log

age author description
Tue, 03 Feb 2009 02:04:06 +0000 mas01mc Added warning about deprecated LSH_N_POINT_BITS coding. Only affects indexes built prior to rev 820 multiprobeLSH
Mon, 02 Feb 2009 17:41:02 +0000 mas01mc DELETED LSH_N_POINT_BITS coding. Tracks and points are now identified with a binary search, std::lower_bound(), on *adb->track_offsets return from LSH retrieval. multiprobeLSH
Sat, 31 Jan 2009 16:25:30 +0000 mas01mc Removed redundant memcpy calls from query_loop_queue(). This changed the order of some memory allocation and de-allocation calls for loaded database data. multiprobeLSH
Fri, 30 Jan 2009 19:44:51 +0000 mas01mc Completed audiodb_datum_qpointers_partial(), fixed bug that was scribbling over pre-computed l2norms when points arrived in original less<PointPair> order, switched to greater<PointPair> order and all is fixed. multiprobeLSH
Fri, 30 Jan 2009 14:58:35 +0000 mas01mc Fixed a problem with inserting into exact_evaluation_queue, duplicate pairs were being inserted when querying multiple hashtables. That is now fixed by adding a unique constraint using set<PointPair> inserts and testing. This should speed things up quite a bit (like, 10-fold with default num tables). multiprobeLSH
Thu, 29 Jan 2009 13:44:47 +0000 mas01mc Added audiodb_datum_qpointers_partial() to make exact evaluation of LSH results more efficient. This routine calculates partial sums only for those vectors needed. multiprobeLSH
Wed, 28 Jan 2009 18:55:46 +0000 mas01mc Added more debugging information: report numPairs returned from LSH bucket retrieval (pre-distance computation), entry/exit from audiodb_query_spec_qpointers(...) multiprobeLSH
Wed, 28 Jan 2009 17:54:18 +0000 mas01mc Set ADB_FIXME_DEFAULT_LSH_N_POINT_BITS to 15 so we can get about 20k vectors per file in the LSH point encoding. This pain and suffering will go away soon, I promise. multiprobeLSH
Wed, 28 Jan 2009 17:18:58 +0000 mas01mc _LSH_DEBUG_ option now reports correct point counts per hashtable read. multiprobeLSH
Wed, 28 Jan 2009 16:02:17 +0000 mas01mc Fixed a bug in LSH hashtable writing to disk that doesn't always sort the t2 entries into strict weak ordering. Now it does. Lots of debugging informational code inserted. multiprobeLSH
Wed, 28 Jan 2009 05:18:14 +0000 mas01mc insert some statistics gathering for hash tables, add _LSH_DEBUG_ macro to output LSH statistics on INDEX and QUERY multiprobeLSH
Tue, 27 Jan 2009 14:52:28 +0000 mas01mc Fixed upper-limit (T) boundary error in MultiProbe::generatePerturbationSets(x, T). Setting this too high spins algorithm1 into infinite heap allocations without possiblity of terminating. This is now silently capped at --lsh_k * 2; the algorithm halts up to this threshold. multiprobeLSH
Tue, 27 Jan 2009 03:49:19 +0000 mas01mc Working version of multiprobe LSH. Passes tests. But needs testing vis-a-vis TP improvement. In the process, found a bug in hash function projections on rare occasions admitting negative values. This only shows for large datasets and was fatal for multiprobe. multiprobeLSH
Mon, 26 Jan 2009 13:19:09 +0000 mas01mc Fixed boundary {-1,+1} order, array syntax in generate_multiprobe_keys(), boundary error in MultiProbe::valid() multiprobeLSH
Mon, 26 Jan 2009 05:42:15 +0000 mas01mc lshlib MultiProbe glue added. Compiles and links against audioDB. Fails LSH tests at the moment. multiprobeLSH
Mon, 26 Jan 2009 02:50:44 +0000 mas01mc Completed initial implementation of LSH MultiProbe class. Now needs to be joined to lshlib.cpp retrieve() method to perform multi-probe queries. multiprobeLSH
Sun, 25 Jan 2009 06:10:38 +0000 mas01mc Completed multiprobe framework for LSH. Requires testing. multiprobeLSH
Sat, 24 Jan 2009 19:51:46 +0000 mas01mc New branch to implement multiprobe LSH. Copy of trunk:802. Added multiprobe.{cpp,h} source files. multiprobeLSH
Sat, 24 Jan 2009 09:39:39 +0000 mas01cr in audioDB::query, open the database file with O_RDONLY.
Sat, 24 Jan 2009 06:16:00 +0000 mas01mc Removed redundant copy in LSH hashfunction computation when #!defined USE_U_FUNCTIONS. Moved LSH macro definitions into lshlib.h
Fri, 23 Jan 2009 21:44:32 +0000 mas01mc Added hook for dumping LSH INDEX tables on load via compile-time switch: -DLSH_DUMP_CORE_TABLES
Fri, 23 Jan 2009 18:45:44 +0000 mas01mc Implemented correct LSH table via compile-time switch -DLSH_DUMP_CORE_TABLES. Dumps on LSH load. memory-leaks
Wed, 21 Jan 2009 21:48:25 +0000 mas01cr Initial checkin of WIP sb-alien bindings.
Thu, 15 Jan 2009 14:23:09 +0000 mas01cr Add missing pointpair.h file.
Thu, 15 Jan 2009 13:57:33 +0000 mas01cr Code rearrangements to tease apart library code from C++ audioDB code.
Tue, 13 Jan 2009 21:37:14 +0000 mas01cr Better soap memory correctness.
Tue, 13 Jan 2009 21:37:10 +0000 mas01cr Free bits of the datum constructed in audioDB::query.
Tue, 13 Jan 2009 21:37:06 +0000 mas01cr Makefile fixup.
Tue, 13 Jan 2009 21:26:33 +0000 mas01cr Remove all those "with library" in the libtests/ short-descriptions
Tue, 13 Jan 2009 21:26:26 +0000 mas01cr Remove vestiges of old allowed_keys implementation of restrict list
Tue, 13 Jan 2009 21:26:21 +0000 mas01cr Makefile improvements from library/application separation.
Sun, 11 Jan 2009 04:24:24 +0000 mas01mc generalized LSH_N_POINT_BITS compile-time option. Set to <=15 or omit for default (backwards-compatible) behaviour. memory-leaks
Sun, 11 Jan 2009 03:57:01 +0000 mas01mc revert QueryADB.py server port to 14475 memory-leaks
Sat, 10 Jan 2009 20:01:53 +0000 mas01mc Fixed bug in LSH_N_POINT_BITS coding when using AudioDB with LSH indexes memory-leaks
Sat, 10 Jan 2009 19:01:01 +0000 mas01mc Changed LSH_N_POINTS_BIT header field to a width of 5 bits malcolm-large-tracks
Sat, 10 Jan 2009 16:47:57 +0000 mas01cr Merge api-inversion branch (-r656:771, but I don't expect to return to
Sat, 10 Jan 2009 16:07:43 +0000 mas01cr More libtests fixups. api-inversion
Sat, 10 Jan 2009 15:33:25 +0000 mas01cr Mostly disentangle API from command-line binary. api-inversion
Sat, 10 Jan 2009 15:33:21 +0000 mas01cr libtests cleanups api-inversion
Sat, 10 Jan 2009 15:33:16 +0000 mas01cr Finish rewriting libtests. api-inversion
Sat, 10 Jan 2009 15:33:11 +0000 mas01cr Even more audiodb_query() deletion api-inversion
Sat, 10 Jan 2009 15:33:07 +0000 mas01cr Yet more audiodb_query() deletions. api-inversion
Sat, 10 Jan 2009 15:33:02 +0000 mas01cr More removal of audiodb_query() api-inversion
Sat, 10 Jan 2009 15:32:58 +0000 mas01cr Continue removing uses of audiodb_query() api-inversion
Sat, 10 Jan 2009 15:32:53 +0000 mas01cr Begin removing uses of audiodb_query() api-inversion
Sat, 10 Jan 2009 15:32:49 +0000 mas01cr Remove dump_query() function. api-inversion
Sat, 10 Jan 2009 15:32:42 +0000 mas01cr Begin cleaning up libtests api-inversion
Sat, 10 Jan 2009 13:38:22 +0000 mas01mc plugged a file descriptor leak in index.cpp memory-leaks
Sat, 10 Jan 2009 13:35:27 +0000 mas01mc NON BACKWARD-COMPATIBLE indexing change: Modified N_POINT_BITS encoding in audioDB database file header. Now requires a 5 bit field, MSB of 32 bit flag field. memory-leaks
Sat, 10 Jan 2009 13:31:10 +0000 mas01mc Added adb.query.req.xml SOAP call to QueryADB.py memory-leaks
Sat, 10 Jan 2009 13:24:36 +0000 mas01mc added virtual keyword to destructors in lshlib.h (for future use) memory-leaks
Sat, 10 Jan 2009 13:23:41 +0000 mas01mc plugged IndexName memory leak in lshlib.cpp memory-leaks
Sat, 10 Jan 2009 13:23:01 +0000 mas01mc added file descriptor close() in lshlib.cpp memory-leaks
Sat, 10 Jan 2009 13:18:24 +0000 mas01mc added destructor call for LSH SINGLETON on server exit memory-leaks
Sat, 10 Jan 2009 13:14:05 +0000 mas01mc added free() call for data_buffer at end of query_loop_points() memory-leaks
Sat, 10 Jan 2009 13:12:52 +0000 mas01mc removed defunct EXACT_MULT from query_loop_points() memory-leaks
Sat, 10 Jan 2009 11:28:51 +0000 mas01mc copy of trunk@746 to new branch memory-leaks
Sat, 10 Jan 2009 11:11:27 +0000 mas01mc copy of trunk to memory-leaks branch pre-api-inversion-merge
Fri, 09 Jan 2009 22:12:57 +0000 mas01mc modified point-bits coding for indexing (BREAKS BACKWARDS COMPATIBILITY). Plugged file-descriptor leak in index.cpp
Fri, 09 Jan 2009 18:05:32 +0000 mas01mc Fixed memory leaks, added WS --no_unit_norming, and removed capping of LSH_N_POINT_BITS to 15 bits, instead allow any number of bits to encode points, remaining bits encode tracks