[MITgcm-devel] Re: [MITgcm-cvs] MITgcm/tools/build_options CVS Commit

[Constantinos Evangelinos]

On Thursday 17 September 2009 05:13:55 am Martin Losch wrote:
> I would be interested in which part of the seaice_evp routine causes
> the slowdown with -fp-model precise, but I do not have access to
> Columbia. Will I be able to reproduce this with "any" ifort compiler
> (i.e. even on simple linux boxes)?

Different processors have different denormal handling slowdown. Itanium2 or 
Alphas have a lot, the x87 unit on Intel/AMD processors has none, the SSEn 
unit on the same processors has some etc. I would try things out on an Itanium 
box with the Intel compilers. We have 6 (up out of 9) on ACES running RHEL3 
with older versions of the Intel compilers and stommel, Carl's Altix for you 
to try.

Constantinos
-- 
Dr. Constantinos Evangelinos
Department of Earth, Atmospheric and Planetary Sciences
Massachusetts Institute of Technology