There are many memory allocations in the write path.  Allocating the
two-phase intermediate buffer outside of the write path might on some
systems make a small difference, especially if there are many collective
I/O calls, or if the system (like Blue Gene) has a small amount of
memory.  Modified from Paul Coffman <pkcoff@us.ibm.com>'s original idea.

For quite some time the barrier here has had the comment 'Why?'.  Since
no one knows, and there are plenty of other syncronization points in
this path, remove it.

bluegene timer code had two "levels" of timing.  that seemed kind of
pointless so lump it all into one level.

this "new" system call (part of POSIX-2001) saves us a system call on
Blue Gene.  Seems to get us back 5 seconds for one workload at small
(half rack) scales.

bglockles uses the common read/write routines for contig read/wrties, so
bluegene timing infrastrucutre wasn't actually timing anything.  Since
this introduces blue gene bits into common code, please do not merge to
master.  Instead, we should rework all the timing bits so that it no
longer times "bluegene" but rather all of ROMIO.  Furthermore, the
locky bits of 'bg:' driver should be yanked anyway, obviating the need
for bglockless.

Protected by an 'ifdef', this BGL-era code bitrotted a bit.  clean it up
and see if it does anything useful today.
- Removes preprocessor guards: the counters and timers do nothing
  expensive unless environment variables are set
- remove the idea of a "level"
- remove barrier from timing collection.
- bugfix: MPI_Wtime() does not necessarily start at zero, so properly initialze
  timers for collective read/write
- report only from I/O aggregators.  when reporting "time spent in i/o"
  vs "time spent communicating" it makes more sense to look only at the
  aggregators.  The non-aggregators are going to skew the results
  because they are spending some communication time actually
  communicating, but some of that time blocked, waiting for aggregators
  to finish.

Kenneth Raffenetti authored Mar 19, 2014 and Pavan Balaji committed Mar 26, 2014

De-dupes executable and host names in the MPIR_proctable by pointing
to an existing copy. Closes #1821

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

The following commits are reverted.

1. "Better checks for VA_ARGS."; commit
de80ec87.

2. "Warning squash for clang."; commit
ccf7f70c

.

The clang warning this was originally trying to solve has been fixed
by the newer versions of clang, AFAICT.

Signed-off-by: Huiwei Lu <huiweilu@mcs.anl.gov>

No reviewer.

ROMIO code assumes all processes will use the same ROMIO driver.  we
were not reaching the "find a common file system" logic when NFS was
enabled, everyone stat-ed the file system without errors, but some
processees found a different file system (like if some processes are
writing to NFS and others to UFS)

See discussion beginning here:
http://lists.mpich.org/pipermail/discuss/2014-March/002403.html



Tested-by: Jeff Squyres <jsquyres@cisco.com>

Signed-off-by: Michael Blocksome <blocksom@us.ibm.com>

Signed-off-by: Michael Blocksome <blocksom@us.ibm.com>

  powerpc64-bgq-linux-gcc v4.7.2

Signed-off-by: Michael Blocksome <blocksom@us.ibm.com>

Signed-off-by: Michael Blocksome <blocksom@us.ibm.com>

Signed-off-by: Michael Blocksome <blocksom@us.ibm.com>

Signed-off-by: Michael Blocksome <blocksom@us.ibm.com>

Test the use of Reduce_local in C++. Use multiple built-in datatypes,
as well as a derived vector.

Junchao Zhang authored Mar 04, 2014 and Kenneth Raffenetti committed Mar 24, 2014

Fixes #2014

Signed-off-by: Ken Raffenetti <raffenet@mcs.anl.gov>

Wesley Bland authored Mar 19, 2014 and Pavan Balaji committed Mar 23, 2014

The constant MPIDI_TAG_UB is used in only one place at the moment, in the
initialization of ch3 (source:src/mpid/ch3/src/mpid_init.c@4b35902a#L131). The
problem is that the value which is being set (MPIR_Process.attrs.tag_ub) is
set differently in pamid (INT_MAX). This leads to weird results when we set
apart a bit in the tag space for failure propagation in non-blocking
collectives (see #2008).

Since this value isn't being referenced anywhere else, there doesn't seem to
be a use for it and it's just leading to confusion. To avoid this, here we
remove this value and just set MPIR_Process.attrs.tag_ub to INT_MAX in both
ch3 and pamid.

See #2009

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

make clean in test should remove all generated files.  This file
was not added to the clean list when it was added to the runlist
output.

Junchao Zhang authored Feb 24, 2014 and Pavan Balaji committed Mar 21, 2014

f77 and f90 were not consistent names with respect to what the code
was providing.  For example, "f90" also used code from "f77".  This
patch changes the naming to "mpif_h" and "use_mpi", which correspond
to using the mpif.h and "use mpi module" conventions specified by the
MPI standard.

This also includes changes to buildiface, autogen.sh and .gitignore to
work with these changes.

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

Fixes #2044

Signed-off-by: Antonio J. Pena <apenya@mcs.anl.gov>

The scenario of the hang is described as follows:

  Assuming the job runs with 4 tasks, task 0 is in a loop  of processing the
  following RMA operations to fetch the displacement, the loop ends if the
  displacement is being updated.

    MPI_Win_get_accumulate( target rank is task 0)
    MPI_Win_flush(task 0)

  task 1 and 3 hang in MPI_Win_flush() waiting for a call to
  MPI_Win_compare_and_swap() to complete. The target rank for this operation is
  task 0.

  task 2 hangs in MPI_Win_flush() waiting for a call to MPI_Accumulate() to
  complete. The target rank for this operation is task 0 as well.

  Task 0 is busy making MPI_Win_get_accumulate() and MPI_Win_flush() calls to
  see if the displacement is being updated, the target rank of the operation is
  task 0 itself which means the operation is local and can be completed without
  a need of making a PAMI dispatcher call.  Meanwhile, the other three tasks
  issue RMA operations to the target task 0 and wait for the completion of the
  operations. Because task 0 is in a loop of making local operations, no PAMI
  dispatcher is called, no progress made for any remote operations which is the
  root cause of the hang.

The fix for the problem is to add a call to PAMI dispatcher in MPI_Win_flush(),
the call is made prior to the check of the condition. Current code checks the
condition first, if the condition is satisfied, then no PAMI dispatcher is called.

The following statement in MPI_Win_flush()

  MPID_PROGRESS_WAIT_WHILE(sync->total != sync->complete)

will be replaced by

  MPID_PROGRESS_WAIT_DO_WHILE(sync->total != sync->complete)

(ibm) D196445

Signed-off-by: Michael Blocksome <blocksom@us.ibm.com>

The suncc compiler on Linux did not like the ternary conditional
operands having different types.

No reviewer.

William Gropp authored Mar 16, 2014 and Kenneth Raffenetti committed Mar 17, 2014

These file had allowed uses of routines that aren't always permitted
according to our coding style.  These edits let the coding style
checker know that these uses are ok (most uses of printf or malloc/free
in MPICH are incorrect - see the style guide).

Signed-off-by: Ken Raffenetti <raffenet@mcs.anl.gov>

William Gropp authored Mar 16, 2014 and Kenneth Raffenetti committed Mar 17, 2014

The coding standards are clear on this; since we can't trust the system
header files not to make definitions that conflict with ours, good
coding practice is to make all MPICH names distinct.  The MPIR_MIN
and MPIR_MAX definitions were made long ago for just this purpose.

Signed-off-by: Ken Raffenetti <raffenet@mcs.anl.gov>

William Gropp authored Mar 16, 2014 and Kenneth Raffenetti committed Mar 17, 2014

The coding checker sometimes needs some hints that certain uses are
permitted.  This adds the necessary information to some of the generated
files.

Signed-off-by: Ken Raffenetti <raffenet@mcs.anl.gov>

Fixes #2048

Signed-off-by: Ken Raffenetti <raffenet@mcs.anl.gov>

Discussion with Michael Raymond: a customer would "like to use larger
than 32-bit stripe units".  Technically, lustre only supports a 32 bit
value for its stripe units -- an *unsigned* 32 bit value.

Since we're here, let's double check that we do not try to jam a larger
value into the lustre hint structure than lustre can support.

Signed-off-by: Michael Raymond <mraymond@sgi.com>

Pathscale C compiler reports symbol redefinition error for mpi_conversion_fn_null__ in such code

extern int mpi_conversion_fn_null__ ( void*v1, MPI_Fint*v2, MPI_Fint*v3, void*v4, MPI_Offset*v5, MPI_Fint *v6, MPI_Fint*v7, MPI_Fint *ierr ) __attribute__((weak,alias("mpi_conversion_fn_null_")));
\#pragma weak mpi_conversion_fn_null__ = mpi_conversion_fn_null_

We should only generate one form code using either "__attribute__" or "#pragma weak"

Fixes #2044

Signed-off-by: Antonio J. Pena <apenya@mcs.anl.gov>

Supports 'make dist', 'make -j' and others.

1. Several tests has been added to test/mpi since the ticket 1720 was
created. These new tests are added to corresponding Makefile.am.

2. There is one minor thing I am not sure in this fix. I don't know the
different of noinst_HEADERS and nodist_noinst_HEADERS, so I just put all
head files under noinst_HEADERS (in test/mpi/Makefile.am).

3. Windows project files like adapt.vcproj were deleted in commit
ba3badff

, they should be removed in
test/mpi/basic/Makefile.am too.

4. MPICH ABI removed support for MPI::Distgraphcomm, so distgraphcxx.cxx
is changed to be not compiled in cxx/topo.

5. One minor change in help description in test/mpi/runtests.in. So when you
run './runtests -batch', you will get help message directing you to
README.

Fixes #1720

Signed-off-by: Antonio J. Pena <apenya@mcs.anl.gov>

if we know the underlying file system supports "resize from one", then
we do not need the non-aggregator processes to open the file

Signed-off-by: Paul Coffman <pkcoff@us.ibm.com>

The Solaris/Sun compilers accept the -Wl, style flags for passing
arguments to the linker in version 8.5 on Linux. We wse this method,
as the compiler no longer passes unrecognized flags to the linker
by default. The matching logic in config.rpath is mostly borrowed
from libtool commit [a55208d5]. Fixes #2045

Signed-off-by: Huiwei Lu <huiweilu@mcs.anl.gov>

Signed-off-by: Junchao Zhang <jczhang@mcs.anl.gov>

--enable-thread-cs=lock-free is not supported yet, it should be tested
in configure process instead of detecting it at make stage.

Fixes #2050

Signed-off-by: Ken Raffenetti <raffenet@mcs.anl.gov>

In MPID_Win_allocate and MPID_Win_allocate_shared, baseptr are defined
as void * and void ** separately, while in MPIDI_Win_fns, both
MPID_Win_allocate and MPID_Win_allocate_shared are registered as
MPIDI_CH3U_Win_allocate, where baseptr is defined as void *.

Fixes #1995

Signed-off-by: Junchao Zhang <jczhang@mcs.anl.gov>

Insert a barrier for every 16 calls to async collectives

(ibm) D192641

Signed-off-by: Michael Blocksome <blocksom@us.ibm.com>

In the unlikley event that someone checks error codes, let them know
they might want to pass a non-null filename to MPI-IO

Signed-off-by: Wesley Bland <wbland@mcs.anl.gov>