| Commit message (Expand) | Author | Age | Files | Lines |
* | x86: h264_intrapred: Don't add the 'd' suffix to the SPLATB_REG macro | Martin Storsjö | 2012-07-06 | 1 | -1/+1 |
* | x86: h264_intrapred: use newly introduced SPLAT* and PSHUFLW macros | Diego Biurrun | 2012-07-05 | 1 | -107/+22 |
* | x86inc: add SPLATB_LOAD, SPLATB_REG, PSHUFLW macros | Loren Merritt | 2012-07-05 | 1 | -21/+0 |
* | x86: h264_intrapred: port to cpuflag macros | Diego Biurrun | 2012-07-05 | 2 | -146/+168 |
* | vp8: Add ifdef guards around the sse2 loopfilter in the sse2slow branch too | Martin Storsjö | 2012-07-05 | 1 | -0/+2 |
* | vp8: loopfilter >=sse2 functions need aligned stack on x86-32. | Martin Storsjö | 2012-07-04 | 1 | -0/+6 |
* | dsputilenc: group yasm and inline asm function pointer assignment. | Ronald S. Bultje | 2012-07-04 | 1 | -1/+1 |
* | dsputilenc_mmx: split assignment of ff_sse16_sse2 to SSE2 section. | Ronald S. Bultje | 2012-06-30 | 1 | -3/+6 |
* | x86: fmtconvert: add special asm for float_to_int16_interleave_misc_* | Ronald S. Bultje | 2012-06-30 | 2 | -5/+85 |
* | x86: vc1: fix and enable optimised loop filter | Mans Rullgard | 2012-06-30 | 2 | -3/+4 |
* | x86: fft: replace call to memcpy by a loop | Christophe Gisquet | 2012-06-27 | 1 | -25/+12 |
* | x86: fft: elf64: fix PIC build | Mans Rullgard | 2012-06-25 | 1 | -0/+4 |
* | x86: fft: win64: fix stack alignment for memcpy() call | Mans Rullgard | 2012-06-25 | 1 | -7/+9 |
* | x86: fft: convert sse inline asm to yasm | Mans Rullgard | 2012-06-25 | 3 | -121/+129 |
* | x86: place some inline asm under #if HAVE_INLINE_ASM | Ronald S. Bultje | 2012-06-25 | 8 | -0/+39 |
* | h264: use asm cabac reader under a generic condition | Mans Rullgard | 2012-06-23 | 1 | -0/+2 |
* | x86: Only use optimizations with cmov if the CPU supports the instruction | Diego Biurrun | 2012-06-23 | 3 | -3/+6 |
* | x86: remove unused inline asm macros from dsputil_mmx.h | Mans Rullgard | 2012-06-23 | 1 | -68/+0 |
* | x86: move some inline asm macros to the only places they are used | Mans Rullgard | 2012-06-23 | 3 | -34/+34 |
* | cosmetics: do not use full path for local headers | Diego Biurrun | 2012-06-22 | 1 | -1/+1 |
* | dwt: remove variable-length arrays | Ronald S. Bultje | 2012-06-17 | 1 | -4/+2 |
* | Add a float DSP framework to libavutil | Justin Ruggles | 2012-06-08 | 2 | -39/+0 |
* | x86: use new schema for ASM macros | Vitor Sessak | 2012-05-29 | 4 | -76/+82 |
* | x86: lavc: use %if HAVE_AVX guards around AVX functions in yasm code. | Justin Ruggles | 2012-05-22 | 1 | -0/+8 |
* | Convert vector_fmul range of functions to YASM and add AVX versions | Kieran Kunhya | 2012-05-21 | 2 | -137/+128 |
* | x86: rv40: Mark rv40_weight functions as MMX2; they use MMX2 instructions. | Michael Kostylev | 2012-05-15 | 2 | -6/+6 |
* | ac3dsp: simplify x86 versions of ac3_max_msb_abs_int16 | Justin Ruggles | 2012-05-15 | 2 | -26/+43 |
* | x86: use more standard construct for setting ASM functions in FFT code | Vitor Sessak | 2012-05-14 | 1 | -22/+23 |
* | x86: vc1: drop MMX loop filter implementation, which uses MMX2 instructions. | Michael Kostylev | 2012-05-12 | 2 | -13/+0 |
* | rv40dsp x86: MMX/MMX2/3DNow/SSE2/SSSE3 implementations of MC | Christophe Gisquet | 2012-05-10 | 4 | -3/+480 |
* | snowdsp: explicitily state instruction size. | Ronald S. Bultje | 2012-05-02 | 1 | -2/+2 |
* | dsputil x86: revert a test back to its previous value | Christophe GISQUET | 2012-04-28 | 1 | -1/+1 |
* | rv34dsp x86: implement MMX2 inverse transform | Christophe Gisquet | 2012-04-28 | 2 | -0/+80 |
* | h264: new assembly version of get_cabac for x86_64 with PIC | Roland Scheidegger | 2012-04-28 | 2 | -23/+120 |
* | h264: use one table instead of several for cabac functions | Roland Scheidegger | 2012-04-28 | 2 | -14/+28 |
* | h264: (trivial) remove unneeded macro argument in x86/cabac.h | Roland Scheidegger | 2012-04-28 | 1 | -3/+3 |
* | Remove lowres video decoding | Mans Rullgard | 2012-04-21 | 1 | -1/+1 |
* | avcodec: remove AVCodecContext.dsp_mask | Mans Rullgard | 2012-04-21 | 1 | -7/+0 |
* | h264: use proper PROLOGUE statement for a function using 8 registers. | Ronald S. Bultje | 2012-04-16 | 1 | -4/+4 |
* | dsputil: fix optimized emu_edge function on Win64. | Ronald S. Bultje | 2012-04-13 | 1 | -1/+1 |
* | ac3dsp: call femms/emms at the end of float_to_fixed24() for 3DNow and SSE | Justin Ruggles | 2012-04-12 | 1 | -2/+4 |
* | h264: fix 10bit biweight functions after recent x86inc.asm fixes. | Ronald S. Bultje | 2012-04-12 | 1 | -4/+4 |
* | build: Consistently handle conditional compilation for all optimization OBJS. | Diego Biurrun | 2012-04-12 | 1 | -3/+2 |
* | x86inc improvements for 64-bit | Henrik Gramner | 2012-04-11 | 10 | -221/+198 |
* | rv40dsp x86: use only one register, for both increment and loop counter | Christophe GISQUET | 2012-04-10 | 1 | -23/+20 |
* | rv40dsp: implement prescaled versions for biweight. | Christophe GISQUET | 2012-04-10 | 2 | -49/+51 |
* | dsputil x86: use SSE float instruction instead of SSE2 integer equivalent | Christophe GISQUET | 2012-04-04 | 2 | -2/+2 |
* | dsputil x86: remove deprecated parameter from scalarproduct_int16 prototype | Christophe GISQUET | 2012-04-04 | 1 | -2/+2 |
* | vp8dsp x86: perform rounding shift with a single instruction | Christophe GISQUET | 2012-04-04 | 1 | -10/+8 |
* | cabac: add overread protection to BRANCHLESS_GET_CABAC(). | Ronald S. Bultje | 2012-03-28 | 2 | -11/+22 |