Hi
I have been running tomawac v6p0 in parallel for a while but recently it has become very slow (slower than running in scalar). Can anyone suggest what I should output to identify the problem/bottleneck please? To run in parallel I am using mpirun with the following set up:
fl01@deepblue:/export/apps/s/mpich2/bin/mpich2version
MPICH2 Version: 1.3.1
MPICH2 Release date: Wed Nov 17 10:48:28 CST 2010
MPICH2 Device: ch3:nemesis
MPICH2 configure: --prefix=/export/apps/s/mpich2 --with-pm=mpd:hydra
MPICH2 CC: gcc -O -O2
MPICH2 CXX: g++ -O -O2
MPICH2 F77: ifort -O -O2
MPICH2 FC: ifort -O2
were I start an mpi ring prior to running tomawac. When running in parallel I have one process running on the headnode and then split the other processes over four or 8 nodes.
Thanks
Fay