1
1
Граф коммитов

26484 Коммитов

Автор SHA1 Сообщение Дата
Alex Mikheev
67d66c2326
oshmem: sshmem: make mmap allocator a default instead of verbs
By default use mmap() to allocate memory for the symmetric heap.
It is safer and more portable choice than sysv and verbs.

Signed-off-by: Alex Mikheev <alexm@mellanox.com>
2016-12-14 13:31:16 +02:00
Nathan Hjelm
8155124adc Merge pull request #2558 from hjelmn/datatype_fix
ompi/datatype: fix bug in darray that causes MPI/IO failures
2016-12-13 14:02:15 -07:00
Yossi
fa6e263821 Merge pull request #2537 from alinask/topic/pml-spml-ucx-api
PML/SPML/UCX: Adapt to the API changes in the UCX lib.
2016-12-13 20:01:47 +02:00
Nathan Hjelm
eb439228b1 ompi/datatype: fix bug in darray that causes MPI/IO failures
This commit fixes errors in the lb and extent of darray datatypes. For
these datatypes the lb should be the start offset of the rank's data
in the array and the extent should be the size of the entire
datatype. In master the lb was always 0 and the extent was always to
small. This commit updates the call to opal_datatype_resize to set the
correct lb and fixes the extent calculation.

Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
2016-12-13 09:25:16 -07:00
Jeff Squyres
f9e8a55a0e Merge pull request #2543 from ggouaillardet/topic/dll_bit_reproducible
ompi/debuggers: make the binary bit reproducible
2016-12-09 06:35:47 -05:00
KAWASHIMA Takahiro
ae056d957c Merge pull request #2545 from kawashima-fj/pr/inactive-persistent-request
ompi/request: Fix a persistent request creation bug
2016-12-09 08:42:31 +09:00
Jeff Squyres
1187212f5d scaling.pl: minor change to perl quoting
Makes emacs syntax hilighting work better.

Signed-off-by: Jeff Squyres <jsquyres@cisco.com>
2016-12-08 09:25:08 -08:00
Ralph Castain
d5a428b646 Scaling test should only launch one proc/node
Signed-off-by: Ralph Castain <rhc@open-mpi.org>
2016-12-08 09:24:22 -08:00
KAWASHIMA Takahiro
6510800c16 ompi/request: Fix a persistent request creation bug
According to the MPI-3.1 p.52 and p.53 (cited below), a request
created by `MPI_*_INIT` but not yet started by `MPI_START` or
`MPI_STARTALL` is inactive therefore `MPI_WAIT` or its friends
must return immediately if such a request is passed.

The current implementation hangs in `MPI_WAIT` and its friends
in such case because a persistent request is initialized as
`req_complete = REQUEST_PENDING`. This commit fixes the
initialization.

Also, this commit fixes internal requests used in `MPI_PROBE`
and `MPI_IPROBE` which was marked wrongly as persistent.

MPI-3.1 p.52:

We shall use the following terminology: A null handle is a handle
with value MPI_REQUEST_NULL. A persistent request and the handle
to it are inactive if the request is not associated with any ongoing
communication (see Section 3.9). A handle is active if it is neither
null nor inactive. An empty status is a status which is set to return
tag = MPI_ANY_TAG, source = MPI_ANY_SOURCE, error = MPI_SUCCESS, and
is also internally configured so that calls to MPI_GET_COUNT,
MPI_GET_ELEMENTS, and MPI_GET_ELEMENTS_X return count = 0 and
MPI_TEST_CANCELLED returns false. We set a status variable to empty
when the value returned by it is not significant. Status is set in
this way so as to prevent errors due to accesses of stale information.

MPI-3.1 p.53:

One is allowed to call MPI_WAIT with a null or inactive request
argument. In this case the operation returns immediately with empty
status.

Signed-off-by: KAWASHIMA Takahiro <t-kawashima@jp.fujitsu.com>
2016-12-08 21:42:05 +09:00
Alina Sklarevich
e9d2d029c6 PML/SPML/UCX: Adapt to the API changes in the UCX lib.
Signed-off-by: Alina Sklarevich <alinas@mellanox.com>
2016-12-08 11:33:29 +02:00
Gilles Gouaillardet
804a784fce Merge pull request #2544 from ggouaillardet/topic/mca_spml_yoda_get
spml/yoda: fix support for BTLs that do not register memory in mca_sp…
2016-12-08 17:26:07 +09:00
Gilles Gouaillardet
062ed9c919 spml/yoda: fix support for BTLs that do not register memory in mca_spml_yoda_get()
Refs open-mpi/ompi#2499

Signed-off-by: Gilles Gouaillardet <gilles@rist.or.jp>
2016-12-08 15:56:25 +09:00
Gilles Gouaillardet
4d8f606420 ompi/debuggers: make the binary bit reproducible
instead of compilation date __DATE__, use a MPI_Get_library_version() like string

Thanks Alastair McKinstry for the report

Fixes open-mpi/ompi#2518

Signed-off-by: Gilles Gouaillardet <gilles@rist.or.jp>
2016-12-08 13:46:43 +09:00
rhc54
341ab683de Merge pull request #2532 from rhc54/topic/pmixptl
Update to latest PMIx master + PTL branch
2016-12-07 17:28:22 -08:00
rhc54
25a3e27b07 Merge pull request #2542 from rhc54/topic/ashley
Correctly cleanup the local children and node map info on remote orte…
2016-12-07 15:53:08 -08:00
Ralph Castain
e1aa7939ef Correctly cleanup the local children and node map info on remote orteds upon job completion. Ensure that register_nspace only includes procs from that job in the proc map
Thanks to Ashley Pittman for the report

Signed-off-by: Ralph Castain <rhc@open-mpi.org>
2016-12-07 13:53:00 -08:00
rhc54
309c967946 Merge pull request #2536 from ggouaillardet/topic/ess_base_update_routing_plan
ess/base: invoke orte_routed.update_routing_plan() earlier
2016-12-07 07:20:20 -08:00
Gilles Gouaillardet
123036dbf8 ess/base: invoke orte_routed.update_routing_plan() earlier
fix an issue that can be evidenced with two nodes
n0$ mpirun --host n1:1 --mca oob_tcp_static_ipv4_ports 1234 -np 1 --mca routed radix --mca oob tcp true

Signed-off-by: Gilles Gouaillardet <gilles@rist.or.jp>
2016-12-07 17:19:25 +09:00
Ralph Castain
fbed2d794a Update to latest PMIx master + PTL branch
Update the usock component to disable it

Signed-off-by: Ralph Castain <rhc@open-mpi.org>
2016-12-06 20:47:44 -08:00
rhc54
f95f11d285 Merge pull request #2534 from rhc54/topic/pmixconfig
Update pmix check headers to support Open BSD
2016-12-06 20:46:34 -08:00
Gilles Gouaillardet
1635c293e6 Merge pull request #2522 from ggouaillardet/topic/misc_asm_fixes
Topic/misc asm fixes
2016-12-07 13:31:49 +09:00
Ralph Castain
d51821cbc7 Update pmix check headers to support Open BSD
Signed-off-by: Ralph Castain <rhc@open-mpi.org>
2016-12-06 19:37:06 -08:00
rhc54
c2f57581df Merge pull request #2527 from rhc54/topic/sig2
Update signal handling to introduce a pause between SIGCONT and SIGT…
2016-12-06 14:02:25 -08:00
Ralph Castain
85a634926b Update signal handling to introduce a pause between SIGCONT and SIGTERM, followed by another pause before SIGKILL. Do this within the odls/kill_local_procs function while we know we are blocked in an event, and before the daemon shuts down the event progress loop
Signed-off-by: Ralph Castain <rhc@open-mpi.org>
2016-12-06 12:34:42 -08:00
alex-mikheev
a3e4c33f0e Merge pull request #2524 from jladd-mlnx/topic/shmemx_h-master
Remove shmemx.h from shmem.h. Add shmem.h to shmemx.h
2016-12-06 09:49:34 +02:00
Gilles Gouaillardet
299a6f8d7c configury: auto-detect armhf and armel architectures on Debian
Thanks Alastair McKinstry for the patch

Fixes open-mpi/ompi#2514

Signed-off-by: Gilles Gouaillardet <gilles@rist.or.jp>
2016-12-06 14:49:54 +09:00
Gilles Gouaillardet
596613c0aa configury: add support for x32 architecture
Thanks Alastair McKinstry for the patch

Fixes open-mpi/ompi#2515

Signed-off-by: Gilles Gouaillardet <gilles@rist.or.jp>
2016-12-06 14:49:37 +09:00
Gilles Gouaillardet
c8b51a2d3b configury: remove some dead code
perl is now mandatory to build Open MPI,
so there is no need to check for it

Signed-off-by: Gilles Gouaillardet <gilles@rist.or.jp>
2016-12-06 14:49:37 +09:00
Joshua Ladd
dc6f4a0feb Remove shmemx.h from shmem.h. Add shmem.h to shmemx.h
Fixes #2483
Signed-off-by: Joshua Ladd <joshual@mellanox.com>
2016-12-06 06:42:26 +02:00
Joshua Ladd
59f40e7cc5 Merge pull request #2500 from vspetrov/hcoll_ctx_free_detection
Detect hcoll_context_free at config
2016-12-05 22:39:40 -05:00
rhc54
7af49eb735 Merge pull request #2521 from rhc54/topic/rml
Resolve a duplicate symbol issue when the rml/ofi component is enabled
2016-12-05 15:33:31 -08:00
Ralph Castain
d8f262e39b Resolve a duplicate symbol issue when the rml/ofi component is enabled
Signed-off-by: Ralph Castain <rhc@open-mpi.org>
2016-12-05 13:41:38 -08:00
Jeff Squyres
40d94fdc5a Merge pull request #2422 from edgargabriel/pr/cycle-buf-default-val
io/ompio: change the default value of mca parameter
2016-12-05 15:33:52 -05:00
Ralph Castain
144a9d267b Update the purge-tab-indents.pl script to avoid resetting permissions
Signed-off-by: Ralph Castain <rhc@open-mpi.org>
2016-12-05 09:49:58 -08:00
Gilles Gouaillardet
5bb3efdc74 configury: check the existence of perl
perl is required by ompi/mpi/man/make_manpage.pl, that is even used in opal.
so simply aborts at configure time if perl is not available

Signed-off-by: Gilles Gouaillardet <gilles@rist.or.jp>
2016-12-05 10:31:51 +09:00
Ralph Castain
f91f8ce494 Protect against NULL param
Signed-off-by: Ralph Castain <rhc@open-mpi.org>
2016-12-03 21:19:32 -08:00
Ralph Castain
f633d5c1b1 Initialize var
Signed-off-by: Ralph Castain <rhc@open-mpi.org>
2016-12-03 21:08:45 -08:00
rhc54
2a01cc853a Merge pull request #2512 from rhc54/topic/dyn
Allow a PMIx tool to spawn a job
2016-12-03 18:15:19 -08:00
Ralph Castain
79cde184ad Allow a PMIx tool to spawn a job
Signed-off-by: Ralph Castain <rhc@open-mpi.org>
2016-12-03 16:00:47 -08:00
rhc54
003f7d308f Merge pull request #2504 from hppritcha/topic/fix_pmix_base_help_file
pmix: Fix pmix base help file.
2016-12-03 11:57:07 -08:00
rhc54
cda9bd3dd9 Merge pull request #2511 from rhc54/topic/nohang
Avoid hanging in show_help if PMIx was unable to initialize
2016-12-03 11:27:55 -08:00
rhc54
fa04ba9ef1 Merge pull request #2510 from rhc54/topic/cleanup
Fix the session directory cleanup - only remove the jobfam session di…
2016-12-03 10:53:35 -08:00
Ralph Castain
a43fae74a5 Avoid hanging in show_help if PMIx was unable to initialize
Signed-off-by: Ralph Castain <rhc@open-mpi.org>
2016-12-03 10:12:58 -08:00
Ralph Castain
af9a55ccf1 Fix the session directory cleanup - only remove the jobfam session dir level if we are the local daemon and are cleaning up our own session directory.
Update the scaling test to run more trials and report the options being tested each time

Signed-off-by: Ralph Castain <rhc@open-mpi.org>
2016-12-03 09:59:18 -08:00
rhc54
b030f9acc0 Merge pull request #2506 from rhc54/topic/pmixext
Now that PMIx has settled on its release strategy and numbering, update the OPAL pmix framework to track
2016-12-03 07:58:42 -08:00
Jeff Squyres
6319332258 Merge pull request #2491 from OMGtechy/master
Swapped use of fprintf for opal_output_verbose
2016-12-03 07:32:03 -05:00
Ralph Castain
a4e3f615e3 Fix executable modes of pmix config scripts
Signed-off-by: Ralph Castain <rhc@open-mpi.org>
2016-12-02 20:50:59 -08:00
Ralph Castain
342dbfcf4e Update
Signed-off-by: Ralph Castain <rhc@open-mpi.org>
2016-12-02 20:33:03 -08:00
Ralph Castain
8e64382edf Update to correct tarball
Signed-off-by: Ralph Castain <rhc@open-mpi.org>
2016-12-02 19:16:07 -08:00
Ralph Castain
1a0bccb536 Now that PMIx has settled on its release strategy and numbering, update the OPAL pmix framework to track
Signed-off-by: Ralph Castain <rhc@open-mpi.org>
2016-12-02 15:44:43 -08:00