public inbox for linux-ia64@vger.kernel.org
 help / color / mirror / Atom feed
* CPU utilization measurement tools/methods for network performance tests
@ 2004-09-27 16:24 Alex Tsariounov
  2004-09-27 22:51 ` Peter Chubb
  0 siblings, 1 reply; 2+ messages in thread
From: Alex Tsariounov @ 2004-09-27 16:24 UTC (permalink / raw)
  To: linux-ia64

Hello:

I am currently using netperf to measure performance of a couple of
network drivers on ia64/2.6.7 and am having a trying time measuring
CPU utilization.

Does anyone have a dependable tool or method to measure CPU
utilization, especially over a minute/two-minute-long measurement
interval?

I'd like to get a little more accurate and dependable result than just
eye-balling top for the run.  Also, something that does not interfere
with the test would be important.

The CPU utilization measurement methods provided by netperf produce
different results depending on which one is used and I'd like to use a
secondary method to confirm them.

Thank you,
Alex

^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: CPU utilization measurement tools/methods for network performance tests
  2004-09-27 16:24 CPU utilization measurement tools/methods for network performance tests Alex Tsariounov
@ 2004-09-27 22:51 ` Peter Chubb
  0 siblings, 0 replies; 2+ messages in thread
From: Peter Chubb @ 2004-09-27 22:51 UTC (permalink / raw)
  To: linux-ia64

>>>>> "Alex" = Alex Tsariounov <alext@fc.hp.com> writes:

Alex> Hello: I am currently using netperf to measure performance of a
Alex> couple of network drivers on ia64/2.6.7 and am having a trying
Alex> time measuring CPU utilization.

I suggest you check out ipbench (http://ipbench.sf.net) which in its
latency measurement has an accurate CPU-usage measurement algorithm,
at least for uniprocessors.  Because so much network handling happens
at interrupt time, the usual system and user times are useless for
this purpose.

There's going to be a new release of ipbench sometime soon now, a
complete rewrite using Python to drive the system, that should be a
lot more stabole and usable than the current version.



-- 
Dr Peter Chubb  http://www.gelato.unsw.edu.au  peterc AT gelato.unsw.edu.au
The technical we do immediately,  the political takes *forever*

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2004-09-27 22:51 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2004-09-27 16:24 CPU utilization measurement tools/methods for network performance tests Alex Tsariounov
2004-09-27 22:51 ` Peter Chubb

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox