[MITgcm-devel] problems with sp4 opt file

[mlosch at awi-bremerhaven.de]

Hi Ed,

this was more a question of mine than a statement. I cannot think 
of a situation where it would be necessary to actually run an 
executable to whether libraries are available, but others most 
likely have more of an overview than I do.

Martin Losch
Alfred Wegener Institute 
Postfach 120161, 27515 Bremerhaven, Germany; 
Tel./Fax: ++49(0471)4831-1872/1797



----- Original Message -----
From: Ed Hill <ed at eh3.com>
Date: Thursday, November 17, 2005 1:30 pm
Subject: Re: [MITgcm-devel] problems with sp4 opt file

> On Thu, 2005-11-17 at 09:24 +0100, Martin Losch wrote:
> > Hallo Ed,
> > 
> > I tried to adopt the sp4+mpi optfile to work on our IBM and I 
ran 
> into 
> > the problem that almost all test for system libraries fail:
> > > ===  Checking system libraries  ===
> > >   Do we have the system() command using mpxlf95_r 
-q64...  yes
> > >   Do we have the fdate() command using mpxlf95_r -q64...  
no
> > >   Do we have the etime() command using mpxlf95_r -q64...  
no
> > >   Can we call simple C routines (here, "cloc()") using 
> mpxlf95_r 
> > > -q64...  no
> > >   Can we unlimit the stack size using mpxlf95_r -q64...  no
> > >   Can we use stat() through C calls...  no
> > >   Can we create NetCDF-enabled binaries...  yes
> > The first three are fine and correct (they do not try to run the 
> > executable), but the next three fail because of the followingvi  
> error 
> > (when they are run after successful compilation):
> > ERROR: 0031-808  Hostfile or pool must be used to request 
nodes
> > because apparently an exectuable generated with mpxlf95 
requires 
> to be 
> > run in an mpi environment, which is not done within 
genmake2. the 
> > netcdf test passes, because the executable is not run (as in 
the 
> case 
> > for the first three test).
> > 
> > So there is an inconsistency there, isn't it? Some of the test 
> require 
> > running the test-executable (test for cloc, unlimit stack size, 
> and 
> > stat) and some don't (netcdf, system, fdate, etime). 
Regardless 
> of how 
> > important or unimportant the passing of any of the tests is, the 
> > testing procedure itself fails if one compiles with MPI. Is it 
> really 
> > necessary to try to run the executables? Or is it enough to 
check 
> > whether the compilation and link-step is successful (as in the 
> case of 
> > netcdf)?
> 
> Hi Martin,
> 
> You're right.  Its best just to check that compiles succeed and 
not to
> run any of the programs.  So I'll remove the run-checks later 
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-devel mailing list
> MITgcm-devel at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-devel
>