On Wed, 27 Jun 2018, Sven Schreiber wrote:
Am 27.06.2018 um 13:49 schrieb Allin Cottrell:
>
> Hmm, this is not MPI-specific. The loop limit "$mpisize" is falling into a
> crack between the name of a scalar variable (like your "mpisize") and an
> expression to be evaluated (as in, for example,
> "1*$mpisize", which would also work). I need to think a little about the
> best fix for this.
It works with $pd though... (outside mpi I mean).
Ah, OK, I now see why that is. Fixed in git and snapshots.
Allin