[FFmpeg-devel] [aarch64] improve performance of ff_hscale_8_to_15_neon

Clément Bœsch u at pkh.me
Wed Nov 27 22:13:35 EET 2019

On Wed, Nov 27, 2019 at 07:36:01PM +0000, Pop, Sebastian wrote:
> Thanks Jean-Baptiste for your review and suggestions on how to improve my patch submission.
> From the git logs I found out that Clément Bœsch wrote the original aarch64 vectorization for that function.

> Maybe Clément could help to review the content of the patch?

Yeah I will by the end of the week. I wrote that a few years ago so I need
to take some time to get back in the context.

BTW, that's quite a huge speed improvement you're bringing in, are you
sure you are always allowed to read up to filter[3]?

Last thing: this same optimization was also written for arm following the
same pattern. You may want to adjust that one as well while waiting for my
review :)


Clément B.

More information about the ffmpeg-devel mailing list