dihedrals bug

breeze at zeneca-ph.co.uk breeze at zeneca-ph.co.uk
Mon Oct 2 06:38:20 EST 1995



Esteemed fellow XPLOR users,

I am encountering an annoying and persistent bug in our IBM RS/6000 3.1
executable, supplied to us by MSI. Every time I do this:

X-PLOR> print threshold=3. cdih

I get this:

  Number of dihedral angle constraints=   41
   overall scale =  200.0000
Segmentation fault

and Xplor dies. It doesn't seem to make any difference whether there are
actually any violations, nor how many restraints are applied. Issuing

X-PLOR> restraints dihedrals ? end

suffers a similar fate. But puzzlingly, the test file 'testcons.inp' runs
fine. I've also had no problems with our SGI executable.
	Does anyone have any idea what's behind this bug, or seen anything
similar (particularly those running on RS/6000s)? I'm running Xplor on
an RS/6000 model 590, but saw the same bug on a model 560. The problem
is an NMR SA refinement of a protein with ~1900 atoms. MSI themselves
were unable to reproduce the problem and could offer no suggestions.

Any information gratefully received,

Alex Breeze

Protein Structure Lab
Zeneca Pharmaceuticals
Mereside, Alderley Park
Cheshire SK10 4TG
UK
email: breeze at zeneca-ph.co.uk





More information about the X-plor mailing list