<div class="gmail_quote">On Mon, Jul 23, 2012 at 10:33 PM, Barry Smith <span dir="ltr"><<a href="mailto:bsmith@mcs.anl.gov" target="_blank">bsmith@mcs.anl.gov</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
So maybe that is the problem. PETSc is forcing affinity to a hardware number core regardless of whether another PETSc program is running and has also forced that same core?</blockquote></div><br><div>Yes. Now what's the right solution? Not set any affinity by default? Inherit the process affinity and subdivide it among threads? There is currently a somewhat crude mechanism in -threadcomm_affinities.</div>