Re: 100% CPU

Jensen Harris (jensen.harris@yale.edu)
Wed, 16 Apr 1997 20:52:39 -0400 (EDT)


On Wed, 16 Apr 1997, Scott McIntyre wrote:

> I got this program called net.medic somewhere. It does network
> analysis etc., all of which is not very relevant. But it also does
> cpu monitoring. When DESchall is running, the graph is pegged at
> 100% CPU utilization. Kinda makes me feel like I'm getting my
> money's worth out of my processor.

It's sort of funny, actually. We were denied the use of 15 Linux
Workstations here by the CS department. Mind you, these workstations are
completely idle all day, just waiting for the next big project (which will
be this summer). But the people in charge ran the client for a day and
then said that the DESCHALL program was too great a tax on the processor.
I'm serious. They said "well, these are the newest computers we have and
we don't want to wear out the processors." I glared at them in disbelief.
"It probably also voids our warranty with HP to run programs like this
because it is an undue strain on the processor." (These are P5-166 HP
workstations). He also said things like "Processors like these are only
'rated' a couple of thousand cycles a minute--going over that is not
something we're about to attempt without study as the effects beforehand."

At this point I was ready to strangle to guy.

"Tell them that if they can make a nice client that doesn't run the
processor too hard that we'd be happy to help."

Note again--he's NOT concerned about the effects of the other users on the
systems (there aren't any). He's concerned about "wearing out the
processors". Like they're going to stop working much more quickly if we
run deschall on them than if they sit idle all day...

If someone with impressive bylines or credentials wants to write a short
note explaining why this may not be too great of a concern (!), I'd
appreciate something I could pass on.

Jensen