Hi,

I've wondered aloud about this before, but nobody commented. Does
anyone have an explanation how dhrystone can report so dramatically
different numbers for the same hardware as 9000 vs. 5000?

Is this just a good example of the inherent difficulty of
benchmarking?

(The 5000 is what I see for my 3/60 under NetBSD 1.2, under SunOS
4.1.1 I got 9000. Both dhrystones were compiled with gcc 2.7.x with
-O to -O3... I know this isn't how dhrystone should be compiled, but
how do I get gcc to compile as pcc -O would?)

Jari Kokko