[FFmpeg-devel] [PATCH] swscale: rgb_to_yuv neon optimizations

[Michael Niedermayer]

On Mon, May 19, 2025 at 09:50:18PM +0200, Dmitriy Kovalenko wrote:
> I've found quite a few ways to optimize existing ffmpeg's rgb to yuv
> subsampled conversion. In this patch stack I'll try to
> improve the performance.
> 
> This particular set of changes is a small improvement to all the
> existing functions and macro. The biggest performance gain is
> coming from post loading increment of the pointer and immediate
> pref etching of the memory blocks and interleaving the multiplication
> shifting operations of
> different registers for better scheduling.
> 
> Also changed a bunch of places where cmp + b.le was used instead
> of one instruction cbnz/tbnz and some other small cleanups.
> 
> Here are checkasm results on the macbook pro with the latest M4 max
> 
> <before>
> 
> bgra_to_uv_1080_c:                                     257.5 ( 1.00x)
> bgra_to_uv_1080_neon:                                  211.9 ( 1.22x)
> bgra_to_uv_1920_c:                                     467.1 ( 1.00x)
> bgra_to_uv_1920_neon:                                  379.3 ( 1.23x)
> bgra_to_uv_half_1080_c:                                198.9 ( 1.00x)
> bgra_to_uv_half_1080_neon:                             125.7 ( 1.58x)
> bgra_to_uv_half_1920_c:                                346.3 ( 1.00x)
> bgra_to_uv_half_1920_neon:                             223.7 ( 1.55x)
> 
> <after>
> 
> bgra_to_uv_1080_c:                                     268.3 ( 1.00x)
> bgra_to_uv_1080_neon:                                  176.0 ( 1.53x)
> bgra_to_uv_1920_c:                                     456.6 ( 1.00x)
> bgra_to_uv_1920_neon:                                  307.7 ( 1.48x)
> bgra_to_uv_half_1080_c:                                193.2 ( 1.00x)
> bgra_to_uv_half_1080_neon:                              96.8 ( 2.00x)
> bgra_to_uv_half_1920_c:                                347.2 ( 1.00x)
> bgra_to_uv_half_1920_neon:                             182.6 ( 1.92x)
> 
> With my proprietary test on IOS it gives around 70% of performance
> improvement converting bgra 1920x1920 image to yuv420p
> 
> On my linux arm cortex-r processing the performance improvement not that
> visible but still consistently faster by 5-10% than the current
> implementation.
> 
> Signed-off-by: Dmitriy Kovalenko <dmtr.kovalenko at outlook.com>
> ---
>  libswscale/aarch64/input.S | 166 +++++++++++++++++++++++++------------
>  1 file changed, 112 insertions(+), 54 deletions(-)
> 
> diff --git a/libswscale/aarch64/input.S b/libswscale/aarch64/input.S
> index c1c0adffc8..ee8eb24c14 100644
> --- a/libswscale/aarch64/input.S
> +++ b/libswscale/aarch64/input.S
> @@ -1,5 +1,4 @@
> -/*
> - * Copyright (c) 2024 Zhao Zhili <quinkblack at foxmail.com>
> +/* Copyright (c) 2024 Zhao Zhili <quinkblack at foxmail.com>
>   *
>   * This file is part of FFmpeg.
>   *
> @@ -57,20 +56,41 @@
>          sqshrn2         \dst\().8h, \dst2\().4s, \right_shift   //
> dst_higher_half = dst2 >> right_shift
>  .endm
>  +// interleaved product version of the rgb to yuv gives slightly better
> performance on non-performant mobile +.macro rgb_to_uv_interleaved_product
> r, g, b, u_coef0, u_coef1, u_coef2, v_coef0, v_coef1, v_coef2, u_dst1,
> u_dst2, v_dst1, v_dst2, u_dst, v_dst, right_shift

error: corrupt patch at line 58
please make sure your line/word wrap settings dont damage patches


[...]
-- 
Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

During times of universal deceit, telling the truth becomes a
revolutionary act. -- George Orwell
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 195 bytes
Desc: not available
URL: <https://ffmpeg.org/pipermail/ffmpeg-devel/attachments/20250520/b9d34cd5/attachment.sig>