[vdr] Re: High CPU usage of vdr-1.1.22

[Henning Holtschneider <hh@holtschneider.com>]

Am Mon, 2003-02-03 um 10.54 schrieb M. Fiegert:

> I used 1.1.17 for some time and had lots of problems. Some days ago I 
> switched to 1.1.22. I lost some recordings right away. I disabled
> the two defines for simultanious recording/playbayck. After that there 
> were no more stability problems. :-)

I'm running VDR with DO_REC_AND_PLAY_ON_PRIMARY_DEVICE disabled but I
still have DO_MULTIPLE_RECORDINGS enabled. The quality of the playback
and the quality (= completeness) of the recorded data depends on the
first setting but one should be able to record on the second card while
replaying on the first just fine. After all, this worked well with VDR
1.0.x and this is not Microsoft ;->
 
> But: I have recordings that seem to be corrupted in some way. The can be 
> viewed without distortions, but sometimes they just get stuck. I need to 
> skip over these moments. Generally fast forward/backword does not work 
> good with these recordings, like the index had been produced with genindex.

Same thing here.

> >But, it uses much more cpu than the old version.  
> >
> Have you tried to disable the two defines? I had high cpu load with my 
> 1.1.17 (with defines) but not with 1.1.22 (without defines). I don`t 
> know about 1.1.22 with defines.
> 
> >The reason may be that the number of interrupts/s hits about 8000-10000.
> >This seems to be excessively high.

My cards trigger 3000-6000 interrupts per second each during recording.
I've never checked the number of interrupts with the old driver and VDR,
though.

hh



-- 
Info:
To unsubscribe send a mail to listar@linuxtv.org with "unsubscribe vdr" as subject.