Re: Oprofile vs. perf vs. gperf
Mahmood N <[email protected]>
| Newsgroups | gmane.linux.oprofile |
|---|---|
| Message-ID | <[email protected]> |
Thanks for the replies. >It would be useful to describe what you are interested in measuring and >then figure out what tools and techniques would best accomplish that. The big picture of what I want to do is to enable the hardware prefetcher, run some programs with the profiler. Then disable the hardware prefetcher, run the programs again and compare the results. Mainly, I am looking for the number of prefetch request sent to the memory, the data cache miss (L1), instructions retired and the processor cycles. Is it enough to use ocount? BTW, my AMD 6274 (15h) has L1, L2 and L3 caches but I don't see any L3 miss/access/requests. The available parameters are http://pastebin.com/Ae1dteRE Regards, Mahmood ------------------------------------------------------------------------------ _______________________________________________ oprofile-list mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/oprofile-list