Subject: Re: probing CPU speed?
To: Chris G. Demetriou <>
From: Brian Stark <>
List: port-i386
Date: 11/19/1998 10:35:18
On 18 Nov 1998, Chris G. Demetriou wrote:

> *sigh*
> And for what purpose would it be nice?

> Who _cares_ what the raw CPU speed is?  It's at best useless, at worst
> seriously deceptive...

Well, it's not useless, but I do agree that it can be deceptive because
getting a precise measurement can be tricky (or so it would appear from
the comments I've read in mhz.c).

anyway, here is an example of how I have found the mhz program (which
prints what it thinks is the CPU speed) to be useful: during the past
few months I have had access to about a dozen different workstations that
I have been using to participate in the RC5-64 contest (see Most of these workstations have very similar
sub-systems (CPU type, SCSI adapters, disk drives, graphic cards, etc...).
The biggest difference between these machines has been the CPU speed.
Since the RC5 client program is CPU-intensive I decided to write some
scripts that first run the mhz program to get an estimate of the CPU speed
and then based on that a length of time is determined to run the RC5
client program. Workstations that have a higher estimated CPU speed will
run the RC5 client for a longer period of time and in the same way, slower
workstations will run the client program for a shorter time. Doing this
allowed my "faster" workstations to run the RC5 client longer than the
slower ones, and I found this to be useful.


| Brian Stark                       | Internet : |
| Siemens PT&D, LLC                 | Voice    : +1 612 536-4697        |
| Power Systems Control Division    | Fax      : +1 612 536-4919        |
| 7225 Northland Drive, Brooklyn Park, Minnesota 55428   USA            |