GitLab

[svn-r10769] Fixed definition of builtin info object MPI_INFO_ENV to be consistent with other builtin objects.  Fixes #1762.  Reviewd by Balaji

Reviewer: buntinas

Added a paremeter to enable/disable the RMA lock-op-unlock merging
optimization.

Reviewer: buntinas

[svn-r10759] Update the trmem routines and macros (defined when malloc took an unsigned int for size) to use size_t instead, as well as some int/size_t cleanup (again for code that correctly used signed int when it was written)

This assertion was made unnecessary by previous cleanups.  It was re-checking
the branch condition, and could thus never fail.

Reviewer: none

No reviewer.

No reviewer.

No reviewer.

No reviewer.

[svn-r10732] Enhanced trmem (tracing memory) code in preparation for changing to size_t from int for sizes

[svn-r10724] Correct implementation of MPI_COUNT, remove MPI_2COMPLEX and MPI_2DOUBLE_COMPLEX (invalid since MPI 1.1), and update generation of Fortran predefined MPI values

to cover the case where an external script closed STDIN before calling
mpiexec.

No reviewer.

for our stdin strategy.  Remove that test.

No reviewer.

No reviewer.

fix this, unless we add some form of flow-control logic between the
proxy and mpiexec.  For the time being, at least throw an error when
we can't keep up and are giving up, instead of hanging.

No reviewer.

block/no-block parameter.

No reviewer.

No reviewer.

No reviewer.

No reviewer.

Fixes tt#1749.

No reviewer.

Reviewed by goodell.

in non-strict MPI or MPICH-specific tests.

No reviewer.

No reviewer.

Reviewed by goodell.

Updated existing request-generating RMA ops test to verify that communication
with PROC_NULL is handled correctly.

Reviewer: goodell

This patch more efficiently handles request-generating operations that target
the local process and MPI_PROC_NULL.  These special cases return completed
generalized requests, which cost less overhead.  This also makes space for a
more efficient implementation when direct, shared memory operations are
implemented.

Reviewer: goodell

Note: Request-generating operations can only be called within a passive target
epoch.

Reviewer: goodell

When called with a target process of PROC_NULL, RMA request-generating
operations did not return a valid request handle.

Reviewer: goodell

Reviewer: goodell

Many RMA tests did not limit their output when an error occurred.  This adds
the simple SQUELCH() macro to these tests to limit per-process error messages.

Reviewer: goodell