colvars fixes

From: Giacomo Fiorin (giacomo.fiorin_at_temple.edu)
Date: Tue Mar 09 2010 - 19:36:08 CST

Hi all, we'd like to notify that three bugs in the collective variables
module (colvars) in NAMD 2.7b2 have been fixed. The updated source files
are in the CVS repository.

- The "angle" component was found to send forces of the opposite sign to its
atoms (namd-l messages on 11/11/2009). If you chose to temporarily fix the
problem by inverting the sign of the force constants, and are now planning
to use a CVS version of NAMD, beware that "angle" forces have now the
correct sign.

- The module was running "empty" (i.e. with no variables defined)
when called from a single NAMD configuration file containing multiple runs
(such as the windows of a FEP run). Now FEP and colvars can be used
together.

- Using atomNameResidueRange gave rise to an initialization error when AMBER
or GROMACS topologies (NAMD exited with the message "no atoms defined for
atom group"). Now atomNameResidueRange can be used also with non-PSF
topologies.

No known bugs are currently in the CVS repository. Thanks to all those who
reported the previous ones.

Support has also been added for:

- Continuing a metadynamics run with useGrids set to "off" into a new run
with useGrids set to "on" (i.e. switch from analytical to interpolated hills
while continuing the same run).

- Scaling the forceConstant of the "harmonic" bias over time.

Have a look to the updated Users Guide for more details.

Giacomo Fiorin

Jérôme Hénin

---- ----
Giacomo Fiorin
  ICMS - Institute for Computational Molecular Science
    Temple University
    1900 N 12 th Street, Philadelphia, PA 19122
work phone: (+1)-215-204-4216
mail: giacomo.fiorin_at_temple.edu
---- ----

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 05:22:48 CST