[Haizea] Final CPU Utilization

[Mehdi Sheikhalishahi]

Interesting issue with utilization.
So you mean because of ramp down effect even
CPUUtilizationProbe.COUNTER_UTILIZATION counter will not be a very good
parameter to show CPU Utilization in general?

Cheers.

On Tue, Jun 15, 2010 at 5:54 PM, Borja Sotomayor <borja at borjanet.com> wrote:

> Hi,
>
> The absolute CPU utilization at the end of a simulation is always
> going to be zero, since all resources have been freed up at the end of
> a simulation. The more interesting value is the final value of the
> average CPU utilization and, even then, the very final value might not
> be too representative because of "ramp down" effects: at the end of a
> simulation, as the final requests are being processed, the CPU
> utilization will drop ("ramp down") because no new requests are being
> submitted.
>
> What I usually do is take the average utilization at the time the last
> lease in the simulation is submitted; however, I usually extract this
> value using R (when analyzing the accounting file), and not using the
> accounting probe infrastructure. I'm actually finishing up my
> dissertation, and I'll be including detailed instructions on how I use
> R to process results, but this information won't be available until
> July or August.
>
>
> > What about slowdown, is there any probe already defined in Haizea?
>
> Yes, the BEProbe will compute the slowdown of best-effort leases. If
> you need to use the slowdown metric with other types of leases, the
> code should be easy to copy to a different probe.
>
> Cheers!
> --
> Borja Sotomayor
> PhD Candidate in Computer Science, University of Chicago
> http://people.cs.uchicago.edu/~borja/
> _______________________________________________
> Haizea mailing list
> Haizea at mailman.cs.uchicago.edu
> https://mailman.cs.uchicago.edu/mailman/listinfo/haizea
>



-- 
Regards,
Mehdi
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.cs.uchicago.edu/pipermail/haizea/attachments/20100615/42f06620/attachment.htm