On 04/25/2011 10:41 AM, Ronald S. Bultje wrote: > Hi, > > On Mon, Apr 25, 2011 at 2:35 PM, Diego Biurrun <di...@biurrun.de> wrote: >> The _fast integer types provide no realworld benefits, but may >> introduce portability issues and are just plain ugly. >> --- >> libavcodec/alpha/simple_idct_alpha.c | 20 +- >> libavcodec/cavs.h | 11 +- >> libavcodec/cavsdata.h | 8 +- >> libavcodec/ffv1.c | 31 ++-- >> libavcodec/jfdctfst.c | 18 +- >> libavcodec/jfdctint.c | 18 +- >> libavcodec/vorbis.c | 14 +- >> libavcodec/vorbis.h | 12 +- >> libavcodec/vorbisdec.c | 330 >> +++++++++++++++++----------------- >> libavcodec/vorbisenc.c | 6 +- >> 10 files changed, 239 insertions(+), 229 deletions(-) > > That's a big statement. I probably agree with you, but now that you're > gonna be our next codec guru anyway, would you mind testing (time > ffmpeg -i vorbis.ogg -f null -v 0 -, 3x before/after patch and take > the average) this and show that there's no significant time > difference? > > If shown to be insignificant, patch OK.
for reference, it would also be good to know what the fast types are defined to on the tested system. -Justin _______________________________________________ libav-devel mailing list libav-devel@libav.org https://lists.libav.org/mailman/listinfo/libav-devel