topic Re: what does LOAD mean when using top? in Operating System - HP-UX

what does LOAD mean when using top?

Richard Mertz — Wed, 11 Oct 2000 13:01:53 GMT

Does anyone know what the LOAD numbers mean that are reported by top?

Re: what does LOAD mean when using top?

Alex Glennie — Wed, 11 Oct 2000 13:04:38 GMT

run "man top" and compare the headings with the descriptions ;)

Re: what does LOAD mean when using top?

Alex Glennie — Wed, 11 Oct 2000 13:12:05 GMT

Ah ... just did and it doesn't mention these ! I suspect they tie up with the same values found under glance which I think does come with descriptions for the various metrics.

Re: what does LOAD mean when using top?

Lasse Knudsen — Wed, 11 Oct 2000 13:14:10 GMT

LOAD means load average and what i actually means is how many processes is in the "run queue" of the scheduler.

It is *not* a very good measure of how loaded a system is. You could easyly have a load avg. of 5 (high) but you CPU does not do anything. This happens when processes are waiting for I/O.

In order to determine how loaded a system is you would use output from a number of tools. I personally favour a combination of 'vmstat' and 'sar'

Re: what does LOAD mean when using top?

CHRIS_ANORUO — Wed, 11 Oct 2000 13:17:11 GMT

It is the average number of jobs on the run queue over 1, 5 , 15 minutes.
You can also run uptime.

Re: what does LOAD mean when using top?

Robert Gamble — Wed, 11 Oct 2000 13:23:12 GMT

Try looking at this older thread:
http://my1.itrc.hp.com/cm/QuestionAnswer/1,1150,0xc7c1119c3420d411b66300108302854d,00.html

Also, look at Document ID: W3564348

Good Luck!

Re: what does LOAD mean when using top?

Alan Riggs — Wed, 11 Oct 2000 13:37:26 GMT

LOAD is an aggregate of running processes, processes in teh run-queue, and processes waiting on I/O.

Re: what does LOAD mean when using top?

James R. Ferguson — Wed, 11 Oct 2000 15:06:49 GMT

Richard:

BTW, Glance (on the second screen) of CPU data gives load too. This metric is derived from a kernel variable (avenrun) which is calculated by summing the number of runnable processes or threads and averaging the samples over the last minute.

...JRF...