ffmpeg & ULE

[Ivan Klymenko]

В Tue, 18 Oct 2011 17:22:29 +0300
Andriy Gapon <avg at FreeBSD.org> пишет:

> on 18/10/2011 17:17 Ivan Klymenko said the following:
> > В Tue, 18 Oct 2011 17:03:51 +0300
> > Andriy Gapon <avg at FreeBSD.org> пишет:
> > 
> >> on 18/10/2011 16:26 Ivan Klymenko said the following:
> >>> В Tue, 18 Oct 2011 16:00:48 +0300
> >>> Andriy Gapon <avg at FreeBSD.org> пишет:
> >>>
> >>>> on 18/10/2011 14:30 Ivan Klymenko said the following:
> >>>>> В Tue, 18 Oct 2011 12:04:36 +0300
> >>>>> Urmas Lett <urmas.lett at eenet.ee> пишет:
> >>>>>
> >>>>>> Hello.
> >>>>>>
> >>>>>> Why is ffmpeg -threads massively slower with ULE than 4BSD?
> >>>>>>
> >>>>>> ffmpeg preset veryfast with sched_bsd:
> >>>>>> real    1m49.407s
> >>>>>> user    6m53.932s
> >>>>>> sys     0m1.700s
> >>>>>>
> >>>>>> ffmpeg preset veryfast with sched_ule:
> >>>>>> real    2m52.711s
> >>>>>> user    6m50.310s
> >>>>>> sys     0m1.582s
> >>>>>>
> >>>>>> #uname -a
> >>>>>> FreeBSD 9.0-RC1 FreeBSD 9.0-RC1 #0: Mon Oct 17 20:32:29 EEST
> >>>>>>
> >>>>>>
> >>>>>
> >>>>> probably because you have a system processor with 2 cores...?
> >>>>> if yes - then use the 4BSD...it is better for the two cores...
> >>>>> IMHO
> >>>>
> >>>> Do you have any facts to substantiate your claim?
> >>>>
> >>>
> >>> well, for example:
> >>> http://img37.imageshack.us/img37/9311/plotj.gif
> >>> http://img850.imageshack.us/img850/584/plot.gif
> >>
> >> Not sure if two pictures is all that it takes.
> >> There is no description of hardware, OS versions, reproducibility
> >> of the results. Also you made a broader claim like "4BSD ... is
> >> better for the two cores", but the pictures demonstrate only that
> >> it is better (by ~10% ? I hate it when the axises do not start at
> >> zero)
> > 
> > http://forum.lissyara.su/viewtopic.php?p=305269#p305269
> > 
> >> only for transactions/s in postgresql sysbench.  There are other
> >> workloads and other important things to measure (like
> >> interactivity, etc).
> >>
> >> Good benchmarking is a real science.
> >>
> > Of course - I understand it.
> > Please tell me what tests should I do?
> > I will give you the test results for review.
> > I also give all the information about the hardware on which the
> > tests were conducted.
> 
> That's a problem for me - I am not an expert in this area and have
> very little experience doing formal performance testing.
> 

So how can I convince you to confirm that I am right?