AMBER Archive (2002)

Subject: Re: Not able to do parallel run (again)

From: Stéphane Teletchéa (steletch_at_biomedicale.univ-paris5.fr)
Date: Mon Jan 07 2002 - 11:08:17 CST

I finally found why my MD run with explicit solvent crashed.
I am using sander_classic for the run and the last lines of my input files
are :
    NMRMAX=0, NTR=0, IEWALD=1, ISOLVP=654,
 &end
61.69234 52.08589 52.06907 90.0 90.0 90.0
64 54 54 4 0 0 0
0.000001

And with these input parameters, the run crashed after two steps (ie 0.1ps of
calculation) without any warning (only MPI errors which were not directly
MPI-related). It worked for one node but not for node>1.

The solution was to put sander format for the Ewald input :
    NMRMAX=0, NTR=0, ISOLVP=654,
 &end
&ewld
A=61.69234, B=52.08589, C=52.06907,
NFFT1=64.0, NFFT2=54.0, NFT3=54.0,
ORDER=4, RSUM_TOL=0.000001,
&end

In this case, the run starts and goes on.

I don't know why there is such a strange behaviour (may be for compatibility
reason), but i hope this mail will help sombody else.

Stef

HAPPY NEW YEAR ! 

-- 
*~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~*
Teletchéa Stéphane - CNRS UMR 8601
Lab. de chimie et biochimie pharmacologiques et toxicologiques
45 rue des Saints-Peres 75270 Paris cedex 06
tel : (33) - 1 42 86 20 86 - fax : (33) - 1 42 86 83 87
mel : steletch_at_biomedicale.univ-paris5.fr
*~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~*