_UNIX_Today!_ - use MUSBUS benchmark

Chris Jankowski chris at yarra.oz.au
Fri Sep 8 12:39:28 AEST 1989


In article <971 at utoday.UUCP> greenber at utoday.UUCP (Ross M. Greenberg) writes:
>
> As I've indicated, we're going to be doing some massive work in the
> benchmarking area shortly.  Part of my responsibility is to figure out
> what those benchmarks really need, and how to compare apples and oranges.
> In the DOS world, it's easy.  In the UNIX world, not so easy.  So. I 
> figure I'll ask the same guys that gave me such a hard time (thanks for
> volunteering! :-) ):  what do *you* think is important in a UNIX
> benchmark?  How do I compare Machine A (with hardware and software
> configuration of 'A') to Machine B with its H/W configuration of 'B'?

This question has been asked many times before. I believe that there is
one correct answer:

	MUSBUS

MUSBUS is a benchmarking suite designed specifically for testing
*multiuser* capabilities of UNIX machines by people at Melbourne
University led by Ken McDonnald.

It meets the following criteria :
	- gauges the whole system performance including disk I/O,
		and terminal I/O.
	- it is in public domain ie. it's free as oposed to AIM 3.
It has other terrific features which may be of less interest to you:
	- it is suitable for benchmarking multiprocessor architectures
		like Pyramid (our MIServer has up to 12 processors).
	- it is known to be a very good tool to check quality of a UNIX
		port ie. it is known to break shoddy UNIX implementations.

By definition as it tests the whole system it is sensitive to configuration.
Adding more memory, more disk space and better tuning typically should beef 
up your results. If it does not it means that you have hit some bottleneck.

So to publish meaningful results you need to quote full configuration 
of the machine and also tuning parameters.
Interpretation of results is also more complex as may be expected from
a real multiuser enviroment.
Running of the benchmark requires some understanding of UNIX.	
The benchmark should be run well past saturation points of the system 
under stress to reveal the system's bottlenecks.
One can also cheat by modifying the benchmark and not telling anybody.

MUSBUS code can certainly be obtained from some repository in the USA.

The whole field of benchmarking is full of religious wars and I do not want
to start another one. If you do not like the benchmark use your own.
MUSBUS is certainly not a suitable benchmark for a heavy user of
a fluid dynamics package on its own single user graphics workstation.
I just really think that using Whetstone and similar benchmarks for 
general purpose multiuser machine (especially multiprocessor one) 
Is non-productive.

      -m-------   Chris Jankowski - Senior Systems Engineer chris at yarra.oz{.au}
    ---mmm-----   Pyramid  Technology  Australia	    fax  +61 3 820 0536
  -----mmmmm---   11th Floor, 14 Queens Road                tel. +61 3 820 0711
-------mmmmmmm-   Melbourne, Victoria, 3004       AUSTRALIA       (03) 820 0711

"Knowing how things work is the basis for appreciation,
and is thus a source of civilized delight."  -- William Safire



More information about the Comp.unix.i386 mailing list