Re-weighting well-tempered simulations with multiple walkers

[Muneeb Sultan]

Hello everyone, 

I had a small technical question regarding WHAM re-weighting with multiple walkers per CV. 

I believe I understand how to reweight to full phase space using WHAM/bin-less WHAM / MBAR when there is a single walker following the tutorial from the website and results of previous questions . However, currently I am running a Multiple walker Bias exchange simulation which is producing a series of HILLS.* files for each sampled coordinate per walker. I think the way to analyze these simulations is to concatenate all the HILLS.* files into a single massive HILLS file and use the same scripts as before. Is that correct or am I missing something very obvious? 

In short, the steps end up being :

1). cat HILLS.* > HILLS.ALL 

2). concatenate all trajectories across all walkers and CVs

3). modify plumed.dat to drop 0 gaussians( PACE=very large number, HEIGHT=0, HILLS=HILLS.ALL)

4). Use the last reported bias as input for the WHAM equations. 

Does this sound like a reasonable plan? I apologize if this has been answered before but I couldn't find it on the google group. Thank you for your help. 

Best,

Muneeb 

tags: Multiple walkers, Bias-Exchange, Well-tempered Metadynamics, reweighting, WHAM, post-processing

[Muneeb Sultan]

Just to be clear Steps 1,3 and 4 would be done for each CV being enhanced so that we get biases along different dimensions as input for WHAM. 

[Giovanni Bussi]

Hi,

are you using file-based multiple walkers? In this case I think what you propose is correct. It's important that the HILLS file that you use for the driver restart (to evaluate bias) apply the same potential that you used during the simulation. You might double check this printing metad.bias during the simulation.

Also, remove the WALKERS keyword when you run driver.

Giovanni

--
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+unsubscribe@googlegroups.com.
To post to this group, send email to plumed...@googlegroups.com.
Visit this group at https://groups.google.com/group/plumed-users.
To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/d991c2e0-1960-4e00-9294-2331b1da8285%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

[Muneeb Sultan]

Thank you for your reply! Glad to know I was not doing something completely un-reasonable. 

Yes, it is using the file-based multiple walkers (i.e. they share a folder but don’t directly communicate via MPI). I have the reported metad.bias for each walker and I will compare the last frames from each walker to the calculated values to see if they line up before running the full concatenated trajectories through the plumed driver. 

> Also, remove the WALKERS keyword when you run driver.

Yep, it causes a spectacular segfault otherwise. 

Thanks again for your help! 

Best,

Muneeb 

On Wednesday, June 28, 2017 at 12:38:21 PM UTC-7, Giovanni Bussi wrote:

Hi,

are you using file-based multiple walkers? In this case I think what you propose is correct. It's important that the HILLS file that you use for the driver restart (to evaluate bias) apply the same potential that you used during the simulation. You might double check this printing metad.bias during the simulation.

Also, remove the WALKERS keyword when you run driver.

Giovanni

On Wed, Jun 28, 2017 at 8:52 PM, Muneeb Sultan <muneeb...@gmail.com> wrote:

Just to be clear Steps 1,3 and 4 would be done for each CV being enhanced so that we get biases along different dimensions as input for WHAM. 

On Wednesday, June 28, 2017 at 11:32:12 AM UTC-7, Muneeb Sultan wrote:

Hello everyone, 

I had a small technical question regarding WHAM re-weighting with multiple walkers per CV. 

I believe I understand how to reweight to full phase space using WHAM/bin-less WHAM / MBAR when there is a single walker following the tutorial from the website and results of previous questions . However, currently I am running a Multiple walker Bias exchange simulation which is producing a series of HILLS.* files for each sampled coordinate per walker. I think the way to analyze these simulations is to concatenate all the HILLS.* files into a single massive HILLS file and use the same scripts as before. Is that correct or am I missing something very obvious? 

In short, the steps end up being :

1). cat HILLS.* > HILLS.ALL 

2). concatenate all trajectories across all walkers and CVs

3). modify plumed.dat to drop 0 gaussians( PACE=very large number, HEIGHT=0, HILLS=HILLS.ALL)

4). Use the last reported bias as input for the WHAM equations. 

Does this sound like a reasonable plan? I apologize if this has been answered before but I couldn't find it on the google group. Thank you for your help. 

Best,

Muneeb 

tags: Multiple walkers, Bias-Exchange, Well-tempered Metadynamics, reweighting, WHAM, post-processing

--
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 post to this group, send email to plumed...@googlegroups.com.
Visit this group at https://groups.google.com/group/plumed-users.