I'm trying to get ADF running under MPICH on a DEC Alpha. When running
the test script run.GO_H2O, things eventually hang in ppcbr4 at the
mpi_allreduce call.
I was putting some debugging statements in this routine, and noticed the
comment: "DO NOT USE this routine, use ppcbnr instead with ipcbop = 4".
Is this really what one is supposed to do?
Tom Spraggins
tas_at_virginia.edu
P.S. On the off chance that anyone else is interested in this platform,
note that libtc/pp/ppmpi.fh needs to be changed -- "status" should be
integer*4, not integer (which defaults to integer*8).
Received on 1999-02-23 15:17:40
This archive was generated by hypermail 2.2.0 : 2006-11-02 07:00:02 CET