libwebp/src/dsp
Vincent Rabaud 8f6df1d0b9 Unroll Predictors 10, 11 and 12.
We see the following speed-ups:
10 -> 13%
11 -> 13%
12 -> 13%

Change-Id: I4734fd388d0f4e508884d0b123976bf2cbe69d2f
2017-06-08 20:37:47 +02:00
..
alpha_processing_mips_dsp_r2.c dsp: add DSP_INIT_STUB 2015-04-02 23:55:35 -07:00
alpha_processing_neon.c dsp: WebPExtractGreen function for alpha decompression 2017-01-17 09:33:25 +00:00
alpha_processing_sse2.c 40% faster ApplyAlphaMultiply_SSE2 2017-01-12 02:33:09 -08:00
alpha_processing_sse41.c SSE4.1 implementation of some alpha-processing functions 2015-06-19 14:17:39 -07:00
alpha_processing.c dsp: WebPExtractGreen function for alpha decompression 2017-01-17 09:33:25 +00:00
argb_mips_dsp_r2.c dsp: add DSP_INIT_STUB 2015-04-02 23:55:35 -07:00
argb_sse2.c dsp: add DSP_INIT_STUB 2015-04-02 23:55:35 -07:00
argb.c make the 'last_cpuinfo_used' variable names unique 2015-01-07 23:38:53 -08:00
common_sse2.h Do not use a register array in SSE. 2016-11-24 13:06:44 +01:00
cost_mips32.c src/{dec,enc,utils}: give filenames a unique suffix 2017-01-19 19:09:48 -08:00
cost_mips_dsp_r2.c src/{dec,enc,utils}: give filenames a unique suffix 2017-01-19 19:09:48 -08:00
cost_sse2.c src/{dec,enc,utils}: give filenames a unique suffix 2017-01-19 19:09:48 -08:00
cost.c src/{dec,enc,utils}: give filenames a unique suffix 2017-01-19 19:09:48 -08:00
cpu.c add a kSlowSSSE3 feature for CPUInfo 2017-01-13 06:19:27 +00:00
dec_clip_tables.c When compiling as C++, avoid narrowing warnings. 2017-01-12 13:39:22 +01:00
dec_mips32.c dsp/mips: cosmetics: add whitespace around XSTR macro 2015-09-03 23:09:13 -07:00
dec_mips_dsp_r2.c dsp/mips: cosmetics: add whitespace around XSTR macro 2015-09-03 23:09:13 -07:00
dec_msa.c Add clang build fix for MSA 2017-02-01 17:45:17 +05:30
dec_neon.c src/{dec,enc,utils}: give filenames a unique suffix 2017-01-19 19:09:48 -08:00
dec_sse2.c src/{dec,enc,utils}: give filenames a unique suffix 2017-01-19 19:09:48 -08:00
dec_sse41.c src/{dec,enc,utils}: give filenames a unique suffix 2017-01-19 19:09:48 -08:00
dec.c src/{dec,enc,utils}: give filenames a unique suffix 2017-01-19 19:09:48 -08:00
dsp.h disable WEBP_USE_XXX optimisations when EMSCRIPTEN is defined 2017-02-08 15:44:20 +00:00
enc_avx2.c dsp: add DSP_INIT_STUB 2015-04-02 23:55:35 -07:00
enc_mips32.c src/{dec,enc,utils}: give filenames a unique suffix 2017-01-19 19:09:48 -08:00
enc_mips_dsp_r2.c src/{dec,enc,utils}: give filenames a unique suffix 2017-01-19 19:09:48 -08:00
enc_msa.c Add clang build fix for MSA 2017-02-01 17:45:17 +05:30
enc_neon.c src/{dec,enc,utils}: give filenames a unique suffix 2017-01-19 19:09:48 -08:00
enc_sse2.c move the SSIM calculation code in ssim.c / ssim_sse2.c 2017-02-21 12:53:35 +01:00
enc_sse41.c src/{dec,enc,utils}: give filenames a unique suffix 2017-01-19 19:09:48 -08:00
enc.c move the SSIM calculation code in ssim.c / ssim_sse2.c 2017-02-21 12:53:35 +01:00
filters_mips_dsp_r2.c MIPS update to new Unfilter API 2016-05-05 15:50:34 +02:00
filters_msa.c cosmetics,dsp/*msa.c: associate '*' with the type 2016-07-15 15:40:41 -07:00
filters_neon.c disable GradientUnfilter_NEON 2017-01-25 16:33:26 -08:00
filters_sse2.c update the Unfilter API in dsp to process one row independently 2016-04-21 08:10:45 +00:00
filters.c NEON: implement alpha-filters (horizontal/vertical/gradient) 2017-01-17 15:44:46 +01:00
lossless_common.h VP8LEnc: remove use of BitsLog2Ceiling() 2017-03-20 02:58:16 -07:00
lossless_enc_mips32.c De-VP8L-ize GetEntropUnrefinedHelper. 2016-09-14 13:55:24 +02:00
lossless_enc_mips_dsp_r2.c dsp: add DSP_INIT_STUB 2015-04-02 23:55:35 -07:00
lossless_enc_msa.c Add MSA optimized color transform functions 2016-07-18 13:49:24 +00:00
lossless_enc_neon.c lossless_enc_neon: add VP8LTransformColor 2015-08-05 00:15:13 -07:00
lossless_enc_sse2.c Get code to fully compile in C++. 2017-01-12 18:03:55 +01:00
lossless_enc_sse41.c Call the C function to finish off lossless SSE loops only when necessary. 2016-12-21 14:25:54 +01:00
lossless_enc.c lossless, VP8LTransformColor_C: make sure no overflow happens with colors. 2017-05-24 11:34:40 +02:00
lossless_mips_dsp_r2.c Remove memcpy in lossless decoding. 2016-11-24 17:45:24 +01:00
lossless_msa.c Add clang build fix for MSA 2017-02-01 17:45:17 +05:30
lossless_neon.c slightly faster Predictor #11 in NEON 2016-12-08 07:32:51 -08:00
lossless_sse2.c Unroll Predictors 10, 11 and 12. 2017-06-08 20:37:47 +02:00
lossless.c src/{dec,enc,utils}: give filenames a unique suffix 2017-01-19 19:09:48 -08:00
lossless.h src/{dec,enc,utils}: give filenames a unique suffix 2017-01-19 19:09:48 -08:00
Makefile.am NEON implementation of some Sharp-YUV420 functions 2017-04-18 19:22:37 +02:00
mips_macro.h dsp/mips: cosmetics: add whitespace around XSTR macro 2015-09-03 23:09:13 -07:00
msa_macro.h Add clang build fix for MSA 2017-02-01 17:45:17 +05:30
neon.h NEON: implement alpha-filters (horizontal/vertical/gradient) 2017-01-17 15:44:46 +01:00
rescaler_mips32.c src/{dec,enc,utils}: give filenames a unique suffix 2017-01-19 19:09:48 -08:00
rescaler_mips_dsp_r2.c src/{dec,enc,utils}: give filenames a unique suffix 2017-01-19 19:09:48 -08:00
rescaler_msa.c src/{dec,enc,utils}: give filenames a unique suffix 2017-01-19 19:09:48 -08:00
rescaler_neon.c src/{dec,enc,utils}: give filenames a unique suffix 2017-01-19 19:09:48 -08:00
rescaler_sse2.c src/{dec,enc,utils}: give filenames a unique suffix 2017-01-19 19:09:48 -08:00
rescaler.c src/{dec,enc,utils}: give filenames a unique suffix 2017-01-19 19:09:48 -08:00
ssim_sse2.c move the SSIM calculation code in ssim.c / ssim_sse2.c 2017-02-21 12:53:35 +01:00
ssim.c ssim.c: remove dead include 2017-02-21 20:17:19 -08:00
upsampling_mips_dsp_r2.c Sync mips32 and dsp_r2 YUV->RGB code with C verison 2016-06-03 10:42:11 +02:00
upsampling_msa.c Add clang build fix for MSA 2017-02-01 17:45:17 +05:30
upsampling_neon.c Merge "NEON: 5% faster conversion to RGB565 and RGBA4444" 2016-12-22 08:39:01 +00:00
upsampling_sse2.c yuv: rationalize the C/SSE2 function naming 2017-04-21 13:52:25 +02:00
upsampling.c yuv: rationalize the C/SSE2 function naming 2017-04-21 13:52:25 +02:00
yuv_mips32.c Sync mips32 and dsp_r2 YUV->RGB code with C verison 2016-06-03 10:42:11 +02:00
yuv_mips_dsp_r2.c Sync mips32 and dsp_r2 YUV->RGB code with C verison 2016-06-03 10:42:11 +02:00
yuv_neon.c NEON: implement ConvertRGB24ToY/BGR24/ARGB/RGBA32ToUV/ARGBToUV 2017-05-09 15:57:20 +02:00
yuv_sse2.c yuv: rationalize the C/SSE2 function naming 2017-04-21 13:52:25 +02:00
yuv.c NEON: implement ConvertRGB24ToY/BGR24/ARGB/RGBA32ToUV/ARGBToUV 2017-05-09 15:57:20 +02:00
yuv.h yuv: rationalize the C/SSE2 function naming 2017-04-21 13:52:25 +02:00