More DPM incompletes in parallel than in serial

Paul

Whenever I do a DPM tracking on my fluent model, I get significantly more (almost 10X more) incomplete particles if I do the DPM in parallel mode, but if I open up the same case and data file in single processor mode I get far less.

Any idea what I can do to take care of this? I'm on a multi-processor computer by the way, its not an ad-hoc parallel network.

I use the shared memory option in the DPM parallel tab.

Thanks.