[smooth-biquad~] CPU consuming


    May 08 2008 | 2:31 pm
    Hi (and especialy hi to cnmat),
    I recently replaced my 64 [biquad~] instances in the patch I'm working
    on by [smooth-biquad~]s. Under some conditions (in that particular
    case, stopping a [groove~] or a [sfplay~], which is the only sound
    source in the patch), the CPU usage starts to raise up to 100%. This
    doesn't happen with the regular [biquad~] object. Turning the DSP off
    and on puts MSP back to a normal state, as does a "clear" to all
    instances. Most of the [smooth-biquad~] are in muted voices of
    [poly~]...
    I know I'm vague but I couldn't reproduce this on small patches. Could
    it be related to this denorm bug on intel? I can't test on a PowerPC
    now, but will as soon as I can.
    _____________________________
    Patrick Delges
    Centre de Recherches et de Formation Musicales de Wallonie asbl

    • May 08 2008 | 2:55 pm
      I have found that I have to mute poly~ voices with a load bang. They
      don't seem to be born muted.
      Cheers
      Gary Lee Nelson
      TIMARA Department
      Oberlin College
      www.timara.oberlin.edu/GaryLeeNelson
      On May 8, 2008, at 4:31 PM, Patrick Delges wrote:
      > Hi (and especialy hi to cnmat),
      >
      > I recently replaced my 64 [biquad~] instances in the patch I'm
      > working on by [smooth-biquad~]s. Under some conditions (in that
      > particular case, stopping a [groove~] or a [sfplay~], which is the
      > only sound source in the patch), the CPU usage starts to raise up
      > to 100%. This doesn't happen with the regular [biquad~] object.
      > Turning the DSP off and on puts MSP back to a normal state, as does
      > a "clear" to all instances. Most of the [smooth-biquad~] are in
      > muted voices of [poly~]...
      > I know I'm vague but I couldn't reproduce this on small patches.
      > Could it be related to this denorm bug on intel? I can't test on a
      > PowerPC now, but will as soon as I can.
      >
      > _____________________________
      > Patrick Delges
      >
      > Centre de Recherches et de Formation Musicales de Wallonie asbl
      > http://www.crfmw.be/max
      >
    • May 08 2008 | 6:59 pm
    • May 09 2008 | 7:53 am
      I had a chance to test my patch on a G5 (OSX.4.11), and there is no
      problem on it. So it seems to be related to intel (core 2 duo, X.5.2,
      Max 4 and Max5).
      p
      On 8 mai 08, at 16:31, Patrick Delges wrote:
      > Hi (and especialy hi to cnmat),
      >
      > I recently replaced my 64 [biquad~] instances in the patch I'm
      > working on by [smooth-biquad~]s. Under some conditions (in that
      > particular case, stopping a [groove~] or a [sfplay~], which is the
      > only sound source in the patch), the CPU usage starts to raise up to
      > 100%. This doesn't happen with the regular [biquad~] object. Turning
      > the DSP off and on puts MSP back to a normal state, as does a
      > "clear" to all instances. Most of the [smooth-biquad~] are in muted
      > voices of [poly~]...
      > I know I'm vague but I couldn't reproduce this on small patches.
      > Could it be related to this denorm bug on intel? I can't test on a
      > PowerPC now, but will as soon as I can.
      _____________________________
      Patrick Delges
      Centre de Recherches et de Formation Musicales de Wallonie asbl