| Commit message (Expand) | Author | Age | Files | Lines |
* | Add d modifier to asm argument to fix nasm compilation. | Reimar Döffinger | 2010-10-07 | 1 | -1/+1 |
* | fft: mark xmm registers as clobbered in ff_imdct_calc_sse | Ramiro Polla | 2010-10-06 | 1 | -0/+1 |
* | MMX, MMX2, SSE2 and SSSE3 optimizations for pred16x16/8x8_plane H264 intra | Ronald S. Bultje | 2010-10-05 | 3 | -0/+557 |
* | snowdsp: Explicitly state the operand sizes | İsmail Dönmez | 2010-10-04 | 1 | -1/+1 |
* | Move static inline function to a macro, so that constant propagation in | Ronald S. Bultje | 2010-09-29 | 1 | -117/+113 |
* | Use sse2 variant of put_pixels16() for no_rnd also. Provides a minor speed | Eli Friedman | 2010-09-29 | 1 | -0/+1 |
* | Merge b_idx and edge variables, and optimize the ASM to directly load variables | Ronald S. Bultje | 2010-09-29 | 1 | -46/+54 |
* | Remove mv_mask variable. Replace the related pand -1/0 instructions by either | Ronald S. Bultje | 2010-09-29 | 1 | -6/+7 |
* | Remove d_idx as a variable, and instead load it as a constant in the asm. | Ronald S. Bultje | 2010-09-29 | 1 | -32/+38 |
* | Unroll inner bidir loop in h264_loop_filter_strength_mmx2(), which gets rid | Ronald S. Bultje | 2010-09-29 | 1 | -5/+19 |
* | Unloop the outer loop in h264_loop_filter_strength_mmx2(), which allows | Ronald S. Bultje | 2010-09-29 | 1 | -25/+29 |
* | Add d suffix to movd target register to make it work with nasm. | Reimar Döffinger | 2010-09-26 | 1 | -2/+2 |
* | Split and then simplify address generation macro. | Reimar Döffinger | 2010-09-26 | 1 | -20/+22 |
* | Remove unused variable. | Ronald S. Bultje | 2010-09-24 | 1 | -1/+0 |
* | Unroll loop in h264_idct_add16intra_sse2(). Basically identical to r25171, this | Ronald S. Bultje | 2010-09-24 | 1 | -28/+28 |
* | Unroll loop in h264_idct_add8_sse2(). This means we can inline scan8[] in the | Ronald S. Bultje | 2010-09-24 | 1 | -29/+20 |
* | x86: disable SSE functions using stack when stack is not aligned | Måns Rullgård | 2010-09-21 | 2 | -2/+4 |
* | x86: remove hack disabling sse2 h264 loop filter with 32-bit icc | Måns Rullgård | 2010-09-18 | 1 | -2/+1 |
* | Don't access upper 32 bits of a 32-bit int on 64-bit systems. | Ronald S. Bultje | 2010-09-17 | 1 | -1/+1 |
* | Properly add HAVE_YASM around yasmified symbols. Should fix compile error | Ronald S. Bultje | 2010-09-17 | 1 | -1/+9 |
* | Move hadamard_diff{,16}_{mmx,mmx2,sse2,ssse3}() from inline asm to yasm, | Ronald S. Bultje | 2010-09-17 | 3 | -212/+291 |
* | Move sse16_sse2() from inline asm to yasm. It is one of the functions causing | Ronald S. Bultje | 2010-09-17 | 3 | -62/+90 |
* | Rename h264_idct_sse2.asm to h264_idct.asm; move inline IDCT asm from | Ronald S. Bultje | 2010-09-14 | 4 | -584/+905 |
* | LGPL SSE2 H.264 iDCT | Jason Garrett-Glaser | 2010-09-10 | 3 | -14/+14 |
* | Move mm_support() from libavcodec to libavutil, make it a public | Stefano Sabatini | 2010-09-08 | 16 | -167/+26 |
* | Use "d" suffix for general-purpose registers used with movd. | Reimar Döffinger | 2010-09-05 | 4 | -30/+30 |
* | Rename FF_MM_ symbols related to CPU features flags as AV_CPU_FLAG_ | Stefano Sabatini | 2010-09-04 | 15 | -106/+108 |
* | Port latest x264 deblock asm (before they moved to using NV12 as internal | Ronald S. Bultje | 2010-09-03 | 5 | -325/+230 |
* | Fix typo in r25019. | Eli Friedman | 2010-09-01 | 1 | -1/+1 |
* | Unscrew breakage after my last commit because of symbol prefixes. | Ronald S. Bultje | 2010-09-01 | 1 | -8/+8 |
* | Rename h264_weight_sse2.asm to h264_weight.asm; add 16x8/8x16/8x4 non-square | Ronald S. Bultje | 2010-09-01 | 4 | -291/+427 |
* | Split h264dsp_mmx.c (which was #included in dsputil_mmx.c) in h264_qpel_mmx.c, | Ronald S. Bultje | 2010-09-01 | 5 | -1319/+1343 |
* | Fix vertical align. | Ronald S. Bultje | 2010-08-31 | 1 | -1/+1 |
* | Fix compilation failure if yasm is disabled (missing vp3 symbols). | Ronald S. Bultje | 2010-08-30 | 1 | -3/+3 |
* | Split intra prediction initialization (i.e. assigning of function pointers) | Ronald S. Bultje | 2010-08-30 | 3 | -83/+103 |
* | Move H264 chroma MC from inline asm to yasm. This fixes VP3/5/6 and VC-1 | Ronald S. Bultje | 2010-08-30 | 7 | -718/+754 |
* | Move VP3 IDCT functions from inline ASM to YASM. This fixes part of the VP3/5/6 | Ronald S. Bultje | 2010-08-30 | 7 | -701/+637 |
* | Put ff_ prefix on non-static {put_signed,put,add}_pixels_clamped_mmx() | Ronald S. Bultje | 2010-08-30 | 8 | -27/+27 |
* | cosmetics in imdct_sse | Loren Merritt | 2010-08-28 | 1 | -25/+20 |
* | Fix typos when converting inline asm to yasm, fixes MMX-only fate-ea-vp61. | Ronald S. Bultje | 2010-08-26 | 1 | -5/+5 |
* | Revert r24931, it broke Win32 and some BSD compiles (yay fate). | Ronald S. Bultje | 2010-08-25 | 1 | -1/+0 |
* | Mark xmm6 and xmm7 as clobbered in ff_vp3_idct_sse2(), which is contributing | Ronald S. Bultje | 2010-08-25 | 1 | -0/+1 |
* | VP6: fix vp6_filter_diag4_mmx/sse on 64-bit | Måns Rullgård | 2010-08-25 | 1 | -0/+3 |
* | Move vp6_filter_diag4() x86 SIMD code from inline ASM to YASM. This should | Ronald S. Bultje | 2010-08-25 | 7 | -270/+178 |
* | Move vp6_filter_diag4() from DSPContext to VP56DSPContext. | Ronald S. Bultje | 2010-08-25 | 3 | -12/+46 |
* | Remove global mm_flags variable | Måns Rullgård | 2010-08-24 | 10 | -10/+14 |
* | Mark xmm registers as clobbered in simple loopfilter. Should fix the last | Ronald S. Bultje | 2010-08-24 | 1 | -11/+11 |
* | imdct/x86: Use "s->mdct_size" instead of "1 << s->mdct_bits". | Alex Converse | 2010-08-23 | 2 | -3/+3 |
* | Fix segfaults in VP8 SIMD code on Win64 (and FATE/win64 failures). | Ronald S. Bultje | 2010-08-23 | 1 | -14/+14 |
* | Convert ff_imdct_half_sse() to yasm. | Alex Converse | 2010-08-22 | 2 | -108/+195 |