Many ways to get to a solution.. how do I measure performance of M4L devices?


    Jan 11 2012 | 3:24 pm
    Hi all - I find myself often making choices between methods of implementing an idea often balancing flexibility vs performance.
    For example, my Ipad controls rack effects. When I change track, it queries all the parameter names for the racks on that track. I could instead have all parameter names for all tracks requested when I start up Live and store them (eg using coll) for retrieval when required. This will probably be faster/better but _how much_ faster and better?
    I love using 'pedant' for simple structures but I don't think it's very useful for measuring M4L devices. What do you guys use?
    Thanks!

    • Jan 11 2012 | 11:30 pm
      I query the names at the beginning. I find it makes about .25 seconds faster when switching patches.
    • Jan 12 2012 | 3:06 am
      Thanks benj373! did you find a way to measure that or is that your estimate from experience?
    • Jan 24 2012 | 12:10 pm
      I always use cpuclock for speed tests (and use an uzi to repeat that same action lots of times to get a representable average time)
    • Jan 24 2012 | 1:48 pm
      This tool is really useful. I put in a bpatcher form, and have it saved as a clipping so I can easily call it up if I want to compare two methods of doing something in a patch.
    • Feb 04 2012 | 4:26 pm
      Thanks David! I use Pedant all the time, but haven't worked out a way to use it properly in Live. I'm curious how you do it.. the code that you compare needs to be inside pedant's subpatcher right?
      I wonder how you manage to use the comparison in a bpatcher.
      @Timo - I'm going to try that!
      Cheers,
      Bas