Mercurial > hg > sv-dependency-builds

annotate src/fftw-3.3.5/doc/FAQ/fftw-faq.html/section5.html @ 83:ae30d91d2ffe

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Replace these with versions built using an older toolset (so as to avoid ABI compatibilities when linking on Ubuntu 14.04 for packaging purposes)
author Chris Cannam
date Fri, 07 Feb 2020 11:51:13 +0000
parents 2cd0e3b3e1fd
children
rev   line source
Chris@42 1 <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
Chris@42 2 <html>
Chris@42 3 <head><title>
Chris@42 4 FFTW FAQ - Section 5
Chris@42 5 </title>
Chris@42 6 <link rev="made" href="mailto:fftw@fftw.org">
Chris@42 7 <link rel="Contents" href="index.html">
Chris@42 8 <link rel="Start" href="index.html">
Chris@42 9 <link rel="Previous" href="section4.html"><link rel="Bookmark" title="FFTW FAQ" href="index.html">
Chris@42 10 </head><body text="#000000" bgcolor="#FFFFFF"><h1>
Chris@42 11 FFTW FAQ - Section 5 <br>
Chris@42 12 Known bugs
Chris@42 13 </h1>
Chris@42 14
Chris@42 15 <ul>
Chris@42 16 <li><a href="#rfftwndbug" rel=subdocument>Q5.1. FFTW 1.1 crashes in rfftwnd on Linux.</a>
Chris@42 17 <li><a href="#fftwmpibug" rel=subdocument>Q5.2. The MPI transforms in FFTW 1.2 give incorrect results/leak
Chris@42 18 memory.</a>
Chris@42 19 <li><a href="#testsingbug" rel=subdocument>Q5.3. The test programs in FFTW 1.2.1 fail when I change FFTW to use single
Chris@42 20 precision.</a>
Chris@42 21 <li><a href="#teststoobig" rel=subdocument>Q5.4. The test program in FFTW 1.2.1 fails for n &gt;
Chris@42 22 46340.</a>
Chris@42 23 <li><a href="#linuxthreads" rel=subdocument>Q5.5. The threaded code fails on Linux Redhat 5.0</a>
Chris@42 24 <li><a href="#bigrfftwnd" rel=subdocument>Q5.6. FFTW 2.0's rfftwnd fails for rank &gt; 1 transforms with a final
Chris@42 25 dimension &gt;= 65536.</a>
Chris@42 26 <li><a href="#primebug" rel=subdocument>Q5.7. FFTW 2.0's complex transforms give the wrong results with prime
Chris@42 27 factors 17 to 97.</a>
Chris@42 28 <li><a href="#mpichbug" rel=subdocument>Q5.8. FFTW 2.1.1's MPI test programs crash with
Chris@42 29 MPICH.</a>
Chris@42 30 <li><a href="#aixthreadbug" rel=subdocument>Q5.9. FFTW 2.1.2's multi-threaded transforms don't work on
Chris@42 31 AIX.</a>
Chris@42 32 <li><a href="#bigprimebug" rel=subdocument>Q5.10. FFTW 2.1.2's complex transforms give incorrect results for large prime
Chris@42 33 sizes.</a>
Chris@42 34 <li><a href="#solaristhreadbug" rel=subdocument>Q5.11. FFTW 2.1.3's multi-threaded transforms don't give any speedup on
Chris@42 35 Solaris.</a>
Chris@42 36 <li><a href="#aixflags" rel=subdocument>Q5.12. FFTW 2.1.3 crashes on AIX.</a>
Chris@42 37 </ul><hr>
Chris@42 38
Chris@42 39 <h2><A name="rfftwndbug">
Chris@42 40 Question 5.1. FFTW 1.1 crashes in rfftwnd on
Chris@42 41 Linux.
Chris@42 42 </A></h2>
Chris@42 43
Chris@42 44 This bug was fixed in FFTW 1.2. There was a bug in
Chris@42 45 <code>rfftwnd</code> causing an incorrect amount of memory to be allocated. The bug showed
Chris@42 46 up in Linux with libc-5.3.12 (and nowhere else that we know of).
Chris@42 47
Chris@42 48 <h2><A name="fftwmpibug">
Chris@42 49 Question 5.2. The MPI transforms in FFTW 1.2 give incorrect
Chris@42 50 results/leak memory.
Chris@42 51 </A></h2>
Chris@42 52
Chris@42 53 These bugs were corrected in FFTW 1.2.1. The MPI transforms (really,
Chris@42 54 just the transpose routines) in FFTW 1.2 had bugs that could cause
Chris@42 55 errors in some situations.
Chris@42 56 <h2><A name="testsingbug">
Chris@42 57 Question 5.3. The test programs in FFTW 1.2.1 fail when I change FFTW
Chris@42 58 to use single precision.
Chris@42 59 </A></h2>
Chris@42 60
Chris@42 61 This bug was fixed in FFTW 1.3. (Older versions of FFTW did
Chris@42 62 work in single precision, but the test programs didn't--the error
Chris@42 63 tolerances in the tests were set for double precision.)
Chris@42 64
Chris@42 65 <h2><A name="teststoobig">
Chris@42 66 Question 5.4. The test program in FFTW 1.2.1 fails for n &gt;
Chris@42 67 46340.
Chris@42 68 </A></h2>
Chris@42 69
Chris@42 70 This bug was fixed in FFTW 1.3. FFTW 1.2.1 produced the right answer,
Chris@42 71 but the test program was wrong. For large n, n*n in the naive
Chris@42 72 transform that we used for comparison overflows 32 bit integer
Chris@42 73 precision, breaking the test.
Chris@42 74 <h2><A name="linuxthreads">
Chris@42 75 Question 5.5. The threaded code fails on Linux Redhat
Chris@42 76 5.0
Chris@42 77 </A></h2>
Chris@42 78
Chris@42 79 We had problems with glibc-2.0.5. The code should work with
Chris@42 80 glibc-2.0.7.
Chris@42 81 <h2><A name="bigrfftwnd">
Chris@42 82 Question 5.6. FFTW 2.0's rfftwnd fails for rank &gt; 1 transforms
Chris@42 83 with a final dimension &gt;= 65536.
Chris@42 84 </A></h2>
Chris@42 85
Chris@42 86 This bug was fixed in FFTW 2.0.1. (There was a 32-bit integer
Chris@42 87 overflow due to a poorly-parenthesized expression.)
Chris@42 88 <h2><A name="primebug">
Chris@42 89 Question 5.7. FFTW 2.0's complex transforms give the wrong results
Chris@42 90 with prime factors 17 to 97.
Chris@42 91 </A></h2>
Chris@42 92
Chris@42 93 There was a bug in the complex transforms that could cause incorrect
Chris@42 94 results under (hopefully rare) circumstances for lengths with
Chris@42 95 intermediate-size prime factors (17-97). This bug was fixed in FFTW
Chris@42 96 2.1.1.
Chris@42 97 <h2><A name="mpichbug">
Chris@42 98 Question 5.8. FFTW 2.1.1's MPI test programs crash with
Chris@42 99 MPICH.
Chris@42 100 </A></h2>
Chris@42 101
Chris@42 102 This bug was fixed in FFTW 2.1.2. The 2.1/2.1.1 MPI test programs
Chris@42 103 crashed when using the MPICH implementation of MPI with the
Chris@42 104 <code>ch_p4</code> device (TCP/IP); the transforms themselves worked fine.
Chris@42 105
Chris@42 106 <h2><A name="aixthreadbug">
Chris@42 107 Question 5.9. FFTW 2.1.2's multi-threaded transforms don't work on
Chris@42 108 AIX.
Chris@42 109 </A></h2>
Chris@42 110
Chris@42 111 This bug was fixed in FFTW 2.1.3. The multi-threaded transforms in
Chris@42 112 previous versions didn't work with AIX's
Chris@42 113 <code>pthreads</code> implementation, which idiosyncratically creates threads in detached
Chris@42 114 (non-joinable) mode by default.
Chris@42 115 <h2><A name="bigprimebug">
Chris@42 116 Question 5.10. FFTW 2.1.2's complex transforms give incorrect results
Chris@42 117 for large prime sizes.
Chris@42 118 </A></h2>
Chris@42 119
Chris@42 120 This bug was fixed in FFTW 2.1.3. FFTW's complex-transform algorithm
Chris@42 121 for prime sizes (in versions 2.0 to 2.1.2) had an integer overflow
Chris@42 122 problem that caused incorrect results for many primes greater than
Chris@42 123 32768 (on 32-bit machines). (Sizes without large prime factors are
Chris@42 124 not affected.)
Chris@42 125 <h2><A name="solaristhreadbug">
Chris@42 126 Question 5.11. FFTW 2.1.3's multi-threaded transforms don't give any
Chris@42 127 speedup on Solaris.
Chris@42 128 </A></h2>
Chris@42 129
Chris@42 130 This bug was fixed in FFTW 2.1.4. (By default, Solaris creates
Chris@42 131 threads that do not parallelize over multiple processors, so one has
Chris@42 132 to request the proper behavior specifically.)
Chris@42 133
Chris@42 134 <h2><A name="aixflags">
Chris@42 135 Question 5.12. FFTW 2.1.3 crashes on AIX.
Chris@42 136 </A></h2>
Chris@42 137
Chris@42 138 The FFTW 2.1.3 <code>configure</code> script picked incorrect compiler flags for the <code>xlc</code> compiler on newer IBM processors. This
Chris@42 139 is fixed in FFTW 2.1.4. <hr>
Chris@42 140 Back: <a href="section4.html" rev=precedes>Internals of FFTW</a>.<br>
Chris@42 141 <a href="index.html" rev=subdocument>Return to contents</a>.<p>
Chris@42 142 <address>
Chris@42 143 <A href="http://www.fftw.org">Matteo Frigo and Steven G. Johnson</A> / <A href="mailto:fftw@fftw.org">fftw@fftw.org</A>
Chris@42 144 - 30 July 2016
Chris@42 145 </address><br>
Chris@42 146 Extracted from FFTW Frequently Asked Questions with Answers,
Chris@42 147 Copyright &copy; 2016 Matteo Frigo and Massachusetts Institute of Technology.
Chris@42 148 </body></html>