[Mitgcm-support] Re: pre38 mpi on SGI

[mitgcm-support at dev.mitgcm.org]

OK problems are solved!

I have an AIM simulation running on the SGI. 

As suggested by Alistair, the problem with the *001.001.data files size was 
simply due to the fact that they were left over from a run done with np=1 
processor. Because they are direct access files, the parts of the files not 
modified when running on multi procs are not erased.

Daniel