diff options
| author | Ronald S. Bultje <rsbultje@gmail.com> | 2016-07-13 12:37:21 -0400 |
|---|---|---|
| committer | Ronald S. Bultje <rsbultje@gmail.com> | 2016-07-26 15:59:07 -0400 |
| commit | 726501a34ea4bf2222ca8e856ba6cf107b48465f (patch) | |
| tree | ad4abf901bace5a7e25a75bf5c2916f8f157476a /libavcodec/x86/vp9dsp_init.c | |
| parent | c6e900e9258ab51714ae02136f6bbbb018277303 (diff) | |
| download | ffmpeg-726501a34ea4bf2222ca8e856ba6cf107b48465f.tar.gz | |
vp9: add 32x32 idct AVX2 implementation.
About 1.8x speedup compared to AVX version for full IDCT. Other
sub-IDCT scenarios also see speedups. Full --bench output for
idct_32x32_add_{bpp}_${subidct}_${opt} (50k cycles):
nop: 16.5
vp9_inv_dct_dct_32x32_add_8_1_c: 2284.4
vp9_inv_dct_dct_32x32_add_8_1_sse2: 145.0
vp9_inv_dct_dct_32x32_add_8_1_ssse3: 137.4
vp9_inv_dct_dct_32x32_add_8_1_avx: 137.1
vp9_inv_dct_dct_32x32_add_8_1_avx2: 73.2
vp9_inv_dct_dct_32x32_add_8_2_c: 14680.8
vp9_inv_dct_dct_32x32_add_8_2_sse2: 2617.2
vp9_inv_dct_dct_32x32_add_8_2_ssse3: 982.9
vp9_inv_dct_dct_32x32_add_8_2_avx: 958.5
vp9_inv_dct_dct_32x32_add_8_2_avx2: 704.2
vp9_inv_dct_dct_32x32_add_8_4_c: 14443.1
vp9_inv_dct_dct_32x32_add_8_4_sse2: 2717.1
vp9_inv_dct_dct_32x32_add_8_4_ssse3: 965.7
vp9_inv_dct_dct_32x32_add_8_4_avx: 1000.7
vp9_inv_dct_dct_32x32_add_8_4_avx2: 717.1
vp9_inv_dct_dct_32x32_add_8_8_c: 14436.4
vp9_inv_dct_dct_32x32_add_8_8_sse2: 2671.8
vp9_inv_dct_dct_32x32_add_8_8_ssse3: 1038.5
vp9_inv_dct_dct_32x32_add_8_8_avx: 983.0
vp9_inv_dct_dct_32x32_add_8_8_avx2: 729.4
vp9_inv_dct_dct_32x32_add_8_16_c: 14614.7
vp9_inv_dct_dct_32x32_add_8_16_sse2: 2701.7
vp9_inv_dct_dct_32x32_add_8_16_ssse3: 1334.4
vp9_inv_dct_dct_32x32_add_8_16_avx: 1276.7
vp9_inv_dct_dct_32x32_add_8_16_avx2: 719.5
vp9_inv_dct_dct_32x32_add_8_32_c: 14363.6
vp9_inv_dct_dct_32x32_add_8_32_sse2: 2575.6
vp9_inv_dct_dct_32x32_add_8_32_ssse3: 2633.9
vp9_inv_dct_dct_32x32_add_8_32_avx: 2539.6
vp9_inv_dct_dct_32x32_add_8_32_avx2: 1395.0
Diffstat (limited to 'libavcodec/x86/vp9dsp_init.c')
| -rw-r--r-- | libavcodec/x86/vp9dsp_init.c | 2 |
1 files changed, 2 insertions, 0 deletions
diff --git a/libavcodec/x86/vp9dsp_init.c b/libavcodec/x86/vp9dsp_init.c index 6fa63ae0ec..10751e5836 100644 --- a/libavcodec/x86/vp9dsp_init.c +++ b/libavcodec/x86/vp9dsp_init.c @@ -115,6 +115,7 @@ itxfm_func(idct, idct, 32, ssse3); itxfm_func(idct, idct, 32, avx); itxfm_func(iwht, iwht, 4, mmx); itxfm_func(idct, idct, 16, avx2); +itxfm_func(idct, idct, 32, avx2); #undef itxfm_func #undef itxfm_funcs @@ -384,6 +385,7 @@ av_cold void ff_vp9dsp_init_x86(VP9DSPContext *dsp, int bpp, int bitexact) if (ARCH_X86_64) { #if ARCH_X86_64 && HAVE_AVX2_EXTERNAL dsp->itxfm_add[TX_16X16][DCT_DCT] = ff_vp9_idct_idct_16x16_add_avx2; + dsp->itxfm_add[TX_32X32][DCT_DCT] = ff_vp9_idct_idct_32x32_add_avx2; init_subpel3_32_64(0, put, 8, avx2); init_subpel3_32_64(1, avg, 8, avx2); #endif |