[PD] PD effiecientcy

[matthew jones]

strangely enough, I tend to run permanently at 100% cpu and get no audio
glitches (winXP).  the gui just slows down and can take a while to respond.
even stranger though, is how it only started doing this after I wiped and
started from scratch (new hard disc) and put more memory in.  before this
clean-reinstall it was exactly the same installation, but I had some
glitches when I moved a sliderbar too fast, for example.  but for sure, the
gui responded a lot faster (before) - the vu meters used to respond really
nicely, now they just crawl.

then again, I am still running 0.35 (!) at home.

matt

-=-=-=-=-=-=-=-=-=-=-=-=-
http://www.loopit.org/
-=-=-=-=-=-=-=-=-=-=-=-=-
----- Original Message -----
From: "Bradon Webb" <amodiuslonodium at yahoo.com>
To: <pd-list at iem.at>
Sent: Wednesday, January 28, 2004 2:58 PM
Subject: [PD] PD effiecientcy


> Hello,
>
> I am running PD on winXP laptop with 1.2 gig 512mB
> Ram,
>
> I don't seem to be having trouble running other audio
> or video applications, with processing so I feel like
> my computer is not necessarily the problem.  I feel
> like I should be able to put more into my PD patches
> before they start cutting out from the audio
> calculation/processing.  I guess the term is underuns?
> when the computer cannot process all the data within
> the specified unit of time to playback smoothly.  I
> seem to be upward 50-70% cpu usage peaking.. my
> questions are thee;
>
> 1.how much (%)CPU usage can I expect to use while
> still maintaining smooth audio output?
>
> 2.If I redesign all my abstractions so there is no GOP
> or visual interface how much faster can I expect the
> patch to run (% cpu).
>
> 3.How do I make the patches more efficient?
>
> 4.How much does running other sound aps in the
> background while PD is running effect/contribute to
> the underruns... It doesn't seem like it is
> contributing much to the CPU problem, becasue I get
> underuns either way with it running, or with it off.
> Should I never do this?
>
> it seems like the only solution right now is to
> somehow cut all the tasks and render them to wav files
> and then use multitracking to put layers back
> together.  which means live performance is pretty much
> out of the question at this point.  I want to get 2x -
> 3x the performance I am getting right now.  But that
> is just an arbitrary feeling of how complex I want the
> patches to be, If the machine can't handle it then
> there is nothing I can do about it, except become a
> better programmer...
>
> Thanks for any advice.
> Bradon
>
> p.s. I know it is kind of abstract to talk this way
> without a specific example, if anyone wants to see a
> patch of mine, I would be more than happy to post for
> a critique.
>
> __________________________________
> Do you Yahoo!?
> Yahoo! SiteBuilder - Free web site building tool. Try it!
> http://webhosting.yahoo.com/ps/sb/
>
> _______________________________________________
> PD-list mailing list
> PD-list at iem.at
> http://iem.at/cgi-bin/mailman/listinfo/pd-list
>