On 07/06/2011 08:29 PM, Ying Zhang wrote: > Hi, > > I've briefly looked through perfmon's document. It seems that > perfmon provides both time-based sampling and event-based sampling. In > the event-based sampling mode, a sampling is made when a preset event > reaches a specific threshold. I think this is exactly what I want. That > is, collect the counter values after every million instructions are > retired. Also, perfmon provides two options: --long-smpl-periods and > --short-smpl-periods. I believe these two options are related to the > sampling period setting. Can anyone generally explain how to do a > sampling in event-based mode? > > Maybe this question is sort of silly, but please, I really have no > idea...
I'd like to reiterate that if you want to use perfmon, you will have to downgrade the kernel that you are using to 2.6.30 or earlier. You can do event-based sampling with perf_events as well, but it's not trivial to describe in an email how to accomplish this with the perf_events kernel API. There is some information in /usr/include/linux/perf_event.h but perhaps not enough to be a HOW TO. There's also some information here, but unfortunately it's more than a little out of date: http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=blob;f=tools/perf/design.txt;h=bd0bb1b1279b8ec258cc40a231fc4a9d289abd96;hb=HEAD - Corey ------------------------------------------------------------------------------ All of the data generated in your IT infrastructure is seriously valuable. Why? It contains a definitive record of application performance, security threats, fraudulent activity, and more. Splunk takes this data and makes sense of it. IT sense. And common sense. http://p.sf.net/sfu/splunk-d2d-c2 _______________________________________________ perfmon2-devel mailing list perfmon2-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/perfmon2-devel
