geq2-rdft2.c -4 +9 cannam@95: cannam@95: Wed Aug 28 14:50:34 EDT 2002 athena cannam@95: * [project @ 2002-08-28 18:50:34 by athena] cannam@95: Use C9x convention for naming (fftwf etc.). Removed installable header cannam@95: files since they will be part of the API. cannam@95: cannam@95: M ./Makefile.am -5 +5 cannam@95: M ./configure.ac -2 +9 cannam@95: M ./kernel/Makefile.am -1 cannam@95: R ./kernel/dfftw3.h cannam@95: R ./kernel/fftw3.h cannam@95: M ./kernel/ifftw.h -2 +17 cannam@95: R ./kernel/lfftw3.h cannam@95: R ./kernel/sfftw3.h cannam@95: M ./tests/Makefile.am -1 +1 cannam@95: cannam@95: Tue Aug 27 23:34:00 EDT 2002 stevenj cannam@95: * [project @ 2002-08-28 03:34:00 by stevenj] cannam@95: allow _1 variants to accept rnk 0 (sz 1) problems cannam@95: cannam@95: M ./rdft/problem.c -3 +3 cannam@95: cannam@95: Tue Aug 27 15:56:09 EDT 2002 stevenj cannam@95: * [project @ 2002-08-27 19:56:09 by stevenj] cannam@95: updated cannam@95: cannam@95: M ./ChangeLog +619 cannam@95: cannam@95: Mon Aug 26 20:14:56 EDT 2002 athena cannam@95: * [project @ 2002-08-27 00:14:56 by athena] cannam@95: Loop unroll is useless cannam@95: cannam@95: M ./dft/rank0.c -28 +6 cannam@95: cannam@95: Mon Aug 26 20:00:41 EDT 2002 athena cannam@95: * [project @ 2002-08-27 00:00:41 by athena] cannam@95: Use indexed addressing cannam@95: cannam@95: M ./dft/ct-ditbuf.c -12 +9 cannam@95: cannam@95: Mon Aug 26 19:46:46 EDT 2002 athena cannam@95: * [project @ 2002-08-26 23:46:46 by athena] cannam@95: Use indexed addressing in transpose routines. (Seems to be cannam@95: slightly better on athlon.) cannam@95: cannam@95: M ./dft/vrank2-transpose.c -16 +8 cannam@95: M ./dft/vrank3-transpose.c -15 +8 cannam@95: cannam@95: Mon Aug 26 12:59:44 EDT 2002 stevenj cannam@95: * [project @ 2002-08-26 16:59:44 by stevenj] cannam@95: added comment about stability cannam@95: cannam@95: M ./reodft/redft00e-r2hc.c -1 +2 cannam@95: M ./reodft/reodft11e-r2hc.c -1 +3 cannam@95: M ./reodft/rodft00e-r2hc.c -1 +2 cannam@95: cannam@95: Mon Aug 26 07:43:53 EDT 2002 athena cannam@95: * [project @ 2002-08-26 11:43:53 by athena] cannam@95: Approximate opcount cannam@95: cannam@95: M ./rdft/rdft2-radix2.c -6 +10 cannam@95: cannam@95: Mon Aug 26 06:38:49 EDT 2002 athena cannam@95: * [project @ 2002-08-26 10:38:49 by athena] cannam@95: Finished rdft2 via dft/rdft cannam@95: cannam@95: M ./dft/rank-geq2.c -2 +2 cannam@95: M ./rdft/rank-geq2.c -2 +2 cannam@95: M ./rdft/rdft2-radix2.c -9 +149 cannam@95: cannam@95: Mon Aug 26 00:15:59 EDT 2002 stevenj cannam@95: * [project @ 2002-08-26 04:15:59 by stevenj] cannam@95: some updates cannam@95: cannam@95: M ./TODO -9 +9 cannam@95: cannam@95: Mon Aug 26 00:05:53 EDT 2002 stevenj cannam@95: * [project @ 2002-08-26 04:05:52 by stevenj] cannam@95: rdft kind is now per-dimension, added rdft/rank-geq2 cannam@95: cannam@95: A ./rdft/rank-geq2.c cannam@95: M ./rdft/Makefile.am -2 +3 cannam@95: M ./rdft/buffered.c -3 +4 cannam@95: M ./rdft/buffered2.c -5 +5 cannam@95: M ./rdft/conf.c -3 +2 cannam@95: M ./rdft/dft-r2hc.c -2 +2 cannam@95: M ./rdft/direct.c -6 +6 cannam@95: M ./rdft/generic.c -5 +5 cannam@95: M ./rdft/hc2hc.c -10 +10 cannam@95: M ./rdft/indirect.c -2 +2 cannam@95: M ./rdft/problem.c -21 +75 cannam@95: M ./rdft/r2hc-hc2r.c -4 +4 cannam@95: M ./rdft/rader-dht.c -4 +4 cannam@95: M ./rdft/rader-hc2hc.c -4 +4 cannam@95: M ./rdft/rank-geq2.c +231 cannam@95: M ./rdft/rdft.h -4 +9 cannam@95: M ./rdft/rdft2-radix2.c -2 +2 cannam@95: M ./rdft/verify.c -5 +5 cannam@95: M ./reodft/redft00e-r2hc.c -3 +3 cannam@95: M ./reodft/reodft010e-r2hc.c -6 +6 cannam@95: M ./reodft/reodft11e-r2hc.c -5 +5 cannam@95: M ./reodft/rodft00e-r2hc.c -3 +3 cannam@95: M ./reodft/verify.c -4 +4 cannam@95: M ./tests/bench.c -2 +4 cannam@95: cannam@95: Sun Aug 25 22:45:38 EDT 2002 stevenj cannam@95: * [project @ 2002-08-26 02:45:38 by stevenj] cannam@95: added note cannam@95: cannam@95: M ./rdft/problem.c -2 +3 cannam@95: cannam@95: Sun Aug 25 22:28:12 EDT 2002 stevenj cannam@95: * [project @ 2002-08-26 02:28:12 by stevenj] cannam@95: must zero real sz cannam@95: cannam@95: M ./rdft/problem.c -2 +4 cannam@95: cannam@95: Sun Aug 25 22:06:52 EDT 2002 stevenj cannam@95: * [project @ 2002-08-26 02:06:52 by stevenj] cannam@95: unified pickdim funcs cannam@95: cannam@95: A ./kernel/pickdim.c cannam@95: M ./dft/rank-geq2.c -32 +10 cannam@95: M ./dft/vrank-geq1.c -45 +3 cannam@95: M ./kernel/Makefile.am -1 +1 cannam@95: M ./kernel/ifftw.h -1 +6 cannam@95: M ./kernel/pickdim.c +82 cannam@95: M ./rdft/vrank-geq1-rdft2.c -45 +3 cannam@95: M ./rdft/vrank-geq1.c -45 +3 cannam@95: M ./tests/bench.c -1 +1 cannam@95: cannam@95: Sun Aug 25 14:10:55 EDT 2002 fftw cannam@95: * [project @ 2002-08-25 18:10:55 by fftw] cannam@95: silence warnings cannam@95: cannam@95: M ./libbench/mp.c -14 cannam@95: M ./rdft/codelet.h -2 +2 cannam@95: M ./rdft/indirect.c -3 +1 cannam@95: M ./rdft/rank-geq2-rdft2.c -5 +2 cannam@95: M ./rdft/verify.c -10 +7 cannam@95: M ./reodft/reodft010e-r2hc.c -2 +2 cannam@95: M ./reodft/reodft11e-r2hc.c -7 +2 cannam@95: M ./reodft/verify.c -2 +2 cannam@95: cannam@95: Sun Aug 25 13:16:49 EDT 2002 athena cannam@95: * [project @ 2002-08-25 17:16:49 by athena] cannam@95: I had to add another planner flag to record whether pointers could cannam@95: become unaligned because of vrank-geq1 solvers (these solvers only cannam@95: plan the first element of a vector problem, but the second element cannam@95: may have a different alignment). This addition is ugly, but I don't cannam@95: see any way around it. cannam@95: cannam@95: M ./dft/codelet.h -3 +5 cannam@95: M ./dft/codelets/n.c -2 +3 cannam@95: M ./dft/codelets/t.c -2 +2 cannam@95: M ./dft/ct-dif.c -3 +3 cannam@95: M ./dft/ct-dit.c -3 +3 cannam@95: M ./dft/ct-ditbuf.c -3 +3 cannam@95: M ./dft/ct-ditf.c -2 +2 cannam@95: M ./dft/direct.c -5 +6 cannam@95: M ./dft/simd/n1b.c -1 +3 cannam@95: M ./dft/simd/n1f.c -1 +3 cannam@95: M ./dft/simd/t1b.c -1 +3 cannam@95: M ./dft/simd/t1f.c -1 +3 cannam@95: M ./dft/vrank-geq1.c -2 +11 cannam@95: M ./kernel/ifftw.h -4 +8 cannam@95: M ./rdft/vrank-geq1-rdft2.c -2 +8 cannam@95: M ./rdft/vrank-geq1.c -2 +7 cannam@95: cannam@95: Sun Aug 25 10:18:25 EDT 2002 athena cannam@95: * [project @ 2002-08-25 14:18:25 by athena] cannam@95: Added thoughts cannam@95: cannam@95: M ./TODO -1 +2 cannam@95: cannam@95: Sun Aug 25 10:08:59 EDT 2002 athena cannam@95: * [project @ 2002-08-25 14:08:59 by athena] cannam@95: Implemented rdft2 via vector rdft + radix2 step cannam@95: cannam@95: A ./rdft/rdft2-radix2.c cannam@95: M ./rdft/Makefile.am -1 +1 cannam@95: M ./rdft/conf.c -2 +2 cannam@95: M ./rdft/rdft.h -2 +2 cannam@95: R ./rdft/rdft2-dft.c cannam@95: M ./rdft/rdft2-radix2.c +325 cannam@95: cannam@95: Sat Aug 24 17:43:54 EDT 2002 athena cannam@95: * [project @ 2002-08-24 21:43:54 by athena] cannam@95: Stylistic changes cannam@95: cannam@95: M ./rdft/rdft2-dft.c -5 +5 cannam@95: cannam@95: Sat Aug 24 11:19:30 EDT 2002 athena cannam@95: * [project @ 2002-08-24 15:19:30 by athena] cannam@95: Simplified mktwiddle interface cannam@95: cannam@95: M ./dft/ct.c -7 +3 cannam@95: M ./dft/generic.c -5 +2 cannam@95: M ./kernel/ifftw.h -1 +3 cannam@95: M ./kernel/twiddle.c -1 +11 cannam@95: M ./rdft/generic.c -6 +3 cannam@95: M ./rdft/rdft2-dft.c -11 +5 cannam@95: M ./reodft/redft00e-r2hc.c -6 +2 cannam@95: M ./reodft/reodft010e-r2hc.c -5 +2 cannam@95: M ./reodft/reodft11e-r2hc.c -8 +3 cannam@95: M ./reodft/rodft00e-r2hc.c -5 +2 cannam@95: cannam@95: Sat Aug 24 11:05:08 EDT 2002 athena cannam@95: * [project @ 2002-08-24 15:05:08 by athena] cannam@95: Unification of certain vector computations. rdft2-dft is now a cannam@95: vector transform. cannam@95: cannam@95: M ./dft/ct-dif.c -2 +2 cannam@95: M ./dft/ct-dit.c -2 +2 cannam@95: M ./dft/ct.c -16 +2 cannam@95: M ./dft/ct.h -4 +1 cannam@95: M ./kernel/ifftw.h -1 +2 cannam@95: M ./kernel/tensor.c -1 +17 cannam@95: M ./rdft/hc2hc-dif.c -2 +2 cannam@95: M ./rdft/hc2hc-dit.c -2 +2 cannam@95: M ./rdft/hc2hc.c -18 +2 cannam@95: M ./rdft/hc2hc.h -4 +1 cannam@95: M ./rdft/rdft2-dft.c -54 +66 cannam@95: cannam@95: Fri Aug 23 20:21:25 EDT 2002 athena cannam@95: * [project @ 2002-08-24 00:21:25 by athena] cannam@95: Intel compiler seems to be still buggy cannam@95: cannam@95: M ./configure.ac +1 cannam@95: M ./simd/sse.c -1 +3 cannam@95: M ./simd/sse2.c -1 +3 cannam@95: cannam@95: Fri Aug 23 16:07:12 EDT 2002 athena cannam@95: * [project @ 2002-08-23 20:07:12 by athena] cannam@95: Streamlined twiddle protocol cannam@95: cannam@95: M ./dft/ct-dif.c -2 +3 cannam@95: M ./dft/ct-dit.c -2 +3 cannam@95: M ./dft/ct-ditbuf.c -2 +2 cannam@95: M ./dft/ct-ditf.c -2 +2 cannam@95: M ./dft/ct.c -13 +6 cannam@95: M ./dft/ct.h -3 +2 cannam@95: M ./dft/generic.c -16 +6 cannam@95: M ./dft/indirect.c -3 +1 cannam@95: M ./kernel/ifftw.h -3 +3 cannam@95: M ./kernel/twiddle.c -7 +14 cannam@95: M ./rdft/generic.c -18 +8 cannam@95: M ./rdft/hc2hc.c -10 +6 cannam@95: M ./rdft/rdft2-dft.c -19 +10 cannam@95: M ./reodft/redft00e-r2hc.c -15 +6 cannam@95: M ./reodft/reodft010e-r2hc.c -19 +10 cannam@95: M ./reodft/reodft11e-r2hc.c -21 +9 cannam@95: M ./reodft/rodft00e-r2hc.c -15 +6 cannam@95: cannam@95: Fri Aug 23 13:22:17 EDT 2002 athena cannam@95: * [project @ 2002-08-23 17:22:17 by athena] cannam@95: Implemented rdft2 via dft (forward only for now) cannam@95: cannam@95: A ./rdft/rdft2-dft.c cannam@95: M ./libbench/verify.c -1 +10 cannam@95: M ./rdft/Makefile.am -1 +1 cannam@95: M ./rdft/conf.c -1 +2 cannam@95: M ./rdft/rdft.h -1 +2 cannam@95: M ./rdft/rdft2-dft.c +213 cannam@95: cannam@95: Thu Aug 22 11:29:29 EDT 2002 athena cannam@95: * [project @ 2002-08-22 15:29:29 by athena] cannam@95: More cleanup of verify cannam@95: cannam@95: M ./kernel/verify-lib.c -21 +13 cannam@95: M ./libbench/verify.c -45 +12 cannam@95: cannam@95: Thu Aug 22 11:16:03 EDT 2002 athena cannam@95: * [project @ 2002-08-22 15:16:03 by athena] cannam@95: Changed error criterion because old one was too strict cannam@95: cannam@95: M ./kernel/verify-lib.c -40 +21 cannam@95: cannam@95: Thu Aug 22 11:15:17 EDT 2002 athena cannam@95: * [project @ 2002-08-22 15:15:17 by athena] cannam@95: Disable shared cannam@95: cannam@95: M ./bootstrap.sh -1 +1 cannam@95: cannam@95: Thu Aug 22 09:19:12 EDT 2002 athena cannam@95: * [project @ 2002-08-22 13:19:12 by athena] cannam@95: Added thoughts cannam@95: cannam@95: M ./TODO +4 cannam@95: cannam@95: Thu Aug 22 09:17:28 EDT 2002 athena cannam@95: * [project @ 2002-08-22 13:17:28 by athena] cannam@95: Oops cannam@95: cannam@95: M ./dft/generic.c -1 +2 cannam@95: cannam@95: Thu Aug 22 09:11:34 EDT 2002 athena cannam@95: * [project @ 2002-08-22 13:11:34 by athena] cannam@95: Do not use inline. Minor changes. cannam@95: cannam@95: M ./dft/generic.c -1 +1 cannam@95: M ./kernel/alloc.c -2 +2 cannam@95: M ./kernel/planner-score.c -2 +2 cannam@95: M ./kernel/tensor.c -3 +3 cannam@95: cannam@95: Wed Aug 21 16:23:26 EDT 2002 stevenj cannam@95: * [project @ 2002-08-21 20:23:26 by stevenj] cannam@95: more commented flags cannam@95: cannam@95: M ./tests/bench.c +2 cannam@95: cannam@95: Tue Aug 20 19:44:43 EDT 2002 stevenj cannam@95: * [project @ 2002-08-20 23:44:43 by stevenj] cannam@95: added DCT-IV and DST-IV cannam@95: cannam@95: A ./reodft/reodft11e-r2hc.c cannam@95: M ./reodft/Makefile.am -1 +2 cannam@95: M ./reodft/conf.c -1 +2 cannam@95: M ./reodft/reodft11e-r2hc.c +297 cannam@95: M ./reodft/verify.c -1 +15 cannam@95: M ./tests/bench.c -1 +5 cannam@95: cannam@95: Tue Aug 20 16:01:36 EDT 2002 athena cannam@95: * [project @ 2002-08-20 20:01:36 by athena] cannam@95: Slight improvement in twiddle scheme cannam@95: cannam@95: M ./genfft/twiddle.ml -2 +9 cannam@95: cannam@95: Tue Aug 20 15:31:54 EDT 2002 stevenj cannam@95: * [project @ 2002-08-20 19:31:54 by stevenj] cannam@95: name fix cannam@95: cannam@95: M ./reodft/conf.c -2 +2 cannam@95: M ./reodft/reodft.h -6 +4 cannam@95: M ./reodft/reodft010e-r2hc.c -2 +2 cannam@95: cannam@95: Tue Aug 20 15:16:48 EDT 2002 stevenj cannam@95: * [project @ 2002-08-20 19:16:48 by stevenj] cannam@95: removed extraneous variable cannam@95: cannam@95: M ./reodft/reodft010e-r2hc.c -5 +1 cannam@95: cannam@95: Tue Aug 20 11:46:29 EDT 2002 athena cannam@95: * [project @ 2002-08-20 15:46:29 by athena] cannam@95: Oops cannam@95: cannam@95: M ./libbench/mp.c -33 +79 cannam@95: M ./libbench/verify.c -33 +25 cannam@95: cannam@95: Tue Aug 20 08:37:45 EDT 2002 athena cannam@95: * [project @ 2002-08-20 12:37:45 by athena] cannam@95: Still playing around cannam@95: cannam@95: M ./genfft/twiddle.ml -35 +62 cannam@95: M ./kernel/trig.c -1 +3 cannam@95: cannam@95: Mon Aug 19 19:56:29 EDT 2002 athena cannam@95: * [project @ 2002-08-19 23:56:29 by athena] cannam@95: Playing around with addition chain cannam@95: cannam@95: M ./TODO -6 +11 cannam@95: M ./genfft/algsimp.ml -3 +4 cannam@95: M ./genfft/expr.ml -1 +6 cannam@95: M ./genfft/expr.mli -1 +2 cannam@95: M ./genfft/twiddle.ml -46 +64 cannam@95: M ./support/addchain.c -8 +25 cannam@95: cannam@95: Mon Aug 19 19:48:56 EDT 2002 stevenj cannam@95: * [project @ 2002-08-19 23:48:56 by stevenj] cannam@95: comments cannam@95: cannam@95: M ./reodft/redft00e-r2hc.c -1 +4 cannam@95: M ./reodft/rodft00e-r2hc.c -1 +4 cannam@95: cannam@95: Mon Aug 19 19:45:35 EDT 2002 stevenj cannam@95: * [project @ 2002-08-19 23:45:35 by stevenj] cannam@95: comment fixes cannam@95: cannam@95: M ./reodft/reodft010e-r2hc.c -5 +2 cannam@95: cannam@95: Mon Aug 19 19:40:18 EDT 2002 stevenj cannam@95: * [project @ 2002-08-19 23:40:18 by stevenj] cannam@95: added reodft stuff cannam@95: cannam@95: A ./reodft/ cannam@95: A ./reodft/Makefile.am cannam@95: A ./reodft/conf.c cannam@95: A ./reodft/redft00e-r2hc.c cannam@95: A ./reodft/reodft.h cannam@95: A ./reodft/reodft010e-r2hc.c cannam@95: A ./reodft/rodft00e-r2hc.c cannam@95: A ./reodft/verify.c cannam@95: M ./Makefile.am -2 +3 cannam@95: M ./configure.ac +2 cannam@95: M ./dft/dft.h -1 +5 cannam@95: M ./rdft/rdft.h -1 +5 cannam@95: M ./reodft/Makefile.am +6 cannam@95: M ./reodft/conf.c +37 cannam@95: M ./reodft/redft00e-r2hc.c +204 cannam@95: M ./reodft/reodft.h +44 cannam@95: M ./reodft/reodft010e-r2hc.c +400 cannam@95: M ./reodft/rodft00e-r2hc.c +199 cannam@95: M ./reodft/verify.c +521 cannam@95: M ./tests/Makefile.am -1 +1 cannam@95: M ./tests/bench.c -2 +11 cannam@95: cannam@95: Sun Aug 18 19:44:14 EDT 2002 athena cannam@95: * [project @ 2002-08-18 23:44:14 by athena] cannam@95: Sync with nbenchfft cannam@95: cannam@95: M ./libbench/Makefile.am +4 cannam@95: M ./libbench/verify.c -4 +3 cannam@95: cannam@95: Sun Aug 18 16:02:37 EDT 2002 athena cannam@95: * [project @ 2002-08-18 20:02:37 by athena] cannam@95: Economy of thought cannam@95: cannam@95: M ./genfft/complex.ml -23 +12 cannam@95: M ./genfft/complex.mli -3 +3 cannam@95: M ./genfft/twiddle.ml -5 +5 cannam@95: cannam@95: Sat Aug 17 15:52:05 EDT 2002 stevenj cannam@95: * [project @ 2002-08-17 19:52:05 by stevenj] cannam@95: distribute addchain.c cannam@95: cannam@95: M ./support/Makefile.am -1 +1 cannam@95: cannam@95: Sat Aug 17 14:09:11 EDT 2002 athena cannam@95: * [project @ 2002-08-17 18:09:11 by athena] cannam@95: Nothing serious cannam@95: cannam@95: M ./support/addchain.c -4 +7 cannam@95: cannam@95: Sat Aug 17 10:47:59 EDT 2002 athena cannam@95: * [project @ 2002-08-17 14:47:59 by athena] cannam@95: New twiddle policy (disabled for now) cannam@95: cannam@95: A ./support/addchain.c cannam@95: M ./genfft/twiddle.ml -1 +98 cannam@95: M ./support/addchain.c +151 cannam@95: cannam@95: Fri Aug 16 23:44:28 EDT 2002 stevenj cannam@95: * [project @ 2002-08-17 03:44:28 by stevenj] cannam@95: bug fix for hc2r (must use inverse dft) cannam@95: cannam@95: M ./rdft/rank-geq2-rdft2.c -2 +2 cannam@95: cannam@95: Fri Aug 16 20:27:10 EDT 2002 athena cannam@95: * [project @ 2002-08-17 00:27:10 by athena] cannam@95: New log3 twiddle policy cannam@95: cannam@95: M ./dft/codelets/inplace/Makefile.am -2 +2 cannam@95: M ./dft/codelets/standard/Makefile.am -1 +1 cannam@95: M ./genfft/twiddle.ml -9 +88 cannam@95: M ./rdft/codelets/hc2r/Makefile.am +6 cannam@95: M ./rdft/codelets/r2hc/Makefile.am -1 +1 cannam@95: cannam@95: Fri Aug 16 18:10:33 EDT 2002 athena cannam@95: * [project @ 2002-08-16 22:10:33 by athena] cannam@95: More verify cleanup cannam@95: cannam@95: M ./dft/verify.c -20 +2 cannam@95: M ./kernel/verify-lib.c -1 +19 cannam@95: M ./kernel/verify.h +3 cannam@95: M ./rdft/verify.c -23 +5 cannam@95: cannam@95: Fri Aug 16 16:31:19 EDT 2002 athena cannam@95: * [project @ 2002-08-16 20:31:19 by athena] cannam@95: Oops cannam@95: cannam@95: M ./rdft/verify.c -3 +3 cannam@95: cannam@95: Fri Aug 16 15:22:36 EDT 2002 athena cannam@95: * [project @ 2002-08-16 19:22:36 by athena] cannam@95: Economy of thought (and code) cannam@95: cannam@95: A ./kernel/verify-lib.c cannam@95: A ./kernel/verify.h cannam@95: M ./dft/verify.c -305 +12 cannam@95: M ./kernel/Makefile.am -2 +2 cannam@95: M ./kernel/verify-lib.c +394 cannam@95: M ./kernel/verify.h +80 cannam@95: M ./rdft/verify.c -364 +23 cannam@95: cannam@95: Fri Aug 16 14:05:45 EDT 2002 athena cannam@95: * [project @ 2002-08-16 18:05:45 by athena] cannam@95: Added comment cannam@95: cannam@95: M ./TODO +2 cannam@95: cannam@95: Fri Aug 16 12:57:43 EDT 2002 athena cannam@95: * [project @ 2002-08-16 16:57:43 by athena] cannam@95: Cleaner rounding algorithm cannam@95: cannam@95: M ./libbench/mp.c -16 +27 cannam@95: cannam@95: Fri Aug 16 11:27:43 EDT 2002 athena cannam@95: * [project @ 2002-08-16 15:27:43 by athena] cannam@95: Can get away with shorter length in bluestein (I think). cannam@95: cannam@95: M ./libbench/mp.c -1 +1 cannam@95: cannam@95: Fri Aug 16 11:08:09 EDT 2002 athena cannam@95: * [project @ 2002-08-16 15:08:09 by athena] cannam@95: Portability improvements cannam@95: cannam@95: M ./libbench/mp.c -11 +4 cannam@95: cannam@95: Fri Aug 16 08:06:31 EDT 2002 athena cannam@95: * [project @ 2002-08-16 12:06:31 by athena] cannam@95: Optionally average accuracy test over many rounds cannam@95: cannam@95: M ./libbench/bench-main.c -2 +9 cannam@95: M ./libbench/bench.h -2 +2 cannam@95: M ./libbench/verify.c -29 +43 cannam@95: cannam@95: Fri Aug 16 07:50:24 EDT 2002 athena cannam@95: * [project @ 2002-08-16 11:50:24 by athena] cannam@95: More accurate formula for trig tables cannam@95: cannam@95: M ./dft/rader.c -3 +3 cannam@95: M ./rdft/rader-dht.c -3 +3 cannam@95: M ./rdft/rader-hc2hc.c -3 +3 cannam@95: cannam@95: Fri Aug 16 06:42:02 EDT 2002 athena cannam@95: * [project @ 2002-08-16 10:42:02 by athena] cannam@95: Implemented accuracy test for all integers cannam@95: cannam@95: M ./libbench/mp.c -6 +123 cannam@95: M ./libbench/verify.c -2 +1 cannam@95: cannam@95: Thu Aug 15 18:54:44 EDT 2002 athena cannam@95: * [project @ 2002-08-15 22:54:44 by athena] cannam@95: inv, neg: make static cannam@95: cannam@95: M ./libbench/mp.c -2 +2 cannam@95: cannam@95: Thu Aug 15 17:25:37 EDT 2002 athena cannam@95: * [project @ 2002-08-15 21:25:37 by athena] cannam@95: Verify was not complete for real transforms cannam@95: cannam@95: M ./libbench/verify.c -9 +34 cannam@95: cannam@95: Thu Aug 15 16:30:03 EDT 2002 athena cannam@95: * [project @ 2002-08-15 20:30:03 by athena] cannam@95: Oops cannam@95: cannam@95: M ./libbench/verify.c -3 +1 cannam@95: cannam@95: Thu Aug 15 16:29:16 EDT 2002 athena cannam@95: * [project @ 2002-08-15 20:29:16 by athena] cannam@95: Fixed hb codelets cannam@95: cannam@95: M ./genfft/gen_hc2hc.ml -3 +5 cannam@95: M ./libbench/verify.c -1 +4 cannam@95: cannam@95: Thu Aug 15 14:10:45 EDT 2002 athena cannam@95: * [project @ 2002-08-15 18:10:45 by athena] cannam@95: Changed twiddle policy cannam@95: cannam@95: M ./dft/codelets/inplace/Makefile.am -2 +2 cannam@95: M ./dft/codelets/standard/Makefile.am -1 +1 cannam@95: M ./rdft/codelets/r2hc/Makefile.am -1 +1 cannam@95: cannam@95: Thu Aug 15 13:32:24 EDT 2002 stevenj cannam@95: * [project @ 2002-08-15 17:32:24 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./rdft/direct2.c -3 +3 cannam@95: cannam@95: Thu Aug 15 11:01:04 EDT 2002 athena cannam@95: * [project @ 2002-08-15 15:01:04 by athena] cannam@95: No point in libbench being a shared library cannam@95: cannam@95: M ./libbench/Makefile.am -2 +2 cannam@95: M ./tests/Makefile.am -1 +1 cannam@95: cannam@95: Thu Aug 15 09:48:37 EDT 2002 athena cannam@95: * [project @ 2002-08-15 13:48:37 by athena] cannam@95: Moved accuracy test to libbench cannam@95: cannam@95: A ./libbench/mp.c cannam@95: M ./libbench/Makefile.am -1 +1 cannam@95: M ./libbench/bench-main.c -4 +13 cannam@95: M ./libbench/bench.h -1 +3 cannam@95: M ./libbench/mp.c +439 cannam@95: M ./libbench/util.c +8 cannam@95: M ./libbench/verify.c -3 +66 cannam@95: M ./tests/Makefile.am -5 +1 cannam@95: R ./tests/accuracy.c cannam@95: R ./tests/mp.c cannam@95: cannam@95: Wed Aug 14 19:48:23 EDT 2002 athena cannam@95: * [project @ 2002-08-14 23:48:23 by athena] cannam@95: Modified accuracy test cannam@95: cannam@95: M ./tests/accuracy.c -33 +20 cannam@95: cannam@95: Wed Aug 14 08:34:26 EDT 2002 athena cannam@95: * [project @ 2002-08-14 12:34:26 by athena] cannam@95: Fixes for long double cannam@95: cannam@95: M ./tests/accuracy.c -2 +3 cannam@95: M ./tests/mp.c -1 +1 cannam@95: cannam@95: Wed Aug 14 08:17:57 EDT 2002 athena cannam@95: * [project @ 2002-08-14 12:17:57 by athena] cannam@95: Normalize input cannam@95: cannam@95: M ./tests/accuracy.c -3 +17 cannam@95: cannam@95: Wed Aug 14 07:26:41 EDT 2002 athena cannam@95: * [project @ 2002-08-14 11:26:41 by athena] cannam@95: Oops cannam@95: cannam@95: M ./tests/accuracy.c +1 cannam@95: cannam@95: Wed Aug 14 07:25:34 EDT 2002 athena cannam@95: * [project @ 2002-08-14 11:25:34 by athena] cannam@95: Also compute relative error cannam@95: cannam@95: M ./tests/accuracy.c -2 +13 cannam@95: cannam@95: Wed Aug 14 07:08:20 EDT 2002 athena cannam@95: * [project @ 2002-08-14 11:08:20 by athena] cannam@95: Loop over N cannam@95: cannam@95: M ./tests/accuracy.c -24 +32 cannam@95: cannam@95: Wed Aug 14 06:54:50 EDT 2002 athena cannam@95: * [project @ 2002-08-14 10:54:50 by athena] cannam@95: simple-minded accuracy test cannam@95: cannam@95: A ./tests/accuracy.c cannam@95: A ./tests/mp.c cannam@95: M ./tests/Makefile.am -1 +6 cannam@95: M ./tests/accuracy.c +48 cannam@95: M ./tests/mp.c +434 cannam@95: cannam@95: Wed Aug 14 03:26:06 EDT 2002 stevenj cannam@95: * [project @ 2002-08-14 07:26:06 by stevenj] cannam@95: whoops cannam@95: cannam@95: A ./rdft/rank-geq2-rdft2.c cannam@95: cannam@95: Tue Aug 13 11:42:41 EDT 2002 athena cannam@95: * [project @ 2002-08-13 15:42:41 by athena] cannam@95: fma() stuff is too nonportable, removed cannam@95: cannam@95: M ./kernel/trig.c -45 +1 cannam@95: cannam@95: Mon Aug 12 14:07:44 EDT 2002 stevenj cannam@95: * [project @ 2002-08-12 18:07:44 by stevenj] cannam@95: slight fix cannam@95: cannam@95: M ./rdft/problem.c -2 +2 cannam@95: cannam@95: Mon Aug 12 14:07:18 EDT 2002 stevenj cannam@95: * [project @ 2002-08-12 18:07:18 by stevenj] cannam@95: use table for rdft_kind_str cannam@95: cannam@95: M ./rdft/problem.c -26 +12 cannam@95: cannam@95: Mon Aug 12 13:43:08 EDT 2002 stevenj cannam@95: * [project @ 2002-08-12 17:43:08 by stevenj] cannam@95: slight fixes cannam@95: cannam@95: M ./rdft/problem2.c -6 +11 cannam@95: cannam@95: Mon Aug 12 13:31:37 EDT 2002 stevenj cannam@95: * [project @ 2002-08-12 17:31:37 by stevenj] cannam@95: multidimensional rdft2 cannam@95: cannam@95: M ./kernel/ifftw.h -2 +3 cannam@95: M ./kernel/planner.c -1 +2 cannam@95: M ./kernel/tensor.c -5 +6 cannam@95: M ./rdft/Makefile.am -2 +2 cannam@95: M ./rdft/buffered2.c -13 +18 cannam@95: M ./rdft/conf.c -1 +2 cannam@95: M ./rdft/direct2.c -9 +14 cannam@95: M ./rdft/nop2.c -2 +12 cannam@95: M ./rdft/problem2.c -36 +41 cannam@95: M ./rdft/rdft.h -7 +9 cannam@95: M ./rdft/vrank-geq1-rdft2.c -6 +23 cannam@95: M ./tests/bench.c -11 +36 cannam@95: cannam@95: Sat Aug 10 19:33:23 EDT 2002 stevenj cannam@95: * [project @ 2002-08-10 23:33:23 by stevenj] cannam@95: use tensor_copy_inplace cannam@95: cannam@95: M ./rdft/indirect.c -13 +5 cannam@95: cannam@95: Sat Aug 10 19:32:03 EDT 2002 stevenj cannam@95: * [project @ 2002-08-10 23:32:03 by stevenj] cannam@95: bugfix, use tensor_copy_inplace cannam@95: cannam@95: M ./dft/rank-geq2.c -4 +10 cannam@95: cannam@95: Sat Aug 10 19:30:39 EDT 2002 stevenj cannam@95: * [project @ 2002-08-10 23:30:39 by stevenj] cannam@95: use tensor_copy_inplace cannam@95: cannam@95: M ./dft/indirect.c -13 +5 cannam@95: cannam@95: Sat Aug 10 19:28:07 EDT 2002 stevenj cannam@95: * [project @ 2002-08-10 23:28:07 by stevenj] cannam@95: added tensor_copy_inplace cannam@95: cannam@95: M ./kernel/ifftw.h -1 +3 cannam@95: M ./kernel/tensor.c -1 +19 cannam@95: cannam@95: Sat Aug 10 19:25:50 EDT 2002 stevenj cannam@95: * [project @ 2002-08-10 23:25:50 by stevenj] cannam@95: fixed trig-function table type cannam@95: cannam@95: M ./kernel/twiddle.c -2 +2 cannam@95: cannam@95: Sat Aug 10 14:41:04 EDT 2002 athena cannam@95: * [project @ 2002-08-10 18:41:04 by athena] cannam@95: Improved trig scheme cannam@95: cannam@95: M ./kernel/trig.c -8 +53 cannam@95: M ./tests/trigtest.c -8 +74 cannam@95: cannam@95: Fri Aug 9 21:05:01 EDT 2002 athena cannam@95: * [project @ 2002-08-10 01:05:01 by athena] cannam@95: Allow for testing using long double instead of pari cannam@95: cannam@95: M ./tests/trigtest.c -15 +27 cannam@95: cannam@95: Fri Aug 9 20:49:32 EDT 2002 athena cannam@95: * [project @ 2002-08-10 00:49:32 by athena] cannam@95: Yet another trig scheme. cannam@95: cannam@95: M ./kernel/trig.c -25 +24 cannam@95: M ./tests/trigtest.c -25 +22 cannam@95: cannam@95: Fri Aug 9 20:38:07 EDT 2002 athena cannam@95: * [project @ 2002-08-10 00:38:07 by athena] cannam@95: Yet another scheme cannam@95: cannam@95: M ./kernel/trig.c -4 +12 cannam@95: M ./tests/trigtest.c +7 cannam@95: cannam@95: Fri Aug 9 20:31:16 EDT 2002 athena cannam@95: * [project @ 2002-08-10 00:31:16 by athena] cannam@95: Careful with overflow cannam@95: cannam@95: M ./kernel/ifftw.h -4 +4 cannam@95: M ./kernel/trig.c -23 +23 cannam@95: M ./tests/trigtest.c -24 +17 cannam@95: cannam@95: Fri Aug 9 20:16:23 EDT 2002 athena cannam@95: * [project @ 2002-08-10 00:16:23 by athena] cannam@95: Avoid overflow cannam@95: cannam@95: M ./kernel/ifftw.h -4 +4 cannam@95: M ./kernel/trig.c -20 +24 cannam@95: M ./tests/trigtest.c -21 +43 cannam@95: cannam@95: Fri Aug 9 19:26:57 EDT 2002 athena cannam@95: * [project @ 2002-08-09 23:26:57 by athena] cannam@95: New(er) trig routines cannam@95: cannam@95: M ./dft/rader.c -8 +6 cannam@95: M ./dft/verify.c -4 +3 cannam@95: M ./kernel/ifftw.h -4 +4 cannam@95: M ./kernel/trig.c -22 +17 cannam@95: M ./kernel/twiddle.c -9 +7 cannam@95: M ./rdft/rader-dht.c -4 +3 cannam@95: M ./rdft/rader-hc2hc.c -8 +6 cannam@95: M ./rdft/verify.c -4 +3 cannam@95: M ./tests/trigtest.c -20 +19 cannam@95: cannam@95: Fri Aug 9 19:25:44 EDT 2002 athena cannam@95: * [project @ 2002-08-09 23:25:44 by athena] cannam@95: Oops cannam@95: cannam@95: M ./tests/bench.c -1 +1 cannam@95: cannam@95: Fri Aug 9 18:49:04 EDT 2002 athena cannam@95: * [project @ 2002-08-09 22:49:04 by athena] cannam@95: New file cannam@95: cannam@95: A ./tests/trigtest.c cannam@95: cannam@95: Fri Aug 9 13:04:00 EDT 2002 athena cannam@95: * [project @ 2002-08-09 17:04:00 by athena] cannam@95: Commented about likely gcc bug cannam@95: cannam@95: M ./TODO +4 cannam@95: cannam@95: Fri Aug 9 13:01:49 EDT 2002 athena cannam@95: * [project @ 2002-08-09 17:01:49 by athena] cannam@95: Improved accuracy of twiddle factors cannam@95: cannam@95: A ./kernel/trig.c cannam@95: M ./dft/rader.c -8 +8 cannam@95: M ./dft/verify.c -5 +4 cannam@95: M ./kernel/Makefile.am -1 +1 cannam@95: M ./kernel/ifftw.h -10 +7 cannam@95: M ./kernel/trig.c +79 cannam@95: M ./kernel/twiddle.c -8 +10 cannam@95: M ./rdft/rader-dht.c -4 +4 cannam@95: M ./rdft/rader-hc2hc.c -8 +8 cannam@95: M ./rdft/verify.c -5 +4 cannam@95: M ./tests/bench.c -1 +1 cannam@95: cannam@95: Thu Aug 8 06:36:23 EDT 2002 athena cannam@95: * [project @ 2002-08-08 10:36:23 by athena] cannam@95: Wrong comment cannam@95: cannam@95: M ./simd/simd-3dnow.h -1 cannam@95: cannam@95: Wed Aug 7 17:14:09 EDT 2002 athena cannam@95: * [project @ 2002-08-07 21:14:09 by athena] cannam@95: Experimental 3dnow port using gcc, to compare it with Stefan's stuff. cannam@95: cannam@95: A ./simd/3dnow.c cannam@95: A ./simd/simd-3dnow.h cannam@95: M ./configure.ac +6 cannam@95: M ./genfft/gen_notw_c.ml -3 +6 cannam@95: M ./genfft/gen_twiddle_c.ml -2 +4 cannam@95: M ./kernel/ifftw.h -2 +2 cannam@95: M ./simd/3dnow.c +66 cannam@95: M ./simd/Makefile.am -2 +2 cannam@95: M ./simd/simd-3dnow.h +164 cannam@95: M ./simd/simd-altivec.h +2 cannam@95: M ./simd/simd-sse.h +3 cannam@95: M ./simd/simd-sse2.h +3 cannam@95: M ./simd/simd.h +4 cannam@95: cannam@95: Wed Aug 7 12:58:10 EDT 2002 athena cannam@95: * [project @ 2002-08-07 16:58:10 by athena] cannam@95: End of AREF experiment cannam@95: cannam@95: M ./genfft/c.ml -2 +2 cannam@95: M ./kernel/ifftw.h -3 +1 cannam@95: cannam@95: Wed Aug 7 07:47:19 EDT 2002 athena cannam@95: * [project @ 2002-08-07 11:47:19 by athena] cannam@95: Oops cannam@95: cannam@95: M ./configure.ac -4 +1 cannam@95: cannam@95: Wed Aug 7 07:46:38 EDT 2002 athena cannam@95: * [project @ 2002-08-07 11:46:38 by athena] cannam@95: Pathetic attempt to reduce size of configure script cannam@95: cannam@95: M ./configure.ac -16 +8 cannam@95: cannam@95: Tue Aug 6 20:38:11 EDT 2002 athena cannam@95: * [project @ 2002-08-07 00:38:11 by athena] cannam@95: Changed array syntax for experiments. cannam@95: cannam@95: M ./genfft/c.ml -8 +10 cannam@95: M ./kernel/ifftw.h -1 +3 cannam@95: cannam@95: Tue Aug 6 19:58:20 EDT 2002 athena cannam@95: * [project @ 2002-08-06 23:58:20 by athena] cannam@95: Fix warning cannam@95: cannam@95: M ./simd/simd-sse2.h +2 cannam@95: cannam@95: Tue Aug 6 13:35:28 EDT 2002 athena cannam@95: * [project @ 2002-08-06 17:35:28 by athena] cannam@95: Move nonportable stuff in one place. cannam@95: cannam@95: M ./dft/problem.c -2 +2 cannam@95: M ./kernel/align.c -1 +8 cannam@95: M ./kernel/ifftw.h -1 +2 cannam@95: M ./rdft/problem.c -2 +2 cannam@95: M ./rdft/problem2.c -2 +2 cannam@95: cannam@95: Tue Aug 6 10:32:53 EDT 2002 athena cannam@95: * [project @ 2002-08-06 14:32:53 by athena] cannam@95: Economy of thought: I didn't like having two algorithms for removing cannam@95: solutions, both correct. At least now we have the same algorithm cannam@95: copied twice. cannam@95: cannam@95: M ./kernel/planner.c -12 +22 cannam@95: cannam@95: Tue Aug 6 09:12:21 EDT 2002 athena cannam@95: * [project @ 2002-08-06 13:12:21 by athena] cannam@95: Added things to do cannam@95: cannam@95: M ./TODO -1 +1 cannam@95: cannam@95: Mon Aug 5 19:54:31 EDT 2002 stevenj cannam@95: * [project @ 2002-08-05 23:54:31 by stevenj] cannam@95: improved interaction of planner with patience flags cannam@95: cannam@95: M ./kernel/ifftw.h -1 +2 cannam@95: M ./kernel/planner.c -12 +34 cannam@95: cannam@95: Mon Aug 5 14:17:58 EDT 2002 stevenj cannam@95: * [project @ 2002-08-05 18:17:58 by stevenj] cannam@95: set up for real-even/odd DFTs, where n is not the size of the data cannam@95: cannam@95: M ./rdft/buffered.c -3 +3 cannam@95: M ./rdft/codelet.h -7 +25 cannam@95: M ./rdft/indirect.c -4 +8 cannam@95: M ./rdft/problem.c -13 +58 cannam@95: M ./rdft/rader-hc2hc.c -1 +1 cannam@95: M ./rdft/rdft.h -1 +3 cannam@95: M ./tests/bench.c -2 +3 cannam@95: cannam@95: Sun Aug 4 23:57:51 EDT 2002 stevenj cannam@95: * [project @ 2002-08-05 03:57:51 by stevenj] cannam@95: DESTROY_INPUT flag cannam@95: cannam@95: M ./dft/ct-dif.c -4 +5 cannam@95: M ./dft/ct-dit.c -3 +5 cannam@95: M ./dft/ct-ditbuf.c -3 +5 cannam@95: M ./dft/ct-ditf.c -3 +5 cannam@95: M ./dft/ct.c -2 +2 cannam@95: M ./dft/ct.h -2 +3 cannam@95: M ./kernel/ifftw.h -2 +3 cannam@95: M ./rdft/hc2hc-buf.c -3 +6 cannam@95: M ./rdft/hc2hc-dif.c -3 +5 cannam@95: M ./rdft/hc2hc-dit.c -3 +5 cannam@95: M ./rdft/hc2hc.c -2 +2 cannam@95: M ./rdft/hc2hc.h -2 +3 cannam@95: M ./rdft/r2hc-hc2r.c -13 +20 cannam@95: M ./tests/bench.c +2 cannam@95: cannam@95: Sun Aug 4 22:50:19 EDT 2002 stevenj cannam@95: * [project @ 2002-08-05 02:50:19 by stevenj] cannam@95: CLASSIC -> IMPATIENT cannam@95: cannam@95: M ./dft/rank-geq2.c -2 +2 cannam@95: M ./dft/vrank-geq1.c -2 +2 cannam@95: M ./kernel/ifftw.h -4 +4 cannam@95: M ./kernel/planner.c -3 +3 cannam@95: M ./rdft/dft-r2hc.c -2 +2 cannam@95: M ./rdft/vrank-geq1-rdft2.c -2 +2 cannam@95: M ./rdft/vrank-geq1.c -2 +2 cannam@95: M ./tests/bench.c -1 +1 cannam@95: cannam@95: Sun Aug 4 19:05:43 EDT 2002 athena cannam@95: * [project @ 2002-08-04 23:05:43 by athena] cannam@95: Require make maintainer-clean to remove the generator, as opposed cannam@95: to make clean. In this way we can type make clean without regenerating cannam@95: all codelets. cannam@95: cannam@95: M ./genfft/Makefile.am -11 +11 cannam@95: M ./genfft-k7/Makefile.am -3 +3 cannam@95: cannam@95: Sun Aug 4 17:34:04 EDT 2002 stevenj cannam@95: * [project @ 2002-08-04 21:34:04 by stevenj] cannam@95: ESTIMATE plans are not blessed cannam@95: cannam@95: M ./kernel/planner.c -10 +13 cannam@95: cannam@95: Sun Aug 4 17:24:37 EDT 2002 stevenj cannam@95: * [project @ 2002-08-04 21:24:37 by stevenj] cannam@95: use flags in wisdom cannam@95: cannam@95: M ./kernel/ifftw.h -2 +4 cannam@95: M ./kernel/planner.c -9 +14 cannam@95: cannam@95: Sun Aug 4 17:03:45 EDT 2002 stevenj cannam@95: * [project @ 2002-08-04 21:03:45 by stevenj] cannam@95: score now takes plnr, not flags, as arg cannam@95: cannam@95: M ./dft/buffered.c -3 +3 cannam@95: M ./dft/ct-dif.c -4 +3 cannam@95: M ./dft/ct-dit.c -3 +3 cannam@95: M ./dft/ct-ditbuf.c -3 +3 cannam@95: M ./dft/ct-ditf.c -3 +3 cannam@95: M ./dft/direct.c -3 +3 cannam@95: M ./dft/generic.c -2 +2 cannam@95: M ./dft/indirect.c -3 +3 cannam@95: M ./dft/nop.c -3 +3 cannam@95: M ./dft/rader.c -4 +4 cannam@95: M ./dft/rank-geq2.c -3 +3 cannam@95: M ./dft/rank0.c -3 +3 cannam@95: M ./dft/vrank-geq1.c -4 +4 cannam@95: M ./dft/vrank2-transpose.c -3 +3 cannam@95: M ./dft/vrank3-transpose.c -3 +3 cannam@95: M ./kernel/ifftw.h -2 +2 cannam@95: M ./kernel/planner-score.c -4 +3 cannam@95: M ./kernel/tensor.c -3 +3 cannam@95: M ./rdft/buffered.c -3 +3 cannam@95: M ./rdft/buffered2.c -3 +3 cannam@95: M ./rdft/dft-r2hc.c -3 +3 cannam@95: M ./rdft/direct.c -3 +3 cannam@95: M ./rdft/direct2.c -3 +3 cannam@95: M ./rdft/generic.c -2 +2 cannam@95: M ./rdft/hc2hc-buf.c -3 +3 cannam@95: M ./rdft/hc2hc-dif.c -3 +3 cannam@95: M ./rdft/hc2hc-dit.c -3 +3 cannam@95: M ./rdft/indirect.c -3 +3 cannam@95: M ./rdft/nop.c -3 +3 cannam@95: M ./rdft/nop2.c -3 +3 cannam@95: M ./rdft/r2hc-hc2r.c -3 +3 cannam@95: M ./rdft/rader-dht.c -2 +2 cannam@95: M ./rdft/rader-hc2hc.c -2 +2 cannam@95: M ./rdft/rank0.c -3 +3 cannam@95: M ./rdft/vrank-geq1-rdft2.c -4 +4 cannam@95: M ./rdft/vrank-geq1.c -4 +4 cannam@95: M ./rdft/vrank2-transpose.c -3 +3 cannam@95: M ./rdft/vrank3-transpose.c -3 +3 cannam@95: M ./tests/bench.c -1 +1 cannam@95: cannam@95: Sun Aug 4 16:37:46 EDT 2002 stevenj cannam@95: * [project @ 2002-08-04 20:37:46 by stevenj] cannam@95: align initial stack in alignment check, which should now pass for gcc 3.1.1 cannam@95: cannam@95: M ./acinclude.m4 -2 +3 cannam@95: cannam@95: Sat Aug 3 20:04:57 EDT 2002 athena cannam@95: * [project @ 2002-08-04 00:04:57 by athena] cannam@95: Detect ultrasparc (sort of) cannam@95: cannam@95: M ./acinclude.m4 +3 cannam@95: cannam@95: Sat Aug 3 19:38:17 EDT 2002 stevenj cannam@95: * [project @ 2002-08-03 23:38:17 by stevenj] cannam@95: added solvtab_rdft_r2r placeholder cannam@95: cannam@95: M ./rdft/codelet.h -1 +3 cannam@95: cannam@95: Sat Aug 3 19:34:49 EDT 2002 athena cannam@95: * [project @ 2002-08-03 23:34:49 by athena] cannam@95: Damn solaris cannam@95: cannam@95: M ./support/Makefile.codelets -2 +2 cannam@95: cannam@95: Sat Aug 3 17:55:44 EDT 2002 stevenj cannam@95: * [project @ 2002-08-03 21:55:44 by stevenj] cannam@95: use E extended precision in solvers cannam@95: cannam@95: M ./rdft/problem.c -7 +9 cannam@95: cannam@95: Sat Aug 3 17:53:29 EDT 2002 stevenj cannam@95: * [project @ 2002-08-03 21:53:29 by stevenj] cannam@95: an alternative notation for D{C,S}T: DXTio, where i/o are {0,1} cannam@95: according to whether the input/output are shifted, respectively. cannam@95: Alternatively, io is the binary representation of the usual cannam@95: DXT-{I,II,III,IV} nomenclature, minus 1. cannam@95: cannam@95: M ./rdft/codelet.h -2 +2 cannam@95: cannam@95: Sat Aug 3 17:49:11 EDT 2002 stevenj cannam@95: * [project @ 2002-08-03 21:49:11 by stevenj] cannam@95: use E extended precision in solvers cannam@95: cannam@95: M ./dft/generic.c -5 +5 cannam@95: M ./dft/rader.c -2 +2 cannam@95: M ./rdft/generic.c -21 +21 cannam@95: M ./rdft/r2hc-hc2r.c -3 +3 cannam@95: M ./rdft/rader-dht.c -4 +4 cannam@95: cannam@95: Sat Aug 3 15:39:49 EDT 2002 athena cannam@95: * [project @ 2002-08-03 19:39:49 by athena] cannam@95: More portability fixes, compiler bugs workarounds, etc. cannam@95: cannam@95: M ./configure.ac -2 +4 cannam@95: M ./kernel/cycle.h -2 +2 cannam@95: M ./kernel/planner.c -2 +2 cannam@95: M ./rdft/problem2.c -2 +4 cannam@95: cannam@95: Sat Aug 3 15:09:56 EDT 2002 athena cannam@95: * [project @ 2002-08-03 19:09:56 by athena] cannam@95: More portability work cannam@95: cannam@95: M ./configure.ac -2 +7 cannam@95: M ./kernel/cycle.h -2 +2 cannam@95: M ./kernel/ifftw.h -9 +2 cannam@95: cannam@95: Sat Aug 3 14:33:40 EDT 2002 athena cannam@95: * [project @ 2002-08-03 18:33:40 by athena] cannam@95: Improved portability, removed gnu make dependencies cannam@95: cannam@95: M ./acinclude.m4 -1 +1 cannam@95: M ./configure.ac -4 +3 cannam@95: M ./kernel/cycle.h -2 +2 cannam@95: M ./kernel/ifftw.h -2 +2 cannam@95: M ./support/Makefile.codelets -14 +6 cannam@95: cannam@95: Sat Aug 3 13:48:53 EDT 2002 athena cannam@95: * [project @ 2002-08-03 17:48:53 by athena] cannam@95: Remember to thank XXX cannam@95: cannam@95: M ./TODO +8 cannam@95: cannam@95: Fri Aug 2 17:38:18 EDT 2002 athena cannam@95: * [project @ 2002-08-02 21:38:18 by athena] cannam@95: Multiplication on altivec requires FMA with -0.0 to be IEEE754 compliant. cannam@95: cannam@95: M ./simd/simd-altivec.h -1 +1 cannam@95: cannam@95: Fri Aug 2 15:26:37 EDT 2002 athena cannam@95: * [project @ 2002-08-02 19:26:37 by athena] cannam@95: Allow for extended precision in codelets cannam@95: cannam@95: M ./genfft/c.ml -3 +4 cannam@95: M ./kernel/ifftw.h -7 +9 cannam@95: cannam@95: Fri Aug 2 08:52:04 EDT 2002 athena cannam@95: * [project @ 2002-08-02 12:52:04 by athena] cannam@95: Shortened names cannam@95: cannam@95: M ./dft/codelets/inplace/Makefile.am -5 +5 cannam@95: cannam@95: Fri Aug 2 03:49:09 EDT 2002 stevenj cannam@95: * [project @ 2002-08-02 07:49:09 by stevenj] cannam@95: added infrastructure for future r2r transforms cannam@95: cannam@95: M ./TODO -3 +1 cannam@95: M ./rdft/codelet.h -3 +33 cannam@95: M ./rdft/problem.c -1 +8 cannam@95: cannam@95: Thu Aug 1 21:29:14 EDT 2002 athena cannam@95: * [project @ 2002-08-02 01:29:05 by athena] cannam@95: Version info cannam@95: cannam@95: M ./Makefile.am +1 cannam@95: M ./configure.ac +2 cannam@95: cannam@95: Thu Aug 1 21:06:22 EDT 2002 athena cannam@95: * [project @ 2002-08-02 01:06:22 by athena] cannam@95: Listened to one customer and added radix-12. Added radix-15 for cannam@95: consistency (whatever that is) cannam@95: cannam@95: M ./dft/codelets/inplace/Makefile.am -1 +1 cannam@95: M ./dft/codelets/standard/Makefile.am -1 +1 cannam@95: M ./dft/k7/codelets/Makefile.am -3 +4 cannam@95: M ./dft/simd/codelets/Makefile.am -2 +2 cannam@95: M ./kernel/align.c -2 +8 cannam@95: M ./rdft/codelets/hc2r/Makefile.am -3 +3 cannam@95: M ./rdft/codelets/r2hc/Makefile.am -3 +3 cannam@95: cannam@95: Thu Aug 1 19:50:53 EDT 2002 stevenj cannam@95: * [project @ 2002-08-01 23:50:53 by stevenj] cannam@95: whoops again, fixed the wrong line cannam@95: cannam@95: M ./kernel/cycle.h -3 +3 cannam@95: cannam@95: Thu Aug 1 19:50:16 EDT 2002 stevenj cannam@95: * [project @ 2002-08-01 23:50:16 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./kernel/cycle.h -2 +2 cannam@95: cannam@95: Thu Aug 1 16:01:15 EDT 2002 stevenj cannam@95: * [project @ 2002-08-01 20:01:15 by stevenj] cannam@95: use new AC_INIT and add VERSION to wisdom cannam@95: cannam@95: M ./configure.ac -3 +4 cannam@95: M ./kernel/planner.c -4 +6 cannam@95: cannam@95: Thu Aug 1 14:56:45 EDT 2002 stevenj cannam@95: * [project @ 2002-08-01 18:56:45 by stevenj] cannam@95: mygetR -> getR cannam@95: cannam@95: M ./kernel/scan.c -3 +3 cannam@95: cannam@95: Thu Aug 1 14:56:02 EDT 2002 stevenj cannam@95: * [project @ 2002-08-01 18:56:02 by stevenj] cannam@95: scanner cleanups: just return 0/1, simplify integer reads cannam@95: cannam@95: M ./dft/problem.c -6 +4 cannam@95: M ./kernel/planner.c -14 +12 cannam@95: M ./kernel/scan.c -78 +42 cannam@95: M ./kernel/tensor.c -9 +8 cannam@95: M ./rdft/problem.c -7 +5 cannam@95: M ./rdft/problem2.c -7 +5 cannam@95: M ./tests/bench.c +7 cannam@95: cannam@95: Thu Aug 1 08:04:01 EDT 2002 athena cannam@95: * [project @ 2002-08-01 12:04:01 by athena] cannam@95: Reverted back to casting pointer to ulong cannam@95: cannam@95: M ./kernel/align.c -2 +2 cannam@95: cannam@95: Thu Aug 1 08:03:46 EDT 2002 athena cannam@95: * [project @ 2002-08-01 12:03:46 by athena] cannam@95: Cast to unsigned long, not long cannam@95: cannam@95: M ./kernel/ifftw.h -6 +6 cannam@95: cannam@95: Thu Aug 1 03:14:50 EDT 2002 stevenj cannam@95: * [project @ 2002-08-01 07:14:50 by stevenj] cannam@95: additional comment cannam@95: cannam@95: M ./kernel/scan.c -2 +4 cannam@95: cannam@95: Thu Aug 1 03:12:37 EDT 2002 stevenj cannam@95: * [project @ 2002-08-01 07:12:37 by stevenj] cannam@95: added comment cannam@95: cannam@95: M ./kernel/scan.c -1 +6 cannam@95: cannam@95: Thu Aug 1 03:03:18 EDT 2002 stevenj cannam@95: * [project @ 2002-08-01 07:03:18 by stevenj] cannam@95: added wisdom import cannam@95: cannam@95: A ./kernel/printers.c cannam@95: A ./kernel/scan.c cannam@95: A ./kernel/scanners.c cannam@95: M ./dft/conf.c -1 +3 cannam@95: M ./dft/dft.h -1 +3 cannam@95: M ./dft/problem.c -2 +31 cannam@95: M ./dft/verify.c -2 +1 cannam@95: M ./kernel/Makefile.am -3 +4 cannam@95: M ./kernel/alloc.c -2 +1 cannam@95: M ./kernel/assert.c -2 +1 cannam@95: M ./kernel/debug.c -9 +2 cannam@95: M ./kernel/ifftw.h -6 +53 cannam@95: M ./kernel/planner.c -22 +106 cannam@95: M ./kernel/print.c -2 +1 cannam@95: M ./kernel/printers.c +80 cannam@95: M ./kernel/problem.c -1 +2 cannam@95: M ./kernel/scan.c +322 cannam@95: M ./kernel/scanners.c +61 cannam@95: M ./kernel/tensor.c -4 +30 cannam@95: M ./kernel/timer.c -2 +1 cannam@95: M ./rdft/conf.c -1 +4 cannam@95: M ./rdft/problem.c -2 +32 cannam@95: M ./rdft/problem2.c -2 +32 cannam@95: M ./rdft/rdft.h -1 +4 cannam@95: M ./rdft/verify.c -2 +1 cannam@95: M ./tests/bench.c -9 +31 cannam@95: cannam@95: Wed Jul 31 23:12:05 EDT 2002 stevenj cannam@95: * [project @ 2002-08-01 03:12:05 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./kernel/align.c -1 +2 cannam@95: cannam@95: Wed Jul 31 22:06:46 EDT 2002 stevenj cannam@95: * [project @ 2002-08-01 02:06:46 by stevenj] cannam@95: use %u for alignment_of cannam@95: cannam@95: M ./dft/problem.c -3 +3 cannam@95: M ./rdft/problem.c -2 +2 cannam@95: M ./rdft/problem2.c -2 +2 cannam@95: cannam@95: Wed Jul 31 21:47:15 EDT 2002 stevenj cannam@95: * [project @ 2002-08-01 01:47:15 by stevenj] cannam@95: ptrdiff_t form cannam@95: cannam@95: M ./kernel/align.c -2 +2 cannam@95: cannam@95: Wed Jul 31 21:33:35 EDT 2002 athena cannam@95: * [project @ 2002-08-01 01:33:35 by athena] cannam@95: Cast to avoid warning from C++ compiler cannam@95: cannam@95: M ./kernel/ifftw.h -2 +2 cannam@95: cannam@95: Wed Jul 31 18:57:04 EDT 2002 athena cannam@95: * [project @ 2002-07-31 22:57:04 by athena] cannam@95: Make problem equality depend on alignments. cannam@95: cannam@95: A ./kernel/align.c cannam@95: M ./dft/problem.c -4 +10 cannam@95: M ./kernel/Makefile.am -5 +6 cannam@95: M ./kernel/align.c +35 cannam@95: M ./kernel/ifftw.h -1 +2 cannam@95: M ./rdft/problem.c -4 +9 cannam@95: M ./rdft/problem2.c -8 +13 cannam@95: M ./simd/simd.h +2 cannam@95: cannam@95: Wed Jul 31 15:45:31 EDT 2002 athena cannam@95: * [project @ 2002-07-31 19:45:31 by athena] cannam@95: Shorter names cannam@95: cannam@95: M ./dft/simd/codelets/Makefile.am -22 +18 cannam@95: cannam@95: Wed Jul 31 14:38:00 EDT 2002 athena cannam@95: * [project @ 2002-07-31 18:38:00 by athena] cannam@95: Oops cannam@95: cannam@95: M ./simd/simd-sse.h -1 +1 cannam@95: cannam@95: Wed Jul 31 14:37:19 EDT 2002 athena cannam@95: * [project @ 2002-07-31 18:37:19 by athena] cannam@95: Fix warning cannam@95: cannam@95: M ./simd/simd-sse.h +2 cannam@95: cannam@95: Wed Jul 31 07:52:53 EDT 2002 athena cannam@95: * [project @ 2002-07-31 11:52:53 by athena] cannam@95: Removed silly abstraction barrier. Also, cons() terminology was cannam@95: no longer appropriate. cannam@95: cannam@95: M ./kernel/alloc.c -2 +2 cannam@95: M ./kernel/ifftw.h -16 +19 cannam@95: M ./kernel/planner-naive.c -2 +2 cannam@95: M ./kernel/planner-score.c -2 +2 cannam@95: M ./kernel/planner.c -26 +13 cannam@95: cannam@95: Tue Jul 30 22:35:24 EDT 2002 stevenj cannam@95: * [project @ 2002-07-31 02:35:24 by stevenj] cannam@95: removed register_registrar and solvtab_exec_reverse hacks cannam@95: cannam@95: M ./kernel/ifftw.h -4 +3 cannam@95: M ./kernel/planner.c -24 +15 cannam@95: M ./kernel/solvtab.c -19 +6 cannam@95: cannam@95: Tue Jul 30 19:54:41 EDT 2002 stevenj cannam@95: * [project @ 2002-07-30 23:54:41 by stevenj] cannam@95: register_registrar doesn't search whole solver list (maybe we should change register_solver instead) cannam@95: cannam@95: M ./kernel/planner.c -1 +3 cannam@95: cannam@95: Tue Jul 30 19:36:37 EDT 2002 stevenj cannam@95: * [project @ 2002-07-30 23:36:37 by stevenj] cannam@95: credit cannam@95: cannam@95: M ./kernel/cycle.h -2 +2 cannam@95: cannam@95: Tue Jul 30 19:34:16 EDT 2002 stevenj cannam@95: * [project @ 2002-07-30 23:34:16 by stevenj] cannam@95: added HP/UX ia64 support, courtesy of Teresa L. Johnson cannam@95: cannam@95: M ./kernel/cycle.h -1 +23 cannam@95: cannam@95: Tue Jul 30 13:28:33 EDT 2002 athena cannam@95: * [project @ 2002-07-30 17:28:33 by athena] cannam@95: Fixed alignment checks cannam@95: cannam@95: M ./dft/simd/n1b.c -1 +1 cannam@95: M ./dft/simd/n1f.c -1 +1 cannam@95: M ./dft/simd/t1b.c -1 +1 cannam@95: M ./dft/simd/t1f.c -1 +1 cannam@95: M ./kernel/alloc.c -1 +2 cannam@95: cannam@95: Tue Jul 30 01:20:11 EDT 2002 stevenj cannam@95: * [project @ 2002-07-30 05:20:11 by stevenj] cannam@95: ugh, wisdom id fixes in exprt_conf cannam@95: cannam@95: M ./kernel/ifftw.h -2 +4 cannam@95: M ./kernel/planner.c -2 +4 cannam@95: M ./kernel/solvtab.c -5 +18 cannam@95: cannam@95: Tue Jul 30 00:41:15 EDT 2002 stevenj cannam@95: * [project @ 2002-07-30 04:41:15 by stevenj] cannam@95: exprt_registrars -> exprt_conf, added missing SOLVTAB_END cannam@95: cannam@95: M ./kernel/ifftw.h -2 +2 cannam@95: M ./kernel/planner.c -5 +6 cannam@95: M ./tests/bench.c -1 +1 cannam@95: cannam@95: Tue Jul 30 00:36:26 EDT 2002 stevenj cannam@95: * [project @ 2002-07-30 04:36:26 by stevenj] cannam@95: exprt_registrars should output self-contained configuration cannam@95: cannam@95: M ./kernel/planner.c -9 +20 cannam@95: cannam@95: Mon Jul 29 23:52:07 EDT 2002 stevenj cannam@95: * [project @ 2002-07-30 03:52:07 by stevenj] cannam@95: added exprt_registrars cannam@95: cannam@95: M ./dft/conf.c -12 +12 cannam@95: M ./kernel/ifftw.h -2 +7 cannam@95: M ./kernel/planner.c -4 +91 cannam@95: M ./kernel/solvtab.c -3 +5 cannam@95: M ./rdft/conf.c -18 +18 cannam@95: M ./support/Makefile.codelets -3 +3 cannam@95: M ./tests/bench.c +2 cannam@95: cannam@95: Mon Jul 29 23:42:27 EDT 2002 stevenj cannam@95: * [project @ 2002-07-30 03:42:27 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./kernel/print.c -2 +3 cannam@95: cannam@95: Mon Jul 29 21:24:51 EDT 2002 athena cannam@95: * [project @ 2002-07-30 01:24:51 by athena] cannam@95: More stringent requirements on strides for SIMD codelets cannam@95: cannam@95: M ./dft/simd/n1b.c -4 +4 cannam@95: M ./dft/simd/n1b.h +3 cannam@95: M ./dft/simd/n1f.c -4 +4 cannam@95: M ./dft/simd/n1f.h +3 cannam@95: M ./dft/simd/t1b.c -2 +2 cannam@95: M ./dft/simd/t1b.h -1 +6 cannam@95: M ./dft/simd/t1f.c -2 +2 cannam@95: M ./dft/simd/t1f.h +5 cannam@95: M ./simd/simd-altivec.h +2 cannam@95: M ./simd/simd-sse.h +2 cannam@95: M ./simd/simd-sse2.h +2 cannam@95: M ./simd/simd.h -2 +3 cannam@95: cannam@95: Mon Jul 29 21:05:49 EDT 2002 stevenj cannam@95: * [project @ 2002-07-30 01:05:49 by stevenj] cannam@95: remove warning cannam@95: cannam@95: M ./rdft/buffered2.c -2 +3 cannam@95: cannam@95: Mon Jul 29 20:51:19 EDT 2002 stevenj cannam@95: * [project @ 2002-07-30 00:51:01 by stevenj] cannam@95: use %td for ptrdiff_t and %T for tensors cannam@95: cannam@95: M ./dft/problem.c -2 +2 cannam@95: M ./kernel/print.c -3 +13 cannam@95: M ./kernel/traverse.c -2 +7 cannam@95: M ./rdft/problem.c -2 +2 cannam@95: M ./rdft/problem2.c -2 +2 cannam@95: cannam@95: Mon Jul 29 16:17:11 EDT 2002 athena cannam@95: * [project @ 2002-07-29 20:17:11 by athena] cannam@95: Fix for SIMD cannam@95: cannam@95: M ./dft/buffered.c -1 +5 cannam@95: cannam@95: Mon Jul 29 15:40:53 EDT 2002 athena cannam@95: * [project @ 2002-07-29 19:40:53 by athena] cannam@95: Missing lfftw_mkstride and lfftw_stride_destroy cannam@95: cannam@95: M ./kernel/ifftw.h -1 +3 cannam@95: cannam@95: Mon Jul 29 14:34:46 EDT 2002 athena cannam@95: * [project @ 2002-07-29 18:34:46 by athena] cannam@95: Implement LDA/STA cannam@95: cannam@95: M ./simd/simd-altivec.h +10 cannam@95: cannam@95: Mon Jul 29 14:19:21 EDT 2002 athena cannam@95: * [project @ 2002-07-29 18:19:21 by athena] cannam@95: More SIMD work cannam@95: cannam@95: M ./dft/simd/n1b.c -4 +4 cannam@95: M ./dft/simd/n1f.c -4 +4 cannam@95: M ./dft/simd/t1b.c -2 +2 cannam@95: M ./dft/simd/t1f.c -2 +2 cannam@95: M ./simd/simd-altivec.h -3 +2 cannam@95: M ./simd/simd-sse.h -3 +2 cannam@95: M ./simd/simd-sse2.h -2 +1 cannam@95: M ./simd/simd.h +3 cannam@95: cannam@95: Mon Jul 29 13:16:12 EDT 2002 athena cannam@95: * [project @ 2002-07-29 17:16:12 by athena] cannam@95: Cleanup cannam@95: cannam@95: M ./simd/simd-altivec.h -9 +9 cannam@95: M ./simd/simd-sse.h -16 +28 cannam@95: M ./simd/simd-sse2.h -13 +25 cannam@95: cannam@95: Mon Jul 29 13:02:38 EDT 2002 stevenj cannam@95: * [project @ 2002-07-29 17:02:38 by stevenj] cannam@95: update cannam@95: cannam@95: M ./ChangeLog +461 cannam@95: cannam@95: Mon Jul 29 12:45:33 EDT 2002 athena cannam@95: * [project @ 2002-07-29 16:45:33 by athena] cannam@95: Also check strides in SIMD codelets cannam@95: cannam@95: M ./dft/simd/n1b.c +4 cannam@95: M ./dft/simd/n1f.c +4 cannam@95: M ./dft/simd/t1b.c +2 cannam@95: M ./dft/simd/t1f.c +2 cannam@95: M ./simd/simd-altivec.h -1 +2 cannam@95: M ./simd/simd-sse.h +2 cannam@95: M ./simd/simd-sse2.h +2 cannam@95: cannam@95: Mon Jul 29 11:26:08 EDT 2002 athena cannam@95: * [project @ 2002-07-29 15:26:08 by athena] cannam@95: Minor changes, mostly for consistency with the big-endian processor cannam@95: cannam@95: M ./simd/simd-altivec.h -18 +18 cannam@95: cannam@95: Mon Jul 29 00:50:06 EDT 2002 stevenj cannam@95: * [project @ 2002-07-29 04:50:06 by stevenj] cannam@95: added comment cannam@95: cannam@95: M ./rdft/rader-dht.c +1 cannam@95: cannam@95: Sun Jul 28 21:19:35 EDT 2002 stevenj cannam@95: * [project @ 2002-07-29 01:19:35 by stevenj] cannam@95: added code for icc's _mm_malloc (memalign replacement) cannam@95: cannam@95: M ./configure.ac -1 +1 cannam@95: M ./kernel/alloc.c -4 +11 cannam@95: cannam@95: Sun Jul 28 17:33:07 EDT 2002 stevenj cannam@95: * [project @ 2002-07-28 21:33:07 by stevenj] cannam@95: slight fixes cannam@95: cannam@95: M ./rdft/problem2.c -6 +11 cannam@95: M ./rdft/verify.c -4 +7 cannam@95: M ./tests/bench.c -1 +2 cannam@95: cannam@95: Sun Jul 28 16:28:43 EDT 2002 stevenj cannam@95: * [project @ 2002-07-28 20:28:43 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./rdft/problem2.c -3 +3 cannam@95: cannam@95: Sun Jul 28 16:13:19 EDT 2002 athena cannam@95: * [project @ 2002-07-28 20:13:19 by athena] cannam@95: Use vec_xor to change sign cannam@95: cannam@95: M ./simd/altivec.c -2 +4 cannam@95: M ./simd/simd-altivec.h -3 +10 cannam@95: cannam@95: Sun Jul 28 16:10:59 EDT 2002 stevenj cannam@95: * [project @ 2002-07-28 20:10:59 by stevenj] cannam@95: added rdft2 cannam@95: cannam@95: A ./rdft/buffered2.c cannam@95: A ./rdft/direct2.c cannam@95: A ./rdft/nop2.c cannam@95: A ./rdft/plan2.c cannam@95: A ./rdft/problem2.c cannam@95: A ./rdft/solve2.c cannam@95: A ./rdft/vrank-geq1-rdft2.c cannam@95: M ./rdft/Makefile.am -2 +4 cannam@95: M ./rdft/buffered2.c +451 cannam@95: M ./rdft/conf.c -1 +7 cannam@95: M ./rdft/direct2.c +221 cannam@95: M ./rdft/khc2r.c -3 +3 cannam@95: M ./rdft/kr2hc.c -3 +3 cannam@95: M ./rdft/nop2.c +88 cannam@95: M ./rdft/plan2.c +33 cannam@95: M ./rdft/problem2.c +175 cannam@95: M ./rdft/rdft.h -1 +53 cannam@95: M ./rdft/solve2.c +31 cannam@95: M ./rdft/verify.c -17 +195 cannam@95: M ./rdft/vrank-geq1-rdft2.c +259 cannam@95: M ./tests/bench.c -6 +35 cannam@95: cannam@95: Sun Jul 28 15:45:54 EDT 2002 athena cannam@95: * [project @ 2002-07-28 19:45:54 by athena] cannam@95: Optimized cannam@95: cannam@95: M ./simd/simd-altivec.h -17 +10 cannam@95: cannam@95: Sun Jul 28 15:11:14 EDT 2002 athena cannam@95: * [project @ 2002-07-28 19:11:14 by athena] cannam@95: Changed ALIGNMENT cannam@95: cannam@95: M ./simd/simd-altivec.h -1 +1 cannam@95: cannam@95: Sun Jul 28 15:09:40 EDT 2002 athena cannam@95: * [project @ 2002-07-28 19:09:40 by athena] cannam@95: alignment := 8 cannam@95: cannam@95: M ./simd/simd-sse.h -1 +1 cannam@95: cannam@95: Sun Jul 28 14:57:22 EDT 2002 athena cannam@95: * [project @ 2002-07-28 18:57:22 by athena] cannam@95: Avoid warning cannam@95: cannam@95: M ./simd/simd-altivec.h -1 +1 cannam@95: cannam@95: Sun Jul 28 14:53:03 EDT 2002 athena cannam@95: * [project @ 2002-07-28 18:53:03 by athena] cannam@95: Oops cannam@95: cannam@95: M ./simd/simd-sse2.h -2 +2 cannam@95: cannam@95: Sun Jul 28 14:50:09 EDT 2002 athena cannam@95: * [project @ 2002-07-28 18:50:09 by athena] cannam@95: New altivec experiment cannam@95: cannam@95: M ./genfft/annotate.ml -5 +5 cannam@95: M ./genfft/genutil.ml -7 +7 cannam@95: M ./genfft/simd.ml -3 +6 cannam@95: M ./genfft/variable.ml -38 +16 cannam@95: M ./genfft/variable.mli -8 +4 cannam@95: M ./simd/simd-altivec.h -15 +15 cannam@95: M ./simd/simd-sse.h -3 +5 cannam@95: M ./simd/simd-sse2.h -2 +2 cannam@95: cannam@95: Sun Jul 28 13:48:20 EDT 2002 athena cannam@95: * [project @ 2002-07-28 17:48:20 by athena] cannam@95: Nothing cannam@95: cannam@95: M ./simd/simd-altivec.h -6 +6 cannam@95: cannam@95: Sun Jul 28 13:47:50 EDT 2002 athena cannam@95: * [project @ 2002-07-28 17:47:50 by athena] cannam@95: Oops cannam@95: cannam@95: M ./simd/simd-altivec.h -26 +19 cannam@95: cannam@95: Sun Jul 28 13:44:28 EDT 2002 athena cannam@95: * [project @ 2002-07-28 17:44:28 by athena] cannam@95: Nothing cannam@95: cannam@95: M ./simd/simd-altivec.h -19 +26 cannam@95: cannam@95: Sun Jul 28 10:38:10 EDT 2002 athena cannam@95: * [project @ 2002-07-28 14:38:10 by athena] cannam@95: Constants are now in separate file. cannam@95: cannam@95: A ./simd/altivec.c cannam@95: M ./simd/Makefile.am -1 +1 cannam@95: M ./simd/altivec.c +38 cannam@95: M ./simd/simd-altivec.h -20 +13 cannam@95: cannam@95: Sun Jul 28 07:58:37 EDT 2002 athena cannam@95: * [project @ 2002-07-28 11:58:37 by athena] cannam@95: More precise comment cannam@95: cannam@95: M ./simd/simd-altivec.h -1 +1 cannam@95: cannam@95: Sun Jul 28 07:56:40 EDT 2002 athena cannam@95: * [project @ 2002-07-28 11:56:40 by athena] cannam@95: gcc-3.1 bug workaround cannam@95: cannam@95: M ./simd/simd-altivec.h -42 +24 cannam@95: cannam@95: Sun Jul 28 01:39:54 EDT 2002 stevenj cannam@95: * [project @ 2002-07-28 05:39:54 by stevenj] cannam@95: slight optimization, and exported zerotens functions cannam@95: cannam@95: M ./dft/buffered.c -4 +4 cannam@95: M ./dft/dft.h -1 +2 cannam@95: M ./dft/problem.c -4 +4 cannam@95: M ./rdft/buffered.c -4 +4 cannam@95: M ./rdft/problem.c -4 +4 cannam@95: M ./rdft/rdft.h -1 +2 cannam@95: cannam@95: Sun Jul 28 00:54:59 EDT 2002 stevenj cannam@95: * [project @ 2002-07-28 04:54:59 by stevenj] cannam@95: should be a plan_dft, not a plan_rdft cannam@95: cannam@95: M ./rdft/dft-r2hc.c -2 +2 cannam@95: cannam@95: Sat Jul 27 21:36:46 EDT 2002 athena cannam@95: * [project @ 2002-07-28 01:36:46 by athena] cannam@95: Optimizations. Make it work with vanilla non-Apple gcc. cannam@95: cannam@95: M ./simd/simd-altivec.h -21 +82 cannam@95: cannam@95: Sat Jul 27 19:20:09 EDT 2002 stevenj cannam@95: * [project @ 2002-07-27 23:20:09 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./rdft/generic.c -5 cannam@95: cannam@95: Sat Jul 27 18:54:01 EDT 2002 stevenj cannam@95: * [project @ 2002-07-27 22:54:01 by stevenj] cannam@95: added hc2r (dif) cannam@95: cannam@95: M ./rdft/generic.c -13 +122 cannam@95: cannam@95: Sat Jul 27 18:31:43 EDT 2002 stevenj cannam@95: * [project @ 2002-07-27 22:31:43 by stevenj] cannam@95: add hc2r (dif) case cannam@95: cannam@95: M ./rdft/rader-hc2hc.c -2 +158 cannam@95: cannam@95: Sat Jul 27 15:09:40 EDT 2002 athena cannam@95: * [project @ 2002-07-27 19:09:40 by athena] cannam@95: Altivec port cannam@95: cannam@95: M ./simd/simd-altivec.h -38 +131 cannam@95: M ./support/Makefile.codelets -3 +3 cannam@95: cannam@95: Sat Jul 27 15:06:21 EDT 2002 athena cannam@95: * [project @ 2002-07-27 19:06:21 by athena] cannam@95: Fixed signed/unsigned bug. cannam@95: cannam@95: M ./kernel/twiddle.c -2 +3 cannam@95: cannam@95: Thu Jul 25 20:11:26 EDT 2002 athena cannam@95: * [project @ 2002-07-26 00:11:26 by athena] cannam@95: Make rank0 unapplicable to in-place problems. cannam@95: cannam@95: M ./dft/rank0.c -1 +2 cannam@95: M ./rdft/rank0.c -1 +2 cannam@95: cannam@95: Thu Jul 25 17:10:52 EDT 2002 stevenj cannam@95: * [project @ 2002-07-25 21:10:52 by stevenj] cannam@95: only works for r odd cannam@95: cannam@95: M ./rdft/generic.c -1 +1 cannam@95: cannam@95: Thu Jul 25 15:30:06 EDT 2002 athena cannam@95: * [project @ 2002-07-25 19:30:06 by athena] cannam@95: Reinserted much better timing-avoidance heuristic cannam@95: cannam@95: M ./kernel/planner-score.c -2 +7 cannam@95: cannam@95: Thu Jul 25 15:21:13 EDT 2002 athena cannam@95: * [project @ 2002-07-25 19:21:13 by athena] cannam@95: Score is now a property of the plan, not of the solver. cannam@95: Revised representation of closures. cannam@95: cannam@95: M ./dft/buffered.c -5 +1 cannam@95: M ./kernel/ifftw.h -2 +7 cannam@95: M ./kernel/plan.c -4 +6 cannam@95: M ./kernel/planner-score.c -17 +30 cannam@95: M ./kernel/traverse.c -17 +16 cannam@95: M ./rdft/buffered.c -5 +1 cannam@95: M ./tests/bench.c -7 +14 cannam@95: cannam@95: Thu Jul 25 06:36:51 EDT 2002 athena cannam@95: * [project @ 2002-07-25 10:36:51 by athena] cannam@95: Cosmetic changes. Added hc2r_128.c cannam@95: cannam@95: M ./genfft/gen_hc2r.ml -3 +3 cannam@95: M ./genfft/gen_r2hc.ml -3 +3 cannam@95: M ./rdft/codelets/hc2r/Makefile.am -1 +1 cannam@95: cannam@95: Thu Jul 25 01:37:53 EDT 2002 stevenj cannam@95: * [project @ 2002-07-25 05:37:53 by stevenj] cannam@95: added hc2r cannam@95: cannam@95: M ./rdft/rader-dht.c -16 +55 cannam@95: cannam@95: Thu Jul 25 00:51:45 EDT 2002 stevenj cannam@95: * [project @ 2002-07-25 04:51:45 by stevenj] cannam@95: added hc2hc-difbuf cannam@95: cannam@95: A ./rdft/hc2hc-buf.c cannam@95: M ./rdft/Makefile.am -3 +3 cannam@95: M ./rdft/hc2hc-buf.c +248 cannam@95: R ./rdft/hc2hc-ditbuf.c cannam@95: M ./rdft/khc2hc-dif.c -3 +1 cannam@95: cannam@95: Thu Jul 25 00:25:06 EDT 2002 stevenj cannam@95: * [project @ 2002-07-25 04:25:06 by stevenj] cannam@95: added rdft-dif cannam@95: cannam@95: A ./rdft/hc2hc-dif.c cannam@95: M ./rdft/Makefile.am -3 +3 cannam@95: M ./rdft/hc2hc-dif.c +128 cannam@95: M ./rdft/hc2hc.c -3 +5 cannam@95: M ./rdft/khc2hc-dif.c -4 +3 cannam@95: M ./rdft/rdft.h -1 +3 cannam@95: cannam@95: Thu Jul 25 00:22:36 EDT 2002 stevenj cannam@95: * [project @ 2002-07-25 04:22:36 by stevenj] cannam@95: whoops, hc2r must be conjugated to have right sign cannam@95: cannam@95: M ./rdft/verify.c -3 +3 cannam@95: cannam@95: Wed Jul 24 23:27:45 EDT 2002 stevenj cannam@95: * [project @ 2002-07-25 03:27:45 by stevenj] cannam@95: slight change cannam@95: cannam@95: M ./dft/ct-dif.c -4 +6 cannam@95: cannam@95: Wed Jul 24 23:24:24 EDT 2002 stevenj cannam@95: * [project @ 2002-07-25 03:24:24 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./rdft/verify.c -3 +2 cannam@95: cannam@95: Wed Jul 24 22:46:39 EDT 2002 stevenj cannam@95: * [project @ 2002-07-25 02:46:39 by stevenj] cannam@95: support hc2r codelets cannam@95: cannam@95: A ./rdft/direct.c cannam@95: M ./rdft/Makefile.am -5 +5 cannam@95: M ./rdft/codelet.h -1 +3 cannam@95: R ./rdft/direct-r2hc.c cannam@95: M ./rdft/direct.c +227 cannam@95: M ./rdft/khc2r.c -5 +2 cannam@95: M ./rdft/rdft.h -2 +2 cannam@95: cannam@95: Wed Jul 24 22:01:53 EDT 2002 stevenj cannam@95: * [project @ 2002-07-25 02:01:53 by stevenj] cannam@95: use vector plan for r/i instead of two separate plans cannam@95: cannam@95: M ./rdft/dft-r2hc.c -36 +20 cannam@95: cannam@95: Wed Jul 24 20:36:34 EDT 2002 stevenj cannam@95: * [project @ 2002-07-25 00:36:34 by stevenj] cannam@95: hack to allow rader/generic to work in-place for small prime sizes, instead of always using buffered cannam@95: cannam@95: M ./dft/buffered.c -1 +5 cannam@95: M ./dft/rader.c -2 +2 cannam@95: M ./kernel/ifftw.h -1 +3 cannam@95: M ./rdft/buffered.c -1 +5 cannam@95: M ./rdft/rader-dht.c -1 +1 cannam@95: M ./rdft/rader-hc2hc.c -1 +1 cannam@95: cannam@95: Wed Jul 24 18:04:41 EDT 2002 stevenj cannam@95: * [project @ 2002-07-24 22:04:41 by stevenj] cannam@95: added rdft-generic cannam@95: cannam@95: A ./rdft/generic.c cannam@95: M ./rdft/Makefile.am -4 +4 cannam@95: M ./rdft/conf.c -2 +2 cannam@95: M ./rdft/generic.c +281 cannam@95: cannam@95: Wed Jul 24 17:27:34 EDT 2002 stevenj cannam@95: * [project @ 2002-07-24 21:27:34 by stevenj] cannam@95: fixed add count cannam@95: cannam@95: M ./dft/generic.c -1 +1 cannam@95: cannam@95: Wed Jul 24 14:52:26 EDT 2002 stevenj cannam@95: * [project @ 2002-07-24 18:52:26 by stevenj] cannam@95: again cannam@95: cannam@95: M ./rdft/rader-hc2hc.c -2 +2 cannam@95: cannam@95: Wed Jul 24 14:51:58 EDT 2002 stevenj cannam@95: * [project @ 2002-07-24 18:51:58 by stevenj] cannam@95: slight fix cannam@95: cannam@95: M ./rdft/rader-hc2hc.c -4 +5 cannam@95: cannam@95: Wed Jul 24 14:51:07 EDT 2002 stevenj cannam@95: * [project @ 2002-07-24 18:51:07 by stevenj] cannam@95: fixed comment cannam@95: cannam@95: M ./rdft/rader-hc2hc.c -3 +3 cannam@95: cannam@95: Wed Jul 24 14:41:24 EDT 2002 stevenj cannam@95: * [project @ 2002-07-24 18:41:24 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./tests/bench.c -2 cannam@95: cannam@95: Wed Jul 24 14:38:15 EDT 2002 stevenj cannam@95: * [project @ 2002-07-24 18:38:15 by stevenj] cannam@95: added rader-hc2hc cannam@95: cannam@95: A ./rdft/rader-hc2hc.c cannam@95: M ./rdft/Makefile.am -2 +2 cannam@95: M ./rdft/conf.c -2 +2 cannam@95: M ./rdft/rader-hc2hc.c +482 cannam@95: M ./rdft/rdft.h -2 +2 cannam@95: M ./tests/bench.c +2 cannam@95: cannam@95: Wed Jul 24 00:07:59 EDT 2002 stevenj cannam@95: * [project @ 2002-07-24 04:07:59 by stevenj] cannam@95: whoops, initialize W cannam@95: cannam@95: M ./dft/rader.c +1 cannam@95: cannam@95: Tue Jul 23 23:03:09 EDT 2002 stevenj cannam@95: * [project @ 2002-07-24 03:03:09 by stevenj] cannam@95: strides should not be unsigned cannam@95: cannam@95: M ./rdft/rader-dht.c -2 +4 cannam@95: cannam@95: Tue Jul 23 23:02:08 EDT 2002 stevenj cannam@95: * [project @ 2002-07-24 03:02:08 by stevenj] cannam@95: more stride sign fixes cannam@95: cannam@95: M ./dft/rader.c -3 +5 cannam@95: cannam@95: Tue Jul 23 23:01:04 EDT 2002 stevenj cannam@95: * [project @ 2002-07-24 03:01:04 by stevenj] cannam@95: strides should not be unsigned! cannam@95: cannam@95: M ./dft/rader.c -1 +2 cannam@95: cannam@95: Tue Jul 23 14:55:25 EDT 2002 stevenj cannam@95: * [project @ 2002-07-23 18:55:25 by stevenj] cannam@95: added comment cannam@95: cannam@95: M ./rdft/dft-r2hc.c -1 +4 cannam@95: cannam@95: Tue Jul 23 14:52:04 EDT 2002 stevenj cannam@95: * [project @ 2002-07-23 18:52:04 by stevenj] cannam@95: another fix to op count cannam@95: cannam@95: M ./rdft/r2hc-hc2r.c -1 +4 cannam@95: cannam@95: Tue Jul 23 14:51:01 EDT 2002 stevenj cannam@95: * [project @ 2002-07-23 18:51:01 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./rdft/r2hc-hc2r.c -2 +2 cannam@95: cannam@95: Tue Jul 23 14:49:43 EDT 2002 stevenj cannam@95: * [project @ 2002-07-23 18:49:43 by stevenj] cannam@95: slight fix to op counts cannam@95: cannam@95: M ./rdft/dft-r2hc.c -3 +3 cannam@95: M ./rdft/r2hc-hc2r.c -3 +3 cannam@95: cannam@95: Tue Jul 23 14:09:19 EDT 2002 stevenj cannam@95: * [project @ 2002-07-23 18:09:18 by stevenj] cannam@95: added dft-r2hc cannam@95: cannam@95: A ./rdft/dft-r2hc.c cannam@95: M ./rdft/Makefile.am -5 +6 cannam@95: M ./rdft/conf.c -2 +3 cannam@95: M ./rdft/dft-r2hc.c +187 cannam@95: M ./rdft/rdft.h -1 +2 cannam@95: cannam@95: Tue Jul 23 02:50:12 EDT 2002 stevenj cannam@95: * [project @ 2002-07-23 06:50:12 by stevenj] cannam@95: better comment and var. name cannam@95: cannam@95: M ./rdft/rader-dht.c -7 +7 cannam@95: cannam@95: Tue Jul 23 02:39:11 EDT 2002 stevenj cannam@95: * [project @ 2002-07-23 06:39:11 by stevenj] cannam@95: fixed tests for hc2r, and added r2hc-hc2r cannam@95: cannam@95: A ./rdft/r2hc-hc2r.c cannam@95: M ./rdft/Makefile.am -3 +3 cannam@95: M ./rdft/conf.c -1 +2 cannam@95: M ./rdft/r2hc-hc2r.c +182 cannam@95: M ./rdft/rdft.h -1 +2 cannam@95: M ./rdft/verify.c -1 +46 cannam@95: M ./tests/bench.c -1 +6 cannam@95: cannam@95: Tue Jul 23 00:45:23 EDT 2002 stevenj cannam@95: * [project @ 2002-07-23 04:45:23 by stevenj] cannam@95: added rader-dht cannam@95: cannam@95: A ./rdft/rader-dht.c cannam@95: M ./rdft/Makefile.am -2 +3 cannam@95: M ./rdft/conf.c -1 +2 cannam@95: M ./rdft/rader-dht.c +433 cannam@95: M ./rdft/rdft.h -1 +2 cannam@95: cannam@95: Mon Jul 22 21:05:12 EDT 2002 athena cannam@95: * [project @ 2002-07-23 01:05:12 by athena] cannam@95: Added r2hc_128, what the hell. cannam@95: cannam@95: M ./rdft/codelets/r2hc/Makefile.am -1 +1 cannam@95: cannam@95: Mon Jul 22 20:48:59 EDT 2002 athena cannam@95: * [project @ 2002-07-23 00:48:59 by athena] cannam@95: Added codelets that compute twiddle factors cannam@95: cannam@95: M ./rdft/codelets/r2hc/Makefile.am -1 +7 cannam@95: cannam@95: Mon Jul 22 19:57:16 EDT 2002 stevenj cannam@95: * [project @ 2002-07-22 23:57:16 by stevenj] cannam@95: added rdft-buffered cannam@95: cannam@95: A ./rdft/buffered.c cannam@95: M ./rdft/Makefile.am -2 +2 cannam@95: M ./rdft/buffered.c +330 cannam@95: M ./rdft/conf.c -2 +2 cannam@95: cannam@95: Mon Jul 22 19:43:39 EDT 2002 stevenj cannam@95: * [project @ 2002-07-22 23:43:39 by stevenj] cannam@95: added hc2hc-ditbuf cannam@95: cannam@95: A ./rdft/hc2hc-ditbuf.c cannam@95: M ./rdft/Makefile.am -3 +3 cannam@95: M ./rdft/hc2hc-ditbuf.c +189 cannam@95: M ./rdft/khc2hc-dit.c -3 +1 cannam@95: cannam@95: Mon Jul 22 14:29:04 EDT 2002 stevenj cannam@95: * [project @ 2002-07-22 18:29:04 by stevenj] cannam@95: use STACK_MALLOC (alloca), since generic radix is always small cannam@95: cannam@95: M ./dft/generic.c -3 +3 cannam@95: cannam@95: Mon Jul 22 14:22:43 EDT 2002 stevenj cannam@95: * [project @ 2002-07-22 18:22:43 by stevenj] cannam@95: small cleanup cannam@95: cannam@95: M ./rdft/hc2hc-dit.c -2 +2 cannam@95: cannam@95: Mon Jul 22 07:42:13 EDT 2002 athena cannam@95: * [project @ 2002-07-22 11:42:13 by athena] cannam@95: What the hell was I thinking? cannam@95: cannam@95: M ./rdft/problem.c -18 +8 cannam@95: cannam@95: Mon Jul 22 07:37:12 EDT 2002 athena cannam@95: * [project @ 2002-07-22 11:37:12 by athena] cannam@95: Reduced code size by using table instead of switch statement. cannam@95: cannam@95: M ./rdft/problem.c -12 +16 cannam@95: cannam@95: Mon Jul 22 07:27:06 EDT 2002 athena cannam@95: * [project @ 2002-07-22 11:27:06 by athena] cannam@95: Changed hash function to avoid collisions with DFT. cannam@95: cannam@95: M ./rdft/problem.c -4 +4 cannam@95: cannam@95: Mon Jul 22 01:37:06 EDT 2002 stevenj cannam@95: * [project @ 2002-07-22 05:37:06 by stevenj] cannam@95: added missing file, whoops cannam@95: cannam@95: A ./rdft/hc2hc-dit.c cannam@95: cannam@95: Mon Jul 22 01:24:17 EDT 2002 stevenj cannam@95: * [project @ 2002-07-22 05:24:17 by stevenj] cannam@95: whoops, generate enough twiddles for odd m cannam@95: cannam@95: M ./rdft/hc2hc.c -2 +2 cannam@95: cannam@95: Mon Jul 22 01:10:21 EDT 2002 stevenj cannam@95: * [project @ 2002-07-22 05:10:21 by stevenj] cannam@95: don't try to verify R2HCII or HC2RIII plans cannam@95: cannam@95: M ./rdft/verify.c -5 +7 cannam@95: cannam@95: Mon Jul 22 01:05:00 EDT 2002 stevenj cannam@95: * [project @ 2002-07-22 05:05:00 by stevenj] cannam@95: recursive case now works, I think cannam@95: cannam@95: M ./rdft/hc2hc.c -9 +4 cannam@95: cannam@95: Mon Jul 22 01:04:40 EDT 2002 stevenj cannam@95: * [project @ 2002-07-22 05:04:40 by stevenj] cannam@95: add extra impulse test for debugging cannam@95: cannam@95: M ./rdft/verify.c -1 +4 cannam@95: cannam@95: Mon Jul 22 01:02:38 EDT 2002 stevenj cannam@95: * [project @ 2002-07-22 05:02:38 by stevenj] cannam@95: whoops, multiply ios offset by stride (and rename to ioffset) cannam@95: cannam@95: M ./rdft/direct-r2hc.c -4 +5 cannam@95: cannam@95: Mon Jul 22 00:22:02 EDT 2002 stevenj cannam@95: * [project @ 2002-07-22 04:22:02 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./rdft/verify.c -2 +2 cannam@95: cannam@95: Sun Jul 21 23:58:14 EDT 2002 stevenj cannam@95: * [project @ 2002-07-22 03:58:14 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./rdft/verify.c -2 +2 cannam@95: cannam@95: Sun Jul 21 23:43:03 EDT 2002 stevenj cannam@95: * [project @ 2002-07-22 03:43:03 by stevenj] cannam@95: added hc2hc-dit cannam@95: cannam@95: M ./genfft/gen_hc2hc.ml -3 +3 cannam@95: M ./rdft/Makefile.am -2 +2 cannam@95: M ./rdft/hc2hc.c -10 +21 cannam@95: M ./rdft/khc2hc-dit.c -4 +3 cannam@95: cannam@95: Sun Jul 21 23:15:12 EDT 2002 stevenj cannam@95: * [project @ 2002-07-22 03:15:12 by stevenj] cannam@95: twiddles can be shared with smaller m's cannam@95: cannam@95: M ./kernel/twiddle.c -4 +4 cannam@95: cannam@95: Sun Jul 21 22:34:28 EDT 2002 stevenj cannam@95: * [project @ 2002-07-22 02:34:28 by stevenj] cannam@95: preparing for recursive rdft... cannam@95: cannam@95: A ./rdft/hc2hc.c cannam@95: A ./rdft/hc2hc.h cannam@95: M ./rdft/Makefile.am -3 +3 cannam@95: M ./rdft/codelet.h -1 +2 cannam@95: M ./rdft/codelets/hfb.c -2 +2 cannam@95: M ./rdft/hc2hc.c +249 cannam@95: M ./rdft/hc2hc.h +69 cannam@95: cannam@95: Sun Jul 21 19:31:22 EDT 2002 stevenj cannam@95: * [project @ 2002-07-21 23:31:22 by stevenj] cannam@95: slight fix, to match libbench/verify.c cannam@95: cannam@95: M ./rdft/verify.c -3 +3 cannam@95: cannam@95: Sun Jul 21 18:43:12 EDT 2002 stevenj cannam@95: * [project @ 2002-07-21 22:43:12 by stevenj] cannam@95: r2hcII has imag parts offset by n-1, not n. We can also allocate fewer strides. cannam@95: cannam@95: M ./rdft/direct-r2hc.c -4 +7 cannam@95: cannam@95: Sun Jul 21 18:27:09 EDT 2002 stevenj cannam@95: * [project @ 2002-07-21 22:27:09 by stevenj] cannam@95: delete unused var cannam@95: cannam@95: M ./rdft/rank0.c -2 +2 cannam@95: cannam@95: Sun Jul 21 02:06:53 EDT 2002 stevenj cannam@95: * [project @ 2002-07-21 06:06:53 by stevenj] cannam@95: added some rdft solvers cannam@95: cannam@95: A ./rdft/direct-r2hc.c cannam@95: A ./rdft/indirect.c cannam@95: A ./rdft/nop.c cannam@95: A ./rdft/rank0.c cannam@95: A ./rdft/vrank-geq1.c cannam@95: A ./rdft/vrank2-transpose.c cannam@95: A ./rdft/vrank3-transpose.c cannam@95: M ./rdft/Makefile.am -2 +4 cannam@95: M ./rdft/codelet.h -3 +8 cannam@95: M ./rdft/codelets/hc2r.c -2 +2 cannam@95: M ./rdft/codelets/r2hc.c -2 +2 cannam@95: M ./rdft/conf.c -4 +4 cannam@95: M ./rdft/direct-r2hc.c +174 cannam@95: M ./rdft/indirect.c +235 cannam@95: R ./rdft/khc2rIII.c cannam@95: M ./rdft/kr2hc.c -5 +2 cannam@95: R ./rdft/kr2hcII.c cannam@95: M ./rdft/nop.c +97 cannam@95: M ./rdft/problem.c -3 +20 cannam@95: M ./rdft/rank0.c +217 cannam@95: M ./rdft/rdft.h -5 +4 cannam@95: M ./rdft/vrank-geq1.c +260 cannam@95: M ./rdft/vrank2-transpose.c +130 cannam@95: M ./rdft/vrank3-transpose.c +171 cannam@95: cannam@95: Sun Jul 21 01:52:54 EDT 2002 stevenj cannam@95: * [project @ 2002-07-21 05:52:54 by stevenj] cannam@95: pass identifier in FFTW() through another macro so that the mangled name cannam@95: can itself be a preprocessor symbol cannam@95: cannam@95: M ./kernel/fftw3.h -4 +5 cannam@95: cannam@95: Sun Jul 21 01:05:21 EDT 2002 stevenj cannam@95: * [project @ 2002-07-21 05:05:21 by stevenj] cannam@95: fix in comment cannam@95: cannam@95: M ./dft/vrank-geq1.c -2 +2 cannam@95: cannam@95: Sun Jul 21 00:47:03 EDT 2002 stevenj cannam@95: * [project @ 2002-07-21 04:47:03 by stevenj] cannam@95: bench tests rdft plans cannam@95: cannam@95: M ./Makefile.am -2 +5 cannam@95: M ./rdft/rdft.h -2 +3 cannam@95: M ./tests/bench.c -16 +36 cannam@95: cannam@95: Sun Jul 21 00:22:14 EDT 2002 stevenj cannam@95: * [project @ 2002-07-21 04:22:14 by stevenj] cannam@95: make rdft.h and dft.h compatible cannam@95: cannam@95: M ./rdft/codelet.h -4 +4 cannam@95: M ./tests/Makefile.am -1 +1 cannam@95: M ./tests/bench.c -1 +1 cannam@95: cannam@95: Sun Jul 21 00:12:19 EDT 2002 stevenj cannam@95: * [project @ 2002-07-21 04:12:19 by stevenj] cannam@95: first-draft rdft verify cannam@95: cannam@95: A ./rdft/verify.c cannam@95: M ./rdft/Makefile.am -1 +1 cannam@95: M ./rdft/problem.c -8 +8 cannam@95: M ./rdft/rdft.h -5 +5 cannam@95: M ./rdft/verify.c +554 cannam@95: cannam@95: Sat Jul 20 22:09:15 EDT 2002 stevenj cannam@95: * [project @ 2002-07-21 02:09:15 by stevenj] cannam@95: got rid of annoying warnings cannam@95: cannam@95: M ./rdft/khc2hc-dif.c -1 +2 cannam@95: M ./rdft/khc2hc-dit.c -1 +2 cannam@95: M ./rdft/khc2r.c -1 +2 cannam@95: M ./rdft/khc2rIII.c -1 +2 cannam@95: M ./rdft/kr2hc.c -1 +2 cannam@95: M ./rdft/kr2hcII.c -1 +2 cannam@95: cannam@95: Sat Jul 20 22:07:37 EDT 2002 stevenj cannam@95: * [project @ 2002-07-21 02:07:37 by stevenj] cannam@95: added stub codelet registration for linking purposes cannam@95: cannam@95: A ./rdft/khc2hc-dif.c cannam@95: A ./rdft/khc2hc-dit.c cannam@95: A ./rdft/khc2r.c cannam@95: A ./rdft/khc2rIII.c cannam@95: A ./rdft/kr2hc.c cannam@95: A ./rdft/kr2hcII.c cannam@95: M ./rdft/Makefile.am -1 +2 cannam@95: M ./rdft/khc2hc-dif.c +31 cannam@95: M ./rdft/khc2hc-dit.c +31 cannam@95: M ./rdft/khc2r.c +31 cannam@95: M ./rdft/khc2rIII.c +31 cannam@95: M ./rdft/kr2hc.c +31 cannam@95: M ./rdft/kr2hcII.c +31 cannam@95: M ./rdft/rdft.h -2 +2 cannam@95: cannam@95: Sat Jul 20 21:46:03 EDT 2002 stevenj cannam@95: * [project @ 2002-07-21 01:46:03 by stevenj] cannam@95: basic rdft stuff cannam@95: cannam@95: A ./rdft/conf.c cannam@95: A ./rdft/plan.c cannam@95: A ./rdft/problem.c cannam@95: A ./rdft/rdft.h cannam@95: A ./rdft/solve.c cannam@95: M ./rdft/Makefile.am -1 +1 cannam@95: M ./rdft/conf.c +47 cannam@95: M ./rdft/plan.c +33 cannam@95: M ./rdft/problem.c +145 cannam@95: M ./rdft/rdft.h +82 cannam@95: M ./rdft/solve.c +31 cannam@95: cannam@95: Sat Jul 20 21:06:50 EDT 2002 stevenj cannam@95: * [project @ 2002-07-21 01:06:50 by stevenj] cannam@95: rdft codelets now compile cannam@95: cannam@95: A ./rdft/Makefile.am cannam@95: A ./rdft/codelet.h cannam@95: A ./rdft/codelets/Makefile.am cannam@95: A ./rdft/codelets/hb.h cannam@95: A ./rdft/codelets/hc2r.c cannam@95: A ./rdft/codelets/hc2r.h cannam@95: A ./rdft/codelets/hc2rIII.h cannam@95: A ./rdft/codelets/hf.h cannam@95: A ./rdft/codelets/hfb.c cannam@95: A ./rdft/codelets/r2hc.c cannam@95: A ./rdft/codelets/r2hc.h cannam@95: A ./rdft/codelets/r2hcII.h cannam@95: M ./Makefile.am -1 +1 cannam@95: M ./configure.ac +2 cannam@95: M ./dft/codelet.h -33 +1 cannam@95: M ./genfft/gen_hc2hc.ml -3 +3 cannam@95: M ./kernel/ifftw.h -1 +35 cannam@95: M ./rdft/Makefile.am +6 cannam@95: M ./rdft/codelet.h +133 cannam@95: M ./rdft/codelets/Makefile.am +7 cannam@95: M ./rdft/codelets/hb.h +23 cannam@95: M ./rdft/codelets/hc2r/Makefile.am -3 +3 cannam@95: M ./rdft/codelets/hc2r.c +44 cannam@95: M ./rdft/codelets/hc2r.h +23 cannam@95: M ./rdft/codelets/hc2rIII.h +23 cannam@95: M ./rdft/codelets/hf.h +23 cannam@95: M ./rdft/codelets/hfb.c +41 cannam@95: M ./rdft/codelets/r2hc/Makefile.am -2 +2 cannam@95: M ./rdft/codelets/r2hc.c +44 cannam@95: M ./rdft/codelets/r2hc.h +23 cannam@95: M ./rdft/codelets/r2hcII.h +23 cannam@95: cannam@95: Sat Jul 20 18:40:31 EDT 2002 athena cannam@95: * [project @ 2002-07-20 22:40:31 by athena] cannam@95: Oops, was generating rdfts instead of hdfts cannam@95: cannam@95: M ./genfft/gen_hc2r.ml -3 +3 cannam@95: cannam@95: Sat Jul 20 18:25:47 EDT 2002 athena cannam@95: * [project @ 2002-07-20 22:25:47 by athena] cannam@95: Added hc2r codelets cannam@95: cannam@95: A ./rdft/codelets/hc2r/ cannam@95: A ./rdft/codelets/hc2r/Makefile.am cannam@95: M ./TODO +4 cannam@95: M ./configure.ac +1 cannam@95: M ./genfft/twiddle.ml -7 +14 cannam@95: M ./genfft-k7/twiddle.ml -12 +7 cannam@95: M ./kernel/twiddle.c -26 +39 cannam@95: M ./rdft/codelets/hc2r/Makefile.am -1 +42 cannam@95: cannam@95: Sat Jul 20 17:54:39 EDT 2002 athena cannam@95: * [project @ 2002-07-20 21:54:39 by athena] cannam@95: return W in hc2hc codelets cannam@95: cannam@95: M ./genfft/gen_hc2hc.ml -3 +4 cannam@95: cannam@95: Sat Jul 20 17:51:06 EDT 2002 athena cannam@95: * [project @ 2002-07-20 21:51:06 by athena] cannam@95: Some work on rdft codelets cannam@95: cannam@95: A ./rdft/ cannam@95: A ./rdft/codelets/ cannam@95: A ./rdft/codelets/r2hc/ cannam@95: A ./rdft/codelets/r2hc/Makefile.am cannam@95: M ./configure.ac +2 cannam@95: M ./dft/codelets/inplace/Makefile.am +4 cannam@95: M ./dft/codelets/standard/Makefile.am +3 cannam@95: M ./dft/simd/codelets/Makefile.am -1 +1 cannam@95: M ./genfft/gen_hc2hc.ml -62 +18 cannam@95: M ./genfft/gen_hc2r.ml -5 +24 cannam@95: M ./genfft/gen_r2hc.ml -5 +21 cannam@95: M ./genfft/trig.ml -22 +22 cannam@95: M ./rdft/codelets/r2hc/Makefile.am -1 +42 cannam@95: M ./support/Makefile.codelets -9 +6 cannam@95: cannam@95: Tue Jul 16 13:55:50 EDT 2002 athena cannam@95: * [project @ 2002-07-16 17:55:50 by athena] cannam@95: fix const cannam@95: cannam@95: M ./kernel/fftw3.h -3 +4 cannam@95: cannam@95: Tue Jul 16 07:00:10 EDT 2002 athena cannam@95: * [project @ 2002-07-16 11:00:10 by athena] cannam@95: Separate CFLAGS in codelets. Fix const in certain places. cannam@95: cannam@95: M ./acinclude.m4 -2 +6 cannam@95: M ./configure.ac +1 cannam@95: M ./dft/codelets/inplace/Makefile.am +1 cannam@95: M ./dft/codelets/standard/Makefile.am +1 cannam@95: M ./dft/indirect.c -2 +2 cannam@95: M ./dft/rank0.c -2 +2 cannam@95: M ./dft/simd/codelets/Makefile.am +1 cannam@95: M ./kernel/version.c -3 +4 cannam@95: M ./tests/bench.c -2 +9 cannam@95: cannam@95: Mon Jul 15 21:10:42 EDT 2002 stevenj cannam@95: * [project @ 2002-07-16 01:10:42 by stevenj] cannam@95: note buffering problem cannam@95: cannam@95: M ./TODO +2 cannam@95: cannam@95: Mon Jul 15 20:27:51 EDT 2002 athena cannam@95: * [project @ 2002-07-16 00:27:51 by athena] cannam@95: Removed unpredictable branch from inner loop cannam@95: cannam@95: M ./dft/generic.c -2 +2 cannam@95: cannam@95: Mon Jul 15 19:35:04 EDT 2002 stevenj cannam@95: * [project @ 2002-07-15 23:35:04 by stevenj] cannam@95: update cannam@95: cannam@95: M ./TODO -2 cannam@95: cannam@95: Mon Jul 15 19:31:39 EDT 2002 stevenj cannam@95: * [project @ 2002-07-15 23:31:39 by stevenj] cannam@95: optimization cannam@95: cannam@95: M ./dft/generic.c -1 +2 cannam@95: cannam@95: Mon Jul 15 19:28:30 EDT 2002 stevenj cannam@95: * [project @ 2002-07-15 23:28:30 by stevenj] cannam@95: added generic dit cannam@95: cannam@95: A ./dft/generic.c cannam@95: M ./dft/Makefile.am -4 +4 cannam@95: M ./dft/conf.c -1 +2 cannam@95: M ./dft/dft.h -1 +2 cannam@95: M ./dft/generic.c +219 cannam@95: M ./kernel/ifftw.h -2 +3 cannam@95: M ./kernel/twiddle.c -1 +13 cannam@95: cannam@95: Mon Jul 15 17:03:53 EDT 2002 stevenj cannam@95: * [project @ 2002-07-15 21:03:53 by stevenj] cannam@95: whoops, mksolver should be static cannam@95: cannam@95: M ./dft/rader.c -2 +2 cannam@95: cannam@95: Mon Jul 15 16:46:36 EDT 2002 athena cannam@95: * [project @ 2002-07-15 20:46:35 by athena] cannam@95: First implementation of gen_hc2hc, probably still buggy. cannam@95: cannam@95: A ./genfft/gen_hc2hc.ml cannam@95: M ./genfft/Makefile.am -9 +15 cannam@95: M ./genfft/algsimp.ml -4 +5 cannam@95: M ./genfft/c.ml -3 +10 cannam@95: M ./genfft/c.mli -1 +2 cannam@95: M ./genfft/gen_hc2hc.ml +220 cannam@95: M ./genfft/gen_hc2r.ml -24 +69 cannam@95: M ./genfft/gen_r2hc.ml -22 +66 cannam@95: M ./genfft/genutil.ml -13 +17 cannam@95: cannam@95: Mon Jul 15 16:40:23 EDT 2002 stevenj cannam@95: * [project @ 2002-07-15 20:40:23 by stevenj] cannam@95: don't count loading of twiddle factors in ops.other, since it isn't cannam@95: counted for the codelets cannam@95: cannam@95: M ./dft/rader.c -1 cannam@95: cannam@95: Mon Jul 15 15:13:19 EDT 2002 stevenj cannam@95: * [project @ 2002-07-15 19:13:19 by stevenj] cannam@95: plan_destroy puts plan to sleep before deallocating it, to eliminate duplicate free calls in solvers cannam@95: cannam@95: M ./dft/ct.c -2 +1 cannam@95: M ./dft/rader.c -2 cannam@95: M ./kernel/plan.c -2 +5 cannam@95: cannam@95: Mon Jul 15 15:07:41 EDT 2002 stevenj cannam@95: * [project @ 2002-07-15 19:07:41 by stevenj] cannam@95: fftw2-like vector recursion flag cannam@95: cannam@95: M ./dft/ct-dif.c -1 +6 cannam@95: M ./dft/ct-dit.c -2 +2 cannam@95: M ./dft/ct-ditbuf.c -2 +2 cannam@95: M ./dft/ct.c -1 +5 cannam@95: M ./dft/vrank-geq1.c -2 +10 cannam@95: M ./kernel/ifftw.h -2 +6 cannam@95: M ./tests/bench.c -2 +1 cannam@95: cannam@95: Sun Jul 14 21:01:44 EDT 2002 athena cannam@95: * [project @ 2002-07-15 01:01:44 by athena] cannam@95: More jokes cannam@95: cannam@95: M ./kernel/planner.c -1 +2 cannam@95: cannam@95: Sun Jul 14 20:36:01 EDT 2002 athena cannam@95: * [project @ 2002-07-15 00:36:01 by athena] cannam@95: Bless plan for testing purposes cannam@95: cannam@95: M ./tests/bench.c +1 cannam@95: cannam@95: Sun Jul 14 20:35:49 EDT 2002 athena cannam@95: * [project @ 2002-07-15 00:35:49 by athena] cannam@95: Canonical linked-list deletion (hope it is right) cannam@95: cannam@95: M ./kernel/planner.c -12 +13 cannam@95: cannam@95: Sun Jul 14 18:26:19 EDT 2002 stevenj cannam@95: * [project @ 2002-07-14 22:26:19 by stevenj] cannam@95: use estimating planner for cld_omega cannam@95: cannam@95: M ./dft/rader.c -1 +1 cannam@95: cannam@95: Sun Jul 14 18:10:56 EDT 2002 stevenj cannam@95: * [project @ 2002-07-14 22:10:56 by stevenj] cannam@95: better internal naming cannam@95: cannam@95: M ./dft/rader.c -3 +3 cannam@95: cannam@95: Sun Jul 14 18:10:01 EDT 2002 stevenj cannam@95: * [project @ 2002-07-14 22:10:01 by stevenj] cannam@95: printing should really be fixed now, grrr cannam@95: cannam@95: M ./dft/rader.c -10 +12 cannam@95: cannam@95: Sun Jul 14 17:57:12 EDT 2002 stevenj cannam@95: * [project @ 2002-07-14 21:57:12 by stevenj] cannam@95: print all distinct child plans cannam@95: cannam@95: M ./dft/rader.c -2 +7 cannam@95: cannam@95: Sun Jul 14 17:49:21 EDT 2002 stevenj cannam@95: * [project @ 2002-07-14 21:49:21 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./tests/bench.c -1 +1 cannam@95: cannam@95: Sun Jul 14 17:45:54 EDT 2002 stevenj cannam@95: * [project @ 2002-07-14 21:45:54 by stevenj] cannam@95: whoops, destroy should delete twiddle/omega from list cannam@95: cannam@95: M ./dft/rader.c -4 +2 cannam@95: cannam@95: Sun Jul 14 17:33:02 EDT 2002 stevenj cannam@95: * [project @ 2002-07-14 21:33:02 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./kernel/planner.c -3 +6 cannam@95: cannam@95: Sun Jul 14 17:12:14 EDT 2002 stevenj cannam@95: * [project @ 2002-07-14 21:12:14 by stevenj] cannam@95: added plan_bless and FORGET_ACCURSED cannam@95: cannam@95: M ./kernel/ifftw.h -2 +6 cannam@95: M ./kernel/plan.c -1 +14 cannam@95: M ./kernel/planner.c -13 +24 cannam@95: cannam@95: Sun Jul 14 16:15:43 EDT 2002 stevenj cannam@95: * [project @ 2002-07-14 20:15:43 by stevenj] cannam@95: further cleanup cannam@95: cannam@95: M ./kernel/traverse.c -6 +4 cannam@95: cannam@95: Sun Jul 14 16:14:15 EDT 2002 stevenj cannam@95: * [project @ 2002-07-14 20:14:15 by stevenj] cannam@95: slight cleanup cannam@95: cannam@95: M ./kernel/traverse.c -8 +7 cannam@95: cannam@95: Sun Jul 14 16:09:17 EDT 2002 stevenj cannam@95: * [project @ 2002-07-14 20:09:17 by stevenj] cannam@95: added traverse_plan via print (ugh) cannam@95: cannam@95: A ./kernel/traverse.c cannam@95: M ./kernel/Makefile.am -3 +3 cannam@95: M ./kernel/ifftw.h -1 +5 cannam@95: M ./kernel/traverse.c +115 cannam@95: M ./tests/bench.c +12 cannam@95: cannam@95: Sun Jul 14 15:08:29 EDT 2002 stevenj cannam@95: * [project @ 2002-07-14 19:08:29 by stevenj] cannam@95: added TW_FULL, and additional n parameter for twiddles cannam@95: cannam@95: M ./dft/ct.c -3 +4 cannam@95: M ./kernel/ifftw.h -5 +5 cannam@95: M ./kernel/twiddle.c -18 +34 cannam@95: cannam@95: Sun Jul 14 15:03:51 EDT 2002 stevenj cannam@95: * [project @ 2002-07-14 19:03:51 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./kernel/planner.c -2 +4 cannam@95: cannam@95: Sun Jul 14 13:49:20 EDT 2002 stevenj cannam@95: * [project @ 2002-07-14 17:49:20 by stevenj] cannam@95: save flags before invoking solver mkplan cannam@95: cannam@95: M ./kernel/ifftw.h -1 +2 cannam@95: M ./kernel/planner-naive.c -2 +2 cannam@95: M ./kernel/planner-score.c -2 +2 cannam@95: M ./kernel/planner.c -3 +10 cannam@95: cannam@95: Sun Jul 14 09:28:37 EDT 2002 athena cannam@95: * [project @ 2002-07-14 13:28:37 by athena] cannam@95: cannam@95: M ./TODO -6 +1 cannam@95: cannam@95: Sat Jul 13 22:17:29 EDT 2002 stevenj cannam@95: * [project @ 2002-07-14 02:17:29 by stevenj] cannam@95: added support for UNICOS _rtc() real-time-clock intrinsic function cannam@95: cannam@95: M ./configure.ac -1 +10 cannam@95: M ./kernel/cycle.h -2 +20 cannam@95: cannam@95: Sat Jul 13 22:06:35 EDT 2002 stevenj cannam@95: * [project @ 2002-07-14 02:06:35 by stevenj] cannam@95: fixed typo: HAVE_TIME_H should include time.h, not sys/time.h cannam@95: cannam@95: M ./kernel/timer.c -2 +2 cannam@95: cannam@95: Sat Jul 13 21:46:02 EDT 2002 stevenj cannam@95: * [project @ 2002-07-14 01:46:02 by stevenj] cannam@95: support AIX read_real_time timer cannam@95: cannam@95: M ./configure.ac +1 cannam@95: M ./kernel/cycle.h -1 +24 cannam@95: cannam@95: Sat Jul 13 17:02:51 EDT 2002 stevenj cannam@95: * [project @ 2002-07-13 21:02:51 by stevenj] cannam@95: use && instead of the (sigh) unportable -a cannam@95: cannam@95: M ./configure.ac -1 +1 cannam@95: cannam@95: Sat Jul 13 16:38:18 EDT 2002 stevenj cannam@95: * [project @ 2002-07-13 20:38:18 by stevenj] cannam@95: use AC_HELP_STRING cannam@95: cannam@95: M ./configure.ac -11 +11 cannam@95: cannam@95: Sat Jul 13 16:05:43 EDT 2002 stevenj cannam@95: * [project @ 2002-07-13 20:05:43 by stevenj] cannam@95: support long-double precision cannam@95: cannam@95: A ./kernel/lfftw3.h cannam@95: M ./configure.ac -2 +22 cannam@95: M ./dft/codelet.h -2 +6 cannam@95: M ./dft/verify.c -10 +29 cannam@95: M ./kernel/Makefile.am -3 +3 cannam@95: M ./kernel/dfftw3.h -1 +2 cannam@95: M ./kernel/fftw3.h -2 +5 cannam@95: M ./kernel/ifftw.h -9 +14 cannam@95: M ./kernel/lfftw3.h +25 cannam@95: M ./kernel/sfftw3.h -1 +2 cannam@95: M ./libbench/bench-main.c -3 +5 cannam@95: M ./libbench/bench-user.h -4 +8 cannam@95: M ./libbench/info.c -2 +4 cannam@95: M ./libbench/verify.c -10 +38 cannam@95: M ./simd/simd-sse2.h -1 +1 cannam@95: cannam@95: Sat Jul 13 15:48:10 EDT 2002 stevenj cannam@95: * [project @ 2002-07-13 19:48:10 by stevenj] cannam@95: whoops whoops cannam@95: cannam@95: M ./dft/rader.c -2 +2 cannam@95: cannam@95: Sat Jul 13 15:47:39 EDT 2002 stevenj cannam@95: * [project @ 2002-07-13 19:47:39 by stevenj] cannam@95: whoops cannam@95: cannam@95: M ./dft/rader.c -1 +1 cannam@95: cannam@95: Sat Jul 13 14:13:42 EDT 2002 stevenj cannam@95: * [project @ 2002-07-13 18:13:42 by stevenj] cannam@95: buffered solver strides have been fixed cannam@95: cannam@95: M ./TODO -7 cannam@95: cannam@95: Sat Jul 13 13:48:13 EDT 2002 stevenj cannam@95: * [project @ 2002-07-13 17:48:13 by stevenj] cannam@95: convention cannam@95: cannam@95: M ./dft/rader.c -1 +1 cannam@95: cannam@95: Sat Jul 13 12:50:06 EDT 2002 stevenj cannam@95: * [project @ 2002-07-13 16:50:06 by stevenj] cannam@95: share twiddle arrays in Rader cannam@95: cannam@95: M ./TODO -2 cannam@95: M ./dft/rader.c -5 +68 cannam@95: cannam@95: Sat Jul 13 12:48:10 EDT 2002 stevenj cannam@95: * [project @ 2002-07-13 16:48:10 by stevenj] cannam@95: call done() after verify cannam@95: cannam@95: M ./libbench/verify.c -1 +2 cannam@95: cannam@95: Fri Jul 12 15:42:04 EDT 2002 stevenj cannam@95: * [project @ 2002-07-12 19:42:04 by stevenj] cannam@95: output planner time with -v cannam@95: cannam@95: M ./tests/bench.c +7 cannam@95: cannam@95: Fri Jul 12 15:40:14 EDT 2002 stevenj cannam@95: * [project @ 2002-07-12 19:40:14 by stevenj] cannam@95: support double outputs cannam@95: cannam@95: M ./kernel/print.c -2 +9 cannam@95: cannam@95: Fri Jul 12 15:09:19 EDT 2002 stevenj cannam@95: * [project @ 2002-07-12 19:09:19 by stevenj] cannam@95: removed extraneous parens cannam@95: cannam@95: M ./dft/vrank-geq1.c -2 +2 cannam@95: cannam@95: Fri Jul 12 15:08:13 EDT 2002 stevenj cannam@95: * [project @ 2002-07-12 19:08:13 by stevenj] cannam@95: increase maxbufsz to 64k; makes a big difference for large 2d transforms cannam@95: cannam@95: M ./dft/buffered.c -2 +2 cannam@95: cannam@95: Fri Jul 12 05:59:26 EDT 2002 athena cannam@95: * [project @ 2002-07-12 09:59:26 by athena] cannam@95: Fix cannam@95: cannam@95: M ./dft/vrank-geq1.c -2 +2 cannam@95: cannam@95: Fri Jul 12 01:22:38 EDT 2002 stevenj cannam@95: * [project @ 2002-07-12 05:22:38 by stevenj] cannam@95: fix comment cannam@95: cannam@95: M ./dft/rank-geq2.c -2 +2 cannam@95: cannam@95: Fri Jul 12 00:59:29 EDT 2002 stevenj cannam@95: * [project @ 2002-07-12 04:59:29 by stevenj] cannam@95: fix in comment cannam@95: cannam@95: M ./kernel/tensor.c -2 +2 cannam@95: cannam@95: Fri Jul 12 00:13:13 EDT 2002 stevenj cannam@95: * [project @ 2002-07-12 04:13:13 by stevenj] cannam@95: updated cannam@95: cannam@95: M ./ChangeLog +873 cannam@95: cannam@95: Thu Jul 11 23:39:27 EDT 2002 stevenj cannam@95: * [project @ 2002-07-12 03:39:27 by stevenj] cannam@95: buffered malloc's buffers cannam@95: cannam@95: M ./TODO -3 cannam@95: cannam@95: Thu Jul 11 23:30:26 EDT 2002 stevenj cannam@95: * [project @ 2002-07-12 03:30:26 by stevenj] cannam@95: share more code between apply and apply_dit in Rader cannam@95: cannam@95: M ./TODO -2 cannam@95: M ./dft/rader.c -69 +30 cannam@95: cannam@95: Mon Jul 8 12:30:34 EDT 2002 athena cannam@95: * [project @ 2002-07-08 16:30:34 by athena] cannam@95: Polished cannam@95: cannam@95: M ./simd/simd-sse.h -18 +28 cannam@95: M ./simd/simd-sse2.h -25 +31 cannam@95: M ./simd/sse.c -4 +2 cannam@95: M ./simd/sse2.c -3 +2 cannam@95: cannam@95: Mon Jul 8 09:47:11 EDT 2002 athena cannam@95: * [project @ 2002-07-08 13:47:11 by athena] cannam@95: cannam@95: M ./support/Makefile.codelets -1 +1 cannam@95: cannam@95: Mon Jul 8 09:42:08 EDT 2002 athena cannam@95: * [project @ 2002-07-08 13:42:08 by athena] cannam@95: SIMD/FMA stuff cannam@95: cannam@95: M ./dft/simd/codelets/Makefile.am -1 +1 cannam@95: M ./genfft/c.ml -7 +2 cannam@95: M ./genfft/gen_notw_c.ml -2 +3 cannam@95: M ./genfft/simd.ml -2 +13 cannam@95: M ./genfft/to_alist.ml -1 +3 cannam@95: M ./genfft/to_alist.mli -1 +2 cannam@95: M ./simd/simd-sse.h +3 cannam@95: M ./simd/simd-sse2.h +3 cannam@95: M ./support/Makefile.codelets -1 +2 cannam@95: cannam@95: Mon Jul 8 07:43:51 EDT 2002 athena cannam@95: * [project @ 2002-07-08 11:43:51 by athena] cannam@95: Avoid code duplication cannam@95: cannam@95: M ./simd/simd-sse.h -12 +6 cannam@95: cannam@95: Sun Jul 7 20:56:15 EDT 2002 athena cannam@95: * [project @ 2002-07-08 00:56:15 by athena] cannam@95: Fixes for FMA+SIMD cannam@95: cannam@95: M ./genfft/Makefile.am -3 +3 cannam@95: M ./genfft/to_alist.ml -7 +14 cannam@95: cannam@95: Sun Jul 7 20:32:01 EDT 2002 athena cannam@95: * [project @ 2002-07-08 00:32:01 by athena] cannam@95: Major changes in SIMD fftw cannam@95: cannam@95: A ./dft/simd/n1b.c cannam@95: A ./dft/simd/n1b.h cannam@95: A ./dft/simd/n1f.c cannam@95: A ./dft/simd/n1f.h cannam@95: A ./dft/simd/t1b.c cannam@95: A ./dft/simd/t1b.h cannam@95: A ./dft/simd/t1f.c cannam@95: A ./dft/simd/t1f.h cannam@95: A ./genfft/gen_notw_c.ml cannam@95: A ./genfft/gen_twiddle_c.ml cannam@95: M ./dft/buffered.c -2 +2 cannam@95: M ./dft/codelets/standard/Makefile.am -6 +5 cannam@95: M ./dft/simd/Makefile.am -3 +1 cannam@95: R ./dft/simd/NAMING cannam@95: M ./dft/simd/codelets/Makefile.am -26 +23 cannam@95: M ./dft/simd/n1b.c +43 cannam@95: M ./dft/simd/n1b.h +24 cannam@95: M ./dft/simd/n1f.c +43 cannam@95: M ./dft/simd/n1f.h +24 cannam@95: R ./dft/simd/n2f.c cannam@95: R ./dft/simd/n2f.h cannam@95: R ./dft/simd/n3f.h cannam@95: R ./dft/simd/n4.c cannam@95: R ./dft/simd/n4.h cannam@95: M ./dft/simd/t1b.c +38 cannam@95: M ./dft/simd/t1b.h +25 cannam@95: M ./dft/simd/t1f.c +38 cannam@95: M ./dft/simd/t1f.h +25 cannam@95: R ./dft/simd/t2f.c cannam@95: R ./dft/simd/t2f.h cannam@95: R ./dft/simd/t3f.h cannam@95: R ./dft/simd/t4.c cannam@95: R ./dft/simd/t4.h cannam@95: M ./genfft/Makefile.am -29 +27 cannam@95: M ./genfft/algsimp.ml -16 +27 cannam@95: M ./genfft/annotate.ml -131 +11 cannam@95: M ./genfft/annotate.mli -17 +2 cannam@95: M ./genfft/c.ml -7 +23 cannam@95: M ./genfft/complex.ml -1 +3 cannam@95: M ./genfft/complex.mli -1 +2 cannam@95: M ./genfft/expr.ml -3 +21 cannam@95: M ./genfft/expr.mli -1 +6 cannam@95: M ./genfft/gen_athnotw.ml -3 +3 cannam@95: M ./genfft/gen_athtw.ml -3 +3 cannam@95: M ./genfft/gen_conv.ml -4 +4 cannam@95: M ./genfft/gen_hc2r.ml -4 +4 cannam@95: M ./genfft/gen_notw.ml -4 +4 cannam@95: M ./genfft/gen_notw_c.ml +155 cannam@95: M ./genfft/gen_r2hc.ml -4 +4 cannam@95: M ./genfft/gen_trig.ml -4 +4 cannam@95: M ./genfft/gen_twiddle.ml -4 +4 cannam@95: M ./genfft/gen_twiddle_c.ml +165 cannam@95: M ./genfft/gen_twidsq.ml -4 +4 cannam@95: M ./genfft/genutil.ml -4 +5 cannam@95: M ./genfft/oracle.ml -1 +2 cannam@95: M ./genfft/simd.ml -177 +95 cannam@95: M ./genfft/simd.mli -2 +2 cannam@95: M ./genfft/simdmagic.ml -47 +1 cannam@95: M ./genfft/to_alist.ml -3 +9 cannam@95: M ./genfft/trig.ml -1 +10 cannam@95: M ./genfft/trig.mli -1 +2 cannam@95: M ./genfft/twiddle.ml -5 +6 cannam@95: M ./kernel/ifftw.h -2 +2 cannam@95: M ./simd/simd-sse.h -61 +88 cannam@95: M ./simd/simd-sse2.h -67 +57 cannam@95: M ./simd/sse.c -1 +4 cannam@95: M ./simd/sse2.c -1 +4 cannam@95: M ./support/Makefile.codelets -67 +2 cannam@95: cannam@95: Fri Jul 5 17:32:09 EDT 2002 athena cannam@95: * [project @ 2002-07-05 21:32:09 by athena] cannam@95: Use unpck instructions instead of shuffles cannam@95: cannam@95: M ./dft/buffered.c -2 +2 cannam@95: M ./simd/simd-altivec.h -13 +13 cannam@95: M ./simd/simd-sse.h -23 +10 cannam@95: cannam@95: Fri Jul 5 15:49:14 EDT 2002 athena cannam@95: * [project @ 2002-07-05 19:49:14 by athena] cannam@95: Minor tweaks cannam@95: cannam@95: M ./dft/codelets/n.c +1 cannam@95: M ./dft/codelets/t.c +1 cannam@95: M ./dft/ct-ditbuf.c -2 +2 cannam@95: M ./dft/verify.c -4 +1 cannam@95: M ./kernel/ifftw.h -10 +9 cannam@95: M ./kernel/planner.c -5 +4 cannam@95: M ./tests/bench.c -2 +2 cannam@95: cannam@95: Fri Jul 5 15:02:54 EDT 2002 athena cannam@95: * [project @ 2002-07-05 19:02:54 by athena] cannam@95: Use score planner cannam@95: cannam@95: M ./tests/bench.c -1 +1 cannam@95: cannam@95: Fri Jul 5 14:49:59 EDT 2002 athena cannam@95: * [project @ 2002-07-05 18:49:59 by athena] cannam@95: Added verifier cannam@95: cannam@95: A ./dft/verify.c cannam@95: A ./kernel/dotens.c cannam@95: A ./kernel/dotens2.c cannam@95: M ./CONVENTIONS +1 cannam@95: M ./dft/Makefile.am -1 +1 cannam@95: M ./dft/dft.h -1 +4 cannam@95: M ./dft/verify.c +428 cannam@95: M ./kernel/Makefile.am -1 +2 cannam@95: M ./kernel/debug.c -2 +2 cannam@95: M ./kernel/dotens.c +47 cannam@95: M ./kernel/dotens2.c +55 cannam@95: M ./kernel/ifftw.h -8 +24 cannam@95: M ./tests/bench.c -8 +14 cannam@95: cannam@95: Wed Jul 3 20:32:28 EDT 2002 athena cannam@95: * [project @ 2002-07-04 00:32:28 by athena] cannam@95: More simd codelets cannam@95: cannam@95: A ./dft/simd/n2f.c cannam@95: A ./dft/simd/n2f.h cannam@95: A ./dft/simd/n3f.h cannam@95: A ./dft/simd/t2f.c cannam@95: A ./dft/simd/t2f.h cannam@95: A ./dft/simd/t3f.h cannam@95: M ./dft/buffered.c -3 +3 cannam@95: M ./dft/codelet.h -1 +3 cannam@95: M ./dft/codelets/n.c +1 cannam@95: M ./dft/codelets/t.c +1 cannam@95: M ./dft/ct-dif.c -1 +6 cannam@95: M ./dft/ct-dit.c -1 +6 cannam@95: M ./dft/ct-ditbuf.c -9 +19 cannam@95: M ./dft/ct.c -11 +17 cannam@95: M ./dft/ct.h -3 +5 cannam@95: M ./dft/simd/Makefile.am -1 +2 cannam@95: M ./dft/simd/NAMING -7 +20 cannam@95: M ./dft/simd/codelets/Makefile.am -4 +27 cannam@95: M ./dft/simd/n2f.c +43 cannam@95: M ./dft/simd/n2f.h +25 cannam@95: M ./dft/simd/n3f.h +1 cannam@95: M ./dft/simd/n4.c -2 +1 cannam@95: M ./dft/simd/t2f.c +40 cannam@95: M ./dft/simd/t2f.h +25 cannam@95: M ./dft/simd/t3f.h +1 cannam@95: M ./dft/simd/t4.c -1 +1 cannam@95: M ./genfft/annotate.ml -3 +3 cannam@95: M ./genfft/gen_notw.ml -9 +23 cannam@95: M ./genfft/gen_twiddle.ml -6 +14 cannam@95: M ./genfft/gen_twidsq.ml -3 +9 cannam@95: M ./genfft/simd.ml -43 +69 cannam@95: M ./genfft/simdmagic.ml -11 +36 cannam@95: M ./genfft-k7/gen_notw.ml -1 +2 cannam@95: M ./genfft-k7/gen_twiddle.ml -1 +2 cannam@95: M ./kernel/alloc.c -5 +2 cannam@95: M ./kernel/ifftw.h -3 +9 cannam@95: M ./simd/simd-sse.h -28 +106 cannam@95: M ./simd/simd-sse2.h -24 +47 cannam@95: cannam@95: Tue Jul 2 16:18:09 EDT 2002 athena cannam@95: * [project @ 2002-07-02 20:18:09 by athena] cannam@95: Oops cannam@95: cannam@95: M ./dft/rank-geq2.c -3 +4 cannam@95: cannam@95: Tue Jul 2 16:13:24 EDT 2002 athena cannam@95: * [project @ 2002-07-02 20:13:24 by athena] cannam@95: Fixed classic mode cannam@95: cannam@95: M ./dft/rank-geq2.c -6 +5 cannam@95: M ./dft/vrank-geq1.c -6 +6 cannam@95: cannam@95: Tue Jul 2 15:38:36 EDT 2002 athena cannam@95: * [project @ 2002-07-02 19:38:36 by athena] cannam@95: Use LDK for constants so that we can play games. cannam@95: cannam@95: M ./genfft/simd.ml -2 +2 cannam@95: M ./simd/simd-altivec.h +1 cannam@95: M ./simd/simd-sse.h -5 +5 cannam@95: M ./simd/simd-sse2.h -2 +5 cannam@95: cannam@95: Tue Jul 2 13:15:58 EDT 2002 athena cannam@95: * [project @ 2002-07-02 17:15:58 by athena] cannam@95: Improved support for fixed strides cannam@95: cannam@95: M ./dft/codelet.h -1 +2 cannam@95: M ./dft/codelets/n.c -1 +2 cannam@95: M ./dft/simd/n4.c +3 cannam@95: M ./dft/simd/t4.c +2 cannam@95: M ./genfft/gen_notw.ml -8 +21 cannam@95: M ./genfft/genutil.ml -1 +6 cannam@95: M ./genfft/simd.ml -7 +7 cannam@95: M ./genfft/simd.mli -3 +3 cannam@95: M ./genfft-k7/gen_notw.ml -1 +2 cannam@95: M ./simd/simd-sse.h -17 +54 cannam@95: cannam@95: Tue Jul 2 10:30:58 EDT 2002 athena cannam@95: * [project @ 2002-07-02 14:30:58 by athena] cannam@95: Changed accounting of flops cannam@95: cannam@95: M ./dft/codelet.h -12 +25 cannam@95: M ./dft/codelets/n.c -4 +5 cannam@95: M ./dft/codelets/n.h -6 +2 cannam@95: M ./dft/codelets/t.c -3 +5 cannam@95: M ./dft/codelets/t.h -3 +2 cannam@95: M ./dft/ct-dif.c -3 +5 cannam@95: M ./dft/ct-dit.c -3 +5 cannam@95: M ./dft/ct-ditbuf.c -3 +4 cannam@95: M ./dft/ct-ditf.c -3 +6 cannam@95: M ./dft/direct.c -5 +5 cannam@95: M ./dft/k7/k7.c -14 +23 cannam@95: M ./dft/simd/n4.c -3 +5 cannam@95: M ./dft/simd/n4.h -5 +2 cannam@95: M ./dft/simd/t4.c -3 +5 cannam@95: M ./dft/simd/t4.h -3 +2 cannam@95: M ./genfft/gen_notw.ml -3 +3 cannam@95: M ./genfft/gen_twiddle.ml -3 +3 cannam@95: M ./genfft/gen_twidsq.ml -3 +3 cannam@95: M ./genfft-k7/gen_notw.ml -2 +2 cannam@95: M ./genfft-k7/gen_twiddle.ml -2 +2 cannam@95: cannam@95: Tue Jul 2 08:51:38 EDT 2002 athena cannam@95: * [project @ 2002-07-02 12:51:38 by athena] cannam@95: Wrong code in non-fma mode cannam@95: cannam@95: M ./genfft/algsimp.ml -11 +15 cannam@95: M ./genfft-k7/algsimp.ml -8 +24 cannam@95: M ./genfft-k7/to_alist.ml -3 +3 cannam@95: M ./simd/simd-sse2.h -4 +9 cannam@95: cannam@95: Mon Jul 1 23:17:06 EDT 2002 athena cannam@95: * [project @ 2002-07-02 03:17:06 by athena] cannam@95: sse2 stuff cannam@95: cannam@95: A ./simd/simd-sse2.h cannam@95: A ./simd/sse2.c cannam@95: M ./genfft/simdmagic.ml -2 +2 cannam@95: M ./kernel/alloc.c -2 +2 cannam@95: M ./simd/Makefile.am -1 +2 cannam@95: M ./simd/simd-sse2.h +125 cannam@95: M ./simd/sse2.c +73 cannam@95: cannam@95: Mon Jul 1 14:05:56 EDT 2002 athena cannam@95: * [project @ 2002-07-01 18:05:56 by athena] cannam@95: Identify CPUs for special codelets cannam@95: cannam@95: A ./simd/sse.c cannam@95: M ./Makefile.am +1 cannam@95: M ./dft/ct.c -4 +4 cannam@95: M ./dft/direct.c -2 +2 cannam@95: M ./dft/k7/k7.c -5 +43 cannam@95: M ./dft/simd/n4.c -1 +1 cannam@95: M ./dft/simd/t4.c -1 +1 cannam@95: M ./kernel/alloc.c -2 +2 cannam@95: M ./simd/Makefile.am -1 +3 cannam@95: M ./simd/simd-altivec.h +2 cannam@95: M ./simd/simd-sse.h +3 cannam@95: M ./simd/sse.c +73 cannam@95: cannam@95: Mon Jul 1 09:26:42 EDT 2002 athena cannam@95: * [project @ 2002-07-01 13:26:42 by athena] cannam@95: Change split problem syntax cannam@95: cannam@95: M ./libbench/problem.c -2 +2 cannam@95: cannam@95: Mon Jul 1 09:11:39 EDT 2002 athena cannam@95: * [project @ 2002-07-01 13:11:39 by athena] cannam@95: Removed -fma flag cannam@95: cannam@95: M ./dft/simd/codelets/Makefile.am -1 +1 cannam@95: cannam@95: Sun Jun 30 20:08:26 EDT 2002 athena cannam@95: * [project @ 2002-07-01 00:08:26 by athena] cannam@95: Work around gcc bug cannam@95: cannam@95: M ./simd/simd-altivec.h -3 +17 cannam@95: cannam@95: Sun Jun 30 18:34:06 EDT 2002 athena cannam@95: * [project @ 2002-06-30 22:34:06 by athena] cannam@95: New simd stuff cannam@95: cannam@95: M ./genfft/algsimp.ml -8 +20 cannam@95: M ./genfft/magic.ml -1 +3 cannam@95: M ./genfft/oracle.ml -2 +2 cannam@95: M ./genfft/simd.ml -2 +4 cannam@95: M ./genfft/to_alist.ml -3 +3 cannam@95: cannam@95: Sun Jun 30 17:00:09 EDT 2002 athena cannam@95: * [project @ 2002-06-30 21:00:09 by athena] cannam@95: Added altivec support cannam@95: cannam@95: A ./simd/simd-altivec.h cannam@95: M ./dft/simd/codelets/Makefile.am -1 +1 cannam@95: M ./simd/Makefile.am -1 +1 cannam@95: M ./simd/simd-altivec.h +66 cannam@95: M ./simd/simd-sse.h -1 cannam@95: cannam@95: Sun Jun 30 14:47:47 EDT 2002 athena cannam@95: * [project @ 2002-06-30 18:47:47 by athena] cannam@95: Forgot file cannam@95: cannam@95: A ./dft/simd/t4.c cannam@95: cannam@95: Sun Jun 30 14:37:55 EDT 2002 athena cannam@95: * [project @ 2002-06-30 18:37:55 by athena] cannam@95: Progress towards simd implementation cannam@95: cannam@95: A ./dft/codelets/f.h cannam@95: A ./dft/codelets/n.c cannam@95: A ./dft/codelets/n.h cannam@95: A ./dft/codelets/q.h cannam@95: A ./dft/codelets/t.c cannam@95: A ./dft/codelets/t.h cannam@95: A ./dft/k7/k7.c cannam@95: A ./dft/simd/ cannam@95: A ./dft/simd/Makefile.am cannam@95: A ./dft/simd/NAMING cannam@95: A ./dft/simd/codelets/ cannam@95: A ./dft/simd/codelets/Makefile.am cannam@95: A ./dft/simd/n4.c cannam@95: A ./dft/simd/n4.h cannam@95: A ./dft/simd/t4.h cannam@95: A ./simd/ cannam@95: A ./simd/Makefile.am cannam@95: A ./simd/README cannam@95: A ./simd/simd-sse.h cannam@95: A ./simd/simd.h cannam@95: M ./Makefile.am -2 +5 cannam@95: M ./configure.ac -6 +32 cannam@95: M ./dft/Makefile.am -3 +3 cannam@95: R ./dft/codelet-k7.h cannam@95: M ./dft/codelet.h -11 +23 cannam@95: M ./dft/codelets/Makefile.am +4 cannam@95: M ./dft/codelets/f.h +1 cannam@95: M ./dft/codelets/inplace/Makefile.am -13 +14 cannam@95: M ./dft/codelets/n.c +34 cannam@95: M ./dft/codelets/n.h +27 cannam@95: M ./dft/codelets/q.h +1 cannam@95: M ./dft/codelets/standard/Makefile.am -13 +14 cannam@95: M ./dft/codelets/t.c +32 cannam@95: M ./dft/codelets/t.h +24 cannam@95: M ./dft/conf.c -2 +5 cannam@95: M ./dft/ct-dif.c -3 +3 cannam@95: M ./dft/ct-dit.c -5 +3 cannam@95: M ./dft/ct-ditbuf.c -2 +2 cannam@95: M ./dft/ct-ditf.c -4 +3 cannam@95: M ./dft/ct.c -6 +6 cannam@95: M ./dft/ct.h -3 +1 cannam@95: M ./dft/direct.c -6 +11 cannam@95: M ./dft/k7/Makefile.am -2 +1 cannam@95: M ./dft/k7/codelets/Makefile.am -33 +29 cannam@95: R ./dft/k7/ct-dif.c cannam@95: R ./dft/k7/ct-dit.c cannam@95: R ./dft/k7/ct-ditbuf.c cannam@95: R ./dft/k7/direct.c cannam@95: M ./dft/k7/k7.c +69 cannam@95: R ./dft/k7/kdft-dif.c cannam@95: R ./dft/k7/kdft-dit.c cannam@95: R ./dft/k7/kdft.c cannam@95: M ./dft/simd/Makefile.am +6 cannam@95: M ./dft/simd/NAMING +17 cannam@95: M ./dft/simd/codelets/Makefile.am +36 cannam@95: M ./dft/simd/n4.c +39 cannam@95: M ./dft/simd/n4.h +28 cannam@95: M ./dft/simd/t4.h +26 cannam@95: M ./genfft/annotate.ml -7 +7 cannam@95: M ./genfft/c.ml -1 +6 cannam@95: M ./genfft/c.mli -1 +2 cannam@95: M ./genfft/gen_notw.ml -12 +20 cannam@95: M ./genfft/gen_twiddle.ml -14 +18 cannam@95: M ./genfft/gen_twidsq.ml -7 +7 cannam@95: M ./genfft/genutil.ml -4 +17 cannam@95: M ./genfft/magic.ml -1 +3 cannam@95: M ./genfft/simd.ml -66 +34 cannam@95: M ./genfft/simd.mli -1 +7 cannam@95: M ./genfft/simdmagic.ml -2 +12 cannam@95: M ./genfft/twiddle.ml -1 +12 cannam@95: M ./genfft/twiddle.mli -1 +2 cannam@95: M ./genfft-k7/genUtil.ml -1 +1 cannam@95: M ./genfft-k7/gen_notw.ml -11 +18 cannam@95: M ./genfft-k7/gen_twiddle.ml -12 +17 cannam@95: M ./kernel/alloc.c -9 +28 cannam@95: M ./kernel/ifftw.h -10 +10 cannam@95: M ./libbench/bench-user.h -1 +2 cannam@95: M ./libbench/problem.c -1 +3 cannam@95: M ./libbench/util.c -1 cannam@95: M ./simd/Makefile.am +1 cannam@95: M ./simd/README +2 cannam@95: M ./simd/simd-sse.h +94 cannam@95: M ./simd/simd.h +33 cannam@95: M ./support/Makefile.codelets -33 +36 cannam@95: M ./tests/Makefile.am +1 cannam@95: M ./tests/bench.c -14 +63 cannam@95: cannam@95: Tue Jun 25 20:23:29 EDT 2002 athena cannam@95: * [project @ 2002-06-26 00:23:29 by athena] cannam@95: Add 128- codelet cannam@95: cannam@95: M ./dft/k7/codelets/Makefile.am -2 +3 cannam@95: cannam@95: Sat Jun 22 20:47:28 EDT 2002 athena cannam@95: * [project @ 2002-06-23 00:47:28 by athena] cannam@95: More simd changes. Ensure proper stack alignment in k7 codelets. cannam@95: cannam@95: M ./configure.ac +1 cannam@95: M ./genfft/c.ml -25 +4 cannam@95: M ./genfft/c.mli -2 +4 cannam@95: M ./genfft/expr.ml -1 +18 cannam@95: M ./genfft/expr.mli -1 +4 cannam@95: M ./genfft/gen_hc2r.ml -3 +3 cannam@95: M ./genfft/gen_notw.ml -3 +3 cannam@95: M ./genfft/gen_r2hc.ml -3 +3 cannam@95: M ./genfft/gen_trig.ml -3 +3 cannam@95: M ./genfft/gen_twiddle.ml -3 +3 cannam@95: M ./genfft/gen_twidsq.ml -3 +3 cannam@95: M ./genfft/genutil.ml -1 +13 cannam@95: M ./genfft/simd.ml -1 +9 cannam@95: M ./genfft/simd.mli -1 +2 cannam@95: M ./genfft-k7/genUtil.ml -34 +6 cannam@95: M ./genfft-k7/gen_notw.ml -2 +2 cannam@95: M ./genfft-k7/gen_twiddle.ml -2 +2 cannam@95: M ./libbench/bench-main.c -1 +11 cannam@95: cannam@95: Sat Jun 22 13:01:33 EDT 2002 athena cannam@95: * [project @ 2002-06-22 17:01:33 by athena] cannam@95: Fixed prototypes cannam@95: cannam@95: M ./kernel/ifftw.h -2 +2 cannam@95: M ./kernel/solvtab.c -2 +2 cannam@95: cannam@95: Sat Jun 22 12:53:26 EDT 2002 athena cannam@95: * [project @ 2002-06-22 16:53:26 by athena] cannam@95: Sparc cycle counter requires v9 cannam@95: cannam@95: M ./kernel/cycle.h -2 +2 cannam@95: cannam@95: Sat Jun 22 11:45:48 EDT 2002 athena cannam@95: * [project @ 2002-06-22 15:45:48 by athena] cannam@95: Minor fixes cannam@95: cannam@95: M ./configure.ac -1 +1 cannam@95: M ./kernel/cycle.h -15 +36 cannam@95: M ./kernel/ifftw.h -1 +5 cannam@95: cannam@95: Sat Jun 22 11:11:46 EDT 2002 athena cannam@95: * [project @ 2002-06-22 15:11:46 by athena] cannam@95: Fixed ev67 detection cannam@95: cannam@95: M ./acinclude.m4 -1 +1 cannam@95: cannam@95: Sat Jun 22 10:19:11 EDT 2002 athena cannam@95: * [project @ 2002-06-22 14:19:11 by athena] cannam@95: Print flops cannam@95: cannam@95: M ./tests/bench.c +2 cannam@95: cannam@95: Sat Jun 22 09:36:47 EDT 2002 athena cannam@95: * [project @ 2002-06-22 13:36:47 by athena] cannam@95: Nothing really cannam@95: cannam@95: M ./genfft/simd.ml -7 +3 cannam@95: cannam@95: Fri Jun 21 22:19:20 EDT 2002 athena cannam@95: * [project @ 2002-06-22 02:19:20 by athena] cannam@95: More simd work cannam@95: cannam@95: M ./dft/codelet-k7.h -2 +2 cannam@95: M ./dft/codelet.h -3 +3 cannam@95: M ./genfft/Makefile.am -2 +2 cannam@95: M ./genfft/c.ml -19 +1 cannam@95: M ./genfft/c.mli -2 +2 cannam@95: M ./genfft/gen_conv.ml -3 +3 cannam@95: M ./genfft/gen_hc2r.ml -3 +3 cannam@95: M ./genfft/gen_notw.ml -3 +3 cannam@95: M ./genfft/gen_r2hc.ml -3 +3 cannam@95: M ./genfft/gen_trig.ml -3 +3 cannam@95: M ./genfft/gen_twiddle.ml -3 +3 cannam@95: M ./genfft/gen_twidsq.ml -3 +3 cannam@95: M ./genfft/genutil.ml -2 +20 cannam@95: M ./genfft/magic.ml -1 +5 cannam@95: M ./genfft/simd.ml -47 +39 cannam@95: M ./genfft/simd.mli -1 +3 cannam@95: M ./genfft/simdmagic.ml -4 +2 cannam@95: M ./genfft/to_alist.ml -2 +5 cannam@95: M ./genfft-k7/magic.ml +4 cannam@95: M ./genfft-k7/to_alist.ml -2 +5 cannam@95: M ./genfft-k7/to_alist.mli -1 +1 cannam@95: M ./kernel/ifftw.h -2 +2 cannam@95: cannam@95: Thu Jun 20 21:22:41 EDT 2002 athena cannam@95: * [project @ 2002-06-21 01:22:41 by athena] cannam@95: More simd work cannam@95: cannam@95: A ./genfft/simd.mli cannam@95: A ./genfft/simdmagic.ml cannam@95: M ./genfft/Makefile.am -7 +8 cannam@95: M ./genfft/annotate.ml -13 +13 cannam@95: M ./genfft/annotate.mli -1 +2 cannam@95: M ./genfft/magic.ml -21 +1 cannam@95: M ./genfft/simd.ml -21 +10 cannam@95: M ./genfft/simd.mli +21 cannam@95: M ./genfft/simdmagic.ml +43 cannam@95: cannam@95: Thu Jun 20 18:51:33 EDT 2002 athena cannam@95: * [project @ 2002-06-20 22:51:33 by athena] cannam@95: More simd work cannam@95: cannam@95: A ./genfft/simd.ml cannam@95: M ./genfft/Makefile.am -23 +7 cannam@95: M ./genfft/annotate.ml -18 +2 cannam@95: M ./genfft/c.ml -6 +1 cannam@95: M ./genfft/c.mli -5 +8 cannam@95: M ./genfft/magic.ml -1 +6 cannam@95: M ./genfft/simd.ml +288 cannam@95: M ./genfft/variable.ml -1 +18 cannam@95: M ./genfft/variable.mli -1 +4 cannam@95: cannam@95: Thu Jun 20 15:04:37 EDT 2002 athena cannam@95: * [project @ 2002-06-20 19:04:37 by athena] cannam@95: Moving towards incorporation of simd stuff cannam@95: cannam@95: M ./genfft/annotate.ml -11 +144 cannam@95: M ./genfft/annotate.mli -2 +16 cannam@95: M ./genfft/gen_athnotw.ml -5 +5 cannam@95: M ./genfft/gen_athtw.ml -6 +6 cannam@95: M ./genfft/gen_conv.ml -3 +3 cannam@95: M ./genfft/gen_hc2r.ml -3 +3 cannam@95: M ./genfft/gen_notw.ml -3 +3 cannam@95: M ./genfft/gen_r2hc.ml -3 +3 cannam@95: M ./genfft/gen_trig.ml -3 +3 cannam@95: M ./genfft/gen_twiddle.ml -4 +4 cannam@95: M ./genfft/gen_twidsq.ml -6 +5 cannam@95: M ./genfft/genutil.ml -26 +17 cannam@95: M ./genfft/magic.ml -2 +17 cannam@95: M ./genfft/twiddle.ml -4 +4 cannam@95: M ./genfft/twiddle.mli -2 +2 cannam@95: M ./genfft/variable.ml -12 +22 cannam@95: M ./genfft/variable.mli -3 +6 cannam@95: cannam@95: Wed Jun 19 18:47:55 EDT 2002 athena cannam@95: * [project @ 2002-06-19 22:47:55 by athena] cannam@95: Reorganized k7 stuff into own directory cannam@95: cannam@95: A ./dft/k7/ cannam@95: A ./dft/k7/Makefile.am cannam@95: A ./dft/k7/codelets/ cannam@95: A ./dft/k7/codelets/Makefile.am cannam@95: A ./dft/k7/ct-dif.c cannam@95: A ./dft/k7/ct-dit.c cannam@95: A ./dft/k7/ct-ditbuf.c cannam@95: A ./dft/k7/direct.c cannam@95: A ./dft/k7/kdft-dif.c cannam@95: A ./dft/k7/kdft-dit.c cannam@95: A ./dft/k7/kdft.c cannam@95: M ./Makefile.am -1 +2 cannam@95: M ./configure.ac -2 +3 cannam@95: M ./dft/Makefile.am -6 +5 cannam@95: M ./dft/codelets/Makefile.am -1 +1 cannam@95: R ./dft/ct-dif-k7.c cannam@95: R ./dft/ct-dit-k7.c cannam@95: R ./dft/ct-ditbuf-k7.c cannam@95: R ./dft/direct-k7.c cannam@95: M ./dft/k7/Makefile.am +7 cannam@95: M ./dft/k7/codelets/Makefile.am +74 cannam@95: M ./dft/k7/ct-dif.c +129 cannam@95: M ./dft/k7/ct-dit.c +137 cannam@95: M ./dft/k7/ct-ditbuf.c +176 cannam@95: M ./dft/k7/direct.c +172 cannam@95: M ./dft/k7/kdft-dif.c +30 cannam@95: M ./dft/k7/kdft-dit.c +31 cannam@95: M ./dft/k7/kdft.c +31 cannam@95: R ./dft/kdft-dif-k7.c cannam@95: R ./dft/kdft-dit-k7.c cannam@95: R ./dft/kdft-k7.c cannam@95: cannam@95: Wed Jun 19 13:21:13 EDT 2002 athena cannam@95: * [project @ 2002-06-19 17:20:37 by athena] cannam@95: Minor experimental stuff cannam@95: cannam@95: M ./genfft/expr.ml -3 +2 cannam@95: M ./genfft/expr.mli -2 +2 cannam@95: M ./genfft/genutil.ml -2 +11 cannam@95: M ./genfft/magic.ml -1 +3 cannam@95: M ./genfft-k7/expr.ml -1 +3 cannam@95: M ./genfft-k7/expr.mli -1 +3 cannam@95: cannam@95: Wed Jun 19 11:20:29 EDT 2002 athena cannam@95: * [project @ 2002-06-19 15:20:29 by athena] cannam@95: Cosmetic changes cannam@95: cannam@95: M ./genfft/expr.ml -1 +4 cannam@95: M ./genfft/expr.mli -1 +3 cannam@95: M ./genfft/genutil.ml -9 +2 cannam@95: cannam@95: Wed Jun 19 01:43:31 EDT 2002 fftw cannam@95: * [project @ 2002-06-19 05:43:31 by fftw] cannam@95: allocate buffers on the fly cannam@95: cannam@95: M ./dft/buffered.c -18 +7 cannam@95: M ./dft/rader.c -17 +20 cannam@95: cannam@95: Tue Jun 18 17:48:41 EDT 2002 athena cannam@95: * [project @ 2002-06-18 21:48:41 by athena] cannam@95: Added ct-ditbuf-k7.c . Major changes required in generator. cannam@95: cannam@95: A ./dft/ct-ditbuf-k7.c cannam@95: M ./dft/Makefile.am -1 +1 cannam@95: M ./dft/codelet-k7.h -1 +2 cannam@95: M ./dft/ct-dif-k7.c -17 +3 cannam@95: M ./dft/ct-dif.c -16 +2 cannam@95: M ./dft/ct-dit-k7.c -18 +10 cannam@95: M ./dft/ct-dit.c -16 +2 cannam@95: M ./dft/ct-ditbuf-k7.c +176 cannam@95: M ./dft/ct-ditbuf.c -16 +2 cannam@95: M ./dft/ct.c -1 +30 cannam@95: M ./dft/ct.h -1 +3 cannam@95: M ./dft/kdft-dit-k7.c -1 +2 cannam@95: M ./dft/rader.c -1 +1 cannam@95: M ./genfft/number.ml -2 +2 cannam@95: M ./genfft/to_alist.ml -2 +2 cannam@95: M ./genfft-k7/Makefile.am -4 +4 cannam@95: M ./genfft-k7/assignmentsToVfpinstrs.ml -1 +4 cannam@95: M ./genfft-k7/gen_twiddle.ml -2 +5 cannam@95: M ./genfft-k7/k7Basics.ml -1 +1 cannam@95: M ./genfft-k7/k7Basics.mli +1 cannam@95: M ./genfft-k7/k7RegisterAllocationBasics.ml +3 cannam@95: M ./genfft-k7/k7RegisterAllocationBasics.mli -2 +4 cannam@95: M ./genfft-k7/k7RegisterAllocator.ml -15 +28 cannam@95: M ./genfft-k7/k7RegisterAllocatorInit.ml -6 +12 cannam@95: M ./genfft-k7/number.ml -2 +2 cannam@95: M ./genfft-k7/to_alist.ml -2 +2 cannam@95: M ./kernel/ifftw.h -2 +2 cannam@95: M ./kernel/planner.c -3 +5 cannam@95: M ./kernel/primes.c -2 +2 cannam@95: cannam@95: Tue Jun 18 11:55:57 EDT 2002 athena cannam@95: * [project @ 2002-06-18 15:55:57 by athena] cannam@95: Nothing, really cannam@95: cannam@95: M ./genfft-k7/gen_twiddle.ml -2 +2 cannam@95: M ./kernel/ifftw.h -18 +19 cannam@95: cannam@95: Tue Jun 18 11:19:59 EDT 2002 athena cannam@95: * [project @ 2002-06-18 15:19:59 by athena] cannam@95: !SINGLE ==> !K7_MODE (for some reason the contrapositive sounds wrong) cannam@95: cannam@95: M ./configure.ac +1 cannam@95: cannam@95: Tue Jun 18 11:07:13 EDT 2002 athena cannam@95: * [project @ 2002-06-18 15:07:13 by athena] cannam@95: Buffer is now symmetric wrt forward/backward transform cannam@95: cannam@95: M ./dft/buffered.c -12 +22 cannam@95: cannam@95: Tue Jun 18 10:33:58 EDT 2002 athena cannam@95: * [project @ 2002-06-18 14:33:58 by athena] cannam@95: Fixed applicable() in indirect.c cannam@95: cannam@95: A ./kernel/debug.c cannam@95: M ./dft/ct-dif.c -2 +1 cannam@95: M ./dft/indirect.c -3 +3 cannam@95: M ./dft/vrank2-transpose.c -10 +14 cannam@95: M ./dft/vrank3-transpose.c -10 +13 cannam@95: M ./kernel/Makefile.am -1 +1 cannam@95: M ./kernel/debug.c +39 cannam@95: M ./kernel/ifftw.h -1 +5 cannam@95: M ./kernel/print.c -4 +10 cannam@95: cannam@95: Tue Jun 18 08:41:18 EDT 2002 athena cannam@95: * [project @ 2002-06-18 12:41:18 by athena] cannam@95: Fixed attempt to free() uninitialized pointer. cannam@95: cannam@95: M ./dft/rader.c -2 +3 cannam@95: cannam@95: Tue Jun 18 08:13:55 EDT 2002 athena cannam@95: * [project @ 2002-06-18 12:13:55 by athena] cannam@95: Added reference counts for awake() cannam@95: cannam@95: M ./CONVENTIONS -1 +1 cannam@95: M ./TODO +5 cannam@95: M ./dft/rader.c -11 +3 cannam@95: M ./kernel/ifftw.h -4 +7 cannam@95: M ./kernel/plan.c -2 +15 cannam@95: M ./kernel/planner.c -3 +3 cannam@95: M ./tests/bench.c -2 +3 cannam@95: cannam@95: Mon Jun 17 20:49:05 EDT 2002 stevenj cannam@95: * [project @ 2002-06-18 00:49:05 by stevenj] cannam@95: updated comment cannam@95: cannam@95: M ./dft/rader.c -2 +2 cannam@95: cannam@95: Mon Jun 17 20:46:45 EDT 2002 stevenj cannam@95: * [project @ 2002-06-18 00:46:45 by stevenj] cannam@95: slight update cannam@95: cannam@95: M ./TODO -1 +3 cannam@95: cannam@95: Mon Jun 17 02:30:16 EDT 2002 fftw cannam@95: * [project @ 2002-06-17 06:30:16 by fftw] cannam@95: moved prime-number stuff into primes.c, so it can be shared with generic codelet and with rfftw rader cannam@95: cannam@95: A ./kernel/primes.c cannam@95: M ./dft/rader.c -104 +5 cannam@95: M ./kernel/Makefile.am -3 +3 cannam@95: M ./kernel/ifftw.h -1 +24 cannam@95: M ./kernel/primes.c +112 cannam@95: cannam@95: Mon Jun 17 02:01:58 EDT 2002 fftw cannam@95: * [project @ 2002-06-17 06:01:58 by fftw] cannam@95: added comment cannam@95: cannam@95: M ./dft/rader.c -6 +9 cannam@95: cannam@95: Mon Jun 17 01:39:55 EDT 2002 fftw cannam@95: * [project @ 2002-06-17 05:39:55 by fftw] cannam@95: added rader-dit cannam@95: cannam@95: M ./dft/rader.c -66 +308 cannam@95: M ./kernel/ifftw.h -1 +10 cannam@95: M ./kernel/twiddle.c -8 +1 cannam@95: cannam@95: Sun Jun 16 23:50:16 EDT 2002 fftw cannam@95: * [project @ 2002-06-17 03:50:16 by fftw] cannam@95: added initial Rader (no DIT yet) cannam@95: cannam@95: A ./dft/rader.c cannam@95: M ./configure.ac +8 cannam@95: M ./dft/Makefile.am -1 +1 cannam@95: M ./dft/conf.c -1 +2 cannam@95: M ./dft/dft.h -1 +2 cannam@95: M ./dft/rader.c +457 cannam@95: cannam@95: Sun Jun 16 22:29:51 EDT 2002 fftw cannam@95: * [project @ 2002-06-17 02:29:51 by fftw] cannam@95: don't warn about long long cannam@95: cannam@95: M ./acinclude.m4 -1 +1 cannam@95: cannam@95: Sun Jun 16 21:30:42 EDT 2002 athena cannam@95: * [project @ 2002-06-17 01:30:42 by athena] cannam@95: Added k7 DIF codelets cannam@95: cannam@95: A ./dft/ct-dif-k7.c cannam@95: A ./dft/kdft-dif-k7.c cannam@95: M ./dft/Makefile.am -5 +6 cannam@95: M ./dft/codelet-k7.h -5 +10 cannam@95: M ./dft/ct-dif-k7.c +143 cannam@95: M ./dft/ct.h -1 +2 cannam@95: M ./dft/kdft-dif-k7.c +30 cannam@95: M ./genfft-k7/gen_twiddle.ml -4 +25 cannam@95: M ./kernel/planner.c -3 +2 cannam@95: M ./tests/bench.c +1 cannam@95: cannam@95: Sun Jun 16 19:13:31 EDT 2002 athena cannam@95: * [project @ 2002-06-16 23:13:31 by athena] cannam@95: Added stuff to do cannam@95: cannam@95: M ./TODO +6 cannam@95: cannam@95: Sun Jun 16 19:05:58 EDT 2002 athena cannam@95: * [project @ 2002-06-16 23:05:58 by athena] cannam@95: Handle dual case R = I + 1 cannam@95: cannam@95: M ./dft/rank0.c -6 +30 cannam@95: cannam@95: Sun Jun 16 18:54:31 EDT 2002 athena cannam@95: * [project @ 2002-06-16 22:54:31 by athena] cannam@95: Removed useless flag cannam@95: cannam@95: M ./bootstrap.sh -1 +1 cannam@95: cannam@95: Sun Jun 16 18:30:32 EDT 2002 athena cannam@95: * [project @ 2002-06-16 22:30:32 by athena] cannam@95: Removed useless file cannam@95: cannam@95: R ./mkdist.sh cannam@95: cannam@95: Sun Jun 16 18:30:18 EDT 2002 athena cannam@95: * [project @ 2002-06-16 22:30:18 by athena] cannam@95: More k7 work. Switched to runtime CLASSIC mode. cannam@95: cannam@95: A ./dft/ct-dit-k7.c cannam@95: A ./dft/kdft-dit-k7.c cannam@95: R ./CLASSIC-MODE cannam@95: M ./Makefile.am -10 +1 cannam@95: M ./configure.ac -21 cannam@95: M ./dft/Makefile.am -5 +5 cannam@95: M ./dft/buffered.c -2 +4 cannam@95: M ./dft/codelet-k7.h -5 +7 cannam@95: M ./dft/codelet.h -1 +2 cannam@95: M ./dft/codelets/inplace/Makefile.am -2 cannam@95: M ./dft/ct-dif.c -3 +4 cannam@95: M ./dft/ct-dit-k7.c +145 cannam@95: M ./dft/ct-dit.c -6 +9 cannam@95: M ./dft/ct-ditbuf.c -9 +11 cannam@95: M ./dft/ct-ditf.c -2 +3 cannam@95: M ./dft/ct.h -5 +3 cannam@95: M ./dft/direct-k7.c -2 +3 cannam@95: M ./dft/direct.c -2 +3 cannam@95: M ./dft/indirect.c -2 +3 cannam@95: M ./dft/kdft-dit-k7.c +30 cannam@95: M ./dft/kdft-dit.c -4 +2 cannam@95: M ./dft/nop.c -2 +3 cannam@95: M ./dft/rank-geq2.c -8 +8 cannam@95: M ./dft/rank0.c -2 +3 cannam@95: M ./dft/vrank-geq1.c -6 +7 cannam@95: M ./dft/vrank2-transpose.c -3 +3 cannam@95: M ./dft/vrank3-transpose.c -3 +3 cannam@95: M ./genfft/gen_twiddle.ml -4 +4 cannam@95: M ./genfft/gen_twidsq.ml -4 +4 cannam@95: M ./genfft/twiddle.ml -7 +28 cannam@95: M ./genfft/twiddle.mli -2 +5 cannam@95: M ./genfft-k7/Makefile.am -25 +24 cannam@95: M ./genfft-k7/genUtil.ml -2 +3 cannam@95: M ./genfft-k7/gen_notw.ml -1 +2 cannam@95: M ./genfft-k7/gen_twiddle.ml -23 +29 cannam@95: M ./genfft-k7/twiddle.ml -7 +36 cannam@95: M ./genfft-k7/twiddle.mli -3 +7 cannam@95: M ./kernel/ifftw.h -15 +13 cannam@95: M ./kernel/planner-naive.c -3 +3 cannam@95: M ./kernel/planner-score.c -5 +6 cannam@95: M ./kernel/planner.c -6 +5 cannam@95: M ./kernel/version.c -7 +2 cannam@95: M ./support/Makefile.codelets +1 cannam@95: M ./tests/bench.c -12 +2 cannam@95: cannam@95: Sun Jun 16 17:15:18 EDT 2002 stevenj cannam@95: * [project @ 2002-06-16 21:15:18 by stevenj] cannam@95: spelling cannam@95: cannam@95: M ./kernel/tensor.c -2 +2 cannam@95: cannam@95: Sun Jun 16 15:51:44 EDT 2002 athena cannam@95: * [project @ 2002-06-16 19:51:44 by athena] cannam@95: Do not compile if not K7_MODE cannam@95: cannam@95: M ./dft/kdft-k7.c -1 +3 cannam@95: cannam@95: Sun Jun 16 15:35:02 EDT 2002 athena cannam@95: * [project @ 2002-06-16 19:35:02 by athena] cannam@95: Do not require K7 definitions to compile cannam@95: cannam@95: M ./dft/codelet-k7.h -1 +2 cannam@95: M ./dft/dft.h -2 +1 cannam@95: cannam@95: Sun Jun 16 08:05:17 EDT 2002 athena cannam@95: * [project @ 2002-06-16 12:05:17 by athena] cannam@95: More k7 stuff cannam@95: cannam@95: A ./dft/codelet-k7.h cannam@95: A ./genfft-k7/gen_twiddle.ml cannam@95: M ./dft/Makefile.am -4 +4 cannam@95: M ./dft/codelet-k7.h +40 cannam@95: M ./dft/codelet.h -20 +5 cannam@95: M ./dft/direct-k7.c -1 +5 cannam@95: M ./genfft-k7/Makefile.am -1 +9 cannam@95: M ./genfft-k7/complex.ml -1 +2 cannam@95: M ./genfft-k7/complex.mli -1 +2 cannam@95: M ./genfft-k7/genUtil.ml -1 +12 cannam@95: M ./genfft-k7/gen_notw.ml -5 +1 cannam@95: M ./genfft-k7/gen_twiddle.ml +149 cannam@95: cannam@95: Sat Jun 15 18:30:43 EDT 2002 athena cannam@95: * [project @ 2002-06-15 22:30:43 by athena] cannam@95: Try to be compatible with automake-1.6 cannam@95: cannam@95: M ./acinclude.m4 -1 +5 cannam@95: cannam@95: Sat Jun 15 18:23:40 EDT 2002 athena cannam@95: * [project @ 2002-06-15 22:23:40 by athena] cannam@95: More merging of Stefan's generator with main genfft branch cannam@95: cannam@95: A ./genfft-k7/algsimp.ml cannam@95: A ./genfft-k7/algsimp.mli cannam@95: A ./genfft-k7/assoctable.ml cannam@95: A ./genfft-k7/assoctable.mli cannam@95: A ./genfft-k7/littlesimp.ml cannam@95: A ./genfft-k7/littlesimp.mli cannam@95: A ./genfft-k7/monads.ml cannam@95: A ./genfft-k7/oracle.ml cannam@95: A ./genfft-k7/oracle.mli cannam@95: A ./genfft-k7/to_alist.ml cannam@95: A ./genfft-k7/to_alist.mli cannam@95: A ./genfft-k7/twiddle.mli cannam@95: A ./support/codelet_asmprelude cannam@95: M ./acinclude.m4 +8 cannam@95: M ./configure.ac +1 cannam@95: M ./genfft/number.ml -7 +8 cannam@95: M ./genfft-k7/Makefile.am -23 +27 cannam@95: M ./genfft-k7/algsimp.ml +517 cannam@95: M ./genfft-k7/algsimp.mli +24 cannam@95: M ./genfft-k7/assignmentsToVfpinstrs.ml -11 +10 cannam@95: M ./genfft-k7/assoctable.ml +66 cannam@95: M ./genfft-k7/assoctable.mli +30 cannam@95: M ./genfft-k7/complex.ml -4 +19 cannam@95: M ./genfft-k7/complex.mli -5 +6 cannam@95: M ./genfft-k7/expr.ml -27 +58 cannam@95: M ./genfft-k7/expr.mli -4 +11 cannam@95: R ./genfft-k7/exprdag.ml cannam@95: R ./genfft-k7/exprdag.mli cannam@95: M ./genfft-k7/genUtil.ml -13 +18 cannam@95: M ./genfft-k7/gen_notw.ml -16 +31 cannam@95: M ./genfft-k7/k7Unparsing.ml -2 +2 cannam@95: M ./genfft-k7/littlesimp.ml +72 cannam@95: M ./genfft-k7/littlesimp.mli +26 cannam@95: M ./genfft-k7/magic.ml -16 +41 cannam@95: M ./genfft-k7/monads.ml +76 cannam@95: M ./genfft-k7/number.ml -9 +5 cannam@95: M ./genfft-k7/number.mli -8 +4 cannam@95: M ./genfft-k7/oracle.ml +138 cannam@95: M ./genfft-k7/oracle.mli +25 cannam@95: M ./genfft-k7/to_alist.ml +258 cannam@95: M ./genfft-k7/to_alist.mli +24 cannam@95: M ./genfft-k7/twiddle.ml -5 +22 cannam@95: M ./genfft-k7/twiddle.mli +29 cannam@95: M ./genfft-k7/util.ml -4 +4 cannam@95: M ./genfft-k7/util.mli -2 +1 cannam@95: M ./genfft-k7/vFpUnparsing.ml -1 +1 cannam@95: M ./genfft-k7/vSimdBasics.ml -2 +3 cannam@95: M ./genfft-k7/vSimdUnparsing.ml -1 +1 cannam@95: M ./genfft-k7/variable.ml +1 cannam@95: M ./genfft-k7/variable.mli +1 cannam@95: M ./support/Makefile.am -1 +1 cannam@95: M ./support/Makefile.codelets -1 +13 cannam@95: M ./support/codelet_asmprelude +8 cannam@95: cannam@95: Sat Jun 15 13:51:39 EDT 2002 athena cannam@95: * [project @ 2002-06-15 17:51:39 by athena] cannam@95: Slowly merging genfft-k7 with main genfft branch cannam@95: cannam@95: A ./genfft-k7/gen_notw.ml cannam@95: M ./genfft/expr.ml -1 +3 cannam@95: M ./genfft/expr.mli -1 +2 cannam@95: M ./genfft/genutil.ml -1 +9 cannam@95: M ./genfft-k7/Makefile.am -26 +23 cannam@95: M ./genfft-k7/complex.ml -34 +92 cannam@95: M ./genfft-k7/complex.mli -1 +11 cannam@95: M ./genfft-k7/expr.ml -1 +19 cannam@95: M ./genfft-k7/expr.mli -4 +3 cannam@95: M ./genfft-k7/exprdag.ml -73 +103 cannam@95: M ./genfft-k7/exprdag.mli -1 cannam@95: M ./genfft-k7/fft.ml -167 +83 cannam@95: M ./genfft-k7/fft.mli -29 +1 cannam@95: M ./genfft-k7/genUtil.ml -56 +33 cannam@95: R ./genfft-k7/gen_hc2hc.ml cannam@95: R ./genfft-k7/gen_hc2real.ml cannam@95: M ./genfft-k7/gen_notw.ml +143 cannam@95: R ./genfft-k7/gen_notwiddle.ml cannam@95: R ./genfft-k7/gen_notwiddle_fixedstride.ml cannam@95: R ./genfft-k7/gen_real2hc.ml cannam@95: R ./genfft-k7/gen_realeven.ml cannam@95: R ./genfft-k7/gen_realeven2.ml cannam@95: R ./genfft-k7/gen_realodd.ml cannam@95: R ./genfft-k7/gen_realodd2.ml cannam@95: R ./genfft-k7/gen_twiddle.ml cannam@95: M ./genfft-k7/magic.ml -11 +35 cannam@95: R ./genfft-k7/symmetry.ml cannam@95: M ./genfft-k7/twiddle.ml -63 +95 cannam@95: M ./genfft-k7/util.ml +15 cannam@95: M ./genfft-k7/util.mli +3 cannam@95: M ./genfft-k7/variable.ml +12 cannam@95: M ./genfft-k7/variable.mli +3 cannam@95: M ./support/Makefile.codelets -1 +1 cannam@95: cannam@95: Fri Jun 14 21:33:02 EDT 2002 athena cannam@95: * [project @ 2002-06-15 01:33:02 by athena] cannam@95: Fixed, really cannam@95: cannam@95: M ./genfft-k7/Makefile.am -1 +1 cannam@95: M ./genfft-k7/genUtil.ml -25 cannam@95: M ./genfft-k7/magic.ml -1 +1 cannam@95: R ./genfft-k7/magic.mli cannam@95: M ./genfft-k7/twiddle.ml -2 +2 cannam@95: M ./support/Makefile.codelets -4 +8 cannam@95: cannam@95: Fri Jun 14 21:27:12 EDT 2002 athena cannam@95: * [project @ 2002-06-15 01:27:12 by athena] cannam@95: Oops... cannam@95: cannam@95: M ./support/Makefile.codelets -1 +1 cannam@95: cannam@95: Fri Jun 14 21:25:34 EDT 2002 athena cannam@95: * [project @ 2002-06-15 01:25:34 by athena] cannam@95: Work properly when $(ALL_CODELETS) = "" cannam@95: cannam@95: M ./support/Makefile.codelets -4 +4 cannam@95: cannam@95: Fri Jun 14 21:11:16 EDT 2002 athena cannam@95: * [project @ 2002-06-15 01:11:16 by athena] cannam@95: Fixed k7 build machinery cannam@95: cannam@95: M ./Makefile.am -1 +2 cannam@95: M ./configure.ac +8 cannam@95: M ./dft/codelet.h -1 +4 cannam@95: M ./dft/codelets/Makefile.am -1 +1 cannam@95: M ./dft/conf.c -1 +4 cannam@95: M ./genfft-k7/gen_notwiddle.ml -1 +3 cannam@95: M ./kernel/ifftw.h -1 +10 cannam@95: M ./support/Makefile.codelets +2 cannam@95: cannam@95: Fri Jun 14 17:42:35 EDT 2002 athena cannam@95: * [project @ 2002-06-14 21:42:35 by athena] cannam@95: More work on k7 stuff cannam@95: cannam@95: M ./Makefile.am -1 +1 cannam@95: M ./configure.ac +1 cannam@95: M ./dft/codelet.h -1 +4 cannam@95: M ./dft/direct-k7.c -3 +3 cannam@95: M ./genfft-k7/Makefile.am -24 +24 cannam@95: R ./genfft-k7/codeletMisc.ml cannam@95: R ./genfft-k7/codeletMisc.mli cannam@95: M ./genfft-k7/genUtil.ml -10 +128 cannam@95: R ./genfft-k7/genUtil.mli cannam@95: R ./genfft-k7/gen_hc2hc.mli cannam@95: R ./genfft-k7/gen_hc2real.mli cannam@95: M ./genfft-k7/gen_notwiddle.ml -3 +33 cannam@95: R ./genfft-k7/gen_notwiddle.mli cannam@95: R ./genfft-k7/gen_real2hc.mli cannam@95: R ./genfft-k7/gen_realeven.mli cannam@95: R ./genfft-k7/gen_realeven2.mli cannam@95: R ./genfft-k7/gen_realodd.mli cannam@95: R ./genfft-k7/gen_realodd2.mli cannam@95: R ./genfft-k7/gen_twiddle.mli cannam@95: R ./genfft-k7/genfft.ml cannam@95: M ./genfft-k7/k7Basics.ml +19 cannam@95: M ./genfft-k7/k7Basics.mli +2 cannam@95: M ./genfft-k7/k7Unparsing.ml -1 +1 cannam@95: M ./genfft-k7/magic.ml -1 cannam@95: M ./genfft-k7/magic.mli -1 cannam@95: M ./kernel/ifftw.h -1 +4 cannam@95: M ./libbench/bench-user.h -1 +3 cannam@95: M ./support/Makefile.am -1 +2 cannam@95: cannam@95: Fri Jun 14 15:54:29 EDT 2002 athena cannam@95: * [project @ 2002-06-14 19:54:29 by athena] cannam@95: More work on k7 stuff cannam@95: cannam@95: A ./dft/direct-k7.c cannam@95: A ./dft/kdft-k7.c cannam@95: M ./dft/Makefile.am -1 +2 cannam@95: M ./dft/codelet.h -1 +13 cannam@95: M ./dft/conf.c -9 +15 cannam@95: M ./dft/dft.h -1 +2 cannam@95: M ./dft/direct-k7.c +167 cannam@95: M ./dft/kdft-k7.c +29 cannam@95: M ./genfft-k7/codeletMisc.ml -91 cannam@95: M ./genfft-k7/codeletMisc.mli -13 +1 cannam@95: M ./genfft-k7/genUtil.ml -24 +19 cannam@95: M ./genfft-k7/gen_notwiddle.ml -2 +44 cannam@95: M ./genfft-k7/genfft.ml +4 cannam@95: M ./genfft-k7/magic.ml +1 cannam@95: M ./genfft-k7/magic.mli +1 cannam@95: M ./support/Makefile.codelets -1 +1 cannam@95: cannam@95: Fri Jun 14 14:18:15 EDT 2002 athena cannam@95: * [project @ 2002-06-14 18:18:15 by athena] cannam@95: Changed my mind again cannam@95: cannam@95: M ./dft/codelet.h -4 +1 cannam@95: M ./dft/direct.c -7 +4 cannam@95: M ./genfft/gen_notw.ml -5 +4 cannam@95: cannam@95: Fri Jun 14 11:53:09 EDT 2002 athena cannam@95: * [project @ 2002-06-14 15:53:09 by athena] cannam@95: Removed some useless stuff. cannam@95: cannam@95: M ./genfft-k7/gen_notwiddle.ml -8 +3 cannam@95: cannam@95: Fri Jun 14 11:01:39 EDT 2002 athena cannam@95: * [project @ 2002-06-14 15:01:39 by athena] cannam@95: Hmm... cannam@95: cannam@95: M ./genfft-k7/gen_notwiddle.ml -3 +8 cannam@95: cannam@95: Fri Jun 14 10:28:12 EDT 2002 athena cannam@95: * [project @ 2002-06-14 14:28:12 by athena] cannam@95: More work in preparation for k7 stuff cannam@95: cannam@95: M ./dft/codelet.h -1 +2 cannam@95: M ./dft/direct.c -3 +4 cannam@95: M ./genfft/gen_notw.ml -4 +5 cannam@95: M ./genfft-k7/gen_notwiddle.ml -18 +13 cannam@95: cannam@95: Fri Jun 14 07:25:28 EDT 2002 athena cannam@95: * [project @ 2002-06-14 11:25:28 by athena] cannam@95: Still preparing to include k7 stuff cannam@95: cannam@95: M ./TODO +4 cannam@95: M ./dft/codelet.h -4 +6 cannam@95: M ./dft/direct.c -1 +5 cannam@95: M ./genfft/gen_notw.ml -7 +9 cannam@95: cannam@95: Fri Jun 14 07:06:02 EDT 2002 athena cannam@95: * [project @ 2002-06-14 11:06:02 by athena] cannam@95: Create .depend cannam@95: cannam@95: M ./bootstrap.sh -2 +2 cannam@95: cannam@95: Fri Jun 14 06:56:15 EDT 2002 athena cannam@95: * [project @ 2002-06-14 10:56:14 by athena] cannam@95: Imported Stefan's K7 generator cannam@95: cannam@95: A ./genfft-k7/ cannam@95: A ./genfft-k7/Makefile.am cannam@95: A ./genfft-k7/assignmentsToVfpinstrs.ml cannam@95: A ./genfft-k7/assignmentsToVfpinstrs.mli cannam@95: A ./genfft-k7/balanceVfpinstrs.ml cannam@95: A ./genfft-k7/balanceVfpinstrs.mli cannam@95: A ./genfft-k7/codeletMisc.ml cannam@95: A ./genfft-k7/codeletMisc.mli cannam@95: A ./genfft-k7/complex.ml cannam@95: A ./genfft-k7/complex.mli cannam@95: A ./genfft-k7/expr.ml cannam@95: A ./genfft-k7/expr.mli cannam@95: A ./genfft-k7/exprdag.ml cannam@95: A ./genfft-k7/exprdag.mli cannam@95: A ./genfft-k7/fft.ml cannam@95: A ./genfft-k7/fft.mli cannam@95: A ./genfft-k7/genUtil.ml cannam@95: A ./genfft-k7/genUtil.mli cannam@95: A ./genfft-k7/gen_hc2hc.ml cannam@95: A ./genfft-k7/gen_hc2hc.mli cannam@95: A ./genfft-k7/gen_hc2real.ml cannam@95: A ./genfft-k7/gen_hc2real.mli cannam@95: A ./genfft-k7/gen_notwiddle.ml cannam@95: A ./genfft-k7/gen_notwiddle.mli cannam@95: A ./genfft-k7/gen_notwiddle_fixedstride.ml cannam@95: A ./genfft-k7/gen_real2hc.ml cannam@95: A ./genfft-k7/gen_real2hc.mli cannam@95: A ./genfft-k7/gen_realeven.ml cannam@95: A ./genfft-k7/gen_realeven.mli cannam@95: A ./genfft-k7/gen_realeven2.ml cannam@95: A ./genfft-k7/gen_realeven2.mli cannam@95: A ./genfft-k7/gen_realodd.ml cannam@95: A ./genfft-k7/gen_realodd.mli cannam@95: A ./genfft-k7/gen_realodd2.ml cannam@95: A ./genfft-k7/gen_realodd2.mli cannam@95: A ./genfft-k7/gen_twiddle.ml cannam@95: A ./genfft-k7/gen_twiddle.mli cannam@95: A ./genfft-k7/genfft.ml cannam@95: A ./genfft-k7/id.ml cannam@95: A ./genfft-k7/id.mli cannam@95: A ./genfft-k7/k7Basics.ml cannam@95: A ./genfft-k7/k7Basics.mli cannam@95: A ./genfft-k7/k7ExecutionModel.ml cannam@95: A ./genfft-k7/k7ExecutionModel.mli cannam@95: A ./genfft-k7/k7FlatInstructionScheduling.ml cannam@95: A ./genfft-k7/k7FlatInstructionScheduling.mli cannam@95: A ./genfft-k7/k7InstructionSchedulingBasics.ml cannam@95: A ./genfft-k7/k7InstructionSchedulingBasics.mli cannam@95: A ./genfft-k7/k7RegisterAllocationBasics.ml cannam@95: A ./genfft-k7/k7RegisterAllocationBasics.mli cannam@95: A ./genfft-k7/k7RegisterAllocator.ml cannam@95: A ./genfft-k7/k7RegisterAllocator.mli cannam@95: A ./genfft-k7/k7RegisterAllocatorEATranslation.ml cannam@95: A ./genfft-k7/k7RegisterAllocatorEATranslation.mli cannam@95: A ./genfft-k7/k7RegisterAllocatorInit.ml cannam@95: A ./genfft-k7/k7RegisterAllocatorInit.mli cannam@95: A ./genfft-k7/k7RegisterReallocation.ml cannam@95: A ./genfft-k7/k7RegisterReallocation.mli cannam@95: A ./genfft-k7/k7Translate.ml cannam@95: A ./genfft-k7/k7Translate.mli cannam@95: A ./genfft-k7/k7Unparsing.ml cannam@95: A ./genfft-k7/k7Unparsing.mli cannam@95: A ./genfft-k7/k7Vectorization.ml cannam@95: A ./genfft-k7/k7Vectorization.mli cannam@95: A ./genfft-k7/magic.ml cannam@95: A ./genfft-k7/magic.mli cannam@95: A ./genfft-k7/memoMonad.ml cannam@95: A ./genfft-k7/memoMonad.mli cannam@95: A ./genfft-k7/nonDetMonad.ml cannam@95: A ./genfft-k7/nonDetMonad.mli cannam@95: A ./genfft-k7/nullVectorization.ml cannam@95: A ./genfft-k7/nullVectorization.mli cannam@95: A ./genfft-k7/number.ml cannam@95: A ./genfft-k7/number.mli cannam@95: A ./genfft-k7/stateMonad.ml cannam@95: A ./genfft-k7/stateMonad.mli cannam@95: A ./genfft-k7/symmetry.ml cannam@95: A ./genfft-k7/twiddle.ml cannam@95: A ./genfft-k7/util.ml cannam@95: A ./genfft-k7/util.mli cannam@95: A ./genfft-k7/vAnnotatedScheduler.ml cannam@95: A ./genfft-k7/vAnnotatedScheduler.mli cannam@95: A ./genfft-k7/vDag.ml cannam@95: A ./genfft-k7/vDag.mli cannam@95: A ./genfft-k7/vFpBasics.ml cannam@95: A ./genfft-k7/vFpBasics.mli cannam@95: A ./genfft-k7/vFpUnparsing.ml cannam@95: A ./genfft-k7/vFpUnparsing.mli cannam@95: A ./genfft-k7/vImproveSchedule.ml cannam@95: A ./genfft-k7/vImproveSchedule.mli cannam@95: A ./genfft-k7/vK7Optimization.ml cannam@95: A ./genfft-k7/vK7Optimization.mli cannam@95: A ./genfft-k7/vScheduler.ml cannam@95: A ./genfft-k7/vScheduler.mli cannam@95: A ./genfft-k7/vSimdBasics.ml cannam@95: A ./genfft-k7/vSimdBasics.mli cannam@95: A ./genfft-k7/vSimdIndexing.ml cannam@95: A ./genfft-k7/vSimdIndexing.mli cannam@95: A ./genfft-k7/vSimdUnparsing.ml cannam@95: A ./genfft-k7/vSimdUnparsing.mli cannam@95: A ./genfft-k7/variable.ml cannam@95: A ./genfft-k7/variable.mli cannam@95: M ./AUTHORS -2 +5 cannam@95: M ./Makefile.am -1 +1 cannam@95: M ./bootstrap.sh +2 cannam@95: M ./configure.ac +1 cannam@95: M ./genfft-k7/Makefile.am +108 cannam@95: M ./genfft-k7/assignmentsToVfpinstrs.ml +225 cannam@95: M ./genfft-k7/assignmentsToVfpinstrs.mli +38 cannam@95: M ./genfft-k7/balanceVfpinstrs.ml +195 cannam@95: M ./genfft-k7/balanceVfpinstrs.mli +29 cannam@95: M ./genfft-k7/codeletMisc.ml +127 cannam@95: M ./genfft-k7/codeletMisc.mli +45 cannam@95: M ./genfft-k7/complex.ml +142 cannam@95: M ./genfft-k7/complex.mli +44 cannam@95: M ./genfft-k7/expr.ml +46 cannam@95: M ./genfft-k7/expr.mli -1 +34 cannam@95: M ./genfft-k7/exprdag.ml +879 cannam@95: M ./genfft-k7/exprdag.mli +42 cannam@95: M ./genfft-k7/fft.ml +311 cannam@95: M ./genfft-k7/fft.mli +50 cannam@95: M ./genfft-k7/genUtil.ml +377 cannam@95: M ./genfft-k7/genUtil.mli +41 cannam@95: M ./genfft-k7/gen_hc2hc.ml +133 cannam@95: M ./genfft-k7/gen_hc2hc.mli +25 cannam@95: M ./genfft-k7/gen_hc2real.ml +77 cannam@95: M ./genfft-k7/gen_hc2real.mli +24 cannam@95: M ./genfft-k7/gen_notwiddle.ml +69 cannam@95: M ./genfft-k7/gen_notwiddle.mli +25 cannam@95: M ./genfft-k7/gen_notwiddle_fixedstride.ml +45 cannam@95: M ./genfft-k7/gen_real2hc.ml +72 cannam@95: M ./genfft-k7/gen_real2hc.mli +24 cannam@95: M ./genfft-k7/gen_realeven.ml +55 cannam@95: M ./genfft-k7/gen_realeven.mli +24 cannam@95: M ./genfft-k7/gen_realeven2.ml +57 cannam@95: M ./genfft-k7/gen_realeven2.mli +24 cannam@95: M ./genfft-k7/gen_realodd.ml +60 cannam@95: M ./genfft-k7/gen_realodd.mli +24 cannam@95: M ./genfft-k7/gen_realodd2.ml +60 cannam@95: M ./genfft-k7/gen_realodd2.mli +24 cannam@95: M ./genfft-k7/gen_twiddle.ml +98 cannam@95: M ./genfft-k7/gen_twiddle.mli +25 cannam@95: M ./genfft-k7/genfft.ml +282 cannam@95: M ./genfft-k7/id.ml +61 cannam@95: M ./genfft-k7/id.mli +79 cannam@95: M ./genfft-k7/k7Basics.ml +524 cannam@95: M ./genfft-k7/k7Basics.mli +219 cannam@95: M ./genfft-k7/k7ExecutionModel.ml +111 cannam@95: M ./genfft-k7/k7ExecutionModel.mli +24 cannam@95: M ./genfft-k7/k7FlatInstructionScheduling.ml +177 cannam@95: M ./genfft-k7/k7FlatInstructionScheduling.mli +22 cannam@95: M ./genfft-k7/k7InstructionSchedulingBasics.ml +205 cannam@95: M ./genfft-k7/k7InstructionSchedulingBasics.mli +74 cannam@95: M ./genfft-k7/k7RegisterAllocationBasics.ml +162 cannam@95: M ./genfft-k7/k7RegisterAllocationBasics.mli +88 cannam@95: M ./genfft-k7/k7RegisterAllocator.ml +516 cannam@95: M ./genfft-k7/k7RegisterAllocator.mli +23 cannam@95: M ./genfft-k7/k7RegisterAllocatorEATranslation.ml +134 cannam@95: M ./genfft-k7/k7RegisterAllocatorEATranslation.mli +27 cannam@95: M ./genfft-k7/k7RegisterAllocatorInit.ml +97 cannam@95: M ./genfft-k7/k7RegisterAllocatorInit.mli +30 cannam@95: M ./genfft-k7/k7RegisterReallocation.ml +259 cannam@95: M ./genfft-k7/k7RegisterReallocation.mli +22 cannam@95: M ./genfft-k7/k7Translate.ml +142 cannam@95: M ./genfft-k7/k7Translate.mli +87 cannam@95: M ./genfft-k7/k7Unparsing.ml +326 cannam@95: M ./genfft-k7/k7Unparsing.mli +53 cannam@95: M ./genfft-k7/k7Vectorization.ml +739 cannam@95: M ./genfft-k7/k7Vectorization.mli +25 cannam@95: M ./genfft-k7/magic.ml +80 cannam@95: M ./genfft-k7/magic.mli +55 cannam@95: M ./genfft-k7/memoMonad.ml +32 cannam@95: M ./genfft-k7/memoMonad.mli +24 cannam@95: M ./genfft-k7/nonDetMonad.ml +136 cannam@95: M ./genfft-k7/nonDetMonad.mli +94 cannam@95: M ./genfft-k7/nullVectorization.ml +78 cannam@95: M ./genfft-k7/nullVectorization.mli +24 cannam@95: M ./genfft-k7/number.ml +158 cannam@95: M ./genfft-k7/number.mli +53 cannam@95: M ./genfft-k7/stateMonad.ml +71 cannam@95: M ./genfft-k7/stateMonad.mli +41 cannam@95: M ./genfft-k7/symmetry.ml +314 cannam@95: M ./genfft-k7/twiddle.ml +116 cannam@95: M ./genfft-k7/util.ml -1 +322 cannam@95: M ./genfft-k7/util.mli +124 cannam@95: M ./genfft-k7/vAnnotatedScheduler.ml +182 cannam@95: M ./genfft-k7/vAnnotatedScheduler.mli +39 cannam@95: M ./genfft-k7/vDag.ml +123 cannam@95: M ./genfft-k7/vDag.mli +52 cannam@95: M ./genfft-k7/vFpBasics.ml +203 cannam@95: M ./genfft-k7/vFpBasics.mli +106 cannam@95: M ./genfft-k7/vFpUnparsing.ml +79 cannam@95: M ./genfft-k7/vFpUnparsing.mli +28 cannam@95: M ./genfft-k7/vImproveSchedule.ml +155 cannam@95: M ./genfft-k7/vImproveSchedule.mli +22 cannam@95: M ./genfft-k7/vK7Optimization.ml +626 cannam@95: M ./genfft-k7/vK7Optimization.mli +22 cannam@95: M ./genfft-k7/vScheduler.ml +166 cannam@95: M ./genfft-k7/vScheduler.mli +31 cannam@95: M ./genfft-k7/vSimdBasics.ml +296 cannam@95: M ./genfft-k7/vSimdBasics.mli +247 cannam@95: M ./genfft-k7/vSimdIndexing.ml +119 cannam@95: M ./genfft-k7/vSimdIndexing.mli +68 cannam@95: M ./genfft-k7/vSimdUnparsing.ml +104 cannam@95: M ./genfft-k7/vSimdUnparsing.mli +33 cannam@95: M ./genfft-k7/variable.ml -1 +119 cannam@95: M ./genfft-k7/variable.mli +53 cannam@95: cannam@95: Thu Jun 13 15:30:41 EDT 2002 athena cannam@95: * [project @ 2002-06-13 19:30:41 by athena] cannam@95: Generator for real->halfcomplex and halfcomplex->real codelets cannam@95: cannam@95: A ./genfft/gen_hc2r.ml cannam@95: A ./genfft/gen_r2hc.ml cannam@95: M ./genfft/Makefile.am -9 +27 cannam@95: M ./genfft/c.ml -2 +2 cannam@95: M ./genfft/complex.ml -1 +8 cannam@95: M ./genfft/complex.mli -1 +3 cannam@95: M ./genfft/gen_hc2r.ml +121 cannam@95: M ./genfft/gen_r2hc.ml +121 cannam@95: M ./genfft/genutil.ml -8 +9 cannam@95: M ./genfft/trig.ml -2 +2 cannam@95: cannam@95: Thu Jun 13 11:54:02 EDT 2002 athena cannam@95: * [project @ 2002-06-13 15:54:02 by athena] cannam@95: Improved hash functions, printers cannam@95: cannam@95: M ./dft/problem.c -5 +10 cannam@95: M ./kernel/planner.c -11 +5 cannam@95: M ./kernel/tensor.c -7 +6 cannam@95: M ./tests/bench.c -1 +1 cannam@95: cannam@95: Thu Jun 13 11:17:31 EDT 2002 athena cannam@95: * [project @ 2002-06-13 15:17:31 by athena] cannam@95: Only regenerate codlist.c in maintainer mode cannam@95: cannam@95: M ./support/Makefile.codelets -18 +18 cannam@95: cannam@95: Thu Jun 13 11:04:24 EDT 2002 athena cannam@95: * [project @ 2002-06-13 15:04:24 by athena] cannam@95: Planner can export solution list cannam@95: cannam@95: M ./dft/problem.c -10 +21 cannam@95: M ./dft/rank-geq2.c -2 +2 cannam@95: M ./kernel/ifftw.h -11 +13 cannam@95: M ./kernel/planner-naive.c -7 +7 cannam@95: M ./kernel/planner-score.c -8 +7 cannam@95: M ./kernel/planner.c -42 +47 cannam@95: M ./kernel/print.c -9 +27 cannam@95: M ./kernel/tensor.c -4 +4 cannam@95: M ./tests/bench.c -5 +6 cannam@95: cannam@95: Thu Jun 13 08:59:53 EDT 2002 athena cannam@95: * [project @ 2002-06-13 12:59:53 by athena] cannam@95: Fixed for intel compiler cannam@95: cannam@95: M ./dft/ct-ditbuf.c -2 +3 cannam@95: M ./dft/dft.h -2 +1 cannam@95: M ./dft/direct.c -3 +1 cannam@95: M ./kernel/cycle.h -2 +2 cannam@95: M ./libbench/bench-user.h -12 +1 cannam@95: cannam@95: Thu Jun 13 08:48:51 EDT 2002 athena cannam@95: * [project @ 2002-06-13 12:48:51 by athena] cannam@95: Revised strategy for constants in codelets cannam@95: cannam@95: M ./dft/codelet.h -13 +2 cannam@95: M ./genfft/c.ml -50 +43 cannam@95: M ./genfft/c.mli -1 +4 cannam@95: M ./genfft/gen_notw.ml -5 +5 cannam@95: M ./genfft/gen_trig.ml -3 +3 cannam@95: M ./genfft/gen_twiddle.ml -3 +3 cannam@95: M ./genfft/gen_twidsq.ml -3 +3 cannam@95: M ./genfft/magic.ml -5 +1 cannam@95: cannam@95: Thu Jun 13 06:21:31 EDT 2002 athena cannam@95: * [project @ 2002-06-13 10:21:31 by athena] cannam@95: Enable score planner in classic mode, naive planner in pro mode. cannam@95: cannam@95: M ./tests/bench.c -3 +3 cannam@95: cannam@95: Wed Jun 12 19:18:18 EDT 2002 athena cannam@95: * [project @ 2002-06-12 23:18:18 by athena] cannam@95: Report classic/pro cannam@95: cannam@95: M ./tests/bench.c -2 +7 cannam@95: cannam@95: Wed Jun 12 19:07:48 EDT 2002 athena cannam@95: * [project @ 2002-06-12 23:07:48 by athena] cannam@95: Fixed behavior of buffered solver for large buffers. cannam@95: cannam@95: M ./dft/buffered.c -5 +5 cannam@95: M ./tests/bench.c -1 +1 cannam@95: cannam@95: Wed Jun 12 18:57:19 EDT 2002 athena cannam@95: * [project @ 2002-06-12 22:57:19 by athena] cannam@95: Make assumption COST(vector) = length * COST(scalar) in classic mode. cannam@95: cannam@95: M ./dft/rank-geq2.c -2 +2 cannam@95: M ./dft/vrank-geq1.c -1 +2 cannam@95: M ./kernel/ifftw.h -4 +7 cannam@95: M ./kernel/planner-naive.c -4 +2 cannam@95: M ./kernel/planner-score.c -5 +3 cannam@95: M ./kernel/planner.c -13 +20 cannam@95: M ./kernel/timer.c -8 +2 cannam@95: M ./libbench/timer.c -2 +2 cannam@95: M ./tests/bench.c -1 +1 cannam@95: cannam@95: Wed Jun 12 18:19:48 EDT 2002 athena cannam@95: * [project @ 2002-06-12 22:19:48 by athena] cannam@95: Revised planner implementation in preparation for wisdom. cannam@95: cannam@95: M ./kernel/ifftw.h -5 +6 cannam@95: M ./kernel/plan.c -1 +2 cannam@95: M ./kernel/planner-naive.c -3 +7 cannam@95: M ./kernel/planner-score.c -3 +7 cannam@95: M ./kernel/planner.c -46 +81 cannam@95: M ./support/Makefile.codelets -19 +19 cannam@95: cannam@95: Wed Jun 12 08:27:36 EDT 2002 athena cannam@95: * [project @ 2002-06-12 12:27:36 by athena] cannam@95: Manually hoist loop invariants. cannam@95: cannam@95: M ./dft/ct-ditbuf.c -5 +11 cannam@95: cannam@95: Wed Jun 12 07:47:41 EDT 2002 athena cannam@95: * [project @ 2002-06-12 11:47:41 by athena] cannam@95: Revised loop to compile better with gcc -O cannam@95: cannam@95: M ./dft/rank-geq2.c -16 +8 cannam@95: M ./dft/rank0.c -23 +22 cannam@95: M ./dft/vrank-geq1.c -2 +2 cannam@95: cannam@95: Tue Jun 11 17:24:09 EDT 2002 athena cannam@95: * [project @ 2002-06-11 21:24:09 by athena] cannam@95: Changed tensor syntax cannam@95: cannam@95: M ./kernel/tensor.c -4 +5 cannam@95: cannam@95: Tue Jun 11 16:39:45 EDT 2002 athena cannam@95: * [project @ 2002-06-11 20:39:45 by athena] cannam@95: Added stuff to do. cannam@95: cannam@95: M ./TODO +15 cannam@95: cannam@95: Tue Jun 11 16:28:14 EDT 2002 athena cannam@95: * [project @ 2002-06-11 20:28:14 by athena] cannam@95: Report classic/pro in version number cannam@95: cannam@95: M ./kernel/version.c -2 +7 cannam@95: cannam@95: Tue Jun 11 14:22:49 EDT 2002 athena cannam@95: * [project @ 2002-06-11 18:22:49 by athena] cannam@95: Renamed versions into classic/pro cannam@95: cannam@95: A ./CLASSIC-MODE cannam@95: M ./Makefile.am -7 +7 cannam@95: R ./RESEARCH-MODE cannam@95: M ./bootstrap.sh -1 +1 cannam@95: M ./configure.ac -13 +13 cannam@95: M ./dft/codelets/inplace/Makefile.am -1 +1 cannam@95: M ./dft/ct-dit.c -2 +2 cannam@95: M ./dft/ct-ditbuf.c -2 +2 cannam@95: M ./dft/kdft-dit.c -2 +2 cannam@95: M ./dft/rank-geq2.c -4 +4 cannam@95: M ./dft/vrank-geq1.c -4 +4 cannam@95: M ./kernel/ifftw.h -3 +3 cannam@95: M ./mkdist.sh -2 +2 cannam@95: M ./tests/bench.c -1 +1 cannam@95: cannam@95: Tue Jun 11 14:06:06 EDT 2002 athena cannam@95: * [project @ 2002-06-11 18:06:06 by athena] cannam@95: Revised planners, estimator cannam@95: cannam@95: M ./kernel/Makefile.am -4 +4 cannam@95: M ./kernel/ifftw.h -6 +7 cannam@95: R ./kernel/planner-estimate.c cannam@95: M ./kernel/planner-naive.c -4 +4 cannam@95: M ./kernel/planner-score.c -4 +4 cannam@95: M ./kernel/planner.c -2 +17 cannam@95: M ./tests/bench.c -5 +5 cannam@95: cannam@95: Tue Jun 11 11:45:41 EDT 2002 athena cannam@95: * [project @ 2002-06-11 15:45:41 by athena] cannam@95: I don't know what I am doing. cannam@95: cannam@95: M ./Makefile.am -2 +2 cannam@95: M ./dft/buffered.c -3 +2 cannam@95: M ./dft/ct-dit.c -4 +5 cannam@95: M ./dft/ct-ditbuf.c -4 +6 cannam@95: M ./dft/kdft-dif.c -3 +2 cannam@95: M ./dft/kdft-difsq.c -3 +2 cannam@95: M ./dft/kdft-dit.c -7 +4 cannam@95: M ./kernel/ifftw.h -3 +2 cannam@95: cannam@95: Tue Jun 11 10:35:52 EDT 2002 athena cannam@95: * [project @ 2002-06-11 14:35:52 by athena] cannam@95: Massive revision of estimator cannam@95: cannam@95: A ./kernel/ops.c cannam@95: M ./Makefile.am -1 +1 cannam@95: M ./dft/buffered.c -8 +7 cannam@95: M ./dft/codelet.h -3 +3 cannam@95: M ./dft/ct-dif.c -5 +4 cannam@95: M ./dft/ct-dit.c -5 +7 cannam@95: M ./dft/ct-ditbuf.c -4 +9 cannam@95: M ./dft/ct-ditf.c -4 +3 cannam@95: M ./dft/ct.c -6 +1 cannam@95: M ./dft/dft.h -1 +3 cannam@95: M ./dft/direct.c -3 +2 cannam@95: M ./dft/indirect.c -3 +2 cannam@95: M ./dft/nop.c -1 +3 cannam@95: M ./dft/rank-geq2.c -3 +7 cannam@95: M ./dft/rank0.c -3 +4 cannam@95: M ./dft/vrank-geq1.c -10 +18 cannam@95: M ./dft/vrank2-transpose.c -3 +3 cannam@95: M ./dft/vrank3-transpose.c -3 +4 cannam@95: M ./genfft/c.ml -2 +2 cannam@95: M ./kernel/Makefile.am -1 +1 cannam@95: M ./kernel/cycle.h -3 +3 cannam@95: R ./kernel/flops.c cannam@95: M ./kernel/ifftw.h -13 +21 cannam@95: M ./kernel/ops.c +58 cannam@95: M ./kernel/plan.c -2 +2 cannam@95: M ./kernel/planner-estimate.c -2 +9 cannam@95: M ./kernel/planner-naive.c -3 +3 cannam@95: M ./kernel/planner-score.c -4 +4 cannam@95: M ./tests/bench.c -1 +3 cannam@95: cannam@95: Tue Jun 11 07:32:20 EDT 2002 athena cannam@95: * [project @ 2002-06-11 11:32:20 by athena] cannam@95: Many changes cannam@95: cannam@95: A ./dft/vrank-geq1.c cannam@95: M ./dft/Makefile.am -1 +1 cannam@95: M ./dft/buffered.c -9 +8 cannam@95: M ./dft/codelets/inplace/Makefile.am -2 +3 cannam@95: M ./dft/codelets/standard/Makefile.am -2 cannam@95: M ./dft/conf.c -2 +2 cannam@95: M ./dft/ct-dit.c -2 +5 cannam@95: M ./dft/ct.c -2 +2 cannam@95: M ./dft/dft.h -6 +4 cannam@95: M ./dft/indirect.c -2 +2 cannam@95: M ./dft/problem.c -1 +13 cannam@95: M ./dft/rank-geq2.c -5 +4 cannam@95: R ./dft/vecloop.c cannam@95: M ./dft/vrank-geq1.c +243 cannam@95: M ./kernel/ifftw.h -2 +6 cannam@95: M ./kernel/planner-estimate.c -2 +2 cannam@95: M ./kernel/planner-naive.c -2 +2 cannam@95: M ./kernel/planner-score.c -4 +17 cannam@95: M ./kernel/planner.c -4 +6 cannam@95: M ./kernel/print.c -6 +27 cannam@95: M ./kernel/tensor.c -1 +15 cannam@95: M ./kernel/timer.c -2 +2 cannam@95: M ./libbench/bench-main.c -3 +6 cannam@95: M ./tests/bench.c -15 +15 cannam@95: cannam@95: Mon Jun 10 21:35:29 EDT 2002 athena cannam@95: * [project @ 2002-06-11 01:35:29 by athena] cannam@95: Keep it simple, stupid. cannam@95: cannam@95: M ./dft/ct-ditbuf.c -21 +4 cannam@95: cannam@95: Mon Jun 10 19:24:28 EDT 2002 athena cannam@95: * [project @ 2002-06-10 23:24:28 by athena] cannam@95: Fixed when #undef PRECOMPUTE_ARRAY_INDICES cannam@95: cannam@95: M ./kernel/ifftw.h -3 +5 cannam@95: cannam@95: Mon Jun 10 17:58:13 EDT 2002 athena cannam@95: * [project @ 2002-06-10 21:58:13 by athena] cannam@95: Minor changes cannam@95: cannam@95: M ./dft/vrank3-transpose.c -5 +3 cannam@95: M ./kernel/print.c -2 +2 cannam@95: cannam@95: Mon Jun 10 16:30:37 EDT 2002 athena cannam@95: * [project @ 2002-06-10 20:30:37 by athena] cannam@95: Added ct-ditbuf.c, many changes everywhere cannam@95: cannam@95: A ./dft/ct-ditbuf.c cannam@95: M ./CONVENTIONS -2 +3 cannam@95: M ./configure.ac +2 cannam@95: M ./dft/Makefile.am -3 +3 cannam@95: M ./dft/buffered.c -11 +11 cannam@95: M ./dft/ct-dif.c -5 +3 cannam@95: M ./dft/ct-dit.c -6 +4 cannam@95: M ./dft/ct-ditbuf.c +187 cannam@95: M ./dft/ct.c -3 +3 cannam@95: M ./dft/ct.h -1 +3 cannam@95: M ./dft/dft.h -1 +2 cannam@95: M ./dft/direct.c -6 +6 cannam@95: M ./dft/indirect.c -6 +6 cannam@95: M ./dft/kdft-dif.c -2 +2 cannam@95: M ./dft/kdft-difsq.c -2 +2 cannam@95: M ./dft/kdft-dit.c -3 +8 cannam@95: M ./dft/kdft.c -2 +2 cannam@95: M ./dft/nop.c -2 +2 cannam@95: M ./dft/rank-geq2.c -6 +6 cannam@95: M ./dft/rank0.c -2 +2 cannam@95: M ./dft/vecloop.c -10 +8 cannam@95: M ./dft/vrank2-transpose.c -2 +2 cannam@95: M ./dft/vrank3-transpose.c -2 +2 cannam@95: M ./kernel/ifftw.h -3 +25 cannam@95: M ./kernel/plan.c -5 +4 cannam@95: M ./kernel/problem.c -2 +2 cannam@95: M ./kernel/timer.c -3 +3 cannam@95: M ./tests/bench.c -7 +8 cannam@95: cannam@95: Mon Jun 10 10:55:40 EDT 2002 athena cannam@95: * [project @ 2002-06-10 14:55:40 by athena] cannam@95: More name mangling cannam@95: cannam@95: M ./kernel/ifftw.h -2 +2 cannam@95: M ./kernel/planner.c -2 +2 cannam@95: M ./tests/bench.c -58 +47 cannam@95: cannam@95: Mon Jun 10 10:08:27 EDT 2002 athena cannam@95: * [project @ 2002-06-10 14:08:27 by athena] cannam@95: Fixed build system for single/double precision cannam@95: cannam@95: M ./Makefile.am -4 +3 cannam@95: M ./acinclude.m4 -13 cannam@95: M ./configure.ac +4 cannam@95: M ./tests/Makefile.am -1 +1 cannam@95: cannam@95: Mon Jun 10 09:04:21 EDT 2002 athena cannam@95: * [project @ 2002-06-10 13:04:21 by athena] cannam@95: Massive renaming to support both single and double precision. cannam@95: (Must recompile everything twice). cannam@95: cannam@95: A ./kernel/dfftw3.h cannam@95: A ./kernel/fftw3.h cannam@95: A ./kernel/sfftw3.h cannam@95: M ./CONVENTIONS -3 +9 cannam@95: M ./configure.ac -1 +1 cannam@95: M ./dft/buffered.c -96 +102 cannam@95: M ./dft/codelet.h -16 +16 cannam@95: M ./dft/codelets/inplace/Makefile.am -1 +1 cannam@95: M ./dft/codelets/standard/Makefile.am -1 +1 cannam@95: M ./dft/conf.c -12 +12 cannam@95: M ./dft/ct-dif.c -31 +33 cannam@95: M ./dft/ct-dit.c -29 +31 cannam@95: M ./dft/ct-ditf.c -27 +29 cannam@95: M ./dft/ct.c -40 +41 cannam@95: M ./dft/ct.h -10 +12 cannam@95: M ./dft/dft.h -25 +27 cannam@95: M ./dft/direct.c -26 +26 cannam@95: M ./dft/indirect.c -50 +63 cannam@95: M ./dft/kdft-dif.c -3 +3 cannam@95: M ./dft/kdft-difsq.c -3 +3 cannam@95: M ./dft/kdft-dit.c -3 +3 cannam@95: M ./dft/kdft.c -3 +3 cannam@95: M ./dft/nop.c -11 +11 cannam@95: M ./dft/plan.c -3 +3 cannam@95: M ./dft/problem.c -39 +40 cannam@95: M ./dft/rank-geq2.c -57 +59 cannam@95: M ./dft/rank0.c -54 +62 cannam@95: M ./dft/solve.c -2 +2 cannam@95: M ./dft/vecloop.c -49 +49 cannam@95: M ./dft/vrank2-transpose.c -27 +28 cannam@95: M ./dft/vrank3-transpose.c -44 +47 cannam@95: M ./genfft/gen_notw.ml -3 +3 cannam@95: M ./genfft/gen_twiddle.ml -4 +4 cannam@95: M ./genfft/gen_twidsq.ml -4 +4 cannam@95: M ./genfft/genutil.ml -2 +2 cannam@95: M ./kernel/Makefile.am -1 +2 cannam@95: M ./kernel/alloc.c -76 +76 cannam@95: M ./kernel/assert.c -2 +2 cannam@95: M ./kernel/awake.c -2 +2 cannam@95: M ./kernel/cycle.h -11 +11 cannam@95: M ./kernel/dfftw3.h +24 cannam@95: R ./kernel/fftw.h cannam@95: M ./kernel/fftw3.h +39 cannam@95: M ./kernel/flops.c -4 +6 cannam@95: M ./kernel/ifftw.h -90 +89 cannam@95: M ./kernel/minmax.c -5 +5 cannam@95: M ./kernel/plan.c -6 +6 cannam@95: M ./kernel/planner-estimate.c -6 +6 cannam@95: M ./kernel/planner-naive.c -7 +7 cannam@95: M ./kernel/planner-score.c -11 +11 cannam@95: M ./kernel/planner.c -73 +71 cannam@95: M ./kernel/print.c -12 +14 cannam@95: M ./kernel/problem.c -5 +5 cannam@95: M ./kernel/sfftw3.h +24 cannam@95: M ./kernel/solver.c -5 +5 cannam@95: M ./kernel/solvtab.c -4 +4 cannam@95: M ./kernel/square.c -2 +2 cannam@95: M ./kernel/stride.c -6 +6 cannam@95: M ./kernel/tensor.c -96 +96 cannam@95: M ./kernel/timer.c -39 +40 cannam@95: M ./kernel/twiddle.c -33 +38 cannam@95: M ./kernel/version.c -3 +4 cannam@95: M ./support/Makefile.codelets -3 +3 cannam@95: M ./tests/bench.c -39 +53 cannam@95: cannam@95: Mon Jun 10 06:49:55 EDT 2002 athena cannam@95: * [project @ 2002-06-10 10:49:55 by athena] cannam@95: Preliminary crude support for vector transforms in benchmark library. cannam@95: cannam@95: M ./libbench/allocate.c -7 +8 cannam@95: M ./libbench/bench-user.h -2 +5 cannam@95: M ./libbench/mflops.c -2 +2 cannam@95: M ./libbench/problem.c -2 +31 cannam@95: M ./tests/bench.c -1 +2 cannam@95: cannam@95: Sun Jun 9 21:11:51 EDT 2002 athena cannam@95: * [project @ 2002-06-10 01:11:51 by athena] cannam@95: Wrong cast cannam@95: cannam@95: M ./kernel/tensor.c -3 +3 cannam@95: cannam@95: Sun Jun 9 16:48:54 EDT 2002 athena cannam@95: * [project @ 2002-06-09 20:48:54 by athena] cannam@95: Added things to do. cannam@95: cannam@95: M ./TODO +1 cannam@95: cannam@95: Sun Jun 9 16:07:12 EDT 2002 athena cannam@95: * [project @ 2002-06-09 20:07:12 by athena] cannam@95: twlen0: make static cannam@95: cannam@95: M ./kernel/twiddle.c -2 +2 cannam@95: cannam@95: Sun Jun 9 15:34:54 EDT 2002 athena cannam@95: * [project @ 2002-06-09 19:34:54 by athena] cannam@95: Nothing cannam@95: cannam@95: M ./dft/buffered.c -3 +3 cannam@95: cannam@95: Sun Jun 9 15:30:13 EDT 2002 athena cannam@95: * [project @ 2002-06-09 19:30:13 by athena] cannam@95: Forgot break in switch statement. cannam@95: cannam@95: M ./kernel/print.c -1 +2 cannam@95: cannam@95: Sun Jun 9 15:27:24 EDT 2002 athena cannam@95: * [project @ 2002-06-09 19:27:24 by athena] cannam@95: Fix for c++ compatibility cannam@95: cannam@95: M ./kernel/print.c -4 +6 cannam@95: cannam@95: Sun Jun 9 15:16:43 EDT 2002 athena cannam@95: * [project @ 2002-06-09 19:16:43 by athena] cannam@95: Added printer, changed everything cannam@95: cannam@95: A ./kernel/print.c cannam@95: M ./TODO -4 cannam@95: M ./dft/buffered.c -26 +22 cannam@95: M ./dft/ct.c -10 +5 cannam@95: M ./dft/direct.c -8 +3 cannam@95: M ./dft/indirect.c -20 +16 cannam@95: M ./dft/nop.c -4 +6 cannam@95: M ./dft/rank-geq2.c -7 +3 cannam@95: M ./dft/rank0.c -6 +3 cannam@95: M ./dft/vecloop.c -5 +3 cannam@95: M ./dft/vrank2-transpose.c -3 +3 cannam@95: M ./dft/vrank3-transpose.c -3 +3 cannam@95: M ./kernel/Makefile.am -2 +2 cannam@95: M ./kernel/ifftw.h -4 +14 cannam@95: M ./kernel/planner.c -3 +10 cannam@95: M ./kernel/print.c +122 cannam@95: M ./tests/bench.c -2 +22 cannam@95: cannam@95: Sun Jun 9 11:37:07 EDT 2002 athena cannam@95: * [project @ 2002-06-09 15:37:07 by athena] cannam@95: Removed redundant nop solver cannam@95: cannam@95: M ./dft/buffered.c -8 +3 cannam@95: M ./dft/nop.c -2 +12 cannam@95: M ./dft/rank0.c -25 +3 cannam@95: M ./tests/bench.c -1 +1 cannam@95: cannam@95: Sun Jun 9 11:06:31 EDT 2002 athena cannam@95: * [project @ 2002-06-09 15:06:31 by athena] cannam@95: More things to do cannam@95: cannam@95: M ./TODO +10 cannam@95: cannam@95: Sun Jun 9 11:01:41 EDT 2002 athena cannam@95: * [project @ 2002-06-09 15:01:41 by athena] cannam@95: Introduced idea of rank -infinity and associated NOP plans cannam@95: cannam@95: A ./TODO cannam@95: A ./dft/nop.c cannam@95: M ./TODO +4 cannam@95: M ./dft/Makefile.am -4 +4 cannam@95: M ./dft/buffered.c -88 +66 cannam@95: M ./dft/conf.c -1 +2 cannam@95: M ./dft/dft.h -1 +2 cannam@95: M ./dft/direct.c -2 +3 cannam@95: M ./dft/indirect.c -1 +2 cannam@95: M ./dft/nop.c +84 cannam@95: M ./dft/problem.c -2 +5 cannam@95: M ./dft/rank0.c -3 +12 cannam@95: M ./dft/vecloop.c -1 +2 cannam@95: M ./kernel/ifftw.h -1 +11 cannam@95: M ./kernel/tensor.c -11 +39 cannam@95: cannam@95: Sun Jun 9 08:36:27 EDT 2002 athena cannam@95: * [project @ 2002-06-09 12:36:27 by athena] cannam@95: Fixed comment cannam@95: cannam@95: M ./dft/buffered.c -2 +2 cannam@95: cannam@95: Sun Jun 9 08:20:13 EDT 2002 athena cannam@95: * [project @ 2002-06-09 12:20:13 by athena] cannam@95: Removed useless assertions. cannam@95: cannam@95: M ./kernel/tensor.c -4 +3 cannam@95: cannam@95: Sun Jun 9 08:19:26 EDT 2002 athena cannam@95: * [project @ 2002-06-09 12:19:26 by athena] cannam@95: Don't malloc(0). cannam@95: cannam@95: M ./kernel/tensor.c -3 +7 cannam@95: cannam@95: Sun Jun 9 08:08:13 EDT 2002 athena cannam@95: * [project @ 2002-06-09 12:08:13 by athena] cannam@95: Fixed signed/unsigned puns cannam@95: cannam@95: M ./dft/buffered.c -11 +11 cannam@95: cannam@95: Sun Jun 9 07:52:22 EDT 2002 athena cannam@95: * [project @ 2002-06-09 11:52:22 by athena] cannam@95: Added buffered.c cannam@95: cannam@95: A ./dft/buffered.c cannam@95: M ./dft/Makefile.am -1 +1 cannam@95: M ./dft/buffered.c +374 cannam@95: M ./dft/conf.c -1 +2 cannam@95: M ./dft/ct-dif.c -2 +2 cannam@95: M ./dft/ct-dit.c -2 +2 cannam@95: M ./dft/ct-ditf.c -2 +2 cannam@95: M ./dft/dft.h -1 +2 cannam@95: M ./dft/direct.c -2 +2 cannam@95: M ./dft/indirect.c -3 +3 cannam@95: M ./dft/rank-geq2.c -2 +2 cannam@95: M ./dft/rank0.c -6 +6 cannam@95: M ./dft/vecloop.c -2 +2 cannam@95: M ./dft/vrank2-transpose.c -2 +2 cannam@95: M ./dft/vrank3-transpose.c -2 +2 cannam@95: M ./libbench/bench-main.c -2 +7 cannam@95: M ./libbench/bench-user.h -1 +3 cannam@95: M ./tests/bench.c -1 +1 cannam@95: cannam@95: Sat Jun 8 16:57:54 EDT 2002 athena cannam@95: * [project @ 2002-06-08 20:57:54 by athena] cannam@95: Fixed printout cannam@95: cannam@95: M ./dft/ct.c -6 +5 cannam@95: cannam@95: Sat Jun 8 16:42:52 EDT 2002 athena cannam@95: * [project @ 2002-06-08 20:42:52 by athena] cannam@95: Fixed comment cannam@95: cannam@95: M ./dft/vrank3-transpose.c -2 +2 cannam@95: cannam@95: Sat Jun 8 16:40:58 EDT 2002 athena cannam@95: * [project @ 2002-06-08 20:40:58 by athena] cannam@95: Added vrank3-transpose, renamed vrank0-transpose -> vrank2-transpose cannam@95: cannam@95: A ./dft/vrank2-transpose.c cannam@95: A ./dft/vrank3-transpose.c cannam@95: M ./dft/Makefile.am -2 +2 cannam@95: M ./dft/conf.c -2 +3 cannam@95: M ./dft/dft.h -1 +3 cannam@95: R ./dft/vrank0-transpose.c cannam@95: M ./dft/vrank2-transpose.c +132 cannam@95: M ./dft/vrank3-transpose.c +173 cannam@95: M ./tests/bench.c -1 +1 cannam@95: cannam@95: Sat Jun 8 15:51:46 EDT 2002 athena cannam@95: * [project @ 2002-06-08 19:51:46 by athena] cannam@95: Added vrank0-transpose cannam@95: cannam@95: A ./dft/rank-geq2.c cannam@95: A ./dft/vrank0-transpose.c cannam@95: M ./bootstrap.sh -1 +1 cannam@95: M ./dft/Makefile.am -1 +2 cannam@95: M ./dft/conf.c -1 +2 cannam@95: M ./dft/direct.c -4 +4 cannam@95: M ./dft/rank-geq2.c +257 cannam@95: M ./dft/rank0.c -2 +2 cannam@95: R ./dft/rank_geq2.c cannam@95: M ./dft/vrank0-transpose.c +132 cannam@95: M ./tests/bench.c -1 +2 cannam@95: cannam@95: Sat Jun 8 15:11:09 EDT 2002 athena cannam@95: * [project @ 2002-06-08 19:11:09 by athena] cannam@95: Added planner-score.c cannam@95: cannam@95: A ./kernel/planner-score.c cannam@95: M ./dft/Makefile.am -2 +2 cannam@95: M ./dft/ct-dif.c -2 +2 cannam@95: M ./dft/ct-dit.c -2 +2 cannam@95: M ./dft/ct-ditf.c -2 +2 cannam@95: M ./dft/direct.c -2 +2 cannam@95: M ./dft/indirect.c -2 +2 cannam@95: M ./dft/rank0.c -2 +2 cannam@95: M ./dft/rank_geq2.c -2 +2 cannam@95: M ./dft/vecloop.c -2 +2 cannam@95: M ./kernel/Makefile.am -3 +3 cannam@95: M ./kernel/ifftw.h -4 +5 cannam@95: M ./kernel/planner-score.c +70 cannam@95: M ./kernel/planner.c -12 +13 cannam@95: M ./tests/bench.c -1 +4 cannam@95: cannam@95: Sat Jun 8 11:10:44 EDT 2002 athena cannam@95: * [project @ 2002-06-08 15:10:44 by athena] cannam@95: Added indirect.c cannam@95: cannam@95: A ./dft/indirect.c cannam@95: M ./dft/Makefile.am -1 +1 cannam@95: M ./dft/conf.c -1 +2 cannam@95: M ./dft/ct-dif.c -2 +2 cannam@95: M ./dft/ct-dit.c -2 +2 cannam@95: M ./dft/ct-ditf.c -4 +6 cannam@95: M ./dft/dft.h -1 +3 cannam@95: M ./dft/indirect.c +227 cannam@95: M ./dft/rank_geq2.c -5 +5 cannam@95: M ./dft/vecloop.c -2 +1 cannam@95: cannam@95: Sat Jun 8 09:34:58 EDT 2002 athena cannam@95: * [project @ 2002-06-08 13:34:58 by athena] cannam@95: dif, ditf solvers cannam@95: cannam@95: A ./mkdist.sh cannam@95: A ./dft/ct-dif.c cannam@95: A ./dft/ct-ditf.c cannam@95: A ./dft/kdft-dif.c cannam@95: A ./dft/kdft-difsq.c cannam@95: M ./Makefile.am -4 +6 cannam@95: M ./dft/Makefile.am -5 +3 cannam@95: M ./dft/codelet.h -2 +3 cannam@95: M ./dft/codelets/Makefile.am -5 cannam@95: M ./dft/codelets/inplace/Makefile.am -2 +4 cannam@95: M ./dft/conf.c -2 +3 cannam@95: M ./dft/ct-dif.c +120 cannam@95: M ./dft/ct-dit.c -2 +1 cannam@95: M ./dft/ct-ditf.c +106 cannam@95: M ./dft/dft.h -1 +2 cannam@95: M ./dft/direct.c -2 +2 cannam@95: M ./dft/kdft-dif.c +29 cannam@95: M ./dft/kdft-difsq.c +29 cannam@95: M ./mkdist.sh +6 cannam@95: M ./tests/Makefile.am -2 +2 cannam@95: cannam@95: Fri Jun 7 18:07:53 EDT 2002 athena cannam@95: * [project @ 2002-06-07 22:07:53 by athena] cannam@95: Implemented rank_geq2. Revised build system cannam@95: cannam@95: A ./RESEARCH-MODE cannam@95: A ./dft/rank_geq2.c cannam@95: M ./Makefile.am -3 +12 cannam@95: M ./RESEARCH-MODE +1 cannam@95: M ./bootstrap.sh -1 +1 cannam@95: M ./configure.ac -1 +10 cannam@95: M ./dft/Makefile.am -1 +3 cannam@95: M ./dft/conf.c -1 +2 cannam@95: M ./dft/dft.h -1 +2 cannam@95: M ./dft/rank_geq2.c +257 cannam@95: M ./dft/vecloop.c -42 +48 cannam@95: M ./kernel/ifftw.h -1 +3 cannam@95: M ./kernel/minmax.c -1 +11 cannam@95: M ./kernel/planner.c -1 +2 cannam@95: M ./support/Makefile.codelets -1 +1 cannam@95: cannam@95: Fri Jun 7 07:12:25 EDT 2002 athena cannam@95: * [project @ 2002-06-07 11:12:25 by athena] cannam@95: Fixed printout cannam@95: cannam@95: M ./kernel/alloc.c -2 +2 cannam@95: cannam@95: Fri Jun 7 07:07:46 EDT 2002 athena cannam@95: * [project @ 2002-06-07 11:07:46 by athena] cannam@95: Added rank0. Revised codelet organization. cannam@95: cannam@95: A ./dft/codelet.h cannam@95: A ./dft/conf.c cannam@95: A ./dft/codelets/ cannam@95: A ./dft/codelets/Makefile.am cannam@95: A ./dft/codelets/inplace/ cannam@95: A ./dft/codelets/inplace/Makefile.am cannam@95: A ./dft/codelets/standard/ cannam@95: A ./dft/codelets/standard/Makefile.am cannam@95: A ./dft/rank0.c cannam@95: M ./Makefile.am -3 +2 cannam@95: M ./bootstrap.sh -5 +1 cannam@95: M ./configure.ac -3 +18 cannam@95: M ./dft/Makefile.am -3 +4 cannam@95: M ./dft/codelet.h +118 cannam@95: M ./dft/codelets/Makefile.am +7 cannam@95: M ./dft/codelets/inplace/Makefile.am -1 +35 cannam@95: M ./dft/codelets/standard/Makefile.am -1 +37 cannam@95: M ./dft/conf.c +30 cannam@95: M ./dft/dft.h -1 +5 cannam@95: M ./dft/rank0.c +263 cannam@95: M ./genfft/gen_notw.ml -3 +3 cannam@95: M ./tests/Makefile.am -4 +5 cannam@95: M ./tests/bench.c -6 +1 cannam@95: cannam@95: Thu Jun 6 18:03:17 EDT 2002 athena cannam@95: * [project @ 2002-06-06 22:03:17 by athena] cannam@95: Added memoization cannam@95: cannam@95: M ./dft/ct.c -2 +2 cannam@95: M ./dft/vecloop.c -2 +2 cannam@95: M ./genfft/trig.ml -5 +5 cannam@95: M ./kernel/ifftw.h -2 +13 cannam@95: M ./kernel/planner-estimate.c -2 +2 cannam@95: M ./kernel/planner-naive.c -2 +2 cannam@95: M ./kernel/planner.c -12 +191 cannam@95: M ./libbench/bench-user.h -1 +2 cannam@95: M ./tests/bench.c -2 +16 cannam@95: cannam@95: Thu Jun 6 08:07:33 EDT 2002 athena cannam@95: * [project @ 2002-06-06 12:07:33 by athena] cannam@95: Added vecloop cannam@95: cannam@95: A ./dft/vecloop.c cannam@95: M ./dft/Makefile.am -1 +1 cannam@95: M ./dft/dft.h -1 +3 cannam@95: M ./dft/direct.c -9 +7 cannam@95: M ./dft/vecloop.c +242 cannam@95: M ./kernel/alloc.c -11 +38 cannam@95: M ./kernel/ifftw.h -2 +4 cannam@95: M ./kernel/planner.c -2 +2 cannam@95: M ./tests/bench.c +5 cannam@95: cannam@95: Wed Jun 5 19:02:56 EDT 2002 athena cannam@95: * [project @ 2002-06-05 23:02:56 by athena] cannam@95: First DIT solver/plan cannam@95: cannam@95: A ./dft/ct-dit.c cannam@95: M ./dft/Makefile.am -2 +2 cannam@95: M ./dft/ct-dit.c +118 cannam@95: M ./dft/ct.c -9 +16 cannam@95: M ./dft/ct.h -3 +14 cannam@95: M ./dft/dft.h -1 +2 cannam@95: M ./dft/direct.c -10 +6 cannam@95: M ./dft/kdft-dit.c -2 +3 cannam@95: M ./kernel/alloc.c -1 +2 cannam@95: M ./kernel/twiddle.c -3 +11 cannam@95: cannam@95: Wed Jun 5 16:03:44 EDT 2002 athena cannam@95: * [project @ 2002-06-05 20:03:44 by athena] cannam@95: More work on ct cannam@95: cannam@95: A ./dft/ct.c cannam@95: A ./dft/ct.h cannam@95: M ./dft/Makefile.am -2 +2 cannam@95: M ./dft/ct.c +166 cannam@95: M ./dft/ct.h +56 cannam@95: M ./kernel/ifftw.h -1 +2 cannam@95: M ./kernel/stride.c -2 +3 cannam@95: M ./kernel/twiddle.c -16 +26 cannam@95: cannam@95: Wed Jun 5 11:28:09 EDT 2002 athena cannam@95: * [project @ 2002-06-05 15:28:09 by athena] cannam@95: Only use cycle counters cannam@95: cannam@95: M ./kernel/ifftw.h -2 +2 cannam@95: M ./kernel/planner-naive.c -2 +2 cannam@95: M ./kernel/timer.c -166 +18 cannam@95: cannam@95: Tue Jun 4 20:22:23 EDT 2002 athena cannam@95: * [project @ 2002-06-05 00:22:23 by athena] cannam@95: Signed/unsigned fixup cannam@95: cannam@95: M ./CONVENTIONS +2 cannam@95: M ./bootstrap.sh -1 +1 cannam@95: M ./kernel/ifftw.h -7 +6 cannam@95: M ./kernel/twiddle.c -28 +16 cannam@95: cannam@95: Tue Jun 4 20:03:56 EDT 2002 athena cannam@95: * [project @ 2002-06-05 00:03:56 by athena] cannam@95: New file twiddle.c cannam@95: cannam@95: A ./kernel/twiddle.c cannam@95: M ./kernel/Makefile.am -1 +1 cannam@95: M ./kernel/ifftw.h -1 +24 cannam@95: M ./kernel/twiddle.c +137 cannam@95: cannam@95: Tue Jun 4 17:49:39 EDT 2002 athena cannam@95: * [project @ 2002-06-04 21:49:39 by athena] cannam@95: Made tensor ranks and vector lengths unsigned. Hopefully fixed cannam@95: all places where it matters. cannam@95: cannam@95: A ./dft/kdft-dit.c cannam@95: M ./configure.ac +1 cannam@95: M ./dft/Makefile.am -1 +2 cannam@95: M ./dft/direct.c -3 +4 cannam@95: M ./dft/kdft-dit.c +28 cannam@95: M ./dft/problem.c -5 +7 cannam@95: M ./genfft/gen_notw.ml -4 +4 cannam@95: M ./genfft/gen_twiddle.ml -4 +4 cannam@95: M ./genfft/gen_twidsq.ml -4 +4 cannam@95: M ./kernel/Makefile.am -2 +2 cannam@95: R ./kernel/codelet.h cannam@95: M ./kernel/flops.c -2 +2 cannam@95: M ./kernel/ifftw.h -17 +20 cannam@95: M ./kernel/tensor.c -35 +38 cannam@95: M ./kernel/timer.c -3 +2 cannam@95: M ./tests/bench.c +1 cannam@95: cannam@95: Tue Jun 4 16:28:58 EDT 2002 athena cannam@95: * [project @ 2002-06-04 20:28:58 by athena] cannam@95: System is in working state now (but very incomplete) cannam@95: cannam@95: A ./dft/kdft.c cannam@95: A ./kernel/planner-estimate.c cannam@95: A ./kernel/solvtab.c cannam@95: A ./tests/ cannam@95: A ./tests/Makefile.am cannam@95: A ./tests/bench.c cannam@95: M ./Makefile.am -1 +1 cannam@95: M ./configure.ac +2 cannam@95: M ./dft/Makefile.am -2 +2 cannam@95: M ./dft/dft.h -1 +10 cannam@95: M ./dft/direct.c -3 +2 cannam@95: M ./dft/kdft.c +29 cannam@95: M ./kernel/Makefile.am -3 +4 cannam@95: M ./kernel/codelet.h -1 +3 cannam@95: M ./kernel/fftw.h -1 +5 cannam@95: M ./kernel/ifftw.h -2 +21 cannam@95: M ./kernel/planner-estimate.c +55 cannam@95: M ./kernel/planner-naive.c -9 +9 cannam@95: M ./kernel/solvtab.c +30 cannam@95: M ./libbench/Makefile.am -2 +2 cannam@95: M ./support/Makefile.codelets +23 cannam@95: M ./tests/Makefile.am -1 +7 cannam@95: M ./tests/bench.c +92 cannam@95: cannam@95: Mon Jun 3 18:10:12 EDT 2002 athena cannam@95: * [project @ 2002-06-03 22:10:12 by athena] cannam@95: Started implementing planners cannam@95: cannam@95: A ./kernel/planner-naive.c cannam@95: A ./kernel/planner.c cannam@95: M ./CONVENTIONS +2 cannam@95: M ./kernel/Makefile.am -1 +1 cannam@95: M ./kernel/ifftw.h -6 +55 cannam@95: M ./kernel/planner-naive.c +58 cannam@95: M ./kernel/planner.c +113 cannam@95: cannam@95: Mon Jun 3 11:44:18 EDT 2002 athena cannam@95: * [project @ 2002-06-03 15:44:18 by athena] cannam@95: Imported libbench from the new benchfft. We will use libbench cannam@95: for benchmarking and testing. cannam@95: cannam@95: A ./libbench/ cannam@95: A ./libbench/Makefile.am cannam@95: A ./libbench/accopy-from.c cannam@95: A ./libbench/accopy-to.c cannam@95: A ./libbench/acopy.c cannam@95: A ./libbench/allocate.c cannam@95: A ./libbench/ascale.c cannam@95: A ./libbench/aset.c cannam@95: A ./libbench/bench-main.c cannam@95: A ./libbench/bench-user.h cannam@95: A ./libbench/bench.h cannam@95: A ./libbench/caadd.c cannam@95: A ./libbench/cacopy.c cannam@95: A ./libbench/can-do.c cannam@95: A ./libbench/cascale.c cannam@95: A ./libbench/caset.c cannam@95: A ./libbench/casub.c cannam@95: A ./libbench/ccopy-from.c cannam@95: A ./libbench/ccopy-to.c cannam@95: A ./libbench/copy-c2c-from.c cannam@95: A ./libbench/copy-c2c-to.c cannam@95: A ./libbench/copy-c2h-1d-fftpack.c cannam@95: A ./libbench/copy-c2h-1d-halfcomplex.c cannam@95: A ./libbench/copy-c2h-1d-packed.c cannam@95: A ./libbench/copy-c2h-1d-unpacked-ri.c cannam@95: A ./libbench/copy-c2h-unpacked.c cannam@95: A ./libbench/copy-c2h.c cannam@95: A ./libbench/copy-c2r-packed.c cannam@95: A ./libbench/copy-c2r-unpacked.c cannam@95: A ./libbench/copy-c2r.c cannam@95: A ./libbench/copy-c2ri.c cannam@95: A ./libbench/copy-h2c-1d-fftpack.c cannam@95: A ./libbench/copy-h2c-1d-halfcomplex.c cannam@95: A ./libbench/copy-h2c-1d-packed.c cannam@95: A ./libbench/copy-h2c-1d-unpacked-ri.c cannam@95: A ./libbench/copy-h2c-unpacked.c cannam@95: A ./libbench/copy-h2c.c cannam@95: A ./libbench/copy-r2c-packed.c cannam@95: A ./libbench/copy-r2c-unpacked.c cannam@95: A ./libbench/copy-r2c.c cannam@95: A ./libbench/copy-ri2c.c cannam@95: A ./libbench/deallocate.c cannam@95: A ./libbench/getopt-utils.c cannam@95: A ./libbench/getopt.c cannam@95: A ./libbench/getopt.h cannam@95: A ./libbench/getopt1.c cannam@95: A ./libbench/info.c cannam@95: A ./libbench/log2.c cannam@95: A ./libbench/main.c cannam@95: A ./libbench/mflops.c cannam@95: A ./libbench/ovtpvt.c cannam@95: A ./libbench/pow2.c cannam@95: A ./libbench/prime.c cannam@95: A ./libbench/problem.c cannam@95: A ./libbench/report.c cannam@95: A ./libbench/speed.c cannam@95: A ./libbench/timer.c cannam@95: A ./libbench/unnormalize.c cannam@95: A ./libbench/util.c cannam@95: A ./libbench/verify.c cannam@95: A ./libbench/zero.c cannam@95: M ./Makefile.am -1 +1 cannam@95: M ./configure.ac -1 +4 cannam@95: M ./libbench/Makefile.am +18 cannam@95: M ./libbench/accopy-from.c +10 cannam@95: M ./libbench/accopy-to.c +10 cannam@95: M ./libbench/acopy.c +11 cannam@95: M ./libbench/allocate.c +40 cannam@95: M ./libbench/ascale.c +13 cannam@95: M ./libbench/aset.c +10 cannam@95: M ./libbench/bench-main.c +218 cannam@95: M ./libbench/bench-user.h +247 cannam@95: M ./libbench/bench.h +66 cannam@95: M ./libbench/caadd.c +15 cannam@95: M ./libbench/cacopy.c +11 cannam@95: M ./libbench/can-do.c +33 cannam@95: M ./libbench/cascale.c +15 cannam@95: M ./libbench/caset.c +10 cannam@95: M ./libbench/casub.c +15 cannam@95: M ./libbench/ccopy-from.c +21 cannam@95: M ./libbench/ccopy-to.c +20 cannam@95: M ./libbench/copy-c2c-from.c +7 cannam@95: M ./libbench/copy-c2c-to.c +7 cannam@95: M ./libbench/copy-c2h-1d-fftpack.c +29 cannam@95: M ./libbench/copy-c2h-1d-halfcomplex.c +29 cannam@95: M ./libbench/copy-c2h-1d-packed.c +40 cannam@95: M ./libbench/copy-c2h-1d-unpacked-ri.c +25 cannam@95: M ./libbench/copy-c2h-unpacked.c +32 cannam@95: M ./libbench/copy-c2h.c +10 cannam@95: M ./libbench/copy-c2r-packed.c +11 cannam@95: M ./libbench/copy-c2r-unpacked.c +24 cannam@95: M ./libbench/copy-c2r.c +6 cannam@95: M ./libbench/copy-c2ri.c +15 cannam@95: M ./libbench/copy-h2c-1d-fftpack.c +31 cannam@95: M ./libbench/copy-h2c-1d-halfcomplex.c +34 cannam@95: M ./libbench/copy-h2c-1d-packed.c +44 cannam@95: M ./libbench/copy-h2c-1d-unpacked-ri.c +35 cannam@95: M ./libbench/copy-h2c-unpacked.c +51 cannam@95: M ./libbench/copy-h2c.c +10 cannam@95: M ./libbench/copy-r2c-packed.c +13 cannam@95: M ./libbench/copy-r2c-unpacked.c +26 cannam@95: M ./libbench/copy-r2c.c +6 cannam@95: M ./libbench/copy-ri2c.c +15 cannam@95: M ./libbench/deallocate.c +14 cannam@95: M ./libbench/getopt-utils.c +104 cannam@95: M ./libbench/getopt.c +1062 cannam@95: M ./libbench/getopt.h +180 cannam@95: M ./libbench/getopt1.c +188 cannam@95: M ./libbench/info.c +57 cannam@95: M ./libbench/log2.c +13 cannam@95: M ./libbench/main.c +40 cannam@95: M ./libbench/mflops.c +22 cannam@95: M ./libbench/ovtpvt.c +13 cannam@95: M ./libbench/pow2.c +6 cannam@95: M ./libbench/prime.c +49 cannam@95: M ./libbench/problem.c +126 cannam@95: M ./libbench/report.c +159 cannam@95: M ./libbench/speed.c +73 cannam@95: M ./libbench/timer.c +218 cannam@95: M ./libbench/unnormalize.c +13 cannam@95: M ./libbench/util.c +189 cannam@95: M ./libbench/verify.c +408 cannam@95: M ./libbench/zero.c +37 cannam@95: cannam@95: Mon Jun 3 09:18:46 EDT 2002 athena cannam@95: * [project @ 2002-06-03 13:18:46 by athena] cannam@95: Removed useless rand.c cannam@95: cannam@95: M ./kernel/Makefile.am -2 +2 cannam@95: M ./kernel/ifftw.h -4 +1 cannam@95: R ./kernel/rand.c cannam@95: M ./kernel/timer.c -2 +2 cannam@95: cannam@95: Mon Jun 3 08:09:05 EDT 2002 athena cannam@95: * [project @ 2002-06-03 12:09:05 by athena] cannam@95: Added timer cannam@95: cannam@95: A ./kernel/cycle.h cannam@95: A ./kernel/timer.c cannam@95: M ./CONVENTIONS +1 cannam@95: M ./dft/problem.c -2 +2 cannam@95: M ./kernel/Makefile.am -1 +1 cannam@95: M ./kernel/alloc.c -2 +2 cannam@95: M ./kernel/cycle.h +204 cannam@95: M ./kernel/ifftw.h -2 +10 cannam@95: M ./kernel/plan.c -2 +4 cannam@95: M ./kernel/timer.c +293 cannam@95: cannam@95: Sun Jun 2 21:03:51 EDT 2002 athena cannam@95: * [project @ 2002-06-03 01:03:51 by athena] cannam@95: Split codelets into standard and inplace cannam@95: cannam@95: M ./configure.ac -2 +4 cannam@95: cannam@95: Sun Jun 2 19:49:03 EDT 2002 athena cannam@95: * [project @ 2002-06-02 23:49:03 by athena] cannam@95: Many many changes cannam@95: cannam@95: A ./CONVENTIONS cannam@95: A ./dft/direct.c cannam@95: A ./dft/plan.c cannam@95: A ./dft/solve.c cannam@95: A ./kernel/awake.c cannam@95: A ./kernel/square.c cannam@95: M ./CONVENTIONS +49 cannam@95: M ./Makefile.am -4 +4 cannam@95: M ./dft/Makefile.am -1 +1 cannam@95: M ./dft/dft.h -3 +22 cannam@95: M ./dft/direct.c +171 cannam@95: M ./dft/plan.c +33 cannam@95: M ./dft/problem.c -10 +10 cannam@95: M ./dft/solve.c +31 cannam@95: M ./kernel/Makefile.am -1 +1 cannam@95: M ./kernel/awake.c +30 cannam@95: M ./kernel/ifftw.h -13 +33 cannam@95: M ./kernel/square.c +28 cannam@95: cannam@95: Sun Jun 2 15:00:11 EDT 2002 athena cannam@95: * [project @ 2002-06-02 19:00:11 by athena] cannam@95: Fixed anachronism cannam@95: cannam@95: M ./kernel/codelet.h -4 +4 cannam@95: cannam@95: Sat Aug 12 21:43:16 EDT 2006 Unknown tagger cannam@95: tagged bar cannam@95: cannam@95: cannam@95: Sun Jun 2 14:42:32 EDT 2002 athena cannam@95: * [project @ 2002-06-02 18:42:32 by athena] cannam@95: Initial revision cannam@95: cannam@95: A ./AUTHORS cannam@95: A ./COPYRIGHT cannam@95: A ./ChangeLog cannam@95: A ./Makefile.am cannam@95: A ./NEWS cannam@95: A ./README cannam@95: A ./acinclude.m4 cannam@95: A ./bootstrap.sh cannam@95: A ./configure.ac cannam@95: A ./dft/ cannam@95: A ./dft/Makefile.am cannam@95: A ./dft/dft.h cannam@95: A ./dft/problem.c cannam@95: A ./genfft/ cannam@95: A ./genfft/Makefile.am cannam@95: A ./genfft/algsimp.ml cannam@95: A ./genfft/algsimp.mli cannam@95: A ./genfft/annotate.ml cannam@95: A ./genfft/annotate.mli cannam@95: A ./genfft/assoctable.ml cannam@95: A ./genfft/assoctable.mli cannam@95: A ./genfft/c.ml cannam@95: A ./genfft/c.mli cannam@95: A ./genfft/complex.ml cannam@95: A ./genfft/complex.mli cannam@95: A ./genfft/conv.ml cannam@95: A ./genfft/conv.mli cannam@95: A ./genfft/dag.ml cannam@95: A ./genfft/dag.mli cannam@95: A ./genfft/expr.ml cannam@95: A ./genfft/expr.mli cannam@95: A ./genfft/fft.ml cannam@95: A ./genfft/fft.mli cannam@95: A ./genfft/gen_athnotw.ml cannam@95: A ./genfft/gen_athtw.ml cannam@95: A ./genfft/gen_conv.ml cannam@95: A ./genfft/gen_notw.ml cannam@95: A ./genfft/gen_trig.ml cannam@95: A ./genfft/gen_twiddle.ml cannam@95: A ./genfft/gen_twidsq.ml cannam@95: A ./genfft/genutil.ml cannam@95: A ./genfft/littlesimp.ml cannam@95: A ./genfft/littlesimp.mli cannam@95: A ./genfft/magic.ml cannam@95: A ./genfft/monads.ml cannam@95: A ./genfft/number.ml cannam@95: A ./genfft/number.mli cannam@95: A ./genfft/oracle.ml cannam@95: A ./genfft/oracle.mli cannam@95: A ./genfft/schedule.ml cannam@95: A ./genfft/schedule.mli cannam@95: A ./genfft/to_alist.ml cannam@95: A ./genfft/to_alist.mli cannam@95: A ./genfft/trig.ml cannam@95: A ./genfft/trig.mli cannam@95: A ./genfft/twiddle.ml cannam@95: A ./genfft/twiddle.mli cannam@95: A ./genfft/unique.ml cannam@95: A ./genfft/unique.mli cannam@95: A ./genfft/util.ml cannam@95: A ./genfft/util.mli cannam@95: A ./genfft/variable.ml cannam@95: A ./genfft/variable.mli cannam@95: A ./kernel/ cannam@95: A ./kernel/Makefile.am cannam@95: A ./kernel/alloc.c cannam@95: A ./kernel/assert.c cannam@95: A ./kernel/codelet.h cannam@95: A ./kernel/fftw.h cannam@95: A ./kernel/flops.c cannam@95: A ./kernel/ifftw.h cannam@95: A ./kernel/minmax.c cannam@95: A ./kernel/plan.c cannam@95: A ./kernel/problem.c cannam@95: A ./kernel/rand.c cannam@95: A ./kernel/solver.c cannam@95: A ./kernel/stride.c cannam@95: A ./kernel/tensor.c cannam@95: A ./kernel/version.c cannam@95: A ./support/ cannam@95: A ./support/Makefile.am cannam@95: A ./support/Makefile.codelets cannam@95: A ./support/codelet_prelude cannam@95: M ./AUTHORS +4 cannam@95: M ./COPYRIGHT +19 cannam@95: M ./Makefile.am +21 cannam@95: M ./acinclude.m4 +356 cannam@95: M ./bootstrap.sh +26 cannam@95: M ./configure.ac +92 cannam@95: M ./dft/Makefile.am +6 cannam@95: M ./dft/dft.h +30 cannam@95: M ./dft/problem.c +121 cannam@95: M ./genfft/Makefile.am +136 cannam@95: M ./genfft/algsimp.ml +517 cannam@95: M ./genfft/algsimp.mli +24 cannam@95: M ./genfft/annotate.ml +180 cannam@95: M ./genfft/annotate.mli +37 cannam@95: M ./genfft/assoctable.ml +66 cannam@95: M ./genfft/assoctable.mli +30 cannam@95: M ./genfft/c.ml +445 cannam@95: M ./genfft/c.mli +65 cannam@95: M ./genfft/complex.ml +198 cannam@95: M ./genfft/complex.mli +61 cannam@95: M ./genfft/conv.ml +130 cannam@95: M ./genfft/conv.mli +23 cannam@95: M ./genfft/dag.ml +110 cannam@95: M ./genfft/dag.mli +44 cannam@95: M ./genfft/expr.ml +93 cannam@95: M ./genfft/expr.mli +38 cannam@95: M ./genfft/fft.ml +227 cannam@95: M ./genfft/fft.mli +24 cannam@95: M ./genfft/gen_athnotw.ml +80 cannam@95: M ./genfft/gen_athtw.ml +108 cannam@95: M ./genfft/gen_conv.ml +89 cannam@95: M ./genfft/gen_notw.ml +126 cannam@95: M ./genfft/gen_trig.ml +159 cannam@95: M ./genfft/gen_twiddle.ml +144 cannam@95: M ./genfft/gen_twidsq.ml +172 cannam@95: M ./genfft/genutil.ml +256 cannam@95: M ./genfft/littlesimp.ml +72 cannam@95: M ./genfft/littlesimp.mli +26 cannam@95: M ./genfft/magic.ml +121 cannam@95: M ./genfft/monads.ml +76 cannam@95: M ./genfft/number.ml +153 cannam@95: M ./genfft/number.mli +49 cannam@95: M ./genfft/oracle.ml +138 cannam@95: M ./genfft/oracle.mli +25 cannam@95: M ./genfft/schedule.ml +186 cannam@95: M ./genfft/schedule.mli +30 cannam@95: M ./genfft/to_alist.ml +258 cannam@95: M ./genfft/to_alist.mli +24 cannam@95: M ./genfft/trig.ml +158 cannam@95: M ./genfft/trig.mli +35 cannam@95: M ./genfft/twiddle.ml +165 cannam@95: M ./genfft/twiddle.mli +29 cannam@95: M ./genfft/unique.ml +39 cannam@95: M ./genfft/unique.mli +25 cannam@95: M ./genfft/util.ml +177 cannam@95: M ./genfft/util.mli +50 cannam@95: M ./genfft/variable.ml +99 cannam@95: M ./genfft/variable.mli +35 cannam@95: M ./kernel/Makefile.am +5 cannam@95: M ./kernel/alloc.c +217 cannam@95: M ./kernel/assert.c +32 cannam@95: M ./kernel/codelet.h +126 cannam@95: M ./kernel/fftw.h +29 cannam@95: M ./kernel/flops.c +41 cannam@95: M ./kernel/ifftw.h +215 cannam@95: M ./kernel/minmax.c +33 cannam@95: M ./kernel/plan.c +50 cannam@95: M ./kernel/problem.c +47 cannam@95: M ./kernel/rand.c +36 cannam@95: M ./kernel/solver.c +43 cannam@95: M ./kernel/stride.c +41 cannam@95: M ./kernel/tensor.c +318 cannam@95: M ./kernel/version.c +26 cannam@95: M ./support/Makefile.am -1 +1 cannam@95: M ./support/Makefile.codelets +99 cannam@95: M ./support/codelet_prelude +8