On Thu, Jun 09, 2005 at 08:52:10PM -0700, Shirley Ma wrote: > > Does rdma_lat have the same issue? > > What does rdma_lat report? > > Latency typical: 0.717812 usec > Latency best : 0.705625 usec > > Latency typical: 1149 cycles > Latency best : 1129.5 cycles >
I'm convinced something is broken and that gettimeofday() works better in your case. Please provide "cat /proc/cpuinfo" output so I can rule out broken firmware and look for something else. > after modifying the per usec cycles > the results are: > > Latency typical: 5.76884 usec > Latency best : 5.66834 usec > > gettimeofday() results are: > > Latency typical: 6.000000 usec > Latency best : 5.500000 usec These result suggest gettimeofday() is not very accurate and less reliable. Note that I currently have patches outstanding on both rdma_bw.c and rdma_lat.c. I'm waiting for Michael to reject, accept, or comment on them. Or can someone else commit changes to those files for me? thanks, grant _______________________________________________ openib-general mailing list [email protected] http://openib.org/mailman/listinfo/openib-general To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
