Guys -- any thoughts on this hg tree / the concept? If I don't hear
anything back within the next day or so, I'm inclined to put this back
into the trunk (I could use this stuff in my MTT runs :-) ).
On Oct 31, 2008, at 1:55 PM, Jeff Squyres (jsquyres) wrote:
> On Oct 31, 2008, at 1:42 PM, Ethan Mallove wrote:
> >> http://www.open-mpi.org/hg/hgwebdir.cgi/jsquyres/mtt-fast-scratch/
> >> What do you guys think?
> > I just ran this, and it seemed to work fine:
> > $ client/mtt --fast-scratch /tmp/mtt-jeff-fast-scratch \
> > --scratch /workspace/mtt-jeff-slow-scratch
> Tim -- what do you guys think? (of both the code patch and the
> concept -- it's a little different than your original idea)
> Also, Ethan and I have put back a few more things into the hg tree,
> mostly dealing with whacky cwd() issues. Not directly related to your
> original patch, but it came up when Ethan and I were discussing the --
> fast-scratch stuff this morning.