Dear all,
I am running a simulation using PBMETAD with Gromacs 4.6.7. Following is my PBMETAD options:
_________
wallcom: UPPER_WALLS ARG=cv4 AT=1.7 KAPPA=150.0
PBMETAD ...
ARG=cv1,cv2,cv3,cv5,cv6,cv7,cv8 SIGMA=0.1,0.1,0.1,0.1,1,1,1 HEIGHT=1.2 PACE=2500 LABEL=pb
FILE=HILLS_1,HILLS_2,HILLS_3,HILLS_5,HILLS_6,HILLS_7,HILLS_8
BIASFACTOR=8 MULTIPLE_WALKERS
GRID_MIN=0,0,0,0,0,0,0 GRID_MAX=3.142,8,1.1,3.142,40,20,20
INTERVAL_MIN=0,0,0,0,0.1,0.1,0.1 INTERVAL_MAX=3.1,8,1,3.1,38,18,19
... PBMETAD
PRINT ARG=cv1,cv2,cv3,cv4,cv5,cv6,cv7,cv8,wallcom.bias STRIDE=1000 FILE=COLVAR
_________
As I am using multiple walkers, I noticed that Plumed prints the value of all the replicas at any timestep into the same HILLS file, so there will be multiple lines for time 0 for example.
My question is, does sum_hills work correctly for these HILLS files? Or do I have to split my HILLS files into multiple files before running sum_hills?
Secondly, as I have not asked plumed to print the pbmetad bias into my COLVAR files, I will have to write a script to calculate the PBMETAD bias from my HILLS files. If my understanding is correct, the PBMETAD bias is just the sum of the exponential of the bias from all the HILLS files (with the Boltzmann factor, of course), right? And to reweight my trajectory I can simply multiple the conformations using the exponential of the PBMETAD bias according to the paper?
Thank you for viewing and I appreciate your help!