| 9aff3a9 | George Bosilca | 05 December 2018, 19:05:47 UTC | Add more details about what is going on. Signed-off-by: George Bosilca <bosilca@icl.utk.edu> | 06 March 2019, 00:43:44 UTC |
| e4aae6b | George Bosilca | 05 December 2018, 04:48:40 UTC | Add a test for very large data. Signed-off-by: George Bosilca <bosilca@icl.utk.edu> | 06 March 2019, 00:43:31 UTC |
| 320a839 | Gilles Gouaillardet | 20 October 2018, 21:50:51 UTC | opal/datatype: correctly handle large datatypes Always use size_t (instead of converting to an uint32_t) in order to correctly support large datatypes. Thanks Ben Menadue for the initial bug report Refs open-mpi/ompi#6016 Signed-off-by: Gilles Gouaillardet <gilles@rist.or.jp> | 06 March 2019, 00:41:39 UTC |
| a2bea96 | Howard Pritchard | 04 March 2019, 14:06:32 UTC | Merge pull request #6445 from rhc54/cmr40x/slurm v4.0.x: Update Slurm PMI configury to account for PMIx | 04 March 2019, 14:06:32 UTC |
| 1675b8e | Ralph Castain | 21 February 2019, 21:28:10 UTC | Ensure we push/pop local AC vars in the right place Signed-off-by: Ralph Castain <rhc@pmix.org> (cherry picked from commit c054d4d1cc063fd22c7111d15d8b14707506bfb7) | 01 March 2019, 16:40:35 UTC |
| 0322ad0 | Ralph Castain | 21 February 2019, 19:33:35 UTC | Update slurm pmi configury to account for pmix When Slurm is built against PMIx, some installations place a copy of the PMIx library that Slurm is linking against in the Slurm PMI location. Current configury ignores that location. The desired behavior is to look for a PMIx lib in that location when --with-pmi is given. If the user also specifies --with-pmix and gives a different location, then override anything previously found and look for it where the user directed. Signed-off-by: Ralph Castain <rhc@pmix.org> (cherry picked from commit cd1b5641beca7f158360983cd31f7297548b0a3c) | 01 March 2019, 16:39:49 UTC |
| c0ee7ad | Geoff Paulsen | 27 February 2019, 18:31:09 UTC | Merge pull request #6359 from gpaulsen/topic/v4.0.x/mpi1removal Fix MPI1 function removal [v4.0.x] Issue 6114 | 27 February 2019, 18:31:09 UTC |
| 6df6a3f | Geoffrey Paulsen | 15 February 2019, 20:19:07 UTC | mpi.h.in: Revamp MPI-1 removed function warnings Refs https://github.com/open-mpi/ompi/issues/6278. This commit is intended to be cherry-picked to v4.0.x and the following commit will ammend to this functionality for master's removal. Changes the prototypes for MPI removed functions in the following ways: There are 4 cases: 1) User wants MPI-1 compatibility (--enable-mpi1-compatibility) MPI_Address (and friends) are declared in mpi.h with deprecation notice 2) User does not want MPI-1 compatibility, and has a C11-capable compiler Declare an MPI_Address (etc.) macro in mpi.h, which will cause a compile-time error using _Static_assert C11 feature 3) User does not want MPI-1 compatibility, and does not have a C11-capable compiler, but the compiler supports error function attributes. Declare an MPI_Address (etc.) macro in mpi.h, which will cause a compile-time error using error function attribute. 4) User does not want MPI-1 compatibility, and does not have a C11-capable compiler, or a compiler that supports error function attributes. Do not declare MPI_Address (etc.) in mpi.h at all. Unless the user is compiling with something like -Werror, this will allow the user's code to compile. We are choosing this because it seems like a losing battle to make some kind of compile time error that is friendly to the user (and doesn't make it look like mpi.h itself is broken). On v4.0.x, this will allow the user code to both compile (albeit with a warning) and link (because the MPI_Address will be in the MPI library because we are preserving ABI back to 3.0.x). On master/v5.0.x, this will allow the user code to compile, but it will fail to link (because the MPI_Address symbol will not be in the MPI library). Signed-off-by: Geoffrey Paulsen <gpaulsen@us.ibm.com> (cherry-picked from 3136a1706cdacb3f7530937da1dcfe15d2febc79) | 27 February 2019, 16:25:23 UTC |
| 5a89295 | Geoff Paulsen | 27 February 2019, 15:07:09 UTC | Merge pull request #6442 from hppritcha/topic/new_update3_v4.0.1 NEWS: another update for 4.0.1 | 27 February 2019, 15:07:09 UTC |
| 8449a00 | Howard Pritchard | 27 February 2019, 15:01:01 UTC | NEWS: another update for 4.0.1 [skip ci] Signed-off-by: Howard Pritchard <howardp@lanl.gov> | 27 February 2019, 15:01:01 UTC |
| 0af10b7 | Howard Pritchard | 26 February 2019, 16:07:19 UTC | Merge pull request #6435 from jsquyres/pr/v4.0.x/fix-hwloc-configury Hold off running hwloc:external feature tests until after we decide i… | 26 February 2019, 16:07:19 UTC |
| 8bf3a86 | Ben Menadue | 25 February 2019, 05:58:11 UTC | Hold off running hwloc:external feature tests until after we decide if we're using the internal or external component. This fixes #6430. Signed-off-by: Ben Menadue <ben.menadue@nci.org.au> (cherry picked from commit 17dcc7041ac272c65eb727f45c5628459cbb6055) | 25 February 2019, 23:09:48 UTC |
| 056d7ad | Howard Pritchard | 25 February 2019, 22:54:15 UTC | Merge pull request #6419 from hppritcha/topic/fix_pgi_usempif08_4.0.x fortran:use mpif08 fix for PGI linking | 25 February 2019, 22:54:15 UTC |
| fd143eb | Howard Pritchard | 25 February 2019, 22:52:34 UTC | Merge pull request #6434 from hppritcha/topic/readme_v4.0.1 README: updates for v4.0.1 release | 25 February 2019, 22:52:34 UTC |
| 6a8dba3 | Howard Pritchard | 25 February 2019, 20:25:39 UTC | README: updates for v4.0.1 release fixes #6425 Signed-off-by: Howard Pritchard <howardp@lanl.gov> | 25 February 2019, 20:25:58 UTC |
| 1920769 | Geoff Paulsen | 22 February 2019, 22:25:38 UTC | Merge pull request #6423 from abouteiller/pr6417to4.0.x v4.x: Cart/Graph create would not run the next_cid algorithm | 22 February 2019, 22:25:38 UTC |
| 4b317e0 | Geoff Paulsen | 22 February 2019, 21:56:17 UTC | Merge pull request #6421 from hppritcha/topic/version_v401 VERSION: update for v4.0.1rc1 release | 22 February 2019, 21:56:17 UTC |
| d6e8d51 | Aurelien Bouteiller | 21 February 2019, 16:40:22 UTC | Cart/Graph create would not run the next_cid algorithm and create disjoint communicator with inconsistent cid. Signed-off-by: Aurelien Bouteiller <bouteill@icl.utk.edu> | 22 February 2019, 20:11:56 UTC |
| a791db2 | Howard Pritchard | 22 February 2019, 20:03:45 UTC | Merge pull request #6420 from hppritcha/topic/license_v401 LICENSE: for v4.0.1 | 22 February 2019, 20:03:45 UTC |
| 8db349f | Howard Pritchard | 22 February 2019, 19:58:04 UTC | VERSION: update for v4.0.1rc1 release Signed-off-by: Howard Pritchard <howardp@lanl.gov> | 22 February 2019, 19:58:37 UTC |
| f38eebb | Howard Pritchard | 22 February 2019, 19:02:18 UTC | LICENSE: for v4.0.1 [skip ci] Signed-off-by: Howard Pritchard <howardp@lanl.gov> | 22 February 2019, 19:02:18 UTC |
| 6596277 | Howard Pritchard | 20 February 2019, 00:39:49 UTC | fortran:use mpif08 fix for PGI linking commit c6070fd2e broke building fortran bindings with PGI compilers. Turns out PGI compilers need to link in the *.o from a module file whether or not there are module subroutines defined or not in the module file. Related to #6411 Signed-off-by: Howard Pritchard <howardp@lanl.gov> (cherry picked from commit 266bc3aced5ff9019f01faef1ed01dd463fafd41) | 22 February 2019, 18:47:40 UTC |
| 7aeb655 | Howard Pritchard | 22 February 2019, 03:29:47 UTC | Merge pull request #6395 from brminich/topic/ucx_net_waddr_4.0.x PML/UCX: Use net worker address for remote peers - v4.0.x | 22 February 2019, 03:29:47 UTC |
| e82523f | Geoff Paulsen | 21 February 2019, 21:48:50 UTC | Merge pull request #6410 from hppritcha/topic/news_for_v4.0.1 update NEWS with a new fix | 21 February 2019, 21:48:50 UTC |
| 7bb728b | Howard Pritchard | 21 February 2019, 15:44:07 UTC | Merge pull request #6399 from hppritcha/topic/excise_ofi_rml rml/ofi: remove | 21 February 2019, 15:44:07 UTC |
| 1c51494 | Mikhail Brinskii | 13 February 2019, 14:23:09 UTC | PML/UCX: Use net worker address for remote peers For remote node peers pack smaller worker address, which contains network device addresses only. This would reduce amount of OOB traffic during startup. Signed-off-by: Mikhail Brinskii <mikhailb@mellanox.com> (cherry picked from commit 751d88192d05edb7e1912bab4e48643c6f9e1574) | 21 February 2019, 14:58:20 UTC |
| 35e3c07 | Howard Pritchard | 20 February 2019, 16:55:07 UTC | update NEWS with a new fix Signed-off-by: Howard Pritchard <howardp@lanl.gov> | 20 February 2019, 16:55:07 UTC |
| 83cb9ca | Howard Pritchard | 20 February 2019, 16:53:50 UTC | Merge pull request #6404 from ggouaillardet/topic/v4.0.x/osc_rdma_self osc/rdma: correctly handle communications to self | 20 February 2019, 16:53:50 UTC |
| f433b64 | Howard Pritchard | 20 February 2019, 16:51:26 UTC | Merge pull request #6405 from ggouaillardet/topic/v4.0.x/man_win_attach_detach man: fix typos in MPI_Win_{attach,detach} | 20 February 2019, 16:51:26 UTC |
| 7b71369 | KAWASHIMA Takahiro | 20 February 2019, 02:22:38 UTC | man: fix more typos in MPI_Win_attach man page Signed-off-by: KAWASHIMA Takahiro <t-kawashima@jp.fujitsu.com> [skip ci] bot:notest (cherry picked from commit open-mpi/ompi@7095ad10a5de1980c1f252f5be1cb72f28f0493b) | 20 February 2019, 04:26:48 UTC |
| 3ab227d | Gilles Gouaillardet | 20 February 2019, 02:09:09 UTC | man: fix typos in MPI_Win_{attach,detach} man pages no code change [skip ci] bot:notest Signed-off-by: Gilles Gouaillardet <gilles@rist.or.jp> (cherry picked from commit open-mpi/ompi@7c0596819b4dd1a4f98d6b92d00f9b568d634438) | 20 February 2019, 04:25:12 UTC |
| 749f518 | Gilles Gouaillardet | 20 February 2019, 00:49:37 UTC | osc/rdma: correctly handle communications to self mark the "self" peer OMPI_OSC_RDMA_PEER_LOCAL_BASE when the window is dynamically created and use_cpu_atomics is set in order to correctly handle communications to self. Thanks Bart Janssens for reporting this issue. Refs. open-mpi/ompi#6394 Signed-off-by: Gilles Gouaillardet <gilles@rist.or.jp> (back-ported from commit open-mpi/ompi@fe05fcc11a5f94fdde9a97423428f9ad779c296c) | 20 February 2019, 04:06:05 UTC |
| 55915c3 | Howard Pritchard | 19 February 2019, 17:27:47 UTC | rml/ofi: remove per discussion at the 2/19/19 devel-core meeting, remove rml/ofi from 4.0.x Signed-off-by: Howard Pritchard <howardp@lanl.gov> | 19 February 2019, 17:27:47 UTC |
| 4fd2c60 | Geoff Paulsen | 15 February 2019, 22:17:25 UTC | Merge pull request #6391 from hppritcha/topic/news_for_v4.0.1 NEWS: update for 4.0.1 release | 15 February 2019, 22:17:25 UTC |
| 4b2c62d | Howard Pritchard | 15 February 2019, 20:43:05 UTC | NEWS: update for 4.0.1 release Signed-off-by: Howard Pritchard <howardp@lanl.gov> | 15 February 2019, 20:43:05 UTC |
| c593b20 | Geoff Paulsen | 15 February 2019, 19:39:12 UTC | Merge pull request #6380 from hppritcha/ggouaillardet-topic/oob_tcp_cross_version_compatibility v4.0.x: oob/tcp: add cross version compatibility support | 15 February 2019, 19:39:12 UTC |
| 40db950 | Howard Pritchard | 15 February 2019, 00:06:47 UTC | Merge pull request #6340 from jsquyres/pr/v4.0.x/make-mpi.h-a-little-friendlier-to-c++ v4.0.x: mpi.h.in: use C++ static_cast<> where appropriate | 15 February 2019, 00:06:47 UTC |
| d2745ad | Howard Pritchard | 15 February 2019, 00:05:32 UTC | Merge pull request #6327 from ggouaillardet/topic/v4.0.x/op ompi/op: fix support of non predefined datatypes with predefined oper… | 15 February 2019, 00:05:32 UTC |
| d82be47 | Howard Pritchard | 15 February 2019, 00:04:53 UTC | Merge pull request #6273 from ggouaillardet/topic/v4.0.x/configury_clang5 v4.0.x: configury: enhance C11 detection | 15 February 2019, 00:04:53 UTC |
| de1dd1c | Howard Pritchard | 13 February 2019, 19:52:11 UTC | oob/tcp: hardwire oob_tcp version string to 4.0.0 Signed-off-by: Howard Pritchard <howardp@lanl.gov> | 13 February 2019, 19:54:03 UTC |
| dd75079 | Gilles Gouaillardet | 06 December 2018, 05:44:39 UTC | oob/tcp: add cross version compatibility support Since we intend to provide cross version compatibility between versions with the same major and minor, use MAJOR.MINOR.0 instead of orte_version_string (e.g. MAJOR.MINOR.RELEASEGREEK). Open MPI 4.0.0 has already been released, so in order to make it compatible with future 4.0.x releases, we have to use 4.0.0 as the version string, that is why we use MAJOR.MINOR.0 instead of MAJOR.MINOR Signed-off-by: Gilles Gouaillardet <gilles@rist.or.jp> | 13 February 2019, 17:21:32 UTC |
| 0b915b7 | Howard Pritchard | 12 February 2019, 16:13:19 UTC | Merge pull request #6333 from jsquyres/pr/v4.0.x/hwloc-macro-conflict-fixes v4.0.x: Various minor hwloc cleanups | 12 February 2019, 16:13:19 UTC |
| e9cef8c | Geoff Paulsen | 11 February 2019, 20:23:24 UTC | Merge pull request #6375 from karasevb/4.0.x_regx_host_ordering_fix v4.0.x/regex: fixed host ordering for different prefixes | 11 February 2019, 20:23:24 UTC |
| 6513b85 | Howard Pritchard | 11 February 2019, 20:16:15 UTC | Merge pull request #6249 from hjelmn/v4.0.x_fix_issue_6201_in_the_v4.0.x_branch v4.0.x: btl/vader: don't try to set reachabilty in add_procs if not requested | 11 February 2019, 20:16:15 UTC |
| 5dd6340 | Howard Pritchard | 11 February 2019, 20:15:52 UTC | Merge pull request #6368 from jsquyres/pr/v4.0.x/fix-ofi-configury v4.0.x: fix OFI configury | 11 February 2019, 20:15:52 UTC |
| 9e306ce | Howard Pritchard | 11 February 2019, 20:14:06 UTC | Merge pull request #6336 from jsquyres/pr/v4.0.x/fix-datatype-destructor-leak v4.0.x: opal/datatype: plug a memory leak in opal_datatype_t destructor | 11 February 2019, 20:14:06 UTC |
| 87c9086 | Boris Karasev | 30 January 2019, 03:23:36 UTC | regx: fixed the order of hosts for ranges with different prefixes Example: For the list of hosts `a01,b00,a00` a regex is generated: `a[2:1.0],b[2:0]`, where `a`-hosts prefixes moved to the begining, it breaks the hosts ordering. This commit fixes regex for that case to `a[2:1],b[2:0],a[2:0]` Signed-off-by: Boris Karasev <karasev.b@gmail.com> (cherry picked from commit 46e38b9193f5554b7a26aa830e98b10d6836acb2) | 11 February 2019, 10:06:49 UTC |
| 62044da | Boris Karasev | 30 January 2019, 03:22:55 UTC | regx/reverse: fixed adding an empty range for no numerical hostnames Example: For the nodelist `jjss,jjss0000001,jjss0000003,jjss0000002` a regular expression was `jjss[0:0],jjss[7:1,3,2]` that led to incorrect unpacking the first host as `jjs0`. This commit fixes an adding empty range for not numeric hostnames. Here is the fixed regex for this exapmle: `jjss,jjss[7:1,3,2]` Signed-off-by: Boris Karasev <karasev.b@gmail.com> (cherry picked from commit 1967e41a71dbfd892513bb02a98fe6bb418777a4) | 11 February 2019, 10:06:34 UTC |
| c154631 | Boris Karasev | 30 January 2019, 03:22:17 UTC | regx/test: update regex test Signed-off-by: Boris Karasev <karasev.b@gmail.com> (cherry picked from commit d1ad90f47e93436d0e1b49e5946da9f6b33aacdf) | 11 February 2019, 10:05:50 UTC |
| 8552d0e | Howard Pritchard | 08 February 2019, 21:44:08 UTC | Merge pull request #6330 from ggouaillardet/topic/v4.0.x/ompi_datatype_set_args ompi/datatype: fix how we compute the space needed for the args | 08 February 2019, 21:44:08 UTC |
| d843220 | Howard Pritchard | 08 February 2019, 21:41:11 UTC | Merge pull request #6307 from uberlinuxguy/v4.0.x-fix-for-6303 Adding changes for issue #6303 for branch v4.0.x. | 08 February 2019, 21:41:11 UTC |
| 85ed3f4 | Howard Pritchard | 08 February 2019, 21:39:39 UTC | Merge pull request #6347 from ggouaillardet/topic/v4.0.x/opal_convertor_raw opal/datatype: fix opal_convertor_raw() | 08 February 2019, 21:39:39 UTC |
| 7513705 | Howard Pritchard | 07 February 2019, 15:44:20 UTC | Merge pull request #6335 from edgargabriel/pr/v4.0.x-floating-point-division-problem Pr/v4.0.x floating point division problem | 07 February 2019, 15:44:20 UTC |
| 7fd62cf | Jeff Squyres | 06 February 2019, 12:53:14 UTC | Remove opal/mca/common/ofi. It never lived up to its purpose (and has caused amorphous indirect errors such as https://github.com/open-mpi/ompi/issues/2519), so delete it. Signed-off-by: Jeff Squyres <jsquyres@cisco.com> (cherry picked from commit dd20174532928e0c9cdbe7b206868e6e4bea9d0b) | 07 February 2019, 14:39:22 UTC |
| 9ad871f | Jeff Squyres | 29 January 2019, 04:10:26 UTC | ofi: revamp OPAL_CHECK_OFI configury Update the OPAL_CHECK_OFI configury macro: - Make it safe to call the macro multiple times: - The checks only execute the first time it is invoked - Subsequent invocations, it just emits a friendly "checking..." message so that configure output is sensible/logical - With the goal of ultimately removing opal/mca/common/ofi, rename the output variables from OPAL_CHECK_OFI to be opal_ofi_{happy|CPPFLAGS|LDFLAGS|LIBS}. - Update btl/usnic and mtl/ofi for these new conventions. - Also, don't use AC_REQUIRE to invoke OPAL_CHECK_OFI because that causes the macro to be invoked at a fairly random time, which makes configure stdout confusing / hard to grok. - Remove a little left-over kruft in OPAL_CHECK_OFI, too (which resulted in an indenting change, making the change to opal_check_ofi.m4 look larger than it really is). Thanks Alastair McKinstry for the report and initial fix. Thanks Rashika Kheria for the reminder. Updated from master cherry pick: the OFI BTL does not exist on the v4.0.x branch. Therefore, did not include the OFI BTL changes on master in this cherry pick. Signed-off-by: Jeff Squyres <jsquyres@cisco.com> (cherry picked from commit f5e1a672ccd5db127e85e1e8f6bcfeb8a8b04527) | 07 February 2019, 14:36:35 UTC |
| 0ae4847 | Gilles Gouaillardet | 01 February 2019, 02:20:13 UTC | opal/datatype: reset ptypes in opal_datatype_clone() Reset ptypes when cloning a datatype in order to prevent a double free() in the opal_datatype_t destructor. This fixes a bug introduced in open-mpi/ompi@7c938f070fa8c906918507dbc78fdadcde324610 Fixes open-mpi/ompi#6346 Signed-off-by: Gilles Gouaillardet <gilles@rist.or.jp> (cherry picked from commit open-mpi/ompi@b395342c9fbdc28f59e01d393d34a6b0c651180e) | 01 February 2019, 05:39:49 UTC |
| 4dfb938 | Howard Pritchard | 01 February 2019, 01:50:05 UTC | Merge pull request #6321 from hppritcha/topic/fix_6236_for_v4.x Topic/fix 6236 for v4.x | 01 February 2019, 01:50:05 UTC |
| 8acdc53 | George Bosilca | 30 January 2019, 00:44:05 UTC | Provide a better fix for #6285. The issue was a little complicated due to the internal stack used in the convertor. The main issue was that in the case where we run out of iov space to save the raw description of the data while hanbdling a repetition (loop), instead of saving the current position and bailing out directly we reading of the next predefined type element. It worked in most cases, except the one identified by the HDF5 test. However, the biggest issue here was the drop in performance for all ensuing calls to the convertor pack/unpack, as instead of handling contiguous loops as a whole (and minimizing the number of memory copies) we copied data description by data description. Signed-off-by: George Bosilca <bosilca@icl.utk.edu> (back-ported from commit open-mpi/ompi@5a82c4fd0769ed007de774ef9115bde73b75de32) | 01 February 2019, 00:28:52 UTC |
| f732773 | Gilles Gouaillardet | 23 January 2019, 05:52:21 UTC | opal/datatype: fix opal_convertor_raw correctly handle the case in which iovec is full and the last accessed element of the datatype is the beginning of a loop Refs. open-mpi/ompi#6285 Thanks Axel Huebl for reporting this Signed-off-by: Gilles Gouaillardet <gilles@rist.or.jp> (back-ported from commit open-mpi/ompi@0832ab5acc2419e670b44ff3857c8c3d21c78e09) | 01 February 2019, 00:26:30 UTC |
| c39426e | Jeff Squyres | 30 January 2019, 12:49:17 UTC | mpi.h.in: use C++ static_cast<> where appropriate When compiling mpi.h with a modern C++ compiler and a high degree of pickyness (e.g., -Wold-style-cast), casting using (void*) in the OMPI_PREDEFINED_GLOBAL and MPI_STATUS*_IGNORE macros will emit warnings. So if we're compiling with a C++ compiler, use C++'s static_cast<> instead of (void*). Thanks to @shadow-fax for identifying the issue. Signed-off-by: Jeff Squyres <jsquyres@cisco.com> (cherry picked from commit 30afdcead915c9c5a62305b93460e2ba8cc6f801) | 31 January 2019, 12:16:07 UTC |
| 90a9c12 | Gilles Gouaillardet | 22 January 2019, 01:57:57 UTC | opal/datatype: plug a memory leak in opal_datatype_t destructor correctly free ptypes if the datatype is not pre-defined. Thanks Axel Huebl for reporting this. Refs. open-mpi/ompi#6291 Signed-off-by: Gilles Gouaillardet <gilles@rist.or.jp> (cherry picked from commit 7c938f070fa8c906918507dbc78fdadcde324610) | 30 January 2019, 18:41:14 UTC |
| e30e5b9 | René Widera | 18 January 2019, 12:55:51 UTC | common/ompio: possible rounding issue Similar to #6286 rounding number of bytes into a single precision floating point value to round up the result of a division is a potential risk due to rounding errors. - remove floating point operations for `round up` - removes floating point conversion for round down (native behavior of integer division) Signed-off-by: René Widera <r.widera@hzdr.de> (cherry picked from commit a91fab80a1e55e1df15f649e18d247e5d4654eb9) | 30 January 2019, 18:31:39 UTC |
| d1e8779 | Edgar Gabriel | 17 January 2019, 17:09:41 UTC | common/ompio: fix a floating point division problem This commit fixes a problem reported on the mailing list with individual writes larger than 512 MB. The culprit is a floating point division of two large, close values. Changing the datatypes from float to double (which is what is being used in the fcoll components) fixes the problem. See issue #6285 and https://forum.hdfgroup.org/t/cannot-write-more-than-512-mb-in-1d/5118 Thanks for Axel Huebl and René Widera for reporting the issue. Signed-off-by: Edgar Gabriel <egabriel@central.uh.edu> (cherry picked from commit c0f8ce0fff4684b670135043dd150abc9d83d988) | 30 January 2019, 18:31:16 UTC |
| a247292 | Gilles Gouaillardet | 30 January 2019, 04:31:44 UTC | topo/treematch: silence a hwloc related warning treematch/km_partitioning.c #include "config.h", but there is no such file when the embedded treematch is used. In order to prevent the embedded treematch from incorrectly using the config.h from the embedded hwloc, generate a dummy config.h. Signed-off-by: Gilles Gouaillardet <gilles@rist.or.jp> (cherry picked from commit 0aeb27f77650d3ee97e17e770c9e5aa487d5e1f5) | 30 January 2019, 12:33:33 UTC |
| c85fd35 | Gilles Gouaillardet | 12 December 2018, 05:18:02 UTC | opal: remove unnecessary #include file opal_config_bottom.h can only be #include'd in opal_config.h, so there is no need to #include "opal_config.h" inside. Signed-off-by: Gilles Gouaillardet <gilles@rist.or.jp> (cherry picked from commit c8790d29de5ef399fd805f27366af6c2dc87ce9e) | 30 January 2019, 12:33:32 UTC |
| f79f14a | Gilles Gouaillardet | 29 January 2019, 15:36:15 UTC | hwloc/base: fix some off-by-one errors Signed-off-by: Gilles Gouaillardet <gilles@rist.or.jp> (cherry picked from commit 73d104f6959c95f676a779826f511826e0b17f6a) | 30 January 2019, 12:33:32 UTC |
| 788c92b | Jeff Squyres | 29 January 2019, 15:36:01 UTC | hwloc/external.h: fix a clash with external HWLOC_VERSION[*] Some macros defined by the embedded hwloc ends up in opal_config.h because hwloc configury m4 files are slurped into Open MPI. These macros are not required here, and they might conflict with an external hwloc install, so simply #undef them in hwloc/external/external.h after including <opal_config.h> but before including the external <hwloc.h>. Signed-off-by: Gilles Gouaillardet <gilles@rist.or.jp> Signed-off-by: Jeff Squyres <jsquyres@cisco.com> (cherry picked from commit f22b7d4f46b03554add3ff2254d1c893359aff84) | 30 January 2019, 12:33:32 UTC |
| fd157a9 | Gilles Gouaillardet | 28 January 2019, 06:24:55 UTC | ompi/datatype: fix how we compute the space needed for the args Refs. open-mpi/ompi#6275 Signed-off-by: Gilles Gouaillardet <gilles@rist.or.jp> (cherry picked from commit open-mpi/ompi@45fb69b2b9e5e06f0ff4f7c97e33110265fc54dc) | 30 January 2019, 02:01:11 UTC |
| f76c81a | Gilles Gouaillardet | 29 January 2019, 00:30:53 UTC | ompi/op: fix support of non predefined datatypes with predefined operators ACCUMULATE, unlike REDUCE, can use with derived datatypes with predefinied operations, with some restrictions outlined in MPI-3:11.3.4. The derived datatype must be composed entierly from one predefined datatype (so you can do all the construction you want, but at the bottom, you can only use one datatype, say, MPI_INT). Refs. open-mpi/ompi#6275 Signed-off-by: Gilles Gouaillardet <gilles@rist.or.jp> (back-ported from commit open-mpi/ompi@bc1cab549883f199d54b302f44da18e430234eeb) | 30 January 2019, 01:29:39 UTC |
| dae71d3 | Ralph Castain | 02 January 2019, 17:03:13 UTC | Correct parsing of ppr directives Needed to apply commit from PR #5778 to get this commit from PR #6238 to apply cleanly. Signed-off-by: Ralph Castain <rhc@open-mpi.org> (cherry picked from commit b19e5edf769858859c96b62d9b2644b44bcc3b03) | 29 January 2019, 18:34:44 UTC |
| 18afb8e | Ralph Castain | 26 September 2018, 17:00:09 UTC | Update mapping system Correctly transfer job-level mapping directives for dynamically spawned jobs to the mapping system. Signed-off-by: Ralph Castain <rhc@open-mpi.org> (cherry picked from commit 45f23ca5c92633f3053569780bea5576eaa1f3a0) | 29 January 2019, 17:04:30 UTC |
| 3d8ddbc | Jason Williams | 25 January 2019, 14:27:33 UTC | Adding changes for issue #6303 for branch v4.0.x. Signed-off-by: Jason Williams <uberlinuxguy@gmail.com> (cherry picked from commit 98d81a5f7a619d5a19615297a6fe8a18d8e3781c) | 29 January 2019, 13:46:59 UTC |
| fb39c7f | Howard Pritchard | 28 January 2019, 21:09:03 UTC | Merge pull request #6271 from rhc54/cmr401/pmix3 v4.0.1: Update to PMIx 3.1.2 | 28 January 2019, 21:09:03 UTC |
| 7c311dc | Howard Pritchard | 28 January 2019, 20:08:20 UTC | Merge pull request #6257 from hoopoepg/topic/get-put-int-uint-8-16-32-v4.0 OSHMEM: added missing API for get/put operations - v4.0 | 28 January 2019, 20:08:20 UTC |
| 335f8c5 | Ralph Castain | 10 January 2019, 18:17:50 UTC | Update to PMIx 3.1.2 Update the OPAL glue configure code to correctly link the opal/pmix3 component to the hwloc used by OMPI instead of defaulting to the system-level hwloc. Required a corresponding update to the PMIx hwloc configure code so we treat hwloc the same way we handle libevent in embedded scenarios. Roll to PMIx v3.1.2 for plugging of memory leaks and addition of faster PMIx_Get response Signed-off-by: Ralph Castain <rhc@pmix.org> | 25 January 2019, 06:58:53 UTC |
| 3ef8a8b | Howard Pritchard | 18 January 2019, 13:34:02 UTC | Merge pull request #6280 from hppritcha/topic/mpool_comp_fix_v40x Update mpool_hugepage_component.c | 18 January 2019, 13:34:02 UTC |
| 2fc5ab7 | heasterday | 11 December 2018, 01:14:16 UTC | Update mpool_hugepage_component.c Signed-off-by: Hunter Easterday <heasterday@lanl.gov> (cherry picked from commit ad0d2c451e63301e5a3b595f9df67bd5c813955e) (cherry picked from commit 509380d99fc7e293f18a2dbb495ef73f9f4cbfef) | 15 January 2019, 15:10:46 UTC |
| c9764f6 | Howard Pritchard | 13 January 2019, 19:31:46 UTC | Merge pull request #6263 from jsquyres/pr/v4.0.x/minor-fortran-valgrind-fix v4.0.x: mpi/fortran: Fix valgrind warnings for type create | 13 January 2019, 19:31:46 UTC |
| cca3caa | Gilles Gouaillardet | 13 January 2019, 03:57:46 UTC | configury: enhance C11 detection clang 5.0 on trusty is busted with respect to C11 atomics This can be evidenced with the simple program below. This test was added into OPAL_PROG_CC_C11_HELPER() and disable C11 atomics if it fails. _Atomic uint32_t a; uint32_t b; atomic_fetch_xor_explicit(&a, b, memory_order_relaxed); Refs. open-mpi/ompi#6264 Signed-off-by: Gilles Gouaillardet <gilles@rist.or.jp> (cherry picked from commit open-mpi/ompi@d1fadebc65a7a7bcfce14641e389f02307f580e2) | 13 January 2019, 05:19:36 UTC |
| bc58e22 | Howard Pritchard | 10 January 2019, 03:10:02 UTC | Merge pull request #6120 from gpaulsen/topic/v4.0.x/re-add-deprecated-oops v4.0.x: Re-add removed deprecate-only MPI-2.0 symbols | 10 January 2019, 03:10:02 UTC |
| 979b401 | Risto Toijala | 08 January 2019, 11:50:33 UTC | mpi/fortran: Fix valgrind warnings for type create Valgrind warns that *newtype is uninitialized when calling from Fortran as e.g. use mpi integer :: t, err call MPI_Type_create_f90_integer(5, t, err) Since newtype is intent(out), this should not happen. There is no reason to convert the type using PMPI_Type_f2c, only to over- write it immediately afterwards. The other type_create_* functions did not convert newtype. The valgrind warnings: ==28441== Conditional jump or move depends on uninitialised value(s) ==28441== at 0x581B555: PMPI_Type_f2c (in [...]/lib/libmpi.so.0.0.0) ==28441== by 0x4E87AB7: MPI_TYPE_CREATE_F90_INTEGER (in [...]/lib/libmpi_mpifh.so.0.0.0) ==28441== by 0x400BA1: MAIN__ (in [...]) ==28441== by 0x400C46: main (in [...]) ==28441== ==28441== Conditional jump or move depends on uninitialised value(s) ==28441== at 0x581B563: PMPI_Type_f2c (in [...]/lib/libmpi.so.0.0.0) ==28441== by 0x4E87AB7: MPI_TYPE_CREATE_F90_INTEGER (in [...]/lib/libmpi_mpifh.so.0.0.0) ==28441== by 0x400BA1: MAIN__ (in [..]) ==28441== by 0x400C46: main (in [...]) ==28441== ==28441== Use of uninitialised value of size 8 ==28441== at 0x581B577: PMPI_Type_f2c (in [...]/lib/libmpi.so.0.0.0) ==28441== by 0x4E87AB7: MPI_TYPE_CREATE_F90_INTEGER (in [...]/lib/libmpi_mpifh.so.0.0.0) ==28441== by 0x400BA1: MAIN__ (in [...]) ==28441== by 0x400C46: main (in [...]) ==28441== Signed-off-by: Risto Toijala <risto.toijala@gmail.com> (cherry picked from commit f14a0f4fc981a488150ac7426683e94645f9fdf7) | 09 January 2019, 15:24:22 UTC |
| ef7ab9b | Sergey Oblomov | 07 December 2018, 12:03:30 UTC | OSHMEM: added missing API for get/put operations - added calls for datatypes int/uint/8/16/32/size/ptrdiff for shmem_g/get/iget/get_nbi/_p/put/iput/put_nbi Signed-off-by: Sergey Oblomov <sergeyo@mellanox.com> (cherry picked from commit cfa9150934481ff19cbd9f2454da6d40a290f08a) | 09 January 2019, 09:52:26 UTC |
| 0258ed5 | Howard Pritchard | 08 January 2019, 16:15:31 UTC | Merge pull request #6252 from ggouaillardet/topic/v4.0.x/pmix-v3.1 pmix/ext3x: fix support for external PMIx v3.1 | 08 January 2019, 16:15:31 UTC |
| b7fbdeb | Nathan Hjelm | 07 January 2019, 21:24:58 UTC | btl/vader: minor correction to match ompi coding style Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov> (cherry picked from commit edaf08bf6d6ed24573187376921fe67a449851b2) | 07 January 2019, 23:40:01 UTC |
| 1408426 | Nathan Hjelm | 07 January 2019, 20:00:48 UTC | btl/vader: don't try to set reachabilty in add_procs if not requested This commit fixes a bug where add_procs can incorrectly return an error when going through the dynamic add_procs path. This doesn't happen normally, only when pml/ob1 is not in use. References #6201 Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov> (cherry picked from commit 30b8336cb40e586e5d926b2b52cd78bf3751e5d3) | 07 January 2019, 21:20:33 UTC |
| 61108b6 | Gilles Gouaillardet | 08 January 2019, 01:09:10 UTC | pmix/ext3x: fix support for external PMIx v3.1 The PMIX_MODEX and PMIX_INFO_ARRAY macros were removed from the PMIx 3.1 standard. Open MPI does not really need them (they are only used to be reported as not supported), so smply #ifdef protect them to support an external PMIx v3.1 The change only need to be done in ext3x/ext3x.c. But since this file is automatically generated from pmix3x/pmix3x.c, we have to update the latter file. Refs. open-mpi/ompi#6247 Signed-off-by: Gilles Gouaillardet <gilles@rist.or.jp> (back-ported from commit open-mpi/ompi@950ba16aa16576582eb71941ec45996239c10a46) | 07 January 2019, 11:27:34 UTC |
| 66c8694 | Howard Pritchard | 03 January 2019, 16:34:35 UTC | Merge pull request #6235 from yosefe/topic/scoll-basic-fix-zero-size-collect-v4.0.x v4.0.x: oshmem/scoll: fix shmem_collect32/64 for zero-size length | 03 January 2019, 16:34:35 UTC |
| cf81394 | Howard Pritchard | 02 January 2019, 16:41:21 UTC | Merge pull request #6218 from jsquyres/pr/v4.0.x/fix-CID-1441826 v4.0.x: odls_base_default_fns.c: remove errant free | 02 January 2019, 16:41:21 UTC |
| d5f5e49 | Howard Pritchard | 02 January 2019, 16:40:04 UTC | Merge pull request #6232 from jsquyres/pr/v4.0.x/distribute-ompi-grequest-header v4.0.x: romio321: ensure to distribute ompi_grequestx.h | 02 January 2019, 16:40:04 UTC |
| ad4b333 | Yossi Itigin | 01 January 2019, 17:50:28 UTC | oshmem/scoll: fix shmem_collect32/64 for zero-size length Fixes scoll_basic failures with shmem_verifier, caused by recent changes in handling of zero-size collectives. - Check for zero-size length only for fixed size collect (shmem_fcollect), but not for variable-size collect (shmem_collect) - Add 'nlong_type' parameter to internal broadcast function, to indicate whether the 'nlong' parameter is valid on non-root PEs, since it's used by shmem_collect algorithm. Before this change, some components assumed it's true (scoll_mpi) while others assumed it's false (scoll_basic). - In scoll_basic, if nlong_type==false, do not exit if nlong==0, since this parameter may not be the same on all PEs. - In scoll_mpi, fallback to scoll_basic if nlong_type==false, since MPI requires the 'count' argument of MPI_Bcast to be valid on all ranks. (Picked from master 939162e) Signed-off-by: Yossi Itigin <yosefe@mellanox.com> | 02 January 2019, 10:15:01 UTC |
| 1a1a932 | Jeff Squyres | 27 December 2018, 23:39:47 UTC | romio321: ensure to distribute ompi_grequestx.h Refs https://github.com/open-mpi/ompi/issues/6227. Thanks to George Marselis for reporting. Signed-off-by: Jeff Squyres <jsquyres@cisco.com> (cherry picked from commit 62321be186dd7d3efcedc2e801f226f6660ea0c4) | 28 December 2018, 21:18:10 UTC |
| 20d231d | Jeff Squyres | 22 December 2018, 11:37:06 UTC | odls_base_default_fns.c: put the free() in the right place Fixes CID 1441826. Signed-off-by: Jeff Squyres <jsquyres@cisco.com> (cherry picked from commit f96c04244db6dbd7f3546d0d1d93e771933f4094) | 22 December 2018, 14:41:33 UTC |
| af7a7f5 | Howard Pritchard | 21 December 2018, 22:33:05 UTC | Merge pull request #6216 from abouteiller/export4x/overspawn v4.0.x: Correctly propagate the oversubscribe flag to the spawnees | 21 December 2018, 22:33:05 UTC |
| fd22597 | Howard Pritchard | 21 December 2018, 22:31:16 UTC | Merge pull request #6167 from ggouaillardet/topic/v4.0.x/btl_uct_fix_warning btl/uct: fix misc warnings | 21 December 2018, 22:31:16 UTC |
| 6331525 | Howard Pritchard | 21 December 2018, 22:30:42 UTC | Merge pull request #6204 from jsquyres/pr/v4.0.x/another-rank-file-fix v4.0.x: another rank file fix | 21 December 2018, 22:30:42 UTC |
| d9b0dad | Aurélien Bouteiller | 24 October 2018, 02:42:35 UTC | Correctly propagate the oversubscribe flag to the spawnees This is a cherry-pick of master (2820aef). The propagation is intended to resolve issue #6130 Signed-off-by: Aurélien Bouteiller <bouteill@icl.utk.edu> | 21 December 2018, 19:53:25 UTC |
| 4aa91e1 | Geoffrey Paulsen | 29 November 2018, 15:15:39 UTC | Return MPI1 function implementations to build list Adding the implementations of the functions that were removed from the MPI standard to the build list, regardless of the state of the OMPI_ENABLE_MPI1_COMPAT. According to the README, we want the OMPI_ENABLE_MPI1_COMPAT configure flag to control which MPI prototypes are exposed in mpi.h, NOT, which are built into the mpi library. Those will remain in the mpi library until a future major release (5.0?) NOTE: for the Fortran implementations, we instead define OMPI_OMIT_MPI1_COMPAT_DECLS to 0 instead of OMPI_ENABLE_MPI1_COMPAT to 1. I'm not sure why, but this seems to work correctly. Also changing the removed MPI_Errhandler_create implementation to use the non removed MPI_Comm_errhandler_function prototype (prototype remains unchanged from MPI_Comm_errhandler_fn) NOTE: This commit is *NOT* a cherry-pick from master, because on master, we are no longer building those symbols by default, but on v4.0.x we _ARE_ still building these symbols by default. This is because the v4.0.x branch is to remain backwards compatible with v3.0.x, while at the same time removing the "removed" symbols from mpi.h (unless the user configures with --enable-mpi1-compatibility) Signed-off-by: Geoffrey Paulsen <gpaulsen@us.ibm.com> | 20 December 2018, 18:22:04 UTC |
| 2d9c936 | Ralph Castain | 18 December 2018, 00:13:55 UTC | If job is fully described, there will be no ppn string to unpack Signed-off-by: Ralph Castain <rhc@open-mpi.org> (cherry picked from commit d7283807410c1d37d966180847debe1d43361e92) | 18 December 2018, 16:27:39 UTC |
| 4be4282 | Howard Pritchard | 17 December 2018, 20:22:19 UTC | Merge pull request #6128 from ggouaillardet/topic/v4.0.x/mpiext_short_path mpiext: keep paths short | 17 December 2018, 20:22:19 UTC |
| 71b83e8 | Howard Pritchard | 17 December 2018, 20:21:21 UTC | Merge pull request #6193 from kawashima-fj/pr/v4.0.x/fix-type-create-f90 v4.0.x: mpi/c: Fix MPI_TYPE_CREATE_F90_{REAL,COMPLEX} | 17 December 2018, 20:21:21 UTC |
