diff options
| author | James Darnley <jdarnley@obe.tv> | 2017-04-05 21:07:47 +0200 |
|---|---|---|
| committer | James Darnley <jdarnley@obe.tv> | 2017-05-15 15:00:20 +0200 |
| commit | 7aa90b4e94147d0512ab28535f6863767b888f19 (patch) | |
| tree | 0409c2abfa881d4268752a22ecde6353b1895ea6 /libavcodec/x86/h264_idct.asm | |
| parent | 27460dfebc296636dec2584e0d74aaa4d48da0b9 (diff) | |
| download | ffmpeg-7aa90b4e94147d0512ab28535f6863767b888f19.tar.gz | |
avcodec/h264: add sse2 versions of previous idct functions
Kaby Lake Pentium:
- ff_h264_idct_add_8_sse2: ~1.18x faster than mmxext
- ff_h264_idct_dc_add_8_sse2: ~1.07x faster than mmxext
Diffstat (limited to 'libavcodec/x86/h264_idct.asm')
| -rw-r--r-- | libavcodec/x86/h264_idct.asm | 11 |
1 files changed, 9 insertions, 2 deletions
diff --git a/libavcodec/x86/h264_idct.asm b/libavcodec/x86/h264_idct.asm index 01078589db..ea91e1ac4d 100644 --- a/libavcodec/x86/h264_idct.asm +++ b/libavcodec/x86/h264_idct.asm @@ -1140,8 +1140,6 @@ IDCT_DC_DEQUANT 0 INIT_MMX sse2 IDCT_DC_DEQUANT 7 -INIT_XMM avx - ; %unmacro STORE_DIFFx2 8 ; remove macro from x86util.asm but yasm doesn't have this yet %macro STORE_DIFFx2 8 ; add1, add2, reg1, reg2, zero, shift, source, stride movd %3, [%7] @@ -1170,6 +1168,10 @@ INIT_XMM avx packuswb m1, m1 %endmacro +%macro IDCT_XMM 1 + +INIT_XMM %1 + cglobal h264_idct_add_8, 3, 3, 8, dst_, block_, stride_ movsxdifnidn stride_q, stride_d IDCT4_ADD dst_q, block_q, stride_q @@ -1182,3 +1184,8 @@ cglobal h264_idct_dc_add_8, 3, 4, 6, dst_, block_, stride_ DC_ADD_INIT r3 DC_ADD_MMXEXT_OP movd, dst_q, stride_q, r3 RET + +%endmacro + +IDCT_XMM sse2 +IDCT_XMM avx |