Having noted the existence of the wondrous Open MPI "statistics" tracker, I feel compelled to commit these changes one file at a time. This will, of course, provide me with wonderful statistics for the number of commits I have done, thus ensuring that those who watch such things become truly impressed by the magnitude of my contribution.
Of course, I will also do a commit for each time I correct a typo in my own software, and each time I add a comment to a file - a comment that, ordinarily, one might expect to have already been in place before the first commit. But then....I wouldn't look as impressive if I did it that way! No, no...far better to add the comments - and do a commit after each one - separately!
So, enjoy all.
Ralph
aka. The longtime Don Quixote crusader against the asinine use of meaningless statistics in place of true performance metrics.
This commit was SVN r2824.
compiled libraries and compononents. Still to do:
- check components
- send mail if it finds anything wrong
- check multiple installation directories (?)
This commit was SVN r2811.
* make hostfile llm properly deal with over subscribe situation. Rather
than returning smaller than requested (which is no longer possible as
it made for a book keeping nightmaer and no one was paying attention
to it anyway), we just over subscribe the nodes. In the future, we
need to add a flag to allocate resources as to whether to allow
over subscription (if the resource allocator permits - clearly rsh
does, rms not so much).
This commit was SVN r2808.
instance per process. Better maps to the multi-pcm design.
* Minor code cleanup in llm and pcm (comments, param checks, etc.)
This commit was SVN r2804.
sets running at once - requires an additional step in spawning to get a
handle (that will contain multiple pcms when we support multi-cell)
* change the selection logic of the pcms to not care about setting threads,
but instead to select based on the selected thread level, since it
would be a little late by the time we did the selection for pcms.
* started the long process of cleaning up the rsh pcm so that it
actually kills processes and things. Still doesn't do anything useful,
but getting to the point where that might be possible
This commit was SVN r2794.
- Make the unified back-end able to send output via fprintf(stderr,
...) (if we're before MPI_INIT or MPI_FINALIZED) and ompi_output(0,
...) if we're between MPI_INIT and MPI_FINALIZED
This commit was SVN r2785.
