Junchao Zhang authored Aug 14, 2013 and Pavan Balaji committed Sep 22, 2013

This commit will increase the ABI version. Fix #1904

Signed-off-by: Pavan Balaji <balaji@mcs.anl.gov>

Junchao Zhang authored Aug 13, 2013 and Pavan Balaji committed Sep 21, 2013

Add missing MPI_T error codes in mpi.h. Will increase the ABI version.

Signed-off-by: Pavan Balaji <balaji@mcs.anl.gov>

In general, added weak symbol support for compilers supporting
__attribute__((weak,alias)) instead of #pragma weak.

Signed-off-by: Pavan Balaji <balaji@mcs.anl.gov>
Signed-off-by: William Gropp <wgropp@illinois.edu>

Probably not too harmful to omit the guards, but better not to jumble
more constants together in the compiler's type-checking machinery.

No reviewer.

Obvious bug.

No reviewer.

In particular, the clang type checking helpers can cause
`MPI_DATATYPE_NULL` to be used before it has been `#define`d.

No reviewer.

This should reduce the amount of work for the compiler, and reduce
the chances of an unintelligent compiler/linker dumping a lot of
these symbols into an executable/library.

Suggested by Jed Brown:
http://lists.mpich.org/pipermail/discuss/2013-January/000230.html

No reviewer.

This says that layout compatibility matching should be used instead of
checking for the exact type (e.g., "mpich_struct_mpi_short_int").

Contributed by Jed Brown:
http://lists.mpich.org/pipermail/discuss/2013-January/000226.html

No reviewer.

Removed duplicate PMPI_Comm_create_group declaration.  Patch provided by Jed
Brown.

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

The dist_graph implementation itself doesn't need any changes, since we
already switched away from an impl where MPI_UNWEIGHTED==NULL.

Reviewed by dinan@.

Reviewed by dinan@.

Reviewed by dinan@.

Forgotten from the previous MPI_Count-related commits (oops).

Reviewed by dinan@.

This is a very lame implementation that creates this object at Init
time and does not populate it with any keys (the MPI standard does not
require that the keys be populated).

Reviewed by buntinas.

arguments are defined.  They are entirely ignored.

Reviewed by buntinas.

is completely ignored.

Reviewed by buntinas.

This commit actually does a few things to the status object:

(1) `int count` becomes `MPI_Count count`;
(2) `count` & `cancelled` have been reordered after the "public"
fields (e.g., MPI_TAG) to ensure that direct access from Fortran is
not somehow broken by the (potentially) non-INTEGER count field;
(3) and a new future-proofing field, `abi_slush_fund`, has been added to
attempt to allow us to add fields to the status object in the future
without having to break ABI compatibility.

There's still too much duplication of the status object's contents/layout
in multiple pieces of the code, but I don't see a reasonable way to fix
this in the near term.

Warning: large count breaks the experimental Fortran binding code
generated for handling the "sizeof(int)!=sizeof(MPI_Fint)" case.

Reviewed by balaji@.

This commit is largely auto-generated boilerplate for the following
"large count" routines:

* MPI_Get_elements_x
* MPI_Type_get_extent_x
* MPI_Type_get_true_extent_x
* MPI_Type_size_x
* MPI_Status_set_elements_x

None of these routines actually _do_ anything (not even assert/error
out) as implemented.  But this makes it very easy to review the logic of
the implementation later on in the succeeding smaller commits.

Reviewed by balaji@.

This commit only contains the build system and mpi.h modifications
necessary for `MPI_Count` to be a valid new type.  It does *not*
actually use this new type anywhere, such as in MPI_Status.

Reviewed by balaji@.

for MPI_Type_create_f90_{integer,real,complex}.

Reviewed by dinan.

Made a cleanup pass on mpi.h.in and the function definitions to make
mpi.h.in consistent with the MPI standard.

1. The usage of "*" vs. "[]" was not consistent with the MPI standard
(aside from the consistency changes added in tickets 125 and 126 in
MPI-3).  This commit makes this notation consistent.

2. Consistently provide function parameter names for all MPI
functions.  Some functions had parameter names, while others did not.

3. Consistently use line-breaks for all function prototypes.  Some
prototypes were on a single line, while others were split across
lines.

4. Reorder prototypes to better place them in sections (e.g., all
collectives go to the same section).  Move all PMPI definitions to the
PMPI section.  Move all MPIX_ extension functions to the same section.

The following prototypes are different from that given in the
standard; we use "indx" instead of "index" or "binding" instead of
"bind", to avoid conflicts with preused variable names: MPI_Waitany,
MPI_Testany, MPI_Graph_create, MPI_Graph_get, MPI_Graph_map,
MPI_T_enum_get_item, MPI_T_cvar_get_info, MPI_T_pvar_get_info.

Reviewed by dinan.

Reviewed by dinan.

This patch adds missing attributes to MPI-3 RMA functions.  Patch provided by
Dmitri Gribenko.

Reviewer: dinan

Updated RMA code to remove trailing "_e" and "_s" on enum and struct type
names to match the MPICH style.

Reviewer: goodell

tickets 125 and 126.

Reviewed by buntinas.

Implementation of MPI-3 window info query routines.

Reviewer: balaji

A slightly older version of this patch was reviewed by goodell.

No reviewer.

No reviewer.

Contributed by Dmitri Gribenko <gribozavr@gmail.com>.

http://lists.mcs.anl.gov/pipermail/mpich2-dev/2012-October/000997.html

MPI_.

A slightly older version of this patch was reviewed by goodell.

No reviewer.

Reviewed by goodell.

even though no device currently takes advantage of this.

No reviewer.

1. Move the new extension FT routines into the src/mpi/comm directory
instead of the separate src/mpix directory, since it's clearly using
the MPI internals.

2. Move the MPIX_mutex functionality to mpi.h.in instead of a separate
mpix.h.  This will allow us to use device-specific mpix_foo.h
functionality that the user can use to buy-in to device-specific
extensions.

3. Move the ARMCI and Mutex functionality to src/armci and src/mutex
directories respectively.

Reviewed by goodell.

Updated window model and flavor attributes in the window object from integers
to enum types.

Reviewer: buntinas

Added the missing disp_unit argument (was added in a later revision of the MPI
3.0 spec) and fixed a bug in base pointer calculations when processes pass a
size of 0.  Added a test case to test MPI-2 ops on shared memory windows.

Reviewer: buntinas

No reviewer.