Bug in T-REMD with colvars

From: Francesco Pietra (chiendarret_at_gmail.com)
Date: Tue Dec 03 2013 - 01:55:14 CST

Bug in T-REMD with colvars could hopefully simply mean bug in information
about T-REMD with colvars.

Some time ago I asked - on general grounds - advice on how to manage to get
colvars state be recognized by restarted replicas. I got no answer. Here, I
am more specific on this issue, working with namd2.10, although same issue
with namd2.9.

Setting in namd config file (alanin_based.namd in the alanin t-remd example
provided by namd)

# constraints
colvars on ; # provide colvarsConfig (and colvarsINput if
colvarsConfig ./P2i_rmsd-colvars.in

gives correct

in replica/0, like in all other 31 replicas of my T-REMD. However, what is
not provided anywhere is what the system asks (see below, please) when I
add a line for colvarsInput into file replica.namd as follows:

if {$i_run} { #restart
  bincoordinates $restart_root.$replica_id.coor
  binvelocities $restart_root.$replica_id.vel
  extendedSystem $restart_root.$replica_id.xsc
  colvarsInput $restart_root.$replica_id.colvars.state
} else {
  temperature $NEWTEMP

Thus, the error arises in the .err file:

REPLICA 23 FATAL ERROR: Error in the collective variables module: exiting.

Looking at the replica 23 log file, one learns that:

colvars: Collective variables biases initialized, 1 in
colvars: Error: in opening restart file

In this scenario, I wish that the line "colvarsInput
$restart_root.$replica_id.colvars.state" I added to replica.namd should be
reformulated, or placed elsewhere (in the namd config file?). That
statement was introduced on imitation of umbrella sampling.

I hope now that developers will be so kind to correct my faults or my
overlooking any T-REMD output, if any, to the benefit of all namd users
too. Or provide anyway a solution to carry out correctly T-REMD with

Thanks a lot for advice

francesco pietra

This archive was generated by hypermail 2.1.6 : Wed Dec 31 2014 - 23:21:57 CST