Thank you all for your suggestions.  Unfortunately I am still stuck with the segmentation fault problem.

 

The swap partition is mounted fine. 

Setting max_chains to a lower number does nothing.

There’s nothing wrong with my selection criteria (0 atoms were selected as rigid groups, 36 for annealing)

 

I tried systematically deleting parts of the pdb input file, then rerunning generate.inp and then anneal.inp.  That worked fine for my ligand alone (as I understand it, given the right toppar inputs, CNS should be able to run torsion dynamics on my ligand – am I right?), but the protein component caused failure again.  I can’t spot a problem by looking so I’m at a bit of a loss.

 

Any more ideas?

 

Sincerely

Jean

 

 

Hi all

 

Apologies for the non-ccp4 question – hopefully some of you who use CNS can help me - I have a problem with anneal.inp in CNS:

 

When I run the anneal script for my structure containing novel ligand, the program fails before the start of the torsion dynamics with no error message - except SEGMENTATION FAULT at the bash commandline.

 

With “quiet” output, the last lines in the output file are:

 

 ASSFIL: file /usr/local/CNS/cns_solve_1.1/libraries/toppar/torsionmdmods opened.

 MESSage=NORM

 EVALUATE: symbol $MESSAGE_OLD_TMOD set to "NORM" (string)

 ECHO=FALSe {OFF}

 EVALUATE: symbol $ECHO_OLD_TMOD set to FALSE (logical)

 NEXTCD: condition evaluated as false

 Program version= 1.1 File version= 1.1

 SELRPN:      0 atoms have been selected out of   4883

 

so the torsionmods routine (and everything before it) ran fine but the torsion dynamics never started.

 

My pdb, mtf, top and par files all work fine for minimization so I don't think there's anything wrong with them.

 

I have been using CNSv1.1 on this Linux machine for some time so it's not a compile problem.

 

I can run the same routine on the same PC, from the same script, with the same protein but with a different ligand and there's no fault.

 

Can anyone suggest a solution?  I've tried just about everything short of running the whole script manually from the command line!

 

Sincerely

Jean Watermeyer