CPU bottleneck on HP-UX 11.31

[Humberto Gámez Rábago]

Hi,

I have a HP-UX 11.31 system running on ia64 hp superdome server SD64B.

Question 1:

Reading the paperdocs for OVPA ovpa-metrics.txt... i can't understand why the GBL_CPU_QUEUE must to be greater than four, to considere there is a
high probability of a CPU bottleneck.?

Question 2:

it is correct the GBL_CPU_QUEUE value shown in the second scenario of the next text?

For example, let's assume we're using a system with eight processors. We start eight CPU intensive threads that consume almost all of the CPU resources. The approximate
values shown for the CPU related queue metrics would be:



GBL_RUN_QUEUE = 1.0

GBL_PRI_QUEUE = 0.1

GBL_CPU_QUEUE = 1.0


Assume we start an additional eight CPU intensive threads. The approximate values now shown are:



GBL_RUN_QUEUE = 2.0

GBL_PRI_QUEUE = 8.0

GBL_CPU_QUEUE = 16.0
<---- ????

Question 3:

For multiprocessors systems maybe the best metric for CPU bottleneck should be the GBL_RUN_QUEUE, it is correct?

[Laurent Menase]

difficult to tell it just like that.
you probably need to have ask for a performance assessment.

[AGUSTIN LIZARRAGA]

Hi Humberto,

The following metric doesn't make sense its use alone by itself:
GBL_CPU_QUEUE
You should take in consideration another meaningfull metric as GBL_RUN_QUEUE which considers the average among *all* processors installed on the systems. That's the metric I wouuld put on my tool bag to access whenever possible and compare its value along the time specially when the server is subject to heavy load. How responsiveness is affected then You should consider other factors as well.

Best Regards.
Agustin Lizarraga
RSAA
GSC AMS ESS

[Hakki Aydin Ucar]

Hi,

There is a link belong to Steven Protter:
http://www.hpux.ws/?p=6