| Commit message (Expand) | Author | Age | Files | Lines |
* | ff_fft_calc_3dn/3dn2/sse: convert intrinsics to inline asm. | Loren Merritt | 2006-08-18 | 3 | -280/+232 |
* | simplify | Michael Niedermayer | 2006-08-18 | 1 | -21/+9 |
* | insufficient alignment | Michael Niedermayer | 2006-08-17 | 1 | -1/+1 |
* | Fix building with --disable-opts but MMX enabled. | Marco Manfredini | 2006-08-13 | 1 | -3/+3 |
* | Support for MacIntel, last part: balign directives | John Dalgliesh | 2006-08-12 | 7 | -36/+36 |
* | vorbis simd tweaks | Loren Merritt | 2006-08-11 | 1 | -51/+67 |
* | convert vector_fmul_reverse_sse2 and vector_fmul_add_add_sse2 to sse | Michael Niedermayer | 2006-08-10 | 1 | -22/+22 |
* | sse and 3dnow implementations of float->int conversion and mdct windowing. | Loren Merritt | 2006-08-10 | 1 | -3/+210 |
* | Fix x86 SIMD asm and pic, patch from Martin von Gagern <Martin.vGagern@gmx.net> | Luca Barbato | 2006-08-10 | 1 | -30/+32 |
* | Support for MacIntel, take xx: '/nop' illegal for old versions of GAS | John Dalgliesh | 2006-08-10 | 1 | -76/+54 |
* | Add support for Mac OS X Intel part 2: Assembler macros in fdct_mmx.c | John Dalgliesh | 2006-08-10 | 1 | -53/+53 |
* | Support for Mac OS X Intel, part 3: binary integer constants: | John Dalgliesh | 2006-08-10 | 1 | -5/+5 |
* | emms -> femms | Loren Merritt | 2006-08-09 | 2 | -2/+2 |
* | gcc 2.95 and 3.4.x on x86 32bit without fomit-frame-pointer can't even find 5... | Loren Merritt | 2006-08-09 | 1 | -10/+16 |
* | slightly faster ff_imdct_calc_3dn2() on amd64. (gcc added a bunch of useless ... | Loren Merritt | 2006-08-08 | 1 | -1/+1 |
* | change vorbis_inverse_coupling_sse2() so it works on sse1 cpus | Michael Niedermayer | 2006-08-08 | 1 | -9/+17 |
* | 3dnow2 implementation of imdct. | Loren Merritt | 2006-08-08 | 1 | -1/+82 |
* | 10l, vorbis_inverse_coupling_sse() was really 3dnow | Loren Merritt | 2006-08-03 | 1 | -3/+3 |
* | sse & sse2 implementations of vorbis channel coupling. | Loren Merritt | 2006-08-03 | 1 | -0/+58 |
* | some MMX optimizations for the CAVS decoder | Stefan Gehrer | 2006-07-29 | 2 | -0/+538 |
* | disable the vp3 mmx and sse2 idcts, their output doesnt match the c idct (tes... | Michael Niedermayer | 2006-07-19 | 1 | -0/+2 |
* | #define SBUTTERFLY outside CONFIG_ENCODERS | Måns Rullgård | 2006-07-05 | 1 | -5/+5 |
* | Move REG_* macros from libavcodec/i386/mmx.h to libavutil/x86_cpu.h | Luca Abeni | 2006-07-03 | 5 | -21/+4 |
* | remove redundant macro definitions | Måns Rullgård | 2006-07-02 | 1 | -6/+0 |
* | kill warning | Måns Rullgård | 2006-07-02 | 1 | -0/+1 |
* | missmatch control for mpeg2 intra dequantization if bitexact=1 | Michael Niedermayer | 2006-04-28 | 1 | -1/+2 |
* | Remove unused and unsupported Cyrix's "Extended MMX", | Zuxy Meng | 2006-04-28 | 1 | -30/+4 |
* | gcc-2.95 compile fix, patch by Wolfram Gloger <wmglo A dent PIS med PIS uni-m... | Wolfram Gloger | 2006-04-18 | 1 | -2/+2 |
* | gmc_mmx tweaks | Loren Merritt | 2006-04-05 | 1 | -13/+11 |
* | mmx implementation of 3-point GMC. (5x faster than C) | Loren Merritt | 2006-04-04 | 1 | -0/+122 |
* | altivec support for snow | Luca Barbato | 2006-03-27 | 1 | -29/+0 |
* | simplified and slightly faster h264_chroma_mc8_mmx | Loren Merritt | 2006-03-25 | 1 | -128/+68 |
* | prefetch pixels for future motion compensation. 2-5% faster h264. | Loren Merritt | 2006-03-23 | 1 | -0/+16 |
* | 10l | Loren Merritt | 2006-03-22 | 1 | -1/+1 |
* | added mmx implementation of h264_chroma_mc2 | Loren Merritt | 2006-03-22 | 3 | -1/+70 |
* | add MMX and SSE versions of ff_snow_inner_add_yblock | Robert Edele | 2006-03-22 | 2 | -0/+297 |
* | snow mmx+sse2 optimizations, part 4 | Robert Edele | 2006-03-21 | 2 | -0/+412 |
* | Add the mmx and sse2 implementations of ff_snow_vertical_compose(). | Robert Edele | 2006-03-20 | 2 | -0/+261 |
* | First part of a series of speed-enchancing patches. | Robert Edele | 2006-03-16 | 1 | -0/+2 |
* | 3DNow! & Extended 3DNow! versions of FFT | Zuxy Meng | 2006-03-08 | 2 | -0/+272 |
* | h264_idct8_add_mmx | Loren Merritt | 2006-03-07 | 2 | -0/+129 |
* | h264_idct_add only needs mmx1 | Loren Merritt | 2006-03-07 | 2 | -4/+6 |
* | use xorps instead of mulps to toggle the sign of a float, as suggested by Sof... | Zuxy Meng | 2006-03-05 | 1 | -10/+10 |
* | gcc2.95 workaround | Loren Merritt | 2006-03-05 | 1 | -22/+24 |
* | remove some useless instructions | Loren Merritt | 2006-03-04 | 1 | -18/+4 |
* | 4% faster h264_qpel_mc | Loren Merritt | 2006-03-02 | 1 | -40/+137 |
* | h264: special case dc-only idct. ~1% faster overall | Loren Merritt | 2006-02-10 | 2 | -0/+83 |
* | 10l in 1.12 | Loren Merritt | 2006-02-09 | 1 | -6/+6 |
* | 10l (x86_64) | Loren Merritt | 2006-02-07 | 1 | -1/+1 |
* | 18% faster put_h264_qpel16_mc[13]2_mmx2 | Loren Merritt | 2006-02-07 | 1 | -14/+87 |