No description
3c9546dfaf
This work is sponsored by, and copyright, Google.
These are ported from the ARM version; thanks to the larger
amount of registers available, we can do the 16x16 and 32x32
transforms in slices 8 pixels wide instead of 4. This gives
a speedup of around 1.4x compared to the 32 bit version.
The fact that aarch64 doesn't have the same d/q register
aliasing makes some of the macros quite a bit simpler as well.
Examples of runtimes vs the 32 bit version, on a Cortex A53:
ARM AArch64
vp9_inv_adst_adst_4x4_add_neon: 90.0 87.7
vp9_inv_adst_adst_8x8_add_neon: 400.0 354.7
vp9_inv_adst_adst_16x16_add_neon: 2526.5 1827.2
vp9_inv_dct_dct_4x4_add_neon: 74.0 72.7
vp9_inv_dct_dct_8x8_add_neon: 271.0 256.7
vp9_inv_dct_dct_16x16_add_neon: 1960.7 1372.7
vp9_inv_dct_dct_32x32_add_neon: 11988.9 8088.3
vp9_inv_wht_wht_4x4_add_neon: 63.0 57.7
The speedup vs C code (2-4x) is smaller than in the 32 bit case,
mostly because the C code ends up significantly faster (around
1.6x faster, with GCC 5.4) when built for aarch64.
Examples of runtimes vs C on a Cortex A57 (for a slightly older version
of the patch):
A57 gcc-5.3 neon
vp9_inv_adst_adst_4x4_add_neon: 152.2 60.0
vp9_inv_adst_adst_8x8_add_neon: 948.2 288.0
vp9_inv_adst_adst_16x16_add_neon: 4830.4 1380.5
vp9_inv_dct_dct_4x4_add_neon: 153.0 58.6
vp9_inv_dct_dct_8x8_add_neon: 789.2 180.2
vp9_inv_dct_dct_16x16_add_neon: 3639.6 917.1
vp9_inv_dct_dct_32x32_add_neon: 20462.1 4985.0
vp9_inv_wht_wht_4x4_add_neon: 91.0 49.8
The asm is around factor 3-4 faster than C on the cortex-a57 and the asm
is around 30-50% faster on the a57 compared to the a53.
Signed-off-by: Martin Storsjö <martin@martin.st>
|
||
|---|---|---|
| compat | ||
| doc | ||
| libavcodec | ||
| libavdevice | ||
| libavfilter | ||
| libavformat | ||
| libavresample | ||
| libavutil | ||
| libswscale | ||
| presets | ||
| tests | ||
| tools | ||
| .gitattributes | ||
| .gitignore | ||
| .travis.yml | ||
| arch.mak | ||
| avconv.c | ||
| avconv.h | ||
| avconv_dxva2.c | ||
| avconv_filter.c | ||
| avconv_opt.c | ||
| avconv_qsv.c | ||
| avconv_vaapi.c | ||
| avconv_vda.c | ||
| avconv_vdpau.c | ||
| avplay.c | ||
| avprobe.c | ||
| Changelog | ||
| cmdutils.c | ||
| cmdutils.h | ||
| cmdutils_common_opts.h | ||
| common.mak | ||
| configure | ||
| COPYING.GPLv2 | ||
| COPYING.GPLv3 | ||
| COPYING.LGPLv2.1 | ||
| COPYING.LGPLv3 | ||
| CREDITS | ||
| INSTALL | ||
| library.mak | ||
| LICENSE | ||
| Makefile | ||
| README | ||
| README.md | ||
| RELEASE | ||
| version.sh | ||
Libav
Libav is a collection of libraries and tools to process multimedia content such as audio, video, subtitles and related metadata.
Libraries
libavcodecprovides implementation of a wider range of codecs.libavformatimplements streaming protocols, container formats and basic I/O access.libavutilincludes hashers, decompressors and miscellaneous utility functions.libavfilterprovides a mean to alter decoded Audio and Video through chain of filters.libavdeviceprovides an abstraction to access capture and playback devices.libavresampleimplements audio mixing and resampling routines.libswscaleimplements color conversion and scaling routines.
Tools
- avconv is a command line toolbox to manipulate, convert and stream multimedia content.
- avplay is a minimalistic multimedia player.
- avprobe is a simple analisys tool to inspect multimedia content.
- Additional small tools such as
aviocat,ismindexandqt-faststart.
Documentation
The offline documentation is available in the doc/ directory.
The online documentation is available in the main website and in the wiki.
Examples
Conding examples are available in the doc/example directory.
License
Libav codebase is mainly LGPL-licensed with optional components licensed under GPL. Please refer to the LICENSE file for detailed information.