AMBER Archive (2004)

Subject: Re: AMBER: segmentation fault of minimization

From: Thomas E. Cheatham, III (cheatham_at_chpc.utah.edu)
Date: Mon Mar 29 2004 - 21:42:01 CST

> parameter (MAX_RSTACK=10000000)
...
> and recompiled it with g77. But I still always get segmentation fault
> when I run this minimization. I think the memory should be ok, right?

You could keep trying to increase MAX_RSTACK, however very telling is
your initial energy from the minimization:

   NSTEP ENERGY RMS GMAX NAME NUMBER
      1 3.1454E+20 4.1536E+20 5.5746E+22 N 1239
 BOND = 7024.1947 ANGLE = 954.1097 DIHED = 560.6102
 VDWAALS = ************* EEL = -54320.3184 HBOND = 0.0000
 1-4 VDW = ************* 1-4 EEL = 6912.4118 CONSTRAINT = 0.0000

Likely you have overlapping atoms as the initial energy and gradient is
extremely large. It is somewhat surprising that you get 1 step and then
an RSTACK overflow, but perhaps you have a number of atoms converging to
common positions and this leads to the overflow...

The program ptraj that comes with AMBER has a command to check overlap
among the atoms. If you are using an earlier version of AMBER
(likely) you can get a newer version at
        www.chpc.utah.edu/~cheatham/software.html

To look for overlap

ptraj prmtop << EOF
trajin inpcrd
checkoverlap
EOF

where prmtop is your prmtop file and inpcrd is you initial coordinates.

Also, rdparm prmtop allows you to inspect bonds/angles/dihedrals
(printbonds) and types (printlj, printtypes) to see if things are OK in
your prmtop.

Let the list know of further troubles,

--tom

-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu