height of parallel bias metadybamics

[Qinghua Liao]

Hello,

I am running a parallel bias metadynamics over four CVs using CP2K and PLUMED v2.8:

I found that the initial height in each of the four HILLS was just 1/4 of the one set in the PLUMED input, is this correct?

Could someone confirm on it? This is my first run with PBMETAD. Thanks a lot!

All the best,

Qinghua

[Jim Pfaendtner]

Hi Qinghua,

This is correct and will depend on the total number of individual bias potentials used in your PB simulation. Please see EQ7-8 and associated discussion in the original paper. 

Jim Pfaendtner

--
You received this message because you are subscribed to the Google Groups "PLUMED users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to plumed-users...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/734b8370-4624-4050-bcdb-ab3f3be9af13n%40googlegroups.com.

[Qinghua Liao]

Hello Jim,

Thanks a lot for your confirmation. Glad that it is the same as the implementation.

I checked the EQ 7-8 of the original paper, but I did not see it talked about the heights of the Gaussians.

Additionally, I have another question, about the Gaussians. Here is my input:

 pb: PBMETAD ARG=diff1,diff2 SIGMA=0.2,0.2 HEIGHT=1.0 PACE=100 TEMP=310 BIASFACTOR=25 GRID_MIN=-4.0,-3.0 GRID_MAX=3.0,3.0 GRID_SPACING=0.04,0.04 FILE=HILLS_diff1,HILLS_diff2

And here are the hills files:

HILLS_diff1

#! FIELDS time diff1 sigma_diff1 height biasf
#! SET multivariate false
#! SET kerneltype stretched-gaussian
    0.04999998387019591     -3.321640053642728                    0.2     0.5208333333333334                     25
    0.09999996774039183     -3.555316179145291                    0.2      0.568729814403156                     25
   ..........
     0.8499997257933305     -3.523509393461763                    0.2      0.255317029768966                     25
     0.8999997096635264     -3.572406096586644                    0.2     0.2962299297846483                     25

HILLS_diff2

#! FIELDS time diff2 sigma_diff2 height biasf
#! SET multivariate false
#! SET kerneltype stretched-gaussian
    0.04999998387019591     -1.372269729574045                    0.2     0.5208333333333334                     25
    0.09999996774039183     -1.513460778984697                    0.2     0.4510857033796234                     25
     .........
     0.8499997257933305     -1.467238822082224                    0.2     0.5381668232184041                     25
     0.8999997096635264     -1.403935378638358                    0.2     0.5021401192377797                     25

I guess the key of parallel bias metadynamics is not adding Gaussians of multiple monodimensional potentials at the same time.

But according to the hills files from my PBMETAD simulation, it seems that all the Gaussians over the two CVs (diff1,diff2) are adding at the same time (every 25 fs).

Is this normal? Thanks a lot!

All the best,

Qinghua