Dear Wesley, https://github.com/mpi-forum/mpi-issues/files/3704357/153_neighbor_mpi-3_bug... includes the full solution, but I have not yet build a pull request. No time for that up to now. I hope that I can do it during or after the meeting. The most important is that all vendors/owners of MPI libraries fix this bug in their MPI implementation, therwise MPI_NEIGHBOR_ALLTOALL/V/W is not usable in cyclic Cartesian applications. (May be that they all already fixed it. Upto now, I've the fixed notice only from OpenMPI and HPE MPT.) Therefore the reading slot is important. Best regards Rolf ----- Original Message -----
From: "Wesley Bland" <[email protected]> To: "Rolf Rabenseifner" <[email protected]>, "Main MPI Forum mailing list" <[email protected]> Sent: Monday, December 2, 2019 2:23:11 PM Subject: Re: [Mpi-forum] Issue 153 / Re: Items for MPI Forum in Albuquerque
Rolf,
So you have a pull request with the text for this proposal? Pull requests are required to be able to merge the text correctly.
Thanks, Wesley
On Dec 2, 2019, at 12:36 AM, Rolf Rabenseifner via mpi-forum <[email protected]> wrote:
Dear Martin and all,
I prepared
https://github.com/mpi-forum/mpi-issues/issues/153 Errata for NEIGHBOR... for discussions and reading in Albuquerque.
Best regards Rolf
-- Dr. Rolf Rabenseifner . . . . . . . . . .. email [email protected] . High Performance Computing Center (HLRS) . phone ++49(0)711/685-65530 . University of Stuttgart . . . . . . . . .. fax ++49(0)711 / 685-65832 . Head of Dpmt Parallel Computing . . . www.hlrs.de/people/rabenseifner . Nobelstr. 19, D-70550 Stuttgart, Germany . . . . (Office: Room 1.307) . _______________________________________________ mpi-forum mailing list [email protected] https://lists.mpi-forum.org/mailman/listinfo/mpi-forum
-- Dr. Rolf Rabenseifner . . . . . . . . . .. email [email protected] . High Performance Computing Center (HLRS) . phone ++49(0)711/685-65530 . University of Stuttgart . . . . . . . . .. fax ++49(0)711 / 685-65832 . Head of Dpmt Parallel Computing . . . www.hlrs.de/people/rabenseifner . Nobelstr. 19, D-70550 Stuttgart, Germany . . . . (Office: Room 1.307) .