Subject: Re: first attempt at LMBENCH run for SunOS-4 vs NetBSD on 3/260...
To: Greg A. Woods <woods@most.weird.com>
From: Simon Burge <simonb@telstra.com.au>
List: port-sun3
Date: 06/07/1996 16:20:18
On Fri, 7 Jun 96 00:21:41 -0400 (EDT) Greg A. Woods wrote:
> I have no idea why yet, but lmbench (mhz) gets the clock speed very
> wrong on these machines.... It claims ~9.85 MHz for the 3/260, and
> about 6.32 MHz for my 3/60 (though the latter was run under a full
> operating X11 environment). Perhaps lmbench just doesn't like the
> mc68k? The code in mhz.c is quite simple, but I've not yet looked at
> the compile output to see what's up.
lmbench just uses a timed loop (of increments) to calculate CPU speed,
and it's not perfect on every type of CPU. For example, I know a i386
takes 2 clock cycles to do an increment, and a i486 or greater only
takes one. It seems like a 68020 is the same as an i386 in this
respect (a 3/60 does run at 20MHz doesn't it?) Also, because it's just
a CPU loop, it's greatly dependent on system load...
Hope this helps to explain your observations.
Simon.
--
Simon Burge simonb@telstra.com.au
UNIX Support, CPR Project, Telstra. +61 3 9634 3974
-------------------------------------------------------------------------------
I've got a can opener at home. Sometimes it works, and other times it doesn't.
Sometimes it only goes half-way around, and then does a lap of honor - Jimeoin.