Hello......
I have come to an important subject..
Please look at the USL methodology here:
http://www.perfdynamics.com/Manifesto/USLscalability.html
Dr. Gunther the author of the USL methodology is using
nonlinear regression on the performance data measurements
to predict scalability?
But i don't understand this nonsense !
Simply because, in a parallel program, a locked region using
locks can escape contention at fewer cores and fewer
threads and can than hit the contention at more cores and more threads..
so how can you than use nonlinear regression to predict
scalability ? this is complete nonsense !
Can you please explain to me why Dr. Gunther the author of USL
is using nonlinear regression to predict scalability ?
Thank you,
Amine Moulay Ramdane.
--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)