On Mar 20, 2009, at 11:06 AM, Eugene Loh wrote:
> > I'm still seeing a very low incidence of the sm segv during
> startup (.
> > 01% -- 23 tests out of ~160k), so let's ship 1.3.1 and roll in
> > Eugene's new sm code for 1.3.2.
> I wanted to join in the fun, but... no go. I'm running an
> job on a single node with np=8. So far about 40K runs with no
> failures. Am I missing a special ingredient?
I wish I knew what it was. :-(
The 160k runs are all my MTT runs. I run a large variety of different
configurations with different compilers, mpirun options, etc.
Although when I poked into this last week, I couldn't find any obvious
pattern as to what exactly was causing the failure.