Re: Performance Monitoring Counters (PMC, PERFCTRS) and SMP?

To: NetBSD Multi-Processor Technical Discussion List <tech-smp%NetBSD.org@localhost>, NetBSD-current Users's Discussion List <current-users%netbsd.org@localhost>
Subject: Re: Performance Monitoring Counters (PMC, PERFCTRS) and SMP?
From: David Laight <david%l8s.co.uk@localhost>
Date: Sun, 16 Oct 2011 12:56:44 +0100

On Sat, Oct 15, 2011 at 08:32:26PM -0700, Greg A. Woods wrote:
> 
> I ask mostly because I'm flustered by getrusage(2) being as good as
> useless for even the most basic of benchmarking purposes...

On NetBSD the sum of the 'sys' and 'user' times is accurate.
As is the 'child cpu time' for reaped zombies.
I've used that from ps (with custom columns) to get high resolution
execution times.

        David

-- 
David Laight: david%l8s.co.uk@localhost

Follow-Ups:
- re: Performance Monitoring Counters (PMC, PERFCTRS) and SMP?
  - From: matthew green

References:
- Performance Monitoring Counters (PMC, PERFCTRS) and SMP?
  - From: Greg A. Woods

Prev by Date: Performance Monitoring Counters (PMC, PERFCTRS) and SMP?
Next by Date: re: Performance Monitoring Counters (PMC, PERFCTRS) and SMP?
Previous by Thread: Performance Monitoring Counters (PMC, PERFCTRS) and SMP?
Next by Thread: re: Performance Monitoring Counters (PMC, PERFCTRS) and SMP?
Indexes:

Home | Main Index | Thread Index | Old Index