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
..
2022-04-22 10:49:46 +03:00
2021-07-03 15:57:53 +02:00
2021-07-03 15:57:53 +02:00
2022-01-04 19:39:22 -06:00
2022-01-22 17:00:45 +01:00
2021-10-03 20:38:29 +02:00
2020-08-08 12:03:42 +02:00
2022-02-24 12:56:49 +01:00
2022-03-16 14:05:26 +02:00
2022-02-24 12:56:49 +01:00
2022-01-04 19:39:22 -06:00
2020-12-12 23:07:22 -05:00
2020-05-15 23:38:17 +03:00
2022-02-24 12:56:49 +01:00
2021-09-20 01:03:52 +02:00
2022-02-24 12:56:49 +01:00
2021-09-06 09:16:52 +02:00
2022-02-24 12:56:49 +01:00
2022-03-19 00:01:57 +02:00
2022-03-16 14:05:26 +02:00
2022-03-16 14:05:26 +02:00
2022-03-16 14:12:46 +02:00
2021-09-20 01:03:52 +02:00
2021-12-03 16:10:51 +01:00