speed testing question/strategy


    Apr 26 2007 | 6:33 pm
    Greetings folks
    I am building a multiple machine setup for
    performance and was doing some testing to try and figure which machine
    to use where - 2 for video and 1 for audio/control
    these are the 3 machines I have access to :
    15" Powerbook G4 1.67ghz, 2 gig RAM, ATI Mobility Radeon 9700
    17" Powerbook G4 1.5ghz, 1.5 gig RAM, ATI Mobility Radeon 9700
    iMac G5, 1.8GHZ, 1GIG RAM, GFORCE FX 5200
    Not having a great idea how to do a basic speed test on these machines I opted to use the CPU and GPU patches from the jitter slab tutorial - #44
    My results were a bit surprising to me:
    15" Powerbook G4 1.67 ghz - 7.8 fps CPU, 30fps GPU
    17" Powerbook G4 1.5 ghz - 7.8 fps CPU, 30fps GPU
    iMac G5, 1.8GHZ - 16.5 FPS CPU, 17.8 FPS GPU
    I was wondering if this seems like a reasonable/ the best way to do this kind of preliminary testing or if there is a better option -
    obviously the best choice is to make the actual patches and then see what really happens
    but these results suggest significant differences in programming stratagies on the iMac vs on the Powerbooks...
    any thoughts?
    thanks
    jamie

    • Apr 26 2007 | 7:03 pm
      The powerbook's graphics cards are better, while the imac's cpu / bus
      is better. given that imac's don't output a 2nd desktop without a
      software hack (that may not matter to you?), its fairly obvious to go
      with the powerbooks using GPU based display (opengl /
      jit.gl.videoplane etc)
      -deKam
      > Greetings folks
      >
      > I am building a multiple machine setup for
      > performance and was doing some testing to try and figure which machine
      > to use where - 2 for video and 1 for audio/control
      >
    • Apr 26 2007 | 7:52 pm
      ya - I agreed with this assessment
      though I am using the software hack
      really though this was a current example for a bigger question about testing strategies...
      - is this how people do this kind of preliminary testing or is there a different /better model?
      just the computer specs don't always tell the whole story -
      for example, in this case the 17" 1.5ghz machine was actually slightly faster than the 1.67ghz 15" machine
      but thanks for the response
      jamie
      ----- Original Message ----
      From: dekam
      Sent: Thursday, April 26, 2007 3:03:57 PM
      Subject: Re: [jitter] speed testing question/strategy
      The powerbook's graphics cards are better, while the imac's cpu / bus
      is better. given that imac's don't output a 2nd desktop without a
      software hack (that may not matter to you?), its fairly obvious to go
      with the powerbooks using GPU based display (opengl /
      jit.gl.videoplane etc)
      -deKam
      > Greetings folks
      >
      > I am building a multiple machine setup for
      > performance and was doing some testing to try and figure which machine
      > to use where - 2 for video and 1 for audio/control
      >