[Openmd-users] A quick poll of OpenMD users
Dan Gezelter
gezelter at nd.edu
Tue Oct 5 17:22:29 EDT 2010
Dear OpenMD users,
We're in the process of rewriting some of the low-level treatment of the non-bonded interactions. The default behavior in OpenMD is currently quite confusing (particularly in regards to how cutoffs are handled) and I'd like to simplify it in the next version. There are some choices to be made, and I'd like to get user input on if this scheme is reasonable. Basically, we need to know what to do when the user doesn't specify a methodology, or specifies incomplete information.
The proposed scheme:
cutoffRadius :
if it was explicitly set, use that value (but do a sanity check)
if it was not explicitly set:
Are there electrostatic atoms? Use 12.0 Angstroms.
No electrostatic atoms? Poll the atom types present in the simulation for suggested cutoff values (e.g. 2.5 * sigma). Use the maximum suggested value that was found.
switchingRadius :
if it was explicitly set, use that value (but do a sanity check)
if it was not explicitly set, use 0.85 * rCut
cutoffMethod : (one of HARD, SWITCHED, SHIFTED_FORCE, SHIFTED_POTENTIAL)
if it was explicitly set, use that choice.
if it was not explicitly set, use SHIFTED_FORCE
electrostaticCutoffMethod : (one of HARD, SWITCHED, SHIFTED_FORCE, SHIFTED_POTENTIAL, REACTION_FIELD, possibly others as we work on this)
if it was explicitly set, use that choice (but do a sanity check)
if it was not explicitly set, default to "cutoffMethod"
electrostaticScreeningMethod : (one of DAMPED, UNDAMPED)
if was explicitly set, use that choice (but do a sanity check)
if it was not explicitly set, choose DAMPED
dampingAlpha : (required for DAMPED)
if it was explicitly set, use that choice,
if not, use 0.425 - cutoffRadius * 0.02, with a minimum value of 0.0.
dielectric : (required for REACTION_FIELD)
if it was explicitly set, use that choice,
if not use 80.0
Any comments before we implement this scheme? It is somewhat different from what the code does now, as we'll be adding the possibility of using separate cutoff methodologies for electrostatics and other interactions.
--Dan
***********************************************
J. Daniel Gezelter
Associate Professor of Chemistry
Director of Graduate Admissions
Department of Chemistry and Biochemistry
251 Nieuwland Science Hall
University of Notre Dame
Notre Dame, IN 46556-5670
phone: +1 (574) 631-7595
fax: +1 (574) 631-6652
e-mail: gezelter at nd.edu
web: http://www.nd.edu/~gezelter
************************************************
More information about the Openmd-users
mailing list