Darxus, > garescorer.c says it uses PGAPack partially because of its ability to use > MPI to distribute load. > This will run garescorer on all 4 of my cores: > mpirun -np 4 ./garescorer -b 10 -e 5500 -t 5.0 > But it appears to do so redundantly, so it doesn't actually run any faster, > it just runs the same processing four times at once. > What's the right way to do this? > I'm using this with libpgapack-mpi1 on Ubuntu Maverick.
If I remember correctly I came to the same conclusion when running GA for 3.3.0, so I just gave up parallel runs without investigating further. Mark
