[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Bug#176267: ITP: mplayer -- Mplayer is a full-featured audioand video player for UN*X like systems



On Mon, Jan 27, 2003 at 09:39:59AM +0100, Gabucino wrote:
> I also tried to play a _one-frame_ VOB with AC3 audio (the xine team - unlike
> us - concentrated very much work on these so called still-frames' support, so I
> expected it to work): it displayed the frame, and tried to play the audio -
> unfortunately the audio was totally skipping.. :)
> Meanwhile my CPU usage was 100% :
> 
>  9922 aaxine                        41.7 48.2  14m root       0:05.41  25   0 R
>  9916 aaxine                        31.5 48.2  14m root       0:05.69  25   0 R
>  9923 aaxine                        20.4 48.2  14m root       0:03.21  16   0 R
> 
> Now this is what I call "perform better" :))))))))) 100% CPU for a jerky
> AC3 playback which should take about 20% CPU anyway.

Are you trolling, or do you just have no understanding of how
scheduling works and what the CPU% value means?

"Total CPU%" is 1 - (unusable[0] cycles / ideal cycles), or "a
measurement of how much available processor time is currently
used". If you proposed this as a performance measure in any serious
benchmarking forum you would be laughed down.

The CPU% value for a *specific* process is a ratio indicating the
average proportion of processing time currently being used by this
process. It is therefore primarily dependant on what other processing
is occuring at the same time. Since it is not even remotely absolute,
it is (also) useless for benchmarking purposes.

[0] Cycles are unusable when the processor doesn't have anything to
do, generally because processes are suspended while waiting for
syscalls to complete.

-- 
  .''`.  ** Debian GNU/Linux ** | Andrew Suffield
 : :' :  http://www.debian.org/ | Dept. of Computing,
 `. `'                          | Imperial College,
   `-             -><-          | London, UK

Attachment: pgpDqBe6euQPP.pgp
Description: PGP signature


Reply to: