Hello All, I am noticing an interesting thing in some simulation results and I was wondering if anyone has an explanation. I have implemented a short time fourier transform algorithm and mapped it across between 1 and 4 processors and simulated it in the M5 simulator. I am noticing that one of the processors will have a very large amount of ticks spent in user mode and the other processors have between 10 and 100 times less ticks spent in USER mode. I am wondering why this time discrepancy is occurring. The way I have split the algorithm each core should have the same amount work to do on an equal amount of data. I have implemented it with Pthreads. Any input?
Aaron -- Aaron S. Williams Graduate Student Arizona State University Department of Electrical Engineering [EMAIL PROTECTED]
_______________________________________________ m5-users mailing list [email protected] http://m5sim.org/cgi-bin/mailman/listinfo/m5-users
