[ESPResSo] Bonded Gay-Berne particles

[Michael Winokur]

Hi All,

I've made some progress in adding new local interactions between GB particles.  
For the moment I've made an adjustment to the dihedral function calls in 
energy.h and forces.h (but it will be better to have a fully separate 
interaction parameter).  At issue is the use of torques in the bonded 
interaction function.  The Gay-Berne potential, up to now, only adds torques 
in the non-bonded interaction function.  As a debugging step I now just add a 
torque of zero to two particles.  
See below:
------------------------------------------------------
 #ifdef ROTATION
     for (j = 0; j < 3; j++) {
            p1->f.torque[j] += 0.00000; 
            p3->f.torque[j] += 0.00000; 
     }
#endif
------------------------------------------------------
and then Espresso breaks somewhere along the line because 
one of the p3 particle coordinates, p3->r.p[2], suddenly becomes "nan".
This doesn't happen straightaway but only after a modest number of iterations.  
Could someone kindly suggest how I should go about troubleshooting this?

Thanks,

Michael


-- 
------------------------------------------------------------------------------------------------------
Michael Winokur
Dept. of Physics, University of Wisconsin
1150 University Ave.
Madison, WI 53706
Phone: 608 263 7475 URL: romano.physics.wisc.edu
-----------------------------------------------------------------------------------------------------