openjpeg/src/lib, branch floating_point_irreversible_encoding Fork of openjpeg library for JPEG2000 encoding and decoding https://git.carlh.net/cgit/openjpeg/atom?h=floating_point_irreversible_encoding 2020-05-24T13:38:48Z T1 encoder: speed-up by aggressive inlining and more cache friendly data organization 2020-05-24T13:38:48Z Even Rouault even.rouault@spatialys.com 2020-05-24T13:38:21Z urn:sha1:1c5627ee7406f84cfb40809b7ac31c63342427df ~ 9% speed improvement seen on 10980x10980 uint16 image, T36JTT_20160914T074612_B02.tif opj_compress time from 17.2s to 15.8s Forward DWT 9-7: major speed up by vectorizing vertical pass 2020-05-22T23:01:05Z Even Rouault even.rouault@spatialys.com 2020-05-22T21:57:51Z urn:sha1:1e931fdb3655c64ab60ea5657f79309331a86485 `bench_dwt -I -encode` times goes from 8.6s to 2.1s Forward DWT 5-3: major speed up by vectorizing vertical pass 2020-05-22T23:01:05Z Even Rouault even.rouault@spatialys.com 2020-05-22T15:50:15Z urn:sha1:a38e970fa59abd796c703ec469e578b09f7ffa33 `bench_dwt -encode` times goes from 7.9s to 1.7s Forward DWT: small code refactoring to allow future improvements for the vertical pass 2020-05-22T14:01:45Z Even Rouault even.rouault@spatialys.com 2020-05-22T13:58:47Z urn:sha1:e69fa09f604bc472f9fbff5c2b2db65e8dbe2418 dwt.c: remove unused typedef 2020-05-22T13:06:29Z Even Rouault even.rouault@spatialys.com 2020-05-22T13:06:29Z urn:sha1:33d3d0de07be710f53940c7548b9f2bd58ff3210 Forward DWT 5x3: performance improvements in horizontal pass, and modest in vertical pass 2020-05-22T13:03:40Z Even Rouault even.rouault@spatialys.com 2020-05-22T13:03:40Z urn:sha1:97b384aecdeea0c286213f5caf6244ec08ea32d7 Forward DWT: small code refactoring to allow future improvements for the horizontal pass 2020-05-22T13:02:33Z Even Rouault even.rouault@spatialys.com 2020-05-21T19:20:19Z urn:sha1:bd5f5ee7dea851aedae630fee094ef3b0ff0c888 Speed-up 9x7 IDWD by ~30% with OPJ_NUM_THREADS=2 2020-05-21T15:21:55Z Even Rouault even.rouault@spatialys.com 2020-05-21T14:54:48Z urn:sha1:45a35223b79dee65a0059f999b690072e829669f "bench_dwt -I" time goes from 2.2s to 1.5s Remove useless + 5U margin in opj_dwt_decode_tile_97() 2020-05-21T13:42:51Z Even Rouault even.rouault@spatialys.com 2020-05-21T09:24:29Z urn:sha1:272b3e0fb2530ca8bffdc8c64f1505a8ff5f6ecc Nothing in code analysis nor test suite shows that this margin is needed. It dates back to commit dbeebe72b9d35f6ff807c21c7f217b569fa894f6 where vector 9x7 decoding was introduced. Speed-up 9x7 IDWD by ~20% 2020-05-21T13:42:51Z Even Rouault even.rouault@spatialys.com 2020-05-21T09:23:00Z urn:sha1:47943daa15983d20ab1ffd0a237125f1493c5658 "bench_dwt -I" time goes from 2.8s to 2.2s