Forum > GPU crunching

[Split] PowerSpectrum Unit Test

<< < (41/62) > >>

Jason G:
How did you do ten runs, while collecting data, on 'that thing' in that timeframe ?  magic ?
[ Oh yeah I set the timer tolerances to do that, I forgot  ::)]

Miep:

--- Quote from: Jason G on 21 Dec 2010, 12:14:18 pm ---How did you do ten runs, while collecting data, on 'that thing' in that timeframe ?  magic ?
[ Oh yeah I set the timer tolerances to do that, I forgot ::)]

--- End quote ---

A run takes some 20 seconds - makes some 5 minutes with graceful rounding. Typing the data into Excel and the calculated values back into the post took about half an hour.  :P

timer tolerances?

Jason G:

--- Quote from: Miep on 21 Dec 2010, 12:23:27 pm ---timer tolerances?

--- End quote ---

Yeah, faster cards probably do 'a few more' runs within the allocated 0.5 seconds per test  ;)

[BTW:] On Opt1, See the difference in the standard deviations of best & worse cases ? , That's memory&bus contention on the worst cases randomising things up a bit  :)

Miep:

--- Quote from: Jason G on 21 Dec 2010, 12:27:14 pm ---Yeah, faster cards probably do 'a few more' runs within the allocated 0.5 seconds per test  ;)
--- End quote ---

Well manual data collection works just as well, only more tedious.


--- Quote ---[BTW:] On Opt1, See the difference in the standard deviations of best & worse cases ? , That's memory&bus contention on the worst cases randomising things up a bit  :)

--- End quote ---

I was wondering more about the apparent lack of variation on the best case. I would have expected a little more fluctuation.

Jason G:

--- Quote from: Miep on 21 Dec 2010, 12:44:19 pm ---I was wondering more about the apparent lack of variation on the best case. I would have expected a little more fluctuation.
--- End quote ---

Best case requires few memory transfers back to the host CPU ( only one best spike & no detections)  ;)

[Edit:] Worst case would be a best signal + numdatapoints/fftlen detections, i.e. not really possible since we're limited to 30 detections, so wouldn't bother transferring more than the first 30 ( ... unlike stock...)

Navigation

[0] Message Index

[#] Next page

[*] Previous page

Go to full version