[MITgcm-support] ifort optimization puzzle

[Martin Losch]

Hi Ed,
this is my ifort

> csysl4::mlosch> ifort -V
> Intel(R) Fortran Compiler for 32-bit applications, Version 8.1    
> Build 20041118Z Package ID: l_fc_pc_8.1.023
> Copyright (C) 1985-2004 Intel Corporation.  All rights reserved.
> FOR NON-COMMERCIAL USE ONLY

So, unfortrunately, I have one even later than 8.1.021. I suspect that 
this is a compiler error, but I cannot see where this happens.

I had hoped that there are some know compiler option combinations that 
are known to be dangerous (appart from -O3),  that some of you would 
know.

The next thing to check is probably what happens when I change the size 
of the problem (change nr or nx,ny). I ASSUME that the problem goes 
away, as soon as I have more than 1 horizontal point. It may also have 
to do with the overlap of (which is 2 > 1).

Martin

On Mar 8, 2005, at 2:24 PM, Ed Hill wrote:

> On Tue, 2005-03-08 at 12:30 +0100, Martin Losch wrote:
>>
>> So MNC is off the hook (o:
>
> Hi Martin,
>
> Excellent!  ;-)
>
> Which version of ifort are you using?  I've read that there were a
> relatively large number of bugs in ifort v8.0 versus the more recent
> v8.1+ releases.  So I'd try an upgrade first.  It'd be a shame to do a
> bunch of careful testing only to (perhaps?) discover a compiler error
> fixed in the latest release.
>
> And if you want, I'll install v8.1.021 on the "authors" network for you
> today.
>
> Ed
>
> -- 
> Edward H. Hill III, PhD
> office:  MIT Dept. of EAPS;  Rm 54-1424;  77 Massachusetts Ave.
>              Cambridge, MA 02139-4307
> emails:  eh3 at mit.edu                ed at eh3.com
> URLs:    http://web.mit.edu/eh3/    http://eh3.com/
> phone:   617-253-0098
> fax:     617-253-4464
>
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://dev.mitgcm.org/mailman/listinfo/mitgcm-support