[FFmpeg-devel] [PATCH v2 0/7] arm64 neon implementation for 8bits functions

[Martin Storsjö]

On Mon, 3 Oct 2022, Grzegorz Bernacki wrote:

> Changes since v1:
>
> - changed tabs to spaces
> - modified branch instruction in vsse8
> - apply Martin's patches with improved instructions scheduling 
>
> Grzegorz Bernacki (4):
>  lavc/aarch64: Add neon implementation for pix_abs8 functions.
>  lavc/aarch64: Provide neon implementation of nsse8
>  lavc/aarch64: Provide optimized implementation of vsse8 for arm64.
>  lavc/aarch64: Add neon implementation for vsse_intra8
>
> Martin Storsjö (3):
>  aarch64: me_cmp: Improve scheduling in ff_pix_abs8_y2_neon
>  aarch64: me_cmp: Fix up the prologue of ff_pix_abs8_xy2_neon
>  aarch64: me_cmp: Improve scheduling in vsse_intra8
>
> libavcodec/aarch64/me_cmp_init_aarch64.c |  33 ++
> libavcodec/aarch64/me_cmp_neon.S         | 414 +++++++++++++++++++++++
> 2 files changed, 447 insertions(+)

Thanks! This mostly looked good to me.

I had actually meant that you would squash my fixes into your patches, 
instead of keeping them as separate ones.

After squashing such changes, it might have been interesting to get 
updated benchmarks in those commit messages (the ones that you have from 
Graviton 3). However in this case, these changes didn't really make much 
difference on out-of-order cores, only on in-order cores, so I guess 
there's not that much value in getting updated benchmarks from Graviton 3 
in this case.

So I went ahead and squashed those patches (and added co-authored-by lines 
where relevant), and pushed them now. Thanks for your contribution!

// Martin