openmpi

comm

Per RFC:

2014-06-01 16:14:10 +00:00

dash_host

orte/util: fix a misc memory leak

2015-06-17 11:17:54 +09:00

hostfile

initialize common symbols from orte

2015-05-08 10:11:58 +09:00

attr.c

Fix bad test - opal_buffer and opal_ptr can support NULL locations

2015-02-17 21:46:23 -08:00

attr.h

Consolidate all the QOS changes into one clean commit

2015-05-06 19:48:42 -07:00

context_fns.c

First part of memory leak cleanups from Gilles

2014-11-24 16:53:33 -08:00

context_fns.h

... Delayed due to notifier commits earlier this day ...

2009-04-29 01:32:14 +00:00

error_strings.c

Rework the OOB selection logic to allow a component (e.g., usock) to direct that it be the sole active component. Remove prior disqualifying code in the oob/tcp component as it was too restrictive - if usock wasn't able to run, it left apps with no way to communicate to their daemon. Have the local daemon check the global modex for the RML URI info of the local procs so it can route messages between them when tcp is the primary channel.

2015-05-08 11:15:21 -07:00

error_strings.h

When we comm_spawn, we really want to respect the original -host directives and not expand the daemon virtual machine unless directed to do so in the comm_spawn command. Otherwise, we will automatically launch daemons on every node in the allocation.

2014-04-30 22:26:18 +00:00

help-regex.txt

check-help-strings cleanup

2014-08-11 03:25:22 +00:00

hnp_contact.c

orte/util: fix memory leaks

2015-03-05 14:06:18 +09:00

hnp_contact.h

As requested by Aurelien at the July design meeting - long time coming, but finally got around to it.

2008-12-10 17:10:39 +00:00

listener.c

Fat SMPs (i.e., systems with nodes containing large numbers of cpus) were failing to start due to connection failures of the opal/pmix support. Root cause was that (a) we were setting the client socket to non-blocking before calling connect, and (b) the server was using the event library to harvest the accepts, and also did the handshake while in that event. So the server would backup beyond the connection backlog limit, and we would fail.

2015-05-29 14:37:14 -07:00

listener.h

Fat SMPs (i.e., systems with nodes containing large numbers of cpus) were failing to start due to connection failures of the opal/pmix support. Root cause was that (a) we were setting the client socket to non-blocking before calling connect, and (b) the server was using the event library to harvest the accepts, and also did the handshake while in that event. So the server would backup beyond the connection backlog limit, and we would fail.

2015-05-29 14:37:14 -07:00

Makefile.am

Fat SMPs (i.e., systems with nodes containing large numbers of cpus) were failing to start due to connection failures of the opal/pmix support. Root cause was that (a) we were setting the client socket to non-blocking before calling connect, and (b) the server was using the event library to harvest the accepts, and also did the handshake while in that event. So the server would backup beyond the connection backlog limit, and we would fail.

2015-05-29 14:37:14 -07:00

name_fns.c

orte/util: fix misc memory leaks

2015-02-17 12:27:23 +09:00

name_fns.h

Per the PR and discussion on today's telecon, extend the process name definition as a two-field struct of uint32_t's down to the OPAL layer. This resolves issues created by prior commits that impacted both heterogeneous and SPARC support. This also simplifies the OMPI code base by removing the need for frequent memcpy's when transitioning between the OMPI/ORTE layers and OPAL.

2014-11-11 17:00:42 -08:00

nidmap.c

Per the PR and discussion on today's telecon, extend the process name definition as a two-field struct of uint32_t's down to the OPAL layer. This resolves issues created by prior commits that impacted both heterogeneous and SPARC support. This also simplifies the OMPI code base by removing the need for frequent memcpy's when transitioning between the OMPI/ORTE layers and OPAL.

2014-11-11 17:00:42 -08:00

nidmap.h

Per the PMIx RFC:

2014-08-21 18:56:47 +00:00

parse_options.c

orte/util: fix misc memory leaks

2015-02-17 12:27:23 +09:00

parse_options.h

Use ports as multicast channels instead of networks so we avoid stepping into reserved spaces.

2011-04-29 18:46:40 +00:00

pre_condition_transports.c

orte/util: fix misc memory leaks

2015-02-17 12:27:23 +09:00

pre_condition_transports.h

In the case of direct-launched processes running under slurm, psm requires that the pre_condition_transports MCA param be set. This is normally computed by mpirun and inserted into each proc's environ, but that doesn't work here.

2011-04-28 13:54:33 +00:00

proc_info.c

Include the FQDN version and non-stripped version of the hostname in our list of aliases as these (plus localhost) are the most common aliases we see.

2015-03-17 06:26:26 -07:00

proc_info.h

Attempt to reduce the RARP traffic during definition of allocations

2015-03-16 16:26:40 -07:00

regex.c

Update the regex to resolve a bug

2014-08-29 22:24:20 +00:00

regex.h

Extend the regular expression parsing support

2014-03-17 21:25:05 +00:00

session_dir.c

For scalability reasons, and to make life easier for the poor Cray-ites, don't bang on the system for the username - we'll just use the uid.

2015-03-19 21:24:13 -07:00

session_dir.h

- Revert r20739

2009-03-05 21:56:03 +00:00

show_help.c

Continue refinement of the DVM operations. Send the spawn request to the right place (it helps) as it isn't a comm_spawn request and has to be treated a little differently. Ensure IO gets forwarded back to the tool. Ensure the tool outputs show_help locally as there is no place to send it.

2015-02-04 06:21:54 -08:00

show_help.h

Remove the old configure option for disabling full rte support - we now use the OMPI rte framework for such purposes

2013-02-28 01:35:55 +00:00