RepoMirrors
/
ffmpeg
mirror of https://git.ffmpeg.org/ffmpeg.git
1
0
Fork 0
Code Issues Projects Releases Wiki Activity
ffmpeg/libswscale/aarch64
History
Martin Storsjö 70db14376c swscale: aarch64: Optimize the final summation in the hscale routine 
Before:                     Cortex A53      A72      A73  Graviton 2  Graviton 3
hscale_8_to_15_width8_neon:     8273.0   4602.5   4289.5      2429.7      1629.1
hscale_8_to_15_width16_neon:   12405.7   6803.0   6359.0      3549.0      2378.4
hscale_8_to_15_width32_neon:   21258.7  11491.7  11469.2      5797.2      3919.6
hscale_8_to_15_width40_neon:   25652.0  14173.7  12488.2      6893.5      4810.4

After:
hscale_8_to_15_width8_neon:     7633.0   3981.5   3350.2      1980.7      1261.1
hscale_8_to_15_width16_neon:   11666.7   5951.0   5512.0      3080.7      2131.4
hscale_8_to_15_width32_neon:   20900.7  10733.2   9481.7      5275.2      3862.1
hscale_8_to_15_width40_neon:   24826.0  13536.2  11502.0      6397.2      4731.9

Thus, this gives overall a 8-29% speedup for the smaller filter
sizes, around 1-8% for the larger filter sizes.

Inspired by a patch by Jonathan Swinney <jswinney@amazon.com>.

Signed-off-by: Martin Storsjö <martin@martin.st>
 		2022-04-22 10:49:46 +03:00
..
Makefile swscale: aarch64: Add a NEON implementation of interleaveBytes 2020-05-15 23:38:17 +03:00
hscale.S swscale: aarch64: Optimize the final summation in the hscale routine 2022-04-22 10:49:46 +03:00
output.S swscale: aarch64: Don't clobber callee-saved registers v8-v15 2020-04-21 23:41:13 +03:00
rgb2rgb.c swscale: aarch64: Add a NEON implementation of interleaveBytes 2020-05-15 23:38:17 +03:00
rgb2rgb_neon.S swscale: aarch64: Add a NEON implementation of interleaveBytes 2020-05-15 23:38:17 +03:00
swscale.c Include attributes.h directly 2021-04-19 14:34:10 +02:00
swscale_unscaled.c sws: rename SwsContext.swscale to convert_unscaled 2021-07-03 15:57:53 +02:00
yuv2rgb_neon.S aarch64/yuv2rgb_neon: fix return value 2020-07-09 10:33:14 +01:00