[PD] Variable CPU?

[cyrille henry]

pd have a -nosleep option, in order to use all available processing power. i think it was made in order to have a dirty solution for this problem.
unfortunately, it look like it's not working on 0.42.5 at least on linux.

maybe you can try to find an older version of pd and start pd -nosleep : it should use 100% cpu every-time, but performances should be good.

Cyrille


Martin Schied a écrit :
> Thibault Walter wrote:
>> Hello
>>
>> I made a little discovery.
>>
>> It works with pd-extended and with pd-vanilla+GEM. Il have a MacBook Pro
>> 2.66 GHz Intel Core 2 duo, with Mac-OS X 10.5.8.
>>
>> When I open one of my video-patch (it works with any one of my video or
>> sound patchs), if I look at the CPU utilisation, I get something like 
>> 7 ou
>> 8. Nothing is running, the GEM window is not even created.
>>
>> If I wait something like 1 or 2 hours without using pd ; or, if I 
>> pause my
>> computer (I close it), wait 30 seconds, and open it, I get something 
>> like 0
>> or 1 as my CPU.
>>   
> This might be a cpu frequency scaling issue. I had similar "phenomena" 
> on Ubuntu once, where the CPU didn't go into its lowest frequency 
> setting anymore after waking up from suspend mode.
> 
> So for example on a fresh reboot your cpu would be able to switch 
> between 800MHz, 1300MHz and 2.6GHz or so, most of the time it's on the 
> lowest setting and therefore your CPU is quite busy (still on 800MHz). 
> Then after the suspend (closing it) it never uses the 800MHz setting but 
> only the highest - and so the cpu cycles percentage where the cpu is 
> used is far less. If you don't want to do the "suspend trick" every time 
> there might be a "performance" setting or so which does the same thing. 
> On ubuntu you can do this using 'cpufreq-set -g performance' on the 
> command line, not sure what to do on OS X.
> 
> - just a guess, might be totally wrong too. -
> 
> cheers
> Martin
> 
>> Then, if I use my patch, when I get normally CPU = 70, and with this 
>> method
>> I get CPU = 20. And of course, the performances are absolutely not the 
>> same.
>>
>> In fact, without this method, my MacBook Pro is not very much more 
>> powerful
>> than my old 1.33GHz PowerPC G4.
>>
>>
>> It works with audio-patchs too.
>>
>> There is an other method : if I turn dsp on, and then off, I get the same
>> result. This method can't work with audio-patchs, because when I turn 
>> on dsp
>> I always get high CPU. So I have to turn dsp on before pausing and 
>> turning
>> on my computer. (I don't know if I'm clear...)
>>
>>
>> Weeell, I realized that it works with my PPC too, but the differences are
>> not so important.
>>
>> Has someone any explanation?
>>
>> excuse me for my english
>>
>> T
>>
>>
>>
>>
>> _______________________________________________
>> Pd-list at iem.at mailing list
>> UNSUBSCRIBE and account-management -> 
>> http://lists.puredata.info/listinfo/pd-list
>>
>>   
> 
> 
> _______________________________________________
> Pd-list at iem.at mailing list
> UNSUBSCRIBE and account-management -> 
> http://lists.puredata.info/listinfo/pd-list
>