openmpi

Jeff Squyres ffddfc5629 Turns out that it's a really Bad Idea(tm) to tm_spawn() and then not

keep the resulting tm_event_t that is generated because the back-end
TM library actually caches a bunch of stuff on it for internal
processing, and doesn't let go of it until tm_poll().

tm_event_t's are similar to (but slightly different than)
MPI_Requests: you can't do a million MPI_Isend()'s on a single
MPI_Request -- a) you need an array of MPI_Request's to fill and b)
you need to keep them around until all the requests have completed.

This commit was SVN r10820.

2006-07-14 22:04:41 +00:00

class

Next step in the project split, mainly source code re-arranging

2006-02-12 01:33:29 +00:00

dss

Fix a bunch of warnings the Sun compilers find:

2006-04-20 15:35:58 +00:00

etc

Next step in the project split, mainly source code re-arranging

2006-02-12 01:33:29 +00:00

include

Per LANL's stated need, add functionality that runs a.out across ALL available process slots if no num_proc is specified on the command line. However, please note the following limitation: we ONLY allow ONE application to be specified on the command line when this feature is invoked. If multiple apps are specified, the user MUST also specify the number to be launched for each and every one of them.