[LAU] Qjackctl 3.2, QT4, KDE 3.5.9, high CPU usage

[hanaghan at starband.net]

Hi Rui,
thanks for the reply.

Ok...just to add, I rebuilt from the SRPM (same spec file though) and no
change. I also installed Gnome and it does it from there as well, just to
further support your point.

I will email thac and see if we can figure it out.

thanks,
R

> hanaghan at starband.net wrote:
>> Hi folks...
>>
>> I still running PCLinuxOS and using Thacś RPMś.
>>
>> Without version numbers of lots of libs, the version numbers in the
>> subject line are relevant to my system. Hopefully thatś enough.
>>
>>
>> I set up a new box from the live CD and added all the audio stuff from
>> apt. Everything worked fine...I updated to Qjackctl 3.2 from and older
>> version even and still no problems. I then did a jumbo update out of
>> Synaptic...
>>
>> When I run Qjackctl or Qsynth in KDE, the CPU usage shoots to the
>> moon. This has happened on another install of the same distro in the
>> the very same sequence and manner. I dont think itś Qjackctl that
>> is the issue..otherwise it would have done it when i first upgraded to
>> 3.2 using QT4. Even though Qsynth shows the same symptoms, pointing
>> toward a possible QT4 issue, that makes no sense either...again, it
>> would have done it before.
>>
>> I suspect something in KDE is doing it. I tried running from Fluxbox,
>> and same thing...but doesn Fluxbox use some KDS libs anyway?
>>
>> I guess I can install Gnome and try that but really dont want the
>> clutter on the machine. Anyone know anything about this? I saw another
>> post on Linuxquestions.org about the same thing on another distro...
>>
>> TIA for any help.
>>
>
> a very similar issue has been experienced before, iirc. it has been
> reported that on debian etch and ubuntu, qjackctl 0.3.x was peaking to
> 100%cpu usage (not %dsp load) and right after jackd is started.
>
> this issues blame has been pointed to libqt4 itself, more specifically
> to the QProcess class wich deals to process interface coordination and
> to the extent of my knowledge a mere qt4 upgrade, from the testing (or
> lenny) repositories, sufficed for the issue go away.
>
> so it seems to be a strict qt4 issue, maybe related to the way it's
> built but not sure (i never had the issue duplicated myself)
>
> hth
>
> cheers
> --
> rncbc aka Rui Nuno Capela
> rncbc at rncbc.org