67922b4ee4
vp9mc/x86: add AVX and AVX2 MC
...
Roughly 25% faster MC than ssse3 for blocksizes 32 and 64.
Reviewed-by: Ronald S. Bultje <rsbultje@gmail.com>
Signed-off-by: James Almer <jamrial@gmail.com>
Signed-off-by: Anton Khirnov <anton@khirnov.net>
9 years ago
3cda179f18
vp9mc/x86: rename ff_* to ff_vp9_*
...
Signed-off-by: Anton Khirnov <anton@khirnov.net>
9 years ago
8be8444d01
vp9mc/x86: rename ff_avg[48]_sse to ff_avg[48]_mmxext
...
pavgb is an sse integer instruction, so the mmxext flag is enough
Signed-off-by: James Almer <jamrial@gmail.com>
Reviewed-by: "Ronald S. Bultje" <rsbultje@gmail.com>
Signed-off-by: Anton Khirnov <anton@khirnov.net>
9 years ago
6ab642d69d
vp9mc/x86: simplify a few inits.
...
Signed-off-by: Anton Khirnov <anton@khirnov.net>
9 years ago
3a09494939
vp9mc/x86: add 16px functions (64bit only).
...
Signed-off-by: Anton Khirnov <anton@khirnov.net>
9 years ago
89466de4ae
vp9/x86: rename vp9dsp to vp9mc
...
It only contains the MC SIMD, other SIMD will go into different files.
9 years ago
3c504bc359
x86: deduplicate some constants
...
Signed-off-by: Anton Khirnov <anton@khirnov.net>
9 years ago
efc9d5c4bc
x86/ttaenc: add ff_ttaenc_filter_process_{ssse3,sse4}
...
Signed-off-by: James Almer <jamrial@gmail.com>
9 years ago
a4edaa0270
vp9: add mxext versions of the single-block (w=8,npx=8) h/v loopfilters.
...
Each takes about 0.1% of runtime in my profiles, and they didn't have
any SIMD yet so far (we only had simd for npx=16 double-block versions).
9 years ago
7ca422bb1b
vp9: add mxext versions of the single-block (w=4,npx=8) h/v loopfilters.
...
Each takes about 0.5% of runtime in my profiles, and they didn't have
any SIMD yet so far (we only had simd for npx=16 double-block versions).
9 years ago
726501a34e
vp9: add 32x32 idct AVX2 implementation.
...
About 1.8x speedup compared to AVX version for full IDCT. Other
sub-IDCT scenarios also see speedups. Full --bench output for
idct_32x32_add_{bpp}_${subidct}_${opt} (50k cycles):
nop: 16.5
vp9_inv_dct_dct_32x32_add_8_1_c: 2284.4
vp9_inv_dct_dct_32x32_add_8_1_sse2: 145.0
vp9_inv_dct_dct_32x32_add_8_1_ssse3: 137.4
vp9_inv_dct_dct_32x32_add_8_1_avx: 137.1
vp9_inv_dct_dct_32x32_add_8_1_avx2: 73.2
vp9_inv_dct_dct_32x32_add_8_2_c: 14680.8
vp9_inv_dct_dct_32x32_add_8_2_sse2: 2617.2
vp9_inv_dct_dct_32x32_add_8_2_ssse3: 982.9
vp9_inv_dct_dct_32x32_add_8_2_avx: 958.5
vp9_inv_dct_dct_32x32_add_8_2_avx2: 704.2
vp9_inv_dct_dct_32x32_add_8_4_c: 14443.1
vp9_inv_dct_dct_32x32_add_8_4_sse2: 2717.1
vp9_inv_dct_dct_32x32_add_8_4_ssse3: 965.7
vp9_inv_dct_dct_32x32_add_8_4_avx: 1000.7
vp9_inv_dct_dct_32x32_add_8_4_avx2: 717.1
vp9_inv_dct_dct_32x32_add_8_8_c: 14436.4
vp9_inv_dct_dct_32x32_add_8_8_sse2: 2671.8
vp9_inv_dct_dct_32x32_add_8_8_ssse3: 1038.5
vp9_inv_dct_dct_32x32_add_8_8_avx: 983.0
vp9_inv_dct_dct_32x32_add_8_8_avx2: 729.4
vp9_inv_dct_dct_32x32_add_8_16_c: 14614.7
vp9_inv_dct_dct_32x32_add_8_16_sse2: 2701.7
vp9_inv_dct_dct_32x32_add_8_16_ssse3: 1334.4
vp9_inv_dct_dct_32x32_add_8_16_avx: 1276.7
vp9_inv_dct_dct_32x32_add_8_16_avx2: 719.5
vp9_inv_dct_dct_32x32_add_8_32_c: 14363.6
vp9_inv_dct_dct_32x32_add_8_32_sse2: 2575.6
vp9_inv_dct_dct_32x32_add_8_32_ssse3: 2633.9
vp9_inv_dct_dct_32x32_add_8_32_avx: 2539.6
vp9_inv_dct_dct_32x32_add_8_32_avx2: 1395.0
9 years ago
7a15cf42ee
x86/diracdsp: make ff_put_signed_rect_clamped_10_sse4 work on x86_32
...
Reviewed-by: Rostislav Pehlivanov <atomnuker@gmail.com>
Signed-off-by: James Almer <jamrial@gmail.com>
9 years ago
d06dfaa5cb
x86: huffyuv: Use EXTERNAL_SSSE3_FAST convenience macro where appropriate
9 years ago
4efab89332
x86: Use *_FAST/*_SLOW CPU feature detection macros where appropriate
9 years ago
0a39c9ac0b
x86: hpeldsp: Don't check for bitexact flag when initializing VP3-specific code
...
That code is only ever initialized with that flag set.
9 years ago
95c1df929b
x86: hpeldsp: Drop unused function parameters
9 years ago
c3e83ad3b7
x86: hpeldsp: Use EXTERNAL_SSE2_FAST where appropriate
9 years ago
1dfc3cf89d
x86: hpeldsp: Split off VP3-specific bits into a separate file
9 years ago
fca3c3b619
hevc: Add AVX2 DC IDCT
...
Originally written by Pierre Edouard Lepere <pierre-edouard.lepere@insa-rennes.fr>.
Integrated to Libav by Josh de Kock <josh@itanimul.li>.
Signed-off-by: Alexandra Hájková <alexandra@khirnov.net>
9 years ago
df1dc52195
diracdsp_init: add missing ARCH_X86_64 check
...
That SIMD is still x86_64 only for now.
Signed-off-by: Rostislav Pehlivanov <atomnuker@gmail.com>
9 years ago
bd61f3c6bf
diracdsp: add SIMD for the 10 bit version of put_signed_rect_clamped
...
Signed-off-by: Rostislav Pehlivanov <rpehlivanov@obe.tv>
9 years ago
80721cc1ff
diracdsp: add dequantization SIMD
...
Currently unused, to be used in the following commits.
Signed-off-by: Rostislav Pehlivanov <rpehlivanov@obe.tv>
9 years ago
f0a2b6249b
vp9: add 16x16 idct avx2 (8-bit).
...
checkasm --bench, 10k runs, for *_add_${bpc}_${sub_idct}_${opt}, shows
that it's about 1.65x as fast as the AVX version for the full IDCT, and
similar speedups for the sub-IDCTs:
nop: 24.6
vp9_inv_dct_dct_16x16_add_8_1_c: 6444.8
vp9_inv_dct_dct_16x16_add_8_1_sse2: 638.6
vp9_inv_dct_dct_16x16_add_8_1_ssse3: 484.4
vp9_inv_dct_dct_16x16_add_8_1_avx: 661.2
vp9_inv_dct_dct_16x16_add_8_1_avx2: 311.5
vp9_inv_dct_dct_16x16_add_8_2_c: 6665.7
vp9_inv_dct_dct_16x16_add_8_2_sse2: 646.9
vp9_inv_dct_dct_16x16_add_8_2_ssse3: 455.2
vp9_inv_dct_dct_16x16_add_8_2_avx: 521.9
vp9_inv_dct_dct_16x16_add_8_2_avx2: 304.3
vp9_inv_dct_dct_16x16_add_8_4_c: 7022.7
vp9_inv_dct_dct_16x16_add_8_4_sse2: 647.4
vp9_inv_dct_dct_16x16_add_8_4_ssse3: 467.1
vp9_inv_dct_dct_16x16_add_8_4_avx: 446.1
vp9_inv_dct_dct_16x16_add_8_4_avx2: 297.0
vp9_inv_dct_dct_16x16_add_8_8_c: 6800.4
vp9_inv_dct_dct_16x16_add_8_8_sse2: 598.6
vp9_inv_dct_dct_16x16_add_8_8_ssse3: 465.7
vp9_inv_dct_dct_16x16_add_8_8_avx: 440.9
vp9_inv_dct_dct_16x16_add_8_8_avx2: 290.2
vp9_inv_dct_dct_16x16_add_8_16_c: 6626.6
vp9_inv_dct_dct_16x16_add_8_16_sse2: 599.5
vp9_inv_dct_dct_16x16_add_8_16_ssse3: 475.0
vp9_inv_dct_dct_16x16_add_8_16_avx: 469.9
vp9_inv_dct_dct_16x16_add_8_16_avx2: 286.4
9 years ago
645489cf90
x86/dcadsp: optimize lfe_fir0_float_fma3 on x86_32
...
About 10% faster.
Signed-off-by: James Almer <jamrial@gmail.com>
9 years ago
293484fa5e
avcodec: add missing xmm/neon clobber test wrappers for the new decode API
...
Reviewed-by: Ronald S. Bultje <rsbultje@gmail.com>
Signed-off-by: James Almer <jamrial@gmail.com>
9 years ago
9eb3da2f99
asm: FF_-prefix internal macros used in inline assembly
...
See merge commit '39d6d3618d48625decaff7d9bdbb45b44ef2a805'.
9 years ago
4a081f224e
libavcodec: fix constness in clobber test avcodec_open2() wrappers
...
Signed-off-by: Martin Storsjö <martin@martin.st>
9 years ago
9df889a5f1
h264: rename h264.[ch] to h264dec.[ch]
...
This is more consistent with the naming of other decoders.
9 years ago
f1a9eee41c
x86: Add missing movsxd for the int stride parameter
...
Signed-off-by: Martin Storsjö <martin@martin.st>
9 years ago
ede4ec1f8f
x86/aacpsdsp: optimize add_squares loop
...
Signed-off-by: James Almer <jamrial@gmail.com>
9 years ago
82dbfccaf0
x86/aacdec: use HADDPS macro
...
Signed-off-by: James Almer <jamrial@gmail.com>
9 years ago
1e9c5bf4c1
asm: FF_-prefix internal macros used in inline assembly
...
These warnings conflict with system macros on Solaris, producing
truckloads of warnings about macro redefinition.
9 years ago
dc40a70c57
Drop unnecessary libavutil/x86/asm.h #includes
9 years ago
a6a750c7ef
tests: Move all test programs to a subdirectory
9 years ago
9630b3fc06
x86: lossless audio: SSE4 madd 32bits
...
The unique user so far is wmalossless 24bits. The few samples tested show an
order of 8, so more unrolling or an avx2 version do not make sense.
Timings: 68 -> 49 cycles
Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
9 years ago
41ed7ab45f
cosmetics: Fix spelling mistakes
...
Signed-off-by: Diego Biurrun <diego@biurrun.de>
9 years ago
01621202aa
build: miscellaneous cosmetics
...
Restore alphabetical order in lists, break overly long lines, do some
prettyprinting, add some explanatory section comments, group parts
together that belong together logically.
9 years ago
305344d89e
avcodec/fft: Add revtab32 for FFTs with more than 65536 samples
...
x86 optimizations are used only for the cases they support (<=65536 samples)
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
9 years ago
ae76b84221
avcodec: Extend fft to size 2^17
...
Asked-for-by: durandal_1707
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
9 years ago
1a094af638
fft: Split MDCT bits off from FFT
9 years ago
e3461197b1
x86/vc1dsp: Split the file into MC and loopfilter
9 years ago
73ff983e8d
fft: x86: cosmetics: Drop silly comments, add comment, whitespace
9 years ago
257b30af8e
x86: hevc: Fix linking with both yasm and optimizations disabled
...
Some optimized functions reference optimized symbols, so the functions
must be explicitly disabled when those symbols are unavailable.
9 years ago
45d3af9059
x86/dcadec: add ff_lfe_fir1_float_{sse3,avx}
...
Reviewed-by: Christophe Gisquet <christophe.gisquet@gmail.com>
Signed-off-by: James Almer <jamrial@gmail.com>
9 years ago
15a24614ae
build: Add vc1dsp component for more fine-grained dependencies
9 years ago
70d685a77f
x86: use the new helper macros where useful
...
Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: James Almer <jamrial@gmail.com>
9 years ago
bcc223523e
x86/vc1dsp: Port vc1_*_hor_16b_shift2 to NASM format
...
Reviewed-by: Christophe Gisquet <christophe.gisquet@gmail.com>
9 years ago
59ebf32bca
huffyuvencdsp: Undefine "i" macro after each use
9 years ago
8ae7447941
x86/dcadec: add ff_lfe_fir0_float_{sse,sse2,avx,fma3}
...
Up to ~4 times faster on x86_64, ~8 times on x86_32 if compiling using x87 fp math.
Reviewed-by: Ronald S. Bultje <rsbultje@gmail.com>
Signed-off-by: James Almer <jamrial@gmail.com>
9 years ago
9fd6ea933f
dirac_dwt: Make x86 files/functions names consistent
9 years ago
James Almer
Clément Bœsch
Ronald S. Bultje
Anton Khirnov
Christophe Gisquet
Diego Biurrun
Rostislav Pehlivanov
Rostislav Pehlivanov
Matthieu Bouron
Martin Storsjö
Vittorio Giovara
Michael Niedermayer
Timothy Gu