Subject: Scientific benchmarking
To: None <netbsd-users@NetBSD.ORG>
From: Jason Brazile <jason@sunra.csci.unt.edu>
List: netbsd-users
Date: 05/29/1997 21:45:56
This doesn't really have much to do with NetBSD in particular (other
than to show that for what the paper concerned itself with, at
least NetBSD didn't seem to cause any noticable problems :-) but
I wanted to mention that I just noticed that the Aaron Brown/Margo
Selzter paper (that was alluded to at USENIX earlier this year) is
now available.

It is an exemplary piece of work in my opinion - solid scientific
principles applied to benchmarking, improving on Larry McVoy's
important "lmbench" contribution:

	http://www.eecs.harvard.edu/~vino/perf/hbench.html

I post this mainly because I have appreciated in the past when other 
people have pointed out such things.

Jason