On Sat, Oct 31, 2015 at 2:30 PM, Nicolas George <geo...@nsup.org> wrote: > Le decadi 10 brumaire, an CCXXIV, Ganesh Ajjanagadde a écrit : >> For the benefit of the thread, here is a benchmark for the find_motion >> function to address this point: >> new: >> 1815531917 decicycles in find_motion, 1024 runs, 0 skips >> >> old: >> 1900282483 decicycles in find_motion, 1024 runs, 0 skips > > So 4-5%. That is nowhere near the values, but that is still a worthy > speedup nonetheless. Thanks for running the test.
The gap is not surprising due to your remark regarding the loop, and is a general caveat regarding benchmarking/profiling that I will try to be more cautious about in the future. For future reference, I have adopted the consistent convention of referring to the block of code in the START, STOP macros so as to make it clear what I have benchmarked. If it is a loop in a function, I will be explicit and put line nos, e.g "235-260:foo" where foo is the function name. > > Regards, > > -- > Nicolas George > > _______________________________________________ > ffmpeg-devel mailing list > ffmpeg-devel@ffmpeg.org > http://ffmpeg.org/mailman/listinfo/ffmpeg-devel > _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel