Commit message (Collapse) | Author | Age | |
---|---|---|---|
* | build: conditionally compile x86 H.264 chroma optimizations | Diego Biurrun | 2011-12-14 |
| | |||
* | x86: Require 7 registers for the cabac asm | Martin Storsjö | 2011-12-12 |
| | | | | | | | The change in 599b4c6ef didn't turn out to work properly on i386 on OS X, where it broke building with PIC enabled. Signed-off-by: Martin Storsjö <martin@martin.st> | ||
* | x86: cabac: replace explicit memory references with "m" operands | Mans Rullgard | 2011-12-11 |
| | | | | | | | | This replaces the explicit offset(reg) memory references with "m" operands for the same locations. As a result, one fewer register operand is needed for these inline asm statements. Signed-off-by: Mans Rullgard <mans@mansr.com> | ||
* | Fix a bunch of common typos. | Diego Biurrun | 2011-12-11 |
| | |||
* | dsputil: use cpuflags in x86 emu_edge_core | Justin Ruggles | 2011-11-22 |
| | | | | avoids passing around the extra argument among all the macros it uses | ||
* | dsputil: use movups instead of movdqu in ff_emu_edge_core_sse() | Justin Ruggles | 2011-11-22 |
| | | | | | This allows emulated_edge_mc_sse() and gmc_sse() to be used under AV_CPU_FLAG_SSE. | ||
* | twinvq: add SSE/AVX optimized sum/difference stereo interleaving | Justin Ruggles | 2011-11-11 |
| | |||
* | Remove redundant filename self-references inside files. | Diego Biurrun | 2011-11-08 |
| | | | | Filenames are brittle across renames and add no useful information. | ||
* | x86: drop pointless ARCH_X86 #ifdef from files in x86 subdirectory | Diego Biurrun | 2011-11-08 |
| | |||
* | dsputil: use cpuflags in x86 versions of vector_clip_int32() | Justin Ruggles | 2011-11-06 |
| | |||
* | h264_weight: remove duplication functions. | Ronald S. Bultje | 2011-11-05 |
| | |||
* | fmtconvert: fix int32_to_float_fmul_scalar() for windows x86_64 | Justin Ruggles | 2011-11-02 |
| | | | | | | | The calling convention only allows 4 non-stack parameter, with each float or int register being skipped if not used. fixes Bug 64 | ||
* | H.264: Cometics to dsputil_mmx.c | Daniel Kang | 2011-10-26 |
| | | | | | | Add whitespace. Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com> | ||
* | h264_weight: initialize "height" function argument properly. | Ronald S. Bultje | 2011-10-22 |
| | | | | | Right now it's not actually initialized on 32-bit, leading to crashes on win32. | ||
* | fmtconvert: port float_to_int16_interleave() 2-channel x86 inline asm to yasm | Justin Ruggles | 2011-10-21 |
| | |||
* | fmtconvert: port int32_to_float_fmul_scalar() x86 inline asm to yasm | Justin Ruggles | 2011-10-21 |
| | |||
* | fmtconvert: check compile-time x86 instruction set flags | Justin Ruggles | 2011-10-21 |
| | |||
* | fmtconvert: port float_to_int16() x86 inline asm to yasm | Justin Ruggles | 2011-10-21 |
| | |||
* | H264: change weight/biweight functions to take a height argument. | Ronald S. Bultje | 2011-10-21 |
| | | | | Neon parts by Mans Rullgard <mans@mansr.com>. | ||
* | Support for lossless and inter H264 4:2:2. | Ronald S. Bultje | 2011-10-21 |
| | |||
* | h264: 4:2:2 intra decoding support | Baptiste Coudurier | 2011-10-21 |
| | | | | | Signed-off-by: Diego Biurrun <diego@biurrun.de> Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com> | ||
* | x86: Move some variable declarations below the appropriat #ifdef. | Diego Biurrun | 2011-10-20 |
| | | | | This avoids some unused variable warnings with YASM disabled. | ||
* | x86: Fix linking of ProRes DSP ASM with YASM disabled. | Diego Biurrun | 2011-10-20 |
| | |||
* | proresdsp: fix function prototypes. | Ronald S. Bultje | 2011-10-14 |
| | | | | Signed-off-by: Janne Grunau <janne-libav@jannau.net> | ||
* | prores: idct sse2/sse4 optimizations. | Ronald S. Bultje | 2011-10-11 |
| | | | | ~3.0-3.5x as fast as original C version, 1.6x as fast overall. | ||
* | fft: avoid a signed overflow | Sean McGovern | 2011-09-23 |
| | | | | | | As a signed integer, 1<<31 overflows, so force it to unsigned. Signed-off-by: Alex Converse <alex.converse@gmail.com> | ||
* | Move clipd macros to x86util.asm. | Ronald S. Bultje | 2011-08-17 |
| | | | | This allows sharing them between multiple .asm files. | ||
* | Fix NASM include directive | Dave Yeo | 2011-08-15 |
| | | | | Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com> | ||
* | dsputil_mmx: Honor HAVE_AMD3DNOW | Alex Converse | 2011-08-15 |
| | |||
* | Move x86util.asm from libavcodec/ to libavutil/. | Ronald S. Bultje | 2011-08-12 |
| | | | | This allows using it in swscale also. | ||
* | Move x86inc.asm to libavutil/. | Ronald S. Bultje | 2011-08-12 |
| | | | | This allows using it in libswscale/ also. | ||
* | Move RV3/4-specific DSP functions into their own context | Kostya Shishkov | 2011-08-11 |
| | | | | Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com> | ||
* | dct32: Add SSE2 ASM optimizations | Vitor Sessak | 2011-08-02 |
| | | | | Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com> | ||
* | H.264: tweak some other x86 asm for Atom | Jason Garrett-Glaser | 2011-07-29 |
| | |||
* | x86: cabac: add operand size suffixes missing from 6c32576 | Mans Rullgard | 2011-07-28 |
| | | | | | | This fixes build with clang. Signed-off-by: Mans Rullgard <mans@mansr.com> | ||
* | x86: cabac: don't load/store context values in asm | Mans Rullgard | 2011-07-28 |
| | | | | | | | | | | | Inspection of compiled code shows gcc handles these fine on its own. Benchmarking also shows no measurable speed difference. Removing the remaining cases in get_cabac_bypass_sign_x86() does cause more substantial changes to the compiled code with uncertain impact. Signed-off-by: Mans Rullgard <mans@mansr.com> | ||
* | H.264: optimize CABAC x86 asm for Atom | Jason Garrett-Glaser | 2011-07-28 |
| | |||
* | x86: fix build with gcc 4.7 | Mans Rullgard | 2011-07-26 |
| | | | | | | | | | | | | | | The upcoming gcc 4.7 has more advanced constant propagation resulting some inline asm operands becoming constants and thus emitted as literals, sometimes in contexts where this results in invalid instructions. This patch changes the constraints of the relevant operands to "rm" thus forcing a valid type. While obviously suboptimal, this is what older gcc versions already did, and there is no change to the code generated with these. Signed-off-by: Mans Rullgard <mans@mansr.com> | ||
* | H.264: Add optimizations to predict x86 assembly. | Daniel Kang | 2011-07-22 |
| | | | | Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com> | ||
* | dnxhd: 10-bit support | Joseph Artsimovich | 2011-07-21 |
| | | | | Signed-off-by: Mans Rullgard <mans@mansr.com> | ||
* | dsputil: update per-arch init funcs for non-h264 high bit depth | Mans Rullgard | 2011-07-21 |
| | | | | Signed-off-by: Mans Rullgard <mans@mansr.com> | ||
* | dsputil: template get_pixels() for different bit depths | Mans Rullgard | 2011-07-21 |
| | | | | Signed-off-by: Mans Rullgard <mans@mansr.com> | ||
* | jfdctint: add 10-bit version | Mans Rullgard | 2011-07-21 |
| | | | | Signed-off-by: Mans Rullgard <mans@mansr.com> | ||
* | simple_idct: add 10-bit version | Mans Rullgard | 2011-07-20 |
| | | | | Signed-off-by: Mans Rullgard <mans@mansr.com> | ||
* | dsputil: remove disabled code | Diego Biurrun | 2011-07-18 |
| | |||
* | x86: Use LOCAL_ALIGNED in mpegvideo_mmx_template | Martin Storsjö | 2011-07-18 |
| | | | | Signed-off-by: Martin Storsjö <martin@martin.st> | ||
* | simple_idct: remove disabled code | Diego Biurrun | 2011-07-17 |
| | |||
* | H.264: Add more x86 assembly for 10-bit H.264 predict functions | Daniel Kang | 2011-07-13 |
| | | | | | | | | Mainly ported from 8-bit H.264 predict. Some code ported from x264. LGPL ok by author. Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com> | ||
* | H.264: add filter_mb_fast support for >8-bit decoding | Jason Garrett-Glaser | 2011-07-11 |
| | | | | Much faster high bit depth deblocking. | ||
* | dsputil: remove ff_emulated_edge_mc macro used in one place | Mans Rullgard | 2011-07-10 |
| | | | | | | | | | | This macro can cause problems in conjunction with the bitdepth template expansion. It was presumably added to keep source compatibility when high bitdepth support was added. However, emulated_edge_mc is a dsputil pointer and should not be called directly, so there is little reason to keep such a macro. Signed-off-by: Mans Rullgard <mans@mansr.com> |