Subject: Re: FPU Performance on DECstation 5000/125
To: Jonathan Stone <jonathan@DSG.Stanford.EDU>
From: Karl Maftoum <>
List: port-pmax
Date: 08/27/1997 20:58:12
On Tue, 26 Aug 1997, Jonathan Stone wrote:

> Gosh, I hope so. If not, I guess the FP emulation is really, really robust.
> Hm. Maybe you're using denorms a lot, and the emulation code isn't as
> well-tuned as Ultrix??  The sure way to find out is to add some
> instrumentation to the kernel trap handlers and interrupt code, in
> interrupt() and trap() in sys/arch/mips/mips/trap.c.

Hmm, this is important because the machines are being used for a 
distributed computing experiment (using pvm) to solve mathematical 
models, so the FPU sorta has to work properly! I'll investigate the 
kernel code. BTW there is a BIG difference .. about 6 times on a float 
multiply, how good is the maths library?

> Rebuild and reinstall /usr/sbin/config.  This is usually due to
> running old versions of config against new config files.  We deleted
> the statically-configured default swap partition; swap partitions (or
> files) are now configured dynamically.

Done, the kernel compiled properly



Karl Maftoum
Computer Engineering Student at the University of Canberra, Australia


I am getting very sick of getting SPAM..... I won't retaliate, I just 
won't buy your product, and will recommend others to do the same.