[mpich-discuss] Why do predefined MPI_Ops function elementwise in MPI_Accumulate, but not in MPI-1 routines?

[Jed Brown]

On Mon, Apr 23, 2012 at 21:02, Jed Brown <jedbrown at mcs.anl.gov> wrote:

> Note that it's not the "hard" implementation that I'm complaining about,
> it's the need to translate between two different MPI_Ops that do exactly
> the same thing, but one that cannot be used with collectives and one that
> cannot be used with one-sided.


Note that this confusion arises in as simple a scenario as using
std::complex<double> from C++ without using the deprecated C++ bindings.

http://trac.mcs.anl.gov/projects/mpich2/ticket/1525
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/mpich-discuss/attachments/20120423/9e068311/attachment.htm>