On Thursday 17 February 2005 21:37, Brad Templeton wrote:
> On Thu, Feb 17, 2005 at 08:52:40PM -0700, Blair wrote:
> > XvMC isupport is compiled into my X stuff, I see it load
> > in /var/log/Xorg.0.log on every X start-up.  I do not see configuration
> > parameters in myth which specifically leverage it though I have been lead
> > to believe by many on the Xorg forums that this is now a transparent
> > process if the XvMX libs are present, and mine are.  So I presume they
> > are in use.
>
> Usually to get xvmc you must compile your myth yourself and enable it,
> and then turn it on (which is the default though) in the setup screens.

If this is the default compile, then I got it, mine is all compiled from 
source.  CVS from two nights ago right at the moment.  I didn't specifically 
set any compile flags though.  

>
> > Hmmm, yes it's hyperthreaded but as I understand it, it is not the IDLE
> > that is misleading but rather the accuracy of specific process/forked
> > operations. IDLE is "supposed" to be accurate (at least where I read ;) 
> > the rest can be off quite a bit.
>
> No, it's the idle that is misleading.  It is reporting idle time that is
> not really there, because the spare ALU is not doing anything but can't
> really do anything as other resources are used up.
>
> > create a disk channel bind.  Watching TV is both writing to, then reading
> > from disk and mythcommflag is clearly paging through the file as fast as
> > it
>
> Hmm.  Well, again this surpises me.  My tests are not watching live tv,
> but watching a recording.   Don't want to complicate it with recording.
>
> I was also watching stuff coming over ethernet, not disk so I effectively
> had no disk IO for the watching, only the commflag.

OK, thats a bit different, I'll run it watching a recording this weekend and 
see what I see.  Ya want the 1080i test too?

-Blair
_______________________________________________
mythtv-users mailing list
[email protected]
http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-users

Reply via email to