[FFmpeg-devel] [PATCH] Separate format conversion DSP functions from DSPContext.
Wed Feb 2 03:49:26 CET 2011
"Ronald S. Bultje" <rsbultje at gmail.com> writes:
> 2011/2/1 M?ns Rullg?rd <mans at mansr.com>:
>> "Ronald S. Bultje" <rsbultje at gmail.com> writes:
>>> On Tue, Feb 1, 2011 at 6:13 PM, Mans Rullgard <mans at mansr.com> wrote:
>>>> From: Justin Ruggles <justin.ruggles at gmail.com>
>>>> This will be beneficial for use with the audio conversion API without
>>>> requiring it to depend on all of dsputil.
>>> I had a quick look at the x86 SIMD, nothing particularly wrong with it.
>>>> + ? ?mov src1q, [srcq+1*gprsize]
>>>> + ? ?mov src2q, [srcq+2*gprsize]
>>>> + ? ?mov src3q, [srcq+3*gprsize]
>>>> + ? ?mov src4q, [srcq+4*gprsize]
>>>> + ? ?mov src5q, [srcq+5*gprsize]
>>> Why do you use the "q" variant of registers? Shouldn't you use the "q"
>>> for x86-64 and "d" for x86-32? Or is "q" automatically "degraded" to
>>> "d" on x86-32?
>> That code is just moved from one file to another. ?I believe the
>> original was written by Loren.
> The patch doesn't remove FLOAT_TO_INT16_INTERLEAVE6 from
> dsputil_yasm.asm then.
I thought the last one did. Did I miss something?
mans at mansr.com
More information about the ffmpeg-devel