Re: Revision 162491 -- fortran -fwhole-file regressions

[Toon Moene <toon at moene dot org>]

Dave Korn wrote:

> On 24/07/2010 19:33, Toon Moene wrote:

>   What I did find out, amusingly enough, is that a number of the controversial
> climate research centers use beowulf clusters based on Xeons or AMD64s, so if
> they didn't use the right compiler flags to get SSE rather than x87 math....
> who knows!

When the Pentium FDIV panic (Nov. '94) broke out, I ran a re-computation 
of some of our most sensitive forecasts, inducing an artificial 10E-4 
relative difference in one grid cell (approximately covering the 
relative appearance of the problem caused by the fault in the Pentium chip).

It had the effect we'd expected: Hard to forecast phenomena (like 
convective precipitation) were changed; large scale weather remained the 
same.

This is on weather forecasting - climate research is an entirely 
different branch of atmospheric research ...

Cheers,

--
Toon Moene - e-mail: toon@moene.org - phone: +31 346 214290
Saturnushof 14, 3738 XG  Maartensdijk, The Netherlands
At home: http://moene.org/~toon/; weather: http://moene.org/~hirlam/
Progress of GNU Fortran: http://gcc.gnu.org/gcc-4.5/changes.html#Fortran