[FFmpeg-devel] [PATCH] avfilter, swresample, swscale: use fabs, fabsf instead of FFABS
gajjanag at mit.edu
Wed Oct 14 13:55:21 CEST 2015
On Wed, Oct 14, 2015 at 7:02 AM, Hendrik Leppkes <h.leppkes at gmail.com> wrote:
> On Wed, Oct 14, 2015 at 12:57 PM, Carl Eugen Hoyos <cehoyos at ag.or.at> wrote:
>> Hendrik Leppkes <h.leppkes <at> gmail.com> writes:
>>> > If you could show that it is faster on any platform
>>> > I would already be happy!
>>> A more important check would be that its not significantly
>>> slower on any other platform. Just because one
>>> compiler/glibc combination manages to produce an efficient
>>> inlined function doesn't necessarily mean that some other
>>> compiler or libc couldn't produce a full function call
>>> with all the overhead that comes with it, becoming
>>> significantly slower.
>> I thought we all agree that such a platform exists
>> somewhere (sometimes) but that a performance gain
>> on x86 Linux is more important, no?
> But for that to be true there must be a clear gain - which this change
> has yet to proof. Testing a significant amount of runs seems to come
> out at about same speed from all the numbers the author himself
Did you try the random number benchmark? That should give a much more
pronounced difference. Furthermore, it may be used to craft audio
files that show significant perf differences (I lack the time or
energy to do that myself at the moment).
> - Hendrik
> ffmpeg-devel mailing list
> ffmpeg-devel at ffmpeg.org
More information about the ffmpeg-devel