[Ffmpeg-devel] Snow slicing support

Guillaume POIRIER poirierg
Tue Apr 11 13:17:31 CEST 2006


On 4/11/06, Loren Merritt <lorenm at u.washington.edu> wrote:
> On Tue, 11 Apr 2006, Guillaume Poirier wrote:
> > Oded Shimon wrote:
> >>
> >> In this case the difference was still obvious, but the results are very
> >> inaccurate. is there a better way for this? maybe START_TIMER around the
> >> whole decode() function?
> >
> > That what I'd do. Since measuring the whole decode() function is likely
> > to be inaccurate due to interruptions, I suggest you run it a number of
> > times (let's say, 100 times) and take the mean value, and the shortest
> > value.
> > This should take care of the measurement fuzzinesss.
> Do you mean to decode a 100-frame snow file, and measure each frame once?
> Then minimum time is meaningless.
> Do you mean decode one frame 100 times? Then cache usage is unrealistic,
> as stuff may be reused from one run to the next that's not reused between
> frames.
> For whole codec optimizations, I get more reproducible results from
> time(1) than from -benchmark.

I was more thinking about about running the whole program several
times (maybe not 100 times, but a good 20 times seams like a minimum),
redirect the ouput of timers to a file, and extract the meaning full
information with some perl or shell script.

... but that's probably overkill if you can get the same effect using
the simple "time" prgm.

I am disillusioned enough to know that no man's opinion on any subject
is worth a damn unless backed up with enough genuine information to
make him really know what he's talking about.

-- H. P. Lovecraft (about the flamewars on FFmpeg and MPlayer-dev mailing lists)

More information about the ffmpeg-devel mailing list