[MITgcm-support] ifort optimization puzzle

[Ed Hill]

On Wed, 2005-03-09 at 11:12 +0100, Martin Losch wrote:
> Ooops, too fast,
> 
> this is what I wanted to say:
> > I don't see anything in  these loops that should really be a problem, 
> > maybe someone with more experience in vektorization can
> have a look.

Hi Martin,

Your tenacity got you the answer--you've found an ifort bug.  Congrats!
Now, *PLEASE* submit a bug report to Intel.  Send them:

 1) a copy of the three problematic subroutines
 2) the exact ifort command you used to compile them
 3) what goes wrong
 4) any other useful info you can provide 
    (eg. "it works fine with -O0")

They (Intel) may help you find a better work-around and/or they may get
this fixed in an upcoming release.

In the mean time, have you tried adding the problematic files to your
optfile in:

  NOOPTFILES="find_rho.F ..."
  NOOPTFLAGS='-O0'

And again, good work on tracking this down!!!

Ed

ps - What are we going to do about the precision issues in kpp?
  Should we go through it and systematically add "_d 0"-type 
  syntax?  Or is that going to cause problems for you and other 
  kpp users?

-- 
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