crash

From: jp2854_at_columbia.edu
Date: Mon Nov 29 2010 - 11:18:24 CST

Hi

I am running an MDFF simulation to fit a crystal structure to a
(.brix) cryoEM density map in solvent. I'm trying to run the
simulation for 4000000 iterations.

The system runs fine for exactly 1728000 iterations and repeatedly
crashes after this cycle. There's no info in the log file; it simply
terminates. The message I get in the shell I launch from is "Charmrun:
error on request socket-- Socket closed before recv."

I am running the job on a 23 node cluster, 4 cpus per node. If I add
"Set INPUTNAME my_file_name" to my parameter file to enable a restart
of the system it runs again and repeatedly crashes at iteration
3456000 which is 2*1728000!

Any idea how to fix this!?

All the best

-Jesper

Jesper Pallesen, PhD
Postdoctoral Research Associate
Department of Biochemistry and Molecular Biophysics
HHMI/Columbia University
650 West 168th Street, P&S Black 2-221
New York, NY 10032
(T) 212 305-9521
(T) 917 587 3365
(F) 212 305-9500
(E) jp2854_at_columbia.edu

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:56:24 CST