- 25 Oct, 2012 3 commits
-
-
Pavan Balaji authored
moved some error messages which should have been mpix to mpi. Reviewed by dinan.
-
Pavan Balaji authored
released. Arguably, we never actually needed them, since these tests are already protected by "mpiversion=3.0" in the testlist. No reviewer.
-
William Gropp authored
-
- 24 Oct, 2012 12 commits
-
-
Darius Buntinas authored
-
Darius Buntinas authored
-
Pavan Balaji authored
Reviewed by goodell.
-
Pavan Balaji authored
A slightly older version of this patch was reviewed by goodell.
-
David Goodell authored
It's not valid for users to call any of the other category routines (well, except maybe "category_changed") as long as this is 0. Reviewed by rgunter@.
-
David Goodell authored
Need to handle NULLs correctly and also return strlen+1 in most cases. Thanks to Ralf for originally spotting this bad behavior. Reviewed by rgunter@.
-
David Goodell authored
This is perfectly valid according to MPI-3. If the caller has no interest in the name or description of a particular variable then they may pass NULL for both the string and corresponding length arguments. Reviewed by rgunter@.
-
David Goodell authored
Attempts to verify that basic "MPI_T_" string handling complies with the MPI-3 standard. Reviewed by rgunter@.
-
David Goodell authored
Without this change, clang gets rather noisy about Valgrind's client request macros whenever -Wunused-value is passed (e.g., when configuring with --enable-strict). No reviewer.
-
David Goodell authored
We incorrectly interchanged the context_id and recvcontext_id. The bug report and fix itself was provided by Devendar Bureddy <bureddy@cse.ohio-state.edu> (thanks!). I wrote the regression test portion of the commit. No reviewer.
-
Pavan Balaji authored
r10444. No reviewer.
-
Pavan Balaji authored
get the nightly snapshots build correctly. No reviewer.
-
- 23 Oct, 2012 13 commits
-
-
Pavan Balaji authored
No reviewer.
-
Pavan Balaji authored
No reviewer.
-
Pavan Balaji authored
MPICH2 --> MPICH1 move. No reviewer.
-
Pavan Balaji authored
Reviewed by robl.
-
Pavan Balaji authored
-
James Dinan authored
Copy/paste oops. Reviewer: none
-
Darius Buntinas authored
-
James Dinan authored
Implemented the local flush operations in terms of the full flush operations. Reviewer: goodell
-
James Dinan authored
Reviewer: goodell
-
James Dinan authored
The RMA op dispatch macro was previously defined in a private header used by the high level routines. This has been moved to mpiimpl.h to join the rest of the RMA op table code. This will allow other routines to dispatch RMA ops at the ADI correctly. Reviewer: goodell.
-
James Dinan authored
Added a new inline function to free the ops list. This is much clearer than how it was done previously and should make it easier to see when/where we can assume the ops list has been totally emptied. Reviewer: goodell
-
James Dinan authored
Refactored RMA ops enqueue/dequeue operations into inline functions to reduce copy/paste code. Added extra safety checks that will be needed as we renovate for MPI-3. Reviewer: goodell
-
James Dinan authored
-
- 22 Oct, 2012 5 commits
-
-
Pavan Balaji authored
No reviewer.
-
Pavan Balaji authored
No reviewer.
-
William Gropp authored
-
William Gropp authored
-
William Gropp authored
-
- 21 Oct, 2012 1 commit
-
-
Pavan Balaji authored
No reviewer.
-
- 20 Oct, 2012 6 commits
-
-
James Dinan authored
The lock packet may not be at the head of the queue, so the trailing get optimization needed to be updated to not skip over the first element. This is a performance, not a correctness fix. Reviewer: none
-
James Dinan authored
This commit implements MPI-3 RMA's flush and flush_all operations. Reviewer: buntinas
-
James Dinan authored
Added a shared lock, third-party communication test for MPI-3 RMA flush. Reviewer: buntinas
-
James Dinan authored
Updated passive target synchronization to allow the RMA lock op to be omitted from the op list when the lock is immediately requested. This is needed to implement MPI-3 RMA flush operations. Reviewer: buntinas
-
James Dinan authored
When enabled, this mode of operation immediately requests the lock when MPI_Win_lock is called. Currently, this is enabled by setting the MPICH_RMA_LOCK_IMMED environment variable. In the future, we can also make this mode of operation available though an info/assert. This capability is needed to implement MPI-3's flush operations. Reviewer: buntinas
-
James Dinan authored
Factored out code to initiate a lock operation and to wait for its completion. This will facilitate future changes, where the lock request can come from several different parts of the code. Reviewer: buntinas
-