ports/openmpi
ports/openmpi
1
1
Форкнуть 0
Код Релизы Активность
24 054 коммитов 3 Ветки 0 Теги
Граф коммитов

8649 Коммитов

Автор SHA1 Сообщение Дата
Edgar Gabriel
9e5ade4e8b argh, a debugging sleep statement got into the source code. 2015-11-16 13:26:57 -06:00
Edgar Gabriel
dbfbcdecd5 make adjustments for the default settings of grouping parameters and the default contiguous group size option. 
minor bug fix in the simple grouping strategy.
 2015-11-16 08:17:27 -06:00
Edgar Gabriel
27628774c7 add a new option for a simple aggregator selection which has zero communication costs. 2015-11-16 08:17:26 -06:00
Edgar Gabriel
66c1ea5fcb change the default value of the grouping option. Also add new grouping option which skips the refinement step in the aggregator selection. 2015-11-16 08:17:23 -06:00
Edgar Gabriel
e8e117503d reduce the communication volume during MPI_File_set_view 2015-11-16 08:17:22 -06:00
yohann
005400a937 mtl/ofi: Make sure the resources are managed by the provider. 2015-11-13 16:16:58 -08:00
Nathan Hjelm
9ef0821856 osc/rdma: fix some threading bugs 
There were two bugs in osc/rdma when using threads:

 - Deadlock is ompi_osc_rdma_start_atomic. This occurs because
   ompi_osc_rdma_frag_alloc is called with the module lock. To fix the
   issue the module lock is now recursive. In the future I will add a
   new lock to protect just the current rdma fragment.

 - Do not drop the lock in ompi_osc_rdma_frag_alloc when calling
   ompi_osc_rdma_frag_complete. Not only is it not needed but dropping
   the lock at this point can cause a competing thread to mess up the
   state.

Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
 2015-11-12 20:25:57 -07:00
Yossi
b750b72a81 Merge pull request #1127 from yosefe/topic/pml-ucx-implement-cancel 
pml_ucx: implement cancel, and add small optimizations.
 2015-11-12 10:50:48 +02:00
yosefe
7becc54d67 pml_ucx: fix typo. 2015-11-12 09:57:41 +02:00
Todd Kordenbrock
b9630f802b Merge pull request #1120 from francois-wellenreiter/mtl_min_mdbind 
mtl-portals4 : remove useless PtlMDBind PtlMDRelease calls for rendez-vous messages
 2015-11-10 14:34:19 -06:00
yosefe
d66b01d380 pml_ucx: implement cancel, and add small optimizations. 2015-11-10 17:40:06 +02:00
Gilles Gouaillardet
d6ff25b9a2 pml/monitoring: initialize common symbols 2015-11-10 13:58:54 +09:00
Jeff Squyres
e89ecac83c bml r2: fix exclusivity comparison 
Fixes open-mpi/ompi#1106
 2015-11-06 13:26:32 -08:00
Francois WELLENREITER
b301b49a40 MTL portals4 : remove useless PtlMDBind PtlMDRelease calls for rendez-vous messages 2015-11-06 15:55:44 +01:00
Ralph Castain
bfdf08ae86 Fix intercomm_create by ensuring that both sides know how to translate jobid to/from nspace 
Return something just to ensure that pack is happy
 2015-11-06 02:19:45 -08:00
Nathan Hjelm
fda5daf453 Merge pull request #1096 from kawashima-fj/pr/fortran-var-type-fix 
Fix Fortran variable types
 2015-11-05 14:27:40 -07:00
Nathan Hjelm
acf3cb9b9b Merge pull request #1095 from kawashima-fj/pr/trivial-fixes 
Some trivial fixes
 2015-11-04 09:45:59 -07:00
yosefe
45c3d04857 pml_ucx: fix request construct/destruct. 
We should invoke OBJ_CONTRUCT/OBJ_DESTRUCT only on regular requests
(which are embedded inside UCX requests) and for the completed request.
Persistent requests are already constructed/destructed by the free list.
This fixes an assertion in ompi_request_destruct.
 2015-11-04 11:03:46 +02:00
KAWASHIMA Takahiro
c09f9f05d3 mpi/tool: Fix an incorrect type cast. 
This bug caused an invalid result value on `MPI_T_cvar_read`
on big-endian machines or for large (>=2Gi) cvar values.
 2015-11-04 11:28:43 +09:00
KAWASHIMA Takahiro
384f4b51d1 fortran: Fix: missing dimension(*) in (I)NEIGHBOR_ALLTOALLW. 2015-11-04 10:38:25 +09:00
KAWASHIMA Takahiro
1092eabfab fortran: Update comment. 
The structure was changed in commit 9c77c6b.
 2015-11-04 10:38:25 +09:00
KAWASHIMA Takahiro
107c0073dd fortran: Fix: MPI_UNWEIGHTED and MPI_WEIGHTS_EMPTY should be arrays. 
Without this modification, gfortran throw the following error
if these variables are used for `MPI_DIST_GRAPH_CREATE_ADJACENT` or
`MPI_DIST_GRAPH_CREATE_ADJACENT`.

  Error: There is no specific subroutine for the generic
  'mpi_dist_graph_create_adjacent' at (1)
 2015-11-04 10:38:25 +09:00
KAWASHIMA Takahiro
d5e1f40a1e fortran: Fix: info should be an integer parameter. 2015-11-04 10:38:24 +09:00
KAWASHIMA Takahiro
9bf93810d7 fortran: Fix: array dimension of MPI_ARGVS_NULL. 
`MPI_ARGVS_NULL` should be a two-dimensional array.
Without this modification, gfortran throw the following error
if `MPI_ARGVS_NULL` is used for `MPI_COMM_SPAWN_MULTIPLE`.

  Error: There is no specific subroutine for the generic
  'mpi_comm_spawn_multiple' at (1)
 2015-11-04 10:38:24 +09:00
George Bosilca
b14212f142 Fix Coverity issue 1338059. 2015-11-02 22:51:52 -05:00
George Bosilca
5c60e76669 Fix Coverity CIDs 1338021, 1338020, 1338019, 1338018. 2015-11-02 17:38:51 -05:00
bosilca
f1a5362f94 Merge pull request #1072 from bosilca/topic/resized 
Fix for the subarray and darray type creation issue.
 2015-11-01 21:17:03 -05:00
George Bosilca
b77c203068 Add more comments and restore the progress, flags, max tag, and max 
context_id from the original PML.
 2015-10-31 17:13:35 -04:00
George Bosilca
3efd494972 Make sure the monitoring infrastructure works well with the 
new dynamic add_procs.
 2015-10-31 17:13:35 -04:00
Guillaume Papauré
86714ad91e change pml_monitoring_messages_count and pml_monitoring_messages_size pvars to use the start/stop features 2015-10-31 17:13:35 -04:00
George Bosilca
a43c2ce529 Fully integrate the monitoring with the MPI_T PVAR. 
Writing to the pml_monitoring_flush variable will set the filename of
the output file.
Stopping a session for the pml_monitoring_flush will force the
generation of the nobitoring output file (as long as the filename
is not NULL).
To reset the monitoring, une has to bind the pml_monitoring_flush to a
session.
 2015-10-31 17:13:35 -04:00
George Bosilca
646a662721 Use the new group interface and add const to the PML send functions. 2015-10-31 17:13:35 -04:00
George Bosilca
5224a7ce4d Allow the pvar to be written by invoking the associated callback. 
Use a PVAR to generate the monitoring dump of the information into a
file.

Use the PVAR to instruct the PML monitoring when to do the dump.
 2015-10-31 17:13:35 -04:00
George Bosilca
df167f4177 Rewrite the close logic to be more clean and cleaner. 2015-10-31 17:13:35 -04:00
George Bosilca
c801ffde86 Use MPI_T variables to handle the flush in a more MPI-blessed way. 
Code cleanup.

Update the monitoring test to use MPI_T variables.
 2015-10-31 17:13:35 -04:00
George Bosilca
4f88c82500 Fix a convertion problem and add a comment about the lack of component 
retain in the new component infrastructure.

Clean Makefile.am to fix "make distcheck".

Update the gitignore rules.
 2015-10-31 17:13:35 -04:00
George Bosilca
80343a0d39 add ability to querry pml monitorinting results with MPI Tools interface 
using performance variables "pml_monitoring_messages_count" and
"pml_monitoring_messages_size"

Per Brice suggestion make all data count and message length be
uint64_t.
 2015-10-31 17:13:35 -04:00
George Bosilca
a47d69202f Add a monitoring PML. This PML track all data exchanges by the processes 
counting or not the collective traffic as a separate entity. The need
for such a PML is simply because the PMPI interface doesn't allow us to
identify the collective generated traffic.
 2015-10-31 17:13:35 -04:00
Rolf vandeVaart
578385ca78 Merge pull request #1079 from rolfv/pr/cuda-require-41 
Make CUDA 4.1 a requirement for CUDA-aware support
 2015-10-29 12:56:22 -04:00
Nathan Hjelm
b1e3936261 Merge pull request #1078 from rolfv/pr/disable-osc-rdma-for-cuda 
Disable the use of osc rdma when we detect a GPU buffer
 2015-10-29 10:03:28 -06:00
Rolf vandeVaart
f2ff6e03ab Make CUDA 4.1 a requirement for CUDA-aware support. 
Remove all related preprocessor conditionals.
 2015-10-29 11:24:02 -04:00
Matias Cabral
8ebcac1b2c Merge pull request #1075 from matcabral/psm2_symbol_rename 
Updated psm2 mtl with new externally exposed symbols of psm2.so 
Fixes open-mpi/ompi#1018
Fixes open-mpi/ompi#1021
 2015-10-28 13:55:45 -07:00
Rolf vandeVaart
87a4cc6118 Disable the use of osc rdma when we detect a GPU buffer as it is not supported in that component. 
This forces a failover to the osc pt2pt component. Fixes #1012
 2015-10-28 14:47:45 -04:00
yosefe
ae738d0434 pml_ucx: add pmi fence in del_procs 2015-10-28 18:34:36 +02:00
Matias A Cabral
ed16d8e1cc Updated psm2 mtl with new externally exposed symbols of psm2.so 
Fixes open-mpi/ompi#1018
Fixes open-mpi/ompi#1021
 2015-10-28 09:12:33 -07:00
yosefe
41b6230be3 pml_ucx: fix debug macros, and initialize mpi request properly. 2015-10-28 10:59:25 +02:00
Ralph Castain
267ca8fcd3 Cleanup the PMIx direct modex support. Add an MCA parameter pmix_base_async_modex that will cause the async modex to be used when set to 1. Default it to 0 for now 
to continue current default behavior.

Also add an MCA param pmix_base_collect_data to direct that the blocking fence shall return all data to each process. Obviously, this param has no effect if async_
modex is used.
 2015-10-27 17:31:56 -07:00
yohann
8bf1c95cdc mtl/ofi: Remove unused help messages. 2015-10-27 09:38:04 -07:00
Nathan Hjelm
69d403d42b Merge pull request #1054 from hjelmn/add_procs_threading 
add_procs: add threading protection for dynamic add_procs
 2015-10-27 09:27:13 -06:00
George Bosilca
679dc9b437 Fix the subarray and darray type creation. Include a 
small patch provided by Gilles.
 2015-10-26 23:44:26 -04:00