[FFmpeg-devel] [PATCH 2/2] lavc/flacdsp: implement wasted32 DSP function for VSX on POWER

[Sean McGovern]

Hi,

On Sat, Jul 6, 2024, 16:18 Rémi Denis-Courmont <remi at remlab.net> wrote:

> Le lauantaina 6. heinäkuuta 2024, 23.00.47 EEST Sean McGovern a écrit :
> > Does wasted32 (and I guess wasted33 by proxy) not have to worry about
> loops
> > tails? I noticed the other vectorized versions don't do anything special
> in
> > that regard.
>
> Frankly, RISC-V vectors (like Arm SVE's) are scalable so I don't need to
> care
> about these details.
>
> It depends if there is padding, and if so if it can be overwritten or only
> read. I don't know about this particular case. Check the x86 code.
>

Naah, the SSE2 version of wasted32 just unrolls it by 4 and call it a day,
so then like you said this is probably either padded or else always a
multiple of 512 bits.


>
> --
> レミ・デニ-クールモン
> http://www.remlab.net/
>
>
>
> _______________________________________________
> ffmpeg-devel mailing list
> ffmpeg-devel at ffmpeg.org
> https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>
> To unsubscribe, visit link above, or email
> ffmpeg-devel-request at ffmpeg.org with subject "unsubscribe".
>

-- Sean McGovern

>