Hi.
Isn't it a bit too high?
On my P3 800 playing an mp3 or ogg takes ~40% CPU. Alsaplayer is about ~5% on the same machine.
Thanks,
Maurizio Umberto Puxeddu.
Hallo, august hat gesagt: // august wrote:
Isn't it a bit too high?
On my P3 800 playing an mp3 or ogg takes ~40% CPU. Alsaplayer is about ~5% on the same machine.
strange. on my p3 600 mhz machine, an mp3 or ogg takes only 1-5%
did you compile libmad and ogg with nasm support?
I also have normal cpu-load, even less than with alsaplayer... ciao
Probably it has nothing to do with this object. I think there are some dark sides of Pure Data CPU usage.
When I measured those numbers I had my current main patch loaded. In that situation I'm pretty sure that starting/stopping readanysf~ on a mp3 made a 40% different in the CPU load.
I shut down PD, loaded only the readanysf~ patch started a ogg and got a normal CPU usage (<5%).
I launched it again with my main patch loaded and starting readanysf~ causes a stable 45-50% of CPU usage *increase* (in fact my patch components were almost all [switch~]ed off so the CPU usage was nearly 0 without readanysf~ anyway).
In past I noticed also other strange behaviours of PD (like its CPU usage not getting down when turning off relevant section of the synth) but this beats them all.
Is there something more about PD that I need to know?
Thanks,
Maurizio Umberto Puxeddu.
On Thu, 2003-05-15 at 14:03, Frank Barknecht wrote:
On my P3 800 playing an mp3 or ogg takes ~40% CPU. Alsaplayer is about ~5% on the same machine.
strange. on my p3 600 mhz machine, an mp3 or ogg takes only 1-5%
did you compile libmad and ogg with nasm support?
I also have normal cpu-load, even less than with alsaplayer... ciao
Hallo, Maurizio Umberto Puxeddu hat gesagt: // Maurizio Umberto Puxeddu wrote:
Probably it has nothing to do with this object. I think there are some dark sides of Pure Data CPU usage.
Might be...
I observed another stange thing with readany, though: To load it for the first time, takes very long - much longer than all other libraries together. Does anyone know the reason for this? I suspect some memory allocation...
ciao
Frank Barknecht _ ______footils.org__
Probably it has nothing to do with this object. I think there are some dark sides of Pure Data CPU usage.
Might be...
I observed another stange thing with readany, though: To load it for the first time, takes very long - much longer than all other libraries together. Does anyone know the reason for this? I suspect some memory allocation...
where do you have readanysf~ installed? maybe pd is looking for it.
otherwise, there is one internal buffer that gets allocated at the beginning. its 16*4*4096 = 262144 bytes large. quite large actually. this could come down to 64k and still be usefull I think. maybe even 32k. probably even best to make it variable according to init paramaters.
ill try and do some testing.
-august.
-
august -
Frank Barknecht -
Maurizio Umberto Puxeddu