Re: performance tuning differences?

[Alan Robertson <alanr@unix.sh>]

Michael E Brown wrote:
> 
> On Wed, 15 Aug 2001, mulyadi wrote:
> 
> > hello all
> >
> > >What I found odd was the same exact floating point intensive code compiled
> > >and run on the same exact hardware performed 20% WORSE under redhat linux
> > >7.1 (20% better using RH 6.2), both straight installs out of the box.
> 
> Changes to signal handling in kernel 2.4 have caused some performance loss
> on Pentium III+ processors. The kernel now saves and restores the SSE
> registers on signals, which incurrs overhead. Congratulations, you have
> found the one area in which performance has gone down from 2.2->2.4. This
> is due to Intel, though, and not specific to Linux.

My guess is that this isn't the problem.

Floating point intensive code rarely does lots of signals - it mainly does
floating point ;-).

I'd suspect a compiler issue...

	-- Alan Robertson
	   alanr@unix.sh

Linux-cluster: generic cluster infrastructure for Linux
Archive:       http://mail.nl.linux.org/linux-cluster/