On Tue, 13 Sep 2022, Hubert Mazur wrote:
Provide optimized implementation for pix_median_abs16 function.
Forgot to update this part of the commit message here too.
Performance comparison tests are shown below. - median_sad_1_c: 273.7 - median_sad_1_neon: 98.2 Benchmarks and tests run with checkasm tool on AWS Graviton 3. Signed-off-by: Hubert Mazur <h...@semihalf.com> --- libavcodec/aarch64/me_cmp_init_aarch64.c | 3 ++ libavcodec/aarch64/me_cmp_neon.S | 65 ++++++++++++++++++++++++ 2 files changed, 68 insertions(+)
The same comments as for patch 1/3; looks reasonble, a bunch of leftover mov instructions which I don't see how they'd be necessary, and please avoid the extra single-lane handling and just do plain vector operations, and extract the single lane at the end.
// Martin _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-devel To unsubscribe, visit link above, or email ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".
