v2.x MPI_Aint functions #667
Conversation
This commit adds support for MPI_Aint_add and MPI_Aint_diff. These functions are implemented as macros in C (explicitly allowed by MPI-3.1). The fortran implementations are a similar mess to the MPI_Wtime implementations. (cherry picked from open-mpi/ompi@062de45) Signed-off-by: Nathan Hjelm <[email protected]>
(cherry picked from commit 8f252e65b02bfe20cacee0ece8c835246eea3bd9) (cherry picked from open-mpi/ompi@258d1aa) Signed-off-by: Nathan Hjelm <[email protected]>
|
note I'm upgrading some plugins on the IU jenkins server so the LANL checks are not going to complete for a while, as the server is draining the current builds in order to safely restart. |
|
Test PASSed. |
|
@hjelmn Same comment as #666 (comment) -- do we have any regression tests for these (particularly the Fortran bindings)? I ask because the Fortran bindings are tricky to get right. :-( |
|
@jsquyres If I remember you did some testing which lead to the second commit in this PR. |
|
I thought Nathan used some of the mpich tests and found bugs in them. |
|
@hppritcha Yes, we found bugs in the fortran that @jsquyres fixed. All the relevant commits are in this PR. |
Signed-off-by: Nathan Hjelm <[email protected]> (cherry picked from open-mpi/ompi@6dd9658)
|
missed a minor commit. added. |
|
Test PASSed. |
|
Test FAILed. |
|
MTT IBM MPI_Aint_* tests coming shortly. |
|
:bot:retest |
|
Test PASSed. |
|
is this ready to merge? |
|
Yes. Tests have been committed to the IBM test suite, so we'll find out if there's any problems. |
Add CUDA MPI extension for Rolf - because @jsquyres said it was OKAY
Opps, forgot to bring these over earlier.
:bot:assign: @jsquyres
:bot🏷️bug
:bot:milestone:v2.0.0