AMBER Archive (2003)

Subject: DelphiLinux V. 4 Segmentation Fault

From: Thomas Steinbrecher (thomas.steinbrecher_at_physchem.uni-freiburg.de)
Date: Wed Mar 19 2003 - 07:10:16 CST

Dear AMBER users,

I try to calculate the electrostatic contribution to the
solvation energy of a protein-ligand-complex with the
Delphi program (Version 4, release 1.1 for Linux-OS) as
part of a MM-PBSA-calculation similar to that of Case et.
al. 1998 in JACS. My complex consists of ~4000 atoms.

I know this question is unrelated to amber, but since many
people seem to use Delphi for MM-PBSA calculations in
Amber, maybe someone here can help me.

When I increase the SCALE parameter in the Delphi input to
4 per A (default is 2) to get a higher grid resolution, the
program fails with the (not very informative) message:

"Segmentation fault". (It runs fine with SCALE=2 or 3)

My delphi input file (generated by mm_pbsa.pl) is:

-----
bndcon=4
scale=4
exdi=80.0
indi=1.0
linit=1000
perfil=80.0
in(crg,file="./amber94_delphi.crg")
in(siz,file="./parse_delphi.siz")
in(pdb)
energy(s)
-----

Is it possible that my system is to big for such a fine
grid resolution?
Has anyone seen the same Delphi failure and/or possible
clues/solutions?

There was a similar question in the archives some time ago
(2001), but I could not find an answer to it.

Thank you in advance for any comments,

Thomas Steinbrecher