On Fri, Apr 19, 2013 at 10:16 PM, Hassan, Ahmad <ahmad.has...@sap.com> wrote:
>>Ok, but the -i is only effective to follow thru the creation of the
>>threads and not when they are
>>already spawned. I see you attaching after program launch. At that
>>time, are all threads already
>>created?
>
> Thanks Stephane. That explains it. I am it giving another go as follows:
>
> (./task -e LLC-LOADS,cycles,instructions  -i START DB) 1>out &
>
> Please can you tell me what is the libpfm4 sampling rate for collecting the 
> stats? My single test will run for 2-4 hours with massive memory traffic and 
> there would be huge number of cycles,instructions and LLC_LOADS. Will that be 
> fine to use libpfm4  with long running application and how representative the 
> statistics would be?
>
You are not sampling here. You are simply counting the events.
Overhead is minimal.
You should really use the perf tool for any serious measurement. The
libpfm4 example
and just demo programs.

> Thanks.
>
> Kind Regards, Ahmad
> SAP
>
>
>
>
>
>
>
>

------------------------------------------------------------------------------
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
_______________________________________________
perfmon2-devel mailing list
perfmon2-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/perfmon2-devel

Reply via email to