FFmpeg

Mirror of https://git.ffmpeg.org/ffmpeg.git https://ffmpeg.org/

Martin Storsjö 388f6e6715 arm: vp9itxfm: Skip empty slices in the first pass of idct_idct 16x16 and 32x32 This work is sponsored by, and copyright, Google. Previously all subpartitions except the eob=1 (DC) case ran with the same runtime: Cortex A7 A8 A9 A53 vp9_inv_dct_dct_16x16_sub16_add_neon: 3188.1 2435.4 2499.0 1969.0 vp9_inv_dct_dct_32x32_sub32_add_neon: 18531.7 16582.3 14207.6 12000.3 By skipping individual 4x16 or 4x32 pixel slices in the first pass, we reduce the runtime of these functions like this: vp9_inv_dct_dct_16x16_sub1_add_neon: 274.6 189.5 211.7 235.8 vp9_inv_dct_dct_16x16_sub2_add_neon: 2064.0 1534.8 1719.4 1248.7 vp9_inv_dct_dct_16x16_sub4_add_neon: 2135.0 1477.2 1736.3 1249.5 vp9_inv_dct_dct_16x16_sub8_add_neon: 2446.7 1828.7 1993.6 1494.7 vp9_inv_dct_dct_16x16_sub12_add_neon: 2832.4 2118.3 2266.5 1735.1 vp9_inv_dct_dct_16x16_sub16_add_neon: 3211.7 2475.3 2523.5 1983.1 vp9_inv_dct_dct_32x32_sub1_add_neon: 756.2 456.7 862.0 553.9 vp9_inv_dct_dct_32x32_sub2_add_neon: 10682.2 8190.4 8539.2 6762.5 vp9_inv_dct_dct_32x32_sub4_add_neon: 10813.5 8014.9 8518.3 6762.8 vp9_inv_dct_dct_32x32_sub8_add_neon: 11859.6 9313.0 9347.4 7514.5 vp9_inv_dct_dct_32x32_sub12_add_neon: 12946.6 10752.4 10192.2 8280.2 vp9_inv_dct_dct_32x32_sub16_add_neon: 14074.6 11946.5 11001.4 9008.6 vp9_inv_dct_dct_32x32_sub20_add_neon: 15269.9 13662.7 11816.1 9762.6 vp9_inv_dct_dct_32x32_sub24_add_neon: 16327.9 14940.1 12626.7 10516.0 vp9_inv_dct_dct_32x32_sub28_add_neon: 17462.7 15776.1 13446.2 11264.7 vp9_inv_dct_dct_32x32_sub32_add_neon: 18575.5 17157.0 14249.3 12015.1 I.e. in general a very minor overhead for the full subpartition case due to the additional loads and cmps, but a significant speedup for the cases when we only need to process a small part of the actual input data. In common VP9 content in a few inspected clips, 70-90% of the non-dc-only 16x16 and 32x32 IDCTs only have nonzero coefficients in the upper left 8x8 or 16x16 subpartitions respectively. This is cherrypicked from libav commit `9c8bc74c2b`. Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>		8 years ago
..
aarch64	Merge commit 'fec76cd430f3c865183a6e5b4caec0743e055605'	9 years ago
arm	Merge commit '7b1ae0e73ab7f7c5eabc70dbe2e579127c6e154f'	8 years ago
x86	x86inc: Add debug symbols indicating sizes of compiled functions	9 years ago
.gitignore	Split global .gitignore file into per-directory files	9 years ago
Makefile	Merge commit 'f8d17d53957056c053a46f9320fa7ae6fe1479a5'	8 years ago
alacdsp.c	checkasm: add alacdsp tests	9 years ago
bswapdsp.c	checkasm: Explicitly declare function prototypes	9 years ago
checkasm.c	Merge commit 'f8d17d53957056c053a46f9320fa7ae6fe1479a5'	8 years ago
checkasm.h	Merge commit 'f8d17d53957056c053a46f9320fa7ae6fe1479a5'	8 years ago
flacdsp.c	checkasm: add flacdsp decorrelate tests	9 years ago
fmtconvert.c	Merge commit '489e6add4478b0f5717dbf644234c6f3a3baf02c'	9 years ago
h264dsp.c	Merge commit 'e48746deec48e9ff195841bc3266b4e153a878cd'	8 years ago
h264pred.c	Merge commit '41ed7ab45fc693f7d7fc35664c0233f4c32d69bb'	9 years ago
h264qpel.c	Merge commit '711781d7a1714ea4eb0217eb1ba04811978c43d1'	9 years ago
jpeg2000dsp.c	checkasm: add jpeg2000dsp rct_int tests	9 years ago
pixblockdsp.c	tests/checkasm/pixblockdsp: Test 8 byte aligned positions	9 years ago
synth_filter.c	Merge commit '568a4323fbde03665b2b23a98068d02b39121812'	9 years ago
v210enc.c	checkasm: v210: Fix array overwrite	9 years ago
vf_blend.c	checkasm: bench each vf_blend mode once	9 years ago
vf_colorspace.c	tests/checkasm/vf_colorspace: Make bpp_mask const	9 years ago
videodsp.c	checkasm: add videodsp emulated_edge_mc test.	9 years ago
vp8dsp.c	Merge commit '80fbb7becae530167373fe5178966b7d7604306e'	8 years ago
vp9dsp.c	arm: vp9itxfm: Skip empty slices in the first pass of idct_idct 16x16 and 32x32	8 years ago