1
1
Граф коммитов

25930 Коммитов

Автор SHA1 Сообщение Дата
Gilles Gouaillardet
184d53a018 oshmem: swap fields of oshmem_proc_data_t to prevent padding
previously, the definition was

struct oshmem_proc_data_t {
    int num_transports;
    char * transport_ids;
};

so in 64 bits arch, the compiler would very likely insert a 4 bytes
padding before the two fields in order to have transport_ids aligned
2016-09-01 14:20:14 +09:00
Gilles Gouaillardet
0a25420dac oshmem: get rid of oshmem_proc_t and use ompi_proc_t instead
store oshmem related per proc data in an oshmem_proc_data_t struct,
that is stored in the padding section of an ompi_proc_t

this data can be accessed via the OSHMEM_PROC_DATA(proc) macro

Fixes open-mpi/ompi#2023
2016-09-01 14:20:14 +09:00
Gilles Gouaillardet
0b8c58298d oob/usock: fix handling of orte_process_name_t *
orte_process_name_t is aligned on 32 bits, so it cannot simply be casted
into an int64_t. use memcpy() instead

Thanks Paul Hargrove for the report
2016-09-01 13:18:02 +09:00
Gilles Gouaillardet
75b7ef97a0 coll/libnbc: fix nbc_ireduce when sendbuf == recvbuf
if sendbuf is equal to recvbuf, that should not be interpreted
as equivalent to MPI_IN_PLACE on the non root rank(s)

Thanks Valentin Petrov for the report
2016-09-01 10:19:05 +09:00
Gilles Gouaillardet
2969235324 libnbc: fix NBC_Copy for predefined datatypes
predefined datatypes such as MPI_LONG_DOUBLE_INT are not really contiguous,
so use span as returned by opal_datatype_span() instead of type extent,
otherwise data might be written above allocated memory.

Thanks Valentin Petrov for the report
2016-09-01 10:18:57 +09:00
Jeff Squyres
16fe18eb7c Merge pull request #2036 from edgargabriel/pr/datatype-refcount-fix
io/ompio: fix the reference count of basic datatypes used as etypes o…
2016-08-31 19:05:49 -04:00
Edgar Gabriel
be183cb3dd io/ompio: fix the reference count of basic datatypes used as etypes or ftypes. 2016-08-31 14:08:26 -05:00
rhc54
39d086e000 Merge pull request #2035 from rhc54/topic/memprofile
Provide a mechanism for obtaining memory profiles of daemons and application profiles for use in studying our memory footprint
2016-08-31 14:06:48 -05:00
Ralph Castain
39992d1ad7 Silence trivial Coverity warnings 2016-08-31 09:42:33 -07:00
Ralph Castain
c1050bc01e Provide a mechanism for obtaining memory profiles of daemons and application profiles for use in studying our memory footprint. Setting OMPI_MEMPROFILE=N causes mpirun to set a timer for N seconds. When the timer fires, mpirun will query each daemon in the job to report its own memory usage plus the average memory usage of its child processes. The Proportional Set Size (PSS) is used for this purpose. 2016-08-31 09:32:07 -07:00
Jeff Squyres
ead9b6389a README: update for new mailman and main web sites
Signed-off-by: Jeff Squyres <jsquyres@cisco.com>
2016-08-31 09:55:36 -04:00
Jeff Squyres
c33aaa5604 Merge pull request #1997 from jsquyres/pr/make-cma-configury-better
opal_check_cma: make consistent with rest of configury
2016-08-31 09:50:24 -04:00
Jeff Squyres
ba0ed2401a Merge pull request #2031 from jsquyres/pr/fix-fortran-runpath-detection
opal_setup_wrappers.m4: fix typo in Fortran rpath detection
2016-08-31 09:28:32 -04:00
rhc54
ed5846038b Merge pull request #2033 from rhc54/topic/state
Ensure that the "running" state is correctly updated
2016-08-31 01:50:38 -05:00
Ralph Castain
9b991bd1f5 Ensure that the "running" state is correctly updated
It is possible that one or more procs could get thru PMIx_Init, and thus be marked as in state "registered", before all local procs have been started. If that happens, then we would report some of the procs in state "running", and the others in state "registered" - which means that the HNP would miss the "running" stage of the state machine.

Thanks to Jingchao Zhang for his patience in tracking this down on the 2.0 branch
2016-08-30 19:24:39 -07:00
Jeff Squyres
1b9e165a4c opal_setup_wrappers.m4: fix typo in Fortran rpath detection
Add missing space in Libtool command for runpath detection.

Signed-off-by: Jeff Squyres <jsquyres@cisco.com>
2016-08-30 16:48:05 -07:00
rhc54
bfe0327f7b Merge pull request #2029 from rhc54/topic/notify
Since we changed storage to pointers in pmix_value_t, we need to allocate space for those values when unpacking
2016-08-29 23:11:23 -05:00
Ralph Castain
cfa784c9a6 Since we changed storage to pointers in pmix_value_t, we need to allocate space for those values when unpacking 2016-08-29 20:22:24 -07:00
Nathan Hjelm
99b26644c1 Merge pull request #2011 from hjelmn/osc_pt2pt_fix
osc/pt2pt: fix possible race in peer locking
2016-08-29 09:17:36 -06:00
Josh Hursey
b0d8638824 Merge pull request #2015 from jjhursey/topic/mixed-hostnames
orte: Expand use of !orte_keep_fqdn_hostnames MCA parameter
2016-08-29 09:14:54 -05:00
George Bosilca
a6d515ba9e Fixes opal_atomic_ll_64. Thanks to Paul Hardgrove for
the report and his patch.

This is an addition to #1140 and should go in 2.x
2016-08-27 12:43:48 -04:00
Nathan Hjelm
d33204b0dc Merge pull request #2021 from hjelmn/xlc_fix
opal/patcher: fix xlc support
2016-08-26 18:15:41 -06:00
rhc54
b90a64e734 Merge pull request #2022 from rhc54/topic/nnodes
Provide the number of nodes in the job
2016-08-26 18:15:24 -05:00
Ralph Castain
2f6e0fec90 Provide the number of nodes in the job 2016-08-26 14:50:41 -07:00
Joshua Hursey
d26dd2c20e orte: Expand the application of !orte_keep_fqdn_hostnames
* Expand the use of the `orte_keep_fqdn_hostnames` MCA parameter when
   it is set to false.
 * If that parameter is set to false (default) then short hostnames
   (e.g., `node01`) will match with the long hostnames (e.g.,
   `node01.mycluster.org`). This allows a user (or resource manager)
    to mix the use of short and long hostnames.
  - Note that this mechanism does _not_ perform a DNS lookup, but
    instead strips off the FQDN by truncating the hostname string at
    the first `.` character (when not an IP address).
     - By default (`false`) the following is true:
       `node01 == node01.mycluster.org == node01.bogus.com`
       since we use `node01` as the hostname.
2016-08-26 16:09:04 -05:00
Jeff Squyres
09ad7e81eb Merge pull request #2007 from jsquyres/pr/usnic-show-local-udp-ports
usnic: show the local UDP ports
2016-08-26 17:03:16 -04:00
Nathan Hjelm
a9bc692d99 opal/patcher: fix xlc support
The xlc compiler seems to behave in a different way that gcc when it
comes the inline asm. There were two problems with the code with xlc:

 - The TOC read in mca_patcher_base_patch_hook used the syntax
   register unsigned long toc asm("r2") to read $r2 (the TOC
   pointer). With gcc this seems to behave as expected but with xlc
   the result in toc is not the same as $r2. I updated the code to use
   asm volatile ("std 2, %0" : "=m" (toc)) to load the TOC pointer.

 - The OPAL_PATCHER_BEGIN macro is meant to be the first thing in a
   hook. On PPC64 it loads the correct TOC pointer (thanks to
   mca_patcher_base_patch_hook) and saves the old one. The
   OPAL_PATCHER_END macro restores the TOC pointer. Because we *need*
   the TOC to be correct before it is accessed in the hook the
   OPAL_PATCHER_BEGIN macro MUST come first. We did this and all was
   well with gcc. With xlc on the other hand there was a TOC access
   before the assembly inserted by OPAL_PATCHER_BEGIN. To fix this
   quickly I broke each hook into a pair of function with the
   OPAL_PATCHER_* macros on the top level functions. This works around
   the issue but is not a clean way to fix this. In the future we
   should 1) either update overwrite to not need this, or 2) figure
   out why xlc is not inserting the asm before the first TOC read.

This fixes open-mpi/ompi#1854

Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
2016-08-26 14:43:03 -06:00
Jeff Squyres
87a5ccc060 usnic: show the local UDP ports
Signed-off-by: Jeff Squyres <jsquyres@cisco.com>
2016-08-26 12:25:18 -07:00
Jeff Squyres
e03a40a0e9 pmix3x: remove generated file
Signed-off-by: Jeff Squyres <jsquyres@cisco.com>
2016-08-26 10:30:47 -07:00
rhc54
03838f275a Merge pull request #2019 from artpol84/fix_schizo
orte/schizo: fix binding detection in slurm component
2016-08-26 09:08:43 -05:00
Edgar Gabriel
b5c757e82c Merge pull request #2014 from edgargabriel/topic/mt-io
Topic/mt io
2016-08-26 08:54:45 -05:00
Jeff Squyres
9ae51a09f2 Merge pull request #1989 from jsquyres/pr/update-usnic-to-libfabric-v1.4
Update usnic BTL to libfabric v1.4
2016-08-26 09:53:07 -04:00
Artem Polyakov
55ac3b0be3 orte/schizo: fix binding detection in slurm component
in SLURM 16.05 the SLURM_CPU_BIND_TYPE is equal to "mask_cpu:"
instead of "mask_cpu". Account for that.
2016-08-26 09:55:52 +03:00
Gilles Gouaillardet
e4bf915e75 pmix3x: remove auto-generated file
remove opal/mca/pmix/pmix3x/pmix/src/include/pmix_config.h.in
.gitignore is correct, so it seems this file was added before .gitignore was updated
2016-08-26 15:00:18 +09:00
rhc54
c0fff60e59 Merge pull request #2017 from rhc54/topic/pmixconfig
Update configury to support multiple PMIx versions
2016-08-25 21:36:34 -05:00
Ralph Castain
af67f16422 Update configury to support multiple PMIx versions, rename pmix2x component to pmix3x for support of PMIx master
Update support for external v1.1.x and v2.x libraries. Minor corrections to the v3.x component
2016-08-25 18:19:05 -07:00
Gilles Gouaillardet
277c319389 opal/util: fix (again and again) incorrect type casting in opal_path_df
and silence CID 1371767

this fixes previous commits :
 - open-mpi/ompi@2eec8970ff
 - open-mpi/ompi@a439afce5b
2016-08-26 09:42:45 +09:00
Nathan Hjelm
89c2f4974c Merge pull request #2016 from hjelmn/wait_sync
opal/wait_sync: add #if protection on header
2016-08-25 15:13:09 -07:00
Nathan Hjelm
f3d4eaeaf7 Merge pull request #2013 from hjelmn/osc_rdma_fix
osc/rdma: fix bug in dynamic memory window tracking code
2016-08-25 13:42:27 -07:00
Nathan Hjelm
de32c779e2 opal/wait_sync: add #if protection on header
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
2016-08-25 14:31:52 -06:00
rhc54
19b0f4db9f Merge pull request #1995 from rhc54/topic/pe-per-rank
Change the behavior of cpus-per-rank.
2016-08-25 14:38:12 -05:00
Edgar Gabriel
1ba03d38ec io/ompio: protect remaining functions in multi-threaded scenarios
protect the remaining functions where necessary by a mutex lock
to avoid problems in multi-threaded executions. Some functions
do not require that in my opinion, and I provided an explanation
in those cases.
2016-08-25 13:45:51 -05:00
Nathan Hjelm
e53de7ecbe osc/rdma: fix bug in dynamic memory window tracking code
This commit fixes an ordering bug in the code that keeps track of all
attached memory windows. The code is intended to keep the memory
regions sorted but was often inserting at the wrong index. Thanks to
Christoph Niethammer for reporting the issue. The reproducer will be
added to nightly MTT testing.

Fixes open-mpi/ompi#2012

Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
2016-08-25 12:08:46 -06:00
Nathan Hjelm
7af138f83b osc/pt2pt: fix possible race in peer locking
It is possible for another thread to process a lock ack before the
peer is set as locked. In this case either setting the locked or the
eager active flag might clobber the other thread. To address this the
flags have been made volatile and are set atomically. Since there is
no a opal_atomic_or or opal_atomic_and function just use cmpset for
now.

Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
2016-08-25 09:28:25 -06:00
Nathan Hjelm
c082068953 Merge pull request #2006 from hjelmn/osc_pt2pt_fix
osc/pt2pt: fix several bugs
2016-08-25 09:19:29 -06:00
rhc54
17a210f7f0 Merge pull request #2008 from rhc54/topic/binding
Correct the binding algorithm to decouple it from oversubscribe.
2016-08-25 09:25:33 -05:00
Edgar Gabriel
1cee83cc1b use the common/ interfaces in file_preallocate instead of the io_ompio_ interfaces.
Necessar for avoiding potential deadlock situations in multi-threaded scenarios.
2016-08-25 08:55:12 -05:00
Jeff Squyres
0d19cc4a13 README: fix a bunch of typos
Thanks to Paul Hargrove for pointing them out.  Really.

Signed-off-by: Jeff Squyres <jsquyres@cisco.com>
2016-08-25 09:15:27 -04:00
Jeff Squyres
f56b16f079 usnic: remove unused variable
Signed-off-by: Jeff Squyres <jsquyres@cisco.com>
2016-08-25 03:53:18 -07:00
Jeff Squyres
9717bcb7e6 btl/usnic: remove stale comment
Signed-off-by: Jeff Squyres <jsquyres@cisco.com>
2016-08-25 03:53:18 -07:00