This problem has been solved (but not necessarily understood).
I was unintentionally using prmtop files that didn't have solvent
pointers, atoms per molecule, or box info. I was able to get the
calculations to run using a fix by tom (tec3):
Edit lines 522 and 2241 of runmd.f (amber11) to read
"if (ifsc == 1 .and. master)" instead of "if (ifsc == 1)"
However, all of this is avoided when using the proper prmtop.
In that case, no errors occurred when using the original amber11 code.
I'm still a little confused about why this runs with the bad prmtops on
two processors but not more ... but overall the problem is resolved.
Thanks,
--Niel
________________________________________
From: steinbrt.rci.rutgers.edu [steinbrt.rci.rutgers.edu]
Sent: Tuesday, February 08, 2011 1:31 AM
To: AMBER Mailing List
Subject: Re: [AMBER] softcore TI with more than 2 procs?
Hi,
the code is certainly tested with more than two processors and should work
fine. You may be using some combination of settings that uncovers an
otherwise unseen bug. Can you run the actual files of tutorial A9 with
more than 2 CPUs? If yes it just comes down to checking what your
simulation does different than the tutorial.
Thomas
Dr. Thomas Steinbrecher
formerly at the
BioMaps Institute
Rutgers University
610 Taylor Rd.
Piscataway, NJ 08854
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Tue Feb 08 2011 - 16:00:03 PST
