[petsc-dev] I hate PetscThreadComm

[Jed Brown]

Barry Smith <bsmith at mcs.anl.gov> writes:

>    Nice, so presumably the RPI guys could potentially use this and toss their custom stuff?

It's not publicly released (though I hope it will be).  And it's a
research implementation (using proxy processes), so there is no reason
to use this if you already have a flat MPI code that works well.  But
it's an interface that we expect will find its way into base MPI
implementations eventually, and provides what I think is a good way to
handle threads (assuming the user chose threads for some reason).
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 818 bytes
Desc: not available
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20150115/752f9381/attachment.sig>