[MITgcm-support] RE: a few MITgcm questions

Dimitris Menemenlis menemenlis at jpl.nasa.gov
Sun Apr 4 12:07:01 EDT 2004


>> Because of g77, for the kernel code it never matters, but
>> there could be some stuff in KPP where its assumed (Dimitris
>> any thoughts?).

Ichiro, Chris, KPP and 1-cpu cube-sphere code work fine with g77.
But I have not tried running on the altix without -r8 option.
I just submitted a small test case without -r8 on kalpana and
kalpana crashed ...  I hope there's no correlation there.

>> the -D_BYTESWAPIO option does not seem to
>> work for sequential binary files.

Sequential binary files are files that have a record length
indicator at the beginning and the end of each record, right?
Those files cannot work with the -D_BYTESWAPIO option since the
record-length would read-in and used before BYTESWAPIO is
called to carry out endian conversion.

A possible work-around is to read-in the file as direct-access
and to discard the record-length information, four bytes at the
beginning and the end of each record.

D.

-- 
Dimitris Menemenlis <menemenlis at jpl.nasa.gov>
Jet Propulsion Lab, California Insitute of Technology
MS 300-323, 4800 Oak Grove Dr, Pasadena CA 91109-8099
tel: 818-354-1656;  fax: 818-393-6720




More information about the MITgcm-support mailing list