On Thu, May 26, 2011 at 6:22 AM, Kostya <[email protected]> wrote: > On Thu, May 26, 2011 at 09:18:29AM -0400, Ronald S. Bultje wrote: >> --- >> libswscale/swscale_template.c | 20 +++----------------- >> 1 files changed, 3 insertions(+), 17 deletions(-) > > looks ok (though LUT may be even better)
Doubt it. on x86, clip is a single arithop with no latency (except in the case the clip actually happens). A LUT is much slower, as it adds latency and uses up a load/store op. Jason _______________________________________________ libav-devel mailing list [email protected] https://lists.libav.org/mailman/listinfo/libav-devel
