[PATCH] Add PAREN_EXPR

[Salvatore Filippone]

Tim Prince wrote:
>It turns out a well known MPI run-time may vary the order of operations in
>MPI_Allreduce, depending on the order of arrival, requiring promotion to
>double precision, or explicitly written substitutes for MPI_Allreduce, to
>enable analyses which otherwise could be done in single.

That's one of the things I always point out to student in lessons and
tutorials, and also that's why if you look at the BLACS, the
communication layer used in ScaLAPACK implemented on top of MPI, you'll
find in the BLACS_GET/BLACS_SET a flag to enforce repeatability. The
alternative is to look at successive runs with different results on same
data and same number of processors, and go on a wild goose chase when no
bug is really there...

Salvatore