max5: waveform~ = cpu hog? what happened to buftime message?


    Apr 27 2008 | 9:37 am
    i haven't done a lot of cpu util. comparison between 4 and 5 so far,
    but this one really seems ridiculous.
    while trying out an old max4 patch (that uses a bunch of waveform~
    objects) in max5 i noticed that my computer started melting pretty
    soon :(
    running the attached test patch on my computer ( ppc pb g4, 1.5GHz,
    os 10.4.11) i got the following results:
    max4
    cpu: 0%
    processor load (activity monitor or top): 30 - 35%
    max 5:
    cpu: 2%
    processor load: 75 - 85%, with occasional spikes to over 90%
    while i know that waveform~ always seemed to draw a lot of power,
    there was the (undocumented) buftime message, which made the decision
    between update rate and cpu util. somewhat manageable.
    unfortunately this message doesn't work in max5 (max is not
    complaining about unknown message, but doesn't change the update rate
    of waveform~ either).
    any comments?
    thanks,
    volker.

    • Apr 27 2008 | 2:47 pm
    • Apr 28 2008 | 9:08 am
      On 27 Apr 2008, at 16:47, Klaas-Jan Govaart wrote:
      >>
      >> running the attached test patch on my computer ( ppc pb g4, 1.5GHz,
      >> os 10.4.11) i got the following results:
      >> max4
      >> cpu: 0%
      >> processor load (activity monitor or top): 30 - 35%
      >>
      >> max 5:
      >> cpu: 2%
      >> processor load: 75 - 85%, with occasional spikes to over 90%
      >>
      >
      > 10.4.11, mbp 2.16GHz:
      >
      > max 4 cpu 0%, activity monitor 19%
      > max 5 cpu 1-2%, activity monitor 83-88%
      thanks for testing, klaas.
      any other takers?
      is someone from cycling looking into it?
      something definitely IS going wrong.
    • Apr 28 2008 | 10:56 am
    • Oct 15 2008 | 1:33 pm
      Hello,
      Any news on this front? Has waveform~'s cpu usage gone down after any of the updates? Or is it like the multislider cpu issue - Inherent to JUCE engine, and not improving anytime soon?
      Any thoughts appreciated.
      kjg
    • Oct 15 2008 | 1:53 pm
      On 15 Oct 2008, at 15:33, Klaas-Jan Govaart wrote:
      >
      > Hello,
      >
      > Any news on this front? Has waveform~'s cpu usage gone down after
      > any of the updates? Or is it like the multislider cpu issue -
      > Inherent to JUCE engine, and not improving anytime soon?
      >
      > Any thoughts appreciated.
      > kjg
      the buftime message is back since 5.0.5 (maybe earlier), which makes
      the update rate of [waveform~] manageable again.
      i haven't made any direct comparisons between 4.6 and 5.0.5, yet, but
      the default setting of 500 ms seems to give reasonable cpu measures
      on my system using max 5.
      volker.
    • Oct 15 2008 | 2:29 pm