import numpy as np
import matplotlib.pyplot as plt
import plumed
import subprocess
import os
from contextlib import contextmanager
import re
import pandas
import pathlib

# small utility to clean up backup files:
def clean():
    print(subprocess.run("rm -f \#*; rm -fr bck.*",shell=True))
# small utility to run with mpiexec
def run_mpi(cmd,nprocs=1):
    return subprocess.run("mpiexec -np {} --oversubscribe {}".format(nprocs,cmd),shell=True)
# change directory and execute commands
# at the end of the context (or in the event of an error),
# previous working directory is restored
@contextmanager
def cd(newdir: os.PathLike):
    prevdir = os.getcwd()
    path=os.path.expanduser(newdir)
    pathlib.Path(path).mkdir(parents=True, exist_ok=True)
    try:
        os.chdir(path)
    except FileNotFoundError:
        try:
            os.mkdir(path)
            os.chdir(path)
        except:
            FileExistsError
    try:
        yield
    finally:
        os.chdir(prevdir)

# this is our standard input file:
!cat grompp.mdp

integrator               = md
dt                       = 0.002
nsteps                   = 500000

nstlog                   = 500
nstenergy                = 500
nstxout-compressed       = 500

continuation             = yes
constraints              = h-bonds
constraint-algorithm     = lincs

cutoff-scheme            = Verlet

coulombtype              = PME
rcoulomb                 = 1.0

vdwtype                  = Cut-off
rvdw                     = 1.0
DispCorr                 = EnerPres

tcoupl                   = V-rescale
tc-grps                  = Protein  SOL
tau-t                    = 0.1      0.1
ref-t                    = 300      300

Notice that we use two temperature groups (protein and solvent). However, when we will do solute tempering, both temperatures will be equal to 300K

Exercise 1a: Test with different temperatures¶

We prepare inputs for a range of temperature and run them for 1 ns (500k steps)

Ts=[300, 400, 500, 600, 700, 800, 900, 1000]

# this box takes a while to run!
for T in Ts:
    with cd("1a/T"+str(T)):
        with open("../../grompp.mdp") as f:
            with open("grompp.mdp","w") as ff:
                for line in f:
                    line=line.strip() # remove newline
                    if re.match("^ *ref-t.*$",line):
                        line=re.sub("300",str(T),line)
                    print(line,file=ff)
        run_mpi("gmx_mpi grompp -c ../../conf.gro -p ../../topol.top")
        run_mpi("gmx_mpi mdrun -nsteps 500000 -pin on -ntomp 12")

               :-) GROMACS - gmx grompp, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx grompp, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/1a/T300
Command line:
  gmx_mpi grompp -c ../../conf.gro -p ../../topol.top

Setting the LD random seed to -167927941
Generated 2145 of the 2145 non-bonded parameter combinations
Generating 1-4 interactions: fudge = 0.5
Generated 2145 of the 2145 1-4 parameter combinations
Excluding 3 bonded neighbours molecule type 'Protein'
Excluding 2 bonded neighbours molecule type 'SOL'
Number of degrees of freedom in T-Coupling group Protein is 53.95
Number of degrees of freedom in T-Coupling group SOL is 3135.05
Estimate for the relative computational load of the PME mesh part: 0.19

GROMACS reminds you: "Predictions can be very difficult - especially about the future." (Niels Bohr)

turning H bonds into constraints...
turning H bonds into constraints...
Analysing residue names:
There are:     3    Protein residues
There are:   523      Water residues
Analysing Protein...
Determining Verlet buffer for a tolerance of 0.005 kJ/mol/ps at 300 K
Calculated rlist for 1x1 atom pair-list as 1.036 nm, buffer size 0.036 nm
Set rlist, assuming 4x4 atom pair-list, to 1.000 nm, buffer size 0.000 nm
Note that mdrun will redetermine rlist based on the actual pair-list setup
Calculating fourier grid dimensions for X Y Z
Using a fourier grid of 24x24x24, spacing 0.118 0.118 0.118
This run will generate roughly 12 Mb of data

                :-) GROMACS - gmx mdrun, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx mdrun, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/1a/T300
Command line:
  gmx_mpi mdrun -nsteps 500000 -pin on -ntomp 12

Compiled SIMD: SSE2, but for this host/run AVX_256 might be better (see log).
The current CPU can measure timings more accurately than the code in
gmx mdrun was configured to use. This might affect your simulation
speed as accurate timings are needed for load-balancing.
Please consider rebuilding gmx mdrun with the GMX_USE_RDTSCP=ON CMake option.
Reading file topol.tpr, VERSION 2020.6-plumed-2.8.0 (single precision)

Overriding nsteps with value passed on the command line: 500000 steps, 1e+03 ps
Changing nstlist from 10 to 50, rlist from 1 to 1.111


Using 1 MPI process
Using 12 OpenMP threads 


Overriding thread affinity set outside gmx mdrun
starting mdrun 'alanine dipeptide in vacuum in water'
500000 steps,   1000.0 ps.

Writing final coordinates.

               Core t (s)   Wall t (s)        (%)
       Time:     6519.324      543.277     1200.0
                 (ns/day)    (hour/ns)
Performance:      159.035        0.151

GROMACS reminds you: "XML is not a language in the sense of a programming language any more than sketches on a napkin are a language." (Charles Simonyi)

               :-) GROMACS - gmx grompp, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx grompp, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/1a/T400
Command line:
  gmx_mpi grompp -c ../../conf.gro -p ../../topol.top

Setting the LD random seed to -21508545
Generated 2145 of the 2145 non-bonded parameter combinations
Generating 1-4 interactions: fudge = 0.5
Generated 2145 of the 2145 1-4 parameter combinations
Excluding 3 bonded neighbours molecule type 'Protein'
Excluding 2 bonded neighbours molecule type 'SOL'
Number of degrees of freedom in T-Coupling group Protein is 53.95
Number of degrees of freedom in T-Coupling group SOL is 3135.05
Estimate for the relative computational load of the PME mesh part: 0.19

GROMACS reminds you: "Never Get a Chance to Kick Ass" (The Amps)

turning H bonds into constraints...
turning H bonds into constraints...
Analysing residue names:
There are:     3    Protein residues
There are:   523      Water residues
Analysing Protein...
Determining Verlet buffer for a tolerance of 0.005 kJ/mol/ps at 400 K
Calculated rlist for 1x1 atom pair-list as 1.046 nm, buffer size 0.046 nm
Set rlist, assuming 4x4 atom pair-list, to 1.004 nm, buffer size 0.004 nm
Note that mdrun will redetermine rlist based on the actual pair-list setup
Calculating fourier grid dimensions for X Y Z
Using a fourier grid of 24x24x24, spacing 0.118 0.118 0.118
This run will generate roughly 12 Mb of data

                :-) GROMACS - gmx mdrun, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx mdrun, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/1a/T400
Command line:
  gmx_mpi mdrun -nsteps 500000 -pin on -ntomp 12

Compiled SIMD: SSE2, but for this host/run AVX_256 might be better (see log).
The current CPU can measure timings more accurately than the code in
gmx mdrun was configured to use. This might affect your simulation
speed as accurate timings are needed for load-balancing.
Please consider rebuilding gmx mdrun with the GMX_USE_RDTSCP=ON CMake option.
Reading file topol.tpr, VERSION 2020.6-plumed-2.8.0 (single precision)

Overriding nsteps with value passed on the command line: 500000 steps, 1e+03 ps
Changing nstlist from 10 to 40, rlist from 1.004 to 1.112


Using 1 MPI process
Using 12 OpenMP threads 


Overriding thread affinity set outside gmx mdrun
starting mdrun 'alanine dipeptide in vacuum in water'
500000 steps,   1000.0 ps.

Writing final coordinates.

               Core t (s)   Wall t (s)        (%)
       Time:     6552.349      546.029     1200.0
                 (ns/day)    (hour/ns)
Performance:      158.234        0.152

GROMACS reminds you: "Stupidity got us into this mess, and stupidity will get us out." (Homer Simpson)

               :-) GROMACS - gmx grompp, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx grompp, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/1a/T500
Command line:
  gmx_mpi grompp -c ../../conf.gro -p ../../topol.top

Setting the LD random seed to -554885641
Generated 2145 of the 2145 non-bonded parameter combinations
Generating 1-4 interactions: fudge = 0.5
Generated 2145 of the 2145 1-4 parameter combinations
Excluding 3 bonded neighbours molecule type 'Protein'
Excluding 2 bonded neighbours molecule type 'SOL'
Number of degrees of freedom in T-Coupling group Protein is 53.95
Number of degrees of freedom in T-Coupling group SOL is 3135.05
Estimate for the relative computational load of the PME mesh part: 0.19

GROMACS reminds you: "Stupidity got us into this mess, and stupidity will get us out." (Homer Simpson)

turning H bonds into constraints...
turning H bonds into constraints...
Analysing residue names:
There are:     3    Protein residues
There are:   523      Water residues
Analysing Protein...
Determining Verlet buffer for a tolerance of 0.005 kJ/mol/ps at 500 K
Calculated rlist for 1x1 atom pair-list as 1.056 nm, buffer size 0.056 nm
Set rlist, assuming 4x4 atom pair-list, to 1.010 nm, buffer size 0.010 nm
Note that mdrun will redetermine rlist based on the actual pair-list setup
Calculating fourier grid dimensions for X Y Z
Using a fourier grid of 24x24x24, spacing 0.118 0.118 0.118
This run will generate roughly 12 Mb of data

                :-) GROMACS - gmx mdrun, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx mdrun, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/1a/T500
Command line:
  gmx_mpi mdrun -nsteps 500000 -pin on -ntomp 12

Compiled SIMD: SSE2, but for this host/run AVX_256 might be better (see log).
The current CPU can measure timings more accurately than the code in
gmx mdrun was configured to use. This might affect your simulation
speed as accurate timings are needed for load-balancing.
Please consider rebuilding gmx mdrun with the GMX_USE_RDTSCP=ON CMake option.
Reading file topol.tpr, VERSION 2020.6-plumed-2.8.0 (single precision)

Overriding nsteps with value passed on the command line: 500000 steps, 1e+03 ps
Changing nstlist from 10 to 40, rlist from 1.01 to 1.127


Using 1 MPI process
Using 12 OpenMP threads 


Overriding thread affinity set outside gmx mdrun
starting mdrun 'alanine dipeptide in vacuum in water'
500000 steps,   1000.0 ps.

Writing final coordinates.

               Core t (s)   Wall t (s)        (%)
       Time:     6613.881      551.157     1200.0
                 (ns/day)    (hour/ns)
Performance:      156.761        0.153

GROMACS reminds you: "This Doesn't Suck, It's a Black Hole !" (K.A. Feenstra)

               :-) GROMACS - gmx grompp, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx grompp, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/1a/T600
Command line:
  gmx_mpi grompp -c ../../conf.gro -p ../../topol.top

Setting the LD random seed to -537952417
Generated 2145 of the 2145 non-bonded parameter combinations
Generating 1-4 interactions: fudge = 0.5
Generated 2145 of the 2145 1-4 parameter combinations
Excluding 3 bonded neighbours molecule type 'Protein'
Excluding 2 bonded neighbours molecule type 'SOL'
Number of degrees of freedom in T-Coupling group Protein is 53.95
Number of degrees of freedom in T-Coupling group SOL is 3135.05
Estimate for the relative computational load of the PME mesh part: 0.19

GROMACS reminds you: "This Doesn't Suck, It's a Black Hole !" (K.A. Feenstra)

turning H bonds into constraints...
turning H bonds into constraints...
Analysing residue names:
There are:     3    Protein residues
There are:   523      Water residues
Analysing Protein...
Determining Verlet buffer for a tolerance of 0.005 kJ/mol/ps at 600 K
Calculated rlist for 1x1 atom pair-list as 1.064 nm, buffer size 0.064 nm
Set rlist, assuming 4x4 atom pair-list, to 1.016 nm, buffer size 0.016 nm
Note that mdrun will redetermine rlist based on the actual pair-list setup
Calculating fourier grid dimensions for X Y Z
Using a fourier grid of 24x24x24, spacing 0.118 0.118 0.118
This run will generate roughly 12 Mb of data

                :-) GROMACS - gmx mdrun, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx mdrun, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/1a/T600
Command line:
  gmx_mpi mdrun -nsteps 500000 -pin on -ntomp 12

Compiled SIMD: SSE2, but for this host/run AVX_256 might be better (see log).
The current CPU can measure timings more accurately than the code in
gmx mdrun was configured to use. This might affect your simulation
speed as accurate timings are needed for load-balancing.
Please consider rebuilding gmx mdrun with the GMX_USE_RDTSCP=ON CMake option.
Reading file topol.tpr, VERSION 2020.6-plumed-2.8.0 (single precision)

Overriding nsteps with value passed on the command line: 500000 steps, 1e+03 ps
Changing nstlist from 10 to 40, rlist from 1.016 to 1.138


Using 1 MPI process
Using 12 OpenMP threads 


Overriding thread affinity set outside gmx mdrun
starting mdrun 'alanine dipeptide in vacuum in water'
500000 steps,   1000.0 ps.

Writing final coordinates.

               Core t (s)   Wall t (s)        (%)
       Time:     6600.518      550.043     1200.0
                 (ns/day)    (hour/ns)
Performance:      157.079        0.153

GROMACS reminds you: "Don't Push Me, Cause I'm Close to the Edge" (Tricky)

               :-) GROMACS - gmx grompp, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx grompp, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/1a/T700
Command line:
  gmx_mpi grompp -c ../../conf.gro -p ../../topol.top

Setting the LD random seed to -1048737
Generated 2145 of the 2145 non-bonded parameter combinations
Generating 1-4 interactions: fudge = 0.5
Generated 2145 of the 2145 1-4 parameter combinations
Excluding 3 bonded neighbours molecule type 'Protein'
Excluding 2 bonded neighbours molecule type 'SOL'
Number of degrees of freedom in T-Coupling group Protein is 53.95
Number of degrees of freedom in T-Coupling group SOL is 3135.05
Estimate for the relative computational load of the PME mesh part: 0.18

GROMACS reminds you: "Religion is a culture of faith; science is a culture of doubt." (Richard Feynman)

turning H bonds into constraints...
turning H bonds into constraints...
Analysing residue names:
There are:     3    Protein residues
There are:   523      Water residues
Analysing Protein...
Determining Verlet buffer for a tolerance of 0.005 kJ/mol/ps at 700 K
Calculated rlist for 1x1 atom pair-list as 1.073 nm, buffer size 0.073 nm
Set rlist, assuming 4x4 atom pair-list, to 1.021 nm, buffer size 0.021 nm
Note that mdrun will redetermine rlist based on the actual pair-list setup
Calculating fourier grid dimensions for X Y Z
Using a fourier grid of 24x24x24, spacing 0.118 0.118 0.118
This run will generate roughly 12 Mb of data

We then generate COLVAR files (with phi,psi time series) and read them in pandas dataframes:

colvar=[]

for T in Ts:
    with cd("1a/T"+str(T)):
        with open("plumed.dat","w") as f:
                  print("""
MOLINFO STRUCTURE=../../ala.pdb 
phi: TORSION ATOMS=@phi-2
psi: TORSION ATOMS=@psi-2
PRINT ARG=phi,psi FILE=COLVAR 
                  """, file=f)
        subprocess.run("plumed driver --ixtc traj_comp.xtc --plumed plumed.dat",shell=True)
        colvar.append(plumed.read_as_pandas("COLVAR"))

PLUMED: PLUMED is starting
PLUMED: Version: 2.8.0 (git: ddb59d16b) compiled on May 17 2022 at 19:07:56
PLUMED: Please cite these papers when using PLUMED [1][2]
PLUMED: For further information see the PLUMED web page at http://www.plumed.org
PLUMED: Root: /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed
PLUMED: For installed feature, see /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed/src/config/config.txt
PLUMED: Molecular dynamics engine: driver
PLUMED: Precision of reals: 8
PLUMED: Running over 1 node
PLUMED: Number of threads: 1
PLUMED: Cache line size: 512
PLUMED: Number of atoms: 1591
PLUMED: File suffix: 
PLUMED: FILE: plumed.dat
PLUMED: Action MOLINFO
PLUMED:   with label @0
PLUMED:   pdb file named ../../ala.pdb contains 1 chains 
PLUMED:   chain named   contains residues 1 to 526 and atoms 1 to 1591 
PLUMED: Action TORSION
PLUMED:   with label phi
PLUMED:   between atoms 5 7 9 15
PLUMED:   using periodic boundary conditions
PLUMED: Action TORSION
PLUMED:   with label psi
PLUMED:   between atoms 7 9 15 17
PLUMED:   using periodic boundary conditions
PLUMED: Action PRINT
PLUMED:   with label @3
PLUMED:   with stride 1
PLUMED:   with arguments phi psi
PLUMED:   on file COLVAR
PLUMED:   with format  %f
PLUMED: END FILE: plumed.dat
PLUMED: Timestep: 1.000000
PLUMED: KbT has not been set by the MD engine
PLUMED: It should be set by hand where needed
PLUMED: Relevant bibliography:
PLUMED:   [1] The PLUMED consortium, Nat. Methods 16, 670 (2019)
PLUMED:   [2] Tribello, Bonomi, Branduardi, Camilloni, and Bussi, Comput. Phys. Commun. 185, 604 (2014)
PLUMED: Please read and cite where appropriate!
PLUMED: Finished setup
PLUMED:                                               Cycles        Total      Average      Minimum      Maximum
PLUMED:                                                    1     0.092092     0.092092     0.092092     0.092092
PLUMED: 1 Prepare dependencies                          1001     0.002127     0.000002     0.000002     0.000012
PLUMED: 2 Sharing data                                  1001     0.010291     0.000010     0.000009     0.000256
PLUMED: 3 Waiting for data                              1001     0.003283     0.000003     0.000003     0.000022
PLUMED: 4 Calculating (forward loop)                    1001     0.011092     0.000011     0.000010     0.000055
PLUMED: 5 Applying (backward loop)                      1001     0.014850     0.000015     0.000014     0.000026
PLUMED: 6 Update                                        1001     0.009367     0.000009     0.000008     0.000060

+++ Loading the PLUMED kernel runtime +++
+++ PLUMED_KERNEL="/net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/libplumedKernel.so" +++

PLUMED: PLUMED is starting
PLUMED: Version: 2.8.0 (git: ddb59d16b) compiled on May 17 2022 at 19:07:56
PLUMED: Please cite these papers when using PLUMED [1][2]
PLUMED: For further information see the PLUMED web page at http://www.plumed.org
PLUMED: Root: /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed
PLUMED: For installed feature, see /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed/src/config/config.txt
PLUMED: Molecular dynamics engine: driver
PLUMED: Precision of reals: 8
PLUMED: Running over 1 node
PLUMED: Number of threads: 1
PLUMED: Cache line size: 512
PLUMED: Number of atoms: 1591
PLUMED: File suffix: 
PLUMED: FILE: plumed.dat
PLUMED: Action MOLINFO
PLUMED:   with label @0
PLUMED:   pdb file named ../../ala.pdb contains 1 chains 
PLUMED:   chain named   contains residues 1 to 526 and atoms 1 to 1591 
PLUMED: Action TORSION
PLUMED:   with label phi
PLUMED:   between atoms 5 7 9 15
PLUMED:   using periodic boundary conditions
PLUMED: Action TORSION
PLUMED:   with label psi
PLUMED:   between atoms 7 9 15 17
PLUMED:   using periodic boundary conditions
PLUMED: Action PRINT
PLUMED:   with label @3
PLUMED:   with stride 1
PLUMED:   with arguments phi psi
PLUMED:   on file COLVAR
PLUMED:   with format  %f
PLUMED: END FILE: plumed.dat
PLUMED: Timestep: 1.000000
PLUMED: KbT has not been set by the MD engine
PLUMED: It should be set by hand where needed
PLUMED: Relevant bibliography:
PLUMED:   [1] The PLUMED consortium, Nat. Methods 16, 670 (2019)
PLUMED:   [2] Tribello, Bonomi, Branduardi, Camilloni, and Bussi, Comput. Phys. Commun. 185, 604 (2014)
PLUMED: Please read and cite where appropriate!
PLUMED: Finished setup
PLUMED:                                               Cycles        Total      Average      Minimum      Maximum
PLUMED:                                                    1     0.092041     0.092041     0.092041     0.092041
PLUMED: 1 Prepare dependencies                          1001     0.002032     0.000002     0.000001     0.000012
PLUMED: 2 Sharing data                                  1001     0.010318     0.000010     0.000009     0.000275
PLUMED: 3 Waiting for data                              1001     0.003435     0.000003     0.000003     0.000021
PLUMED: 4 Calculating (forward loop)                    1001     0.011163     0.000011     0.000010     0.000044
PLUMED: 5 Applying (backward loop)                      1001     0.014768     0.000015     0.000014     0.000034
PLUMED: 6 Update                                        1001     0.009524     0.000010     0.000008     0.000060

+++ Loading the PLUMED kernel runtime +++
+++ PLUMED_KERNEL="/net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/libplumedKernel.so" +++

PLUMED: PLUMED is starting
PLUMED: Version: 2.8.0 (git: ddb59d16b) compiled on May 17 2022 at 19:07:56
PLUMED: Please cite these papers when using PLUMED [1][2]
PLUMED: For further information see the PLUMED web page at http://www.plumed.org
PLUMED: Root: /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed
PLUMED: For installed feature, see /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed/src/config/config.txt
PLUMED: Molecular dynamics engine: driver
PLUMED: Precision of reals: 8
PLUMED: Running over 1 node
PLUMED: Number of threads: 1
PLUMED: Cache line size: 512
PLUMED: Number of atoms: 1591
PLUMED: File suffix: 
PLUMED: FILE: plumed.dat
PLUMED: Action MOLINFO
PLUMED:   with label @0
PLUMED:   pdb file named ../../ala.pdb contains 1 chains 
PLUMED:   chain named   contains residues 1 to 526 and atoms 1 to 1591 
PLUMED: Action TORSION
PLUMED:   with label phi
PLUMED:   between atoms 5 7 9 15
PLUMED:   using periodic boundary conditions
PLUMED: Action TORSION
PLUMED:   with label psi
PLUMED:   between atoms 7 9 15 17
PLUMED:   using periodic boundary conditions
PLUMED: Action PRINT
PLUMED:   with label @3
PLUMED:   with stride 1
PLUMED:   with arguments phi psi
PLUMED:   on file COLVAR
PLUMED:   with format  %f
PLUMED: END FILE: plumed.dat
PLUMED: Timestep: 1.000000
PLUMED: KbT has not been set by the MD engine
PLUMED: It should be set by hand where needed
PLUMED: Relevant bibliography:
PLUMED:   [1] The PLUMED consortium, Nat. Methods 16, 670 (2019)
PLUMED:   [2] Tribello, Bonomi, Branduardi, Camilloni, and Bussi, Comput. Phys. Commun. 185, 604 (2014)
PLUMED: Please read and cite where appropriate!
PLUMED: Finished setup
PLUMED:                                               Cycles        Total      Average      Minimum      Maximum
PLUMED:                                                    1     0.092507     0.092507     0.092507     0.092507
PLUMED: 1 Prepare dependencies                          1001     0.002068     0.000002     0.000001     0.000013
PLUMED: 2 Sharing data                                  1001     0.010422     0.000010     0.000010     0.000318
PLUMED: 3 Waiting for data                              1001     0.003303     0.000003     0.000003     0.000022
PLUMED: 4 Calculating (forward loop)                    1001     0.011159     0.000011     0.000010     0.000047
PLUMED: 5 Applying (backward loop)                      1001     0.014961     0.000015     0.000014     0.000032
PLUMED: 6 Update                                        1001     0.009591     0.000010     0.000008     0.000066

+++ Loading the PLUMED kernel runtime +++
+++ PLUMED_KERNEL="/net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/libplumedKernel.so" +++

PLUMED: PLUMED is starting
PLUMED: Version: 2.8.0 (git: ddb59d16b) compiled on May 17 2022 at 19:07:56
PLUMED: Please cite these papers when using PLUMED [1][2]
PLUMED: For further information see the PLUMED web page at http://www.plumed.org
PLUMED: Root: /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed
PLUMED: For installed feature, see /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed/src/config/config.txt
PLUMED: Molecular dynamics engine: driver
PLUMED: Precision of reals: 8
PLUMED: Running over 1 node
PLUMED: Number of threads: 1
PLUMED: Cache line size: 512
PLUMED: Number of atoms: 1591
PLUMED: File suffix: 
PLUMED: FILE: plumed.dat
PLUMED: Action MOLINFO
PLUMED:   with label @0
PLUMED:   pdb file named ../../ala.pdb contains 1 chains 
PLUMED:   chain named   contains residues 1 to 526 and atoms 1 to 1591 
PLUMED: Action TORSION
PLUMED:   with label phi
PLUMED:   between atoms 5 7 9 15
PLUMED:   using periodic boundary conditions
PLUMED: Action TORSION
PLUMED:   with label psi
PLUMED:   between atoms 7 9 15 17
PLUMED:   using periodic boundary conditions
PLUMED: Action PRINT
PLUMED:   with label @3
PLUMED:   with stride 1
PLUMED:   with arguments phi psi
PLUMED:   on file COLVAR
PLUMED:   with format  %f
PLUMED: END FILE: plumed.dat
PLUMED: Timestep: 1.000000
PLUMED: KbT has not been set by the MD engine
PLUMED: It should be set by hand where needed
PLUMED: Relevant bibliography:
PLUMED:   [1] The PLUMED consortium, Nat. Methods 16, 670 (2019)
PLUMED:   [2] Tribello, Bonomi, Branduardi, Camilloni, and Bussi, Comput. Phys. Commun. 185, 604 (2014)
PLUMED: Please read and cite where appropriate!
PLUMED: Finished setup
PLUMED:                                               Cycles        Total      Average      Minimum      Maximum
PLUMED:                                                    1     0.092000     0.092000     0.092000     0.092000
PLUMED: 1 Prepare dependencies                          1001     0.002023     0.000002     0.000001     0.000013
PLUMED: 2 Sharing data                                  1001     0.010264     0.000010     0.000009     0.000320
PLUMED: 3 Waiting for data                              1001     0.003399     0.000003     0.000003     0.000022
PLUMED: 4 Calculating (forward loop)                    1001     0.011264     0.000011     0.000010     0.000047
PLUMED: 5 Applying (backward loop)                      1001     0.014655     0.000015     0.000014     0.000026
PLUMED: 6 Update                                        1001     0.009479     0.000009     0.000008     0.000066

+++ Loading the PLUMED kernel runtime +++
+++ PLUMED_KERNEL="/net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/libplumedKernel.so" +++

PLUMED: PLUMED is starting
PLUMED: Version: 2.8.0 (git: ddb59d16b) compiled on May 17 2022 at 19:07:56
PLUMED: Please cite these papers when using PLUMED [1][2]
PLUMED: For further information see the PLUMED web page at http://www.plumed.org
PLUMED: Root: /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed
PLUMED: For installed feature, see /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed/src/config/config.txt
PLUMED: Molecular dynamics engine: driver
PLUMED: Precision of reals: 8
PLUMED: Running over 1 node
PLUMED: Number of threads: 1
PLUMED: Cache line size: 512
PLUMED: Number of atoms: 1591
PLUMED: File suffix: 
PLUMED: FILE: plumed.dat
PLUMED: Action MOLINFO
PLUMED:   with label @0
PLUMED:   pdb file named ../../ala.pdb contains 1 chains 
PLUMED:   chain named   contains residues 1 to 526 and atoms 1 to 1591 
PLUMED: Action TORSION
PLUMED:   with label phi
PLUMED:   between atoms 5 7 9 15
PLUMED:   using periodic boundary conditions
PLUMED: Action TORSION
PLUMED:   with label psi
PLUMED:   between atoms 7 9 15 17
PLUMED:   using periodic boundary conditions
PLUMED: Action PRINT
PLUMED:   with label @3
PLUMED:   with stride 1
PLUMED:   with arguments phi psi
PLUMED:   on file COLVAR
PLUMED:   with format  %f
PLUMED: END FILE: plumed.dat
PLUMED: Timestep: 1.000000
PLUMED: KbT has not been set by the MD engine
PLUMED: It should be set by hand where needed
PLUMED: Relevant bibliography:
PLUMED:   [1] The PLUMED consortium, Nat. Methods 16, 670 (2019)
PLUMED:   [2] Tribello, Bonomi, Branduardi, Camilloni, and Bussi, Comput. Phys. Commun. 185, 604 (2014)
PLUMED: Please read and cite where appropriate!
PLUMED: Finished setup
PLUMED:                                               Cycles        Total      Average      Minimum      Maximum
PLUMED:                                                    1     0.092450     0.092450     0.092450     0.092450
PLUMED: 1 Prepare dependencies                          1001     0.002074     0.000002     0.000002     0.000011
PLUMED: 2 Sharing data                                  1001     0.010134     0.000010     0.000009     0.000276
PLUMED: 3 Waiting for data                              1001     0.003320     0.000003     0.000003     0.000021
PLUMED: 4 Calculating (forward loop)                    1001     0.011425     0.000011     0.000010     0.000050
PLUMED: 5 Applying (backward loop)                      1001     0.014631     0.000015     0.000014     0.000033
PLUMED: 6 Update                                        1001     0.009758     0.000010     0.000009     0.000062

+++ Loading the PLUMED kernel runtime +++
+++ PLUMED_KERNEL="/net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/libplumedKernel.so" +++

We then plot the resulting time series for phi

for i in range(len(Ts)):
    plt.title("T={}".format(Ts[i]))
    plt.plot(colvar[i].time,colvar[i].phi,".")
    plt.show()

As you can see, the number of transitions to the higher energy state is increasing with temperature. Notice that we have two simultaneous effects:

The population of state B is increasing as we increase the temperature
Also the rates (both A->B and B->A are increasign when we increase the temperature

Exercise 1b: Run a parallel tempering simulation¶

Based on the analysis above, we decide to use T=800 at maximum temperature. This decision is partly arbitrary, but at least we should be able to get an approximate value with an analysis similar to what we have done above. Also notice that parallel tempering (and also solute tempering) simulations are not optimal for systems with such a high barrier on a single degree of freedom. In cases where we, for instance, combine parallel tempering (or solute tempering) with metadynamics or other enhanced sampling methods, determining the optimal temperature for the hottest replica should be might be more difficult.

Tmax=800
Ts=np.geomspace(300,Tmax,32)

We now run a parallel tempering simulations with a range of temperatures

# this box takes a while to run!
with cd("1b/32reps"):
    for i in range(len(Ts)):
        with cd("replica{}".format(i)):
            with open("../../../grompp.mdp") as f:
                with open("grompp.mdp","w") as ff:
                    for line in f:
                        line=line.strip() # remove newline
                        if re.match("^ *ref-t.*$",line):
                            line=re.sub("300",str(Ts[i]),line)
                        print(line,file=ff)
            run_mpi("gmx_mpi grompp -c ../../../conf.gro -p ../../../topol.top")
    run_mpi("gmx_mpi mdrun -multidir replica? replica?? -replex 100 -nsteps 500000 -ntomp 1",len(Ts))

               :-) GROMACS - gmx grompp, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx grompp, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/1b/16reps/replica0
Command line:
  gmx_mpi grompp -c ../../../conf.gro -p ../../../topol.top

Setting the LD random seed to -73433428
Generated 2145 of the 2145 non-bonded parameter combinations
Generating 1-4 interactions: fudge = 0.5
Generated 2145 of the 2145 1-4 parameter combinations
Excluding 3 bonded neighbours molecule type 'Protein'
Excluding 2 bonded neighbours molecule type 'SOL'
Number of degrees of freedom in T-Coupling group Protein is 53.95
Number of degrees of freedom in T-Coupling group SOL is 3135.05
Estimate for the relative computational load of the PME mesh part: 0.19

GROMACS reminds you: "In the Meantime, Take Care of Yourself aaand Eachother" (J. Springer)

turning H bonds into constraints...
turning H bonds into constraints...
Analysing residue names:
There are:     3    Protein residues
There are:   523      Water residues
Analysing Protein...
Determining Verlet buffer for a tolerance of 0.005 kJ/mol/ps at 300 K
Calculated rlist for 1x1 atom pair-list as 1.036 nm, buffer size 0.036 nm
Set rlist, assuming 4x4 atom pair-list, to 1.000 nm, buffer size 0.000 nm
Note that mdrun will redetermine rlist based on the actual pair-list setup
Calculating fourier grid dimensions for X Y Z
Using a fourier grid of 24x24x24, spacing 0.118 0.118 0.118
This run will generate roughly 12 Mb of data

               :-) GROMACS - gmx grompp, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx grompp, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/1b/16reps/replica1
Command line:
  gmx_mpi grompp -c ../../../conf.gro -p ../../../topol.top

Setting the LD random seed to -1489045617
Generated 2145 of the 2145 non-bonded parameter combinations
Generating 1-4 interactions: fudge = 0.5
Generated 2145 of the 2145 1-4 parameter combinations
Excluding 3 bonded neighbours molecule type 'Protein'
Excluding 2 bonded neighbours molecule type 'SOL'
Number of degrees of freedom in T-Coupling group Protein is 53.95
Number of degrees of freedom in T-Coupling group SOL is 3135.05
Estimate for the relative computational load of the PME mesh part: 0.19

GROMACS reminds you: "In the Meantime, Take Care of Yourself aaand Eachother" (J. Springer)

turning H bonds into constraints...
turning H bonds into constraints...
Analysing residue names:
There are:     3    Protein residues
There are:   523      Water residues
Analysing Protein...
Determining Verlet buffer for a tolerance of 0.005 kJ/mol/ps at 320.272 K
Calculated rlist for 1x1 atom pair-list as 1.038 nm, buffer size 0.038 nm
Set rlist, assuming 4x4 atom pair-list, to 1.000 nm, buffer size 0.000 nm
Note that mdrun will redetermine rlist based on the actual pair-list setup
Calculating fourier grid dimensions for X Y Z
Using a fourier grid of 24x24x24, spacing 0.118 0.118 0.118
This run will generate roughly 12 Mb of data

               :-) GROMACS - gmx grompp, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx grompp, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/1b/16reps/replica2
Command line:
  gmx_mpi grompp -c ../../../conf.gro -p ../../../topol.top

Setting the LD random seed to 1119739901
Generated 2145 of the 2145 non-bonded parameter combinations
Generating 1-4 interactions: fudge = 0.5
Generated 2145 of the 2145 1-4 parameter combinations
Excluding 3 bonded neighbours molecule type 'Protein'
Excluding 2 bonded neighbours molecule type 'SOL'
Number of degrees of freedom in T-Coupling group Protein is 53.95
Number of degrees of freedom in T-Coupling group SOL is 3135.05
Estimate for the relative computational load of the PME mesh part: 0.19

GROMACS reminds you: "In the Meantime, Take Care of Yourself aaand Eachother" (J. Springer)

turning H bonds into constraints...
turning H bonds into constraints...
Analysing residue names:
There are:     3    Protein residues
There are:   523      Water residues
Analysing Protein...
Determining Verlet buffer for a tolerance of 0.005 kJ/mol/ps at 341.914 K
Calculated rlist for 1x1 atom pair-list as 1.040 nm, buffer size 0.040 nm
Set rlist, assuming 4x4 atom pair-list, to 1.000 nm, buffer size 0.000 nm
Note that mdrun will redetermine rlist based on the actual pair-list setup
Calculating fourier grid dimensions for X Y Z
Using a fourier grid of 24x24x24, spacing 0.118 0.118 0.118
This run will generate roughly 12 Mb of data

               :-) GROMACS - gmx grompp, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx grompp, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/1b/16reps/replica3
Command line:
  gmx_mpi grompp -c ../../../conf.gro -p ../../../topol.top

Setting the LD random seed to -302713875
Generated 2145 of the 2145 non-bonded parameter combinations
Generating 1-4 interactions: fudge = 0.5
Generated 2145 of the 2145 1-4 parameter combinations
Excluding 3 bonded neighbours molecule type 'Protein'
Excluding 2 bonded neighbours molecule type 'SOL'
Number of degrees of freedom in T-Coupling group Protein is 53.95
Number of degrees of freedom in T-Coupling group SOL is 3135.05
Estimate for the relative computational load of the PME mesh part: 0.19

GROMACS reminds you: "In the Meantime, Take Care of Yourself aaand Eachother" (J. Springer)

turning H bonds into constraints...
turning H bonds into constraints...
Analysing residue names:
There are:     3    Protein residues
There are:   523      Water residues
Analysing Protein...
Determining Verlet buffer for a tolerance of 0.005 kJ/mol/ps at 365.019 K
Calculated rlist for 1x1 atom pair-list as 1.042 nm, buffer size 0.042 nm
Set rlist, assuming 4x4 atom pair-list, to 1.002 nm, buffer size 0.002 nm
Note that mdrun will redetermine rlist based on the actual pair-list setup
Calculating fourier grid dimensions for X Y Z
Using a fourier grid of 24x24x24, spacing 0.118 0.118 0.118
This run will generate roughly 12 Mb of data

               :-) GROMACS - gmx grompp, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx grompp, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/1b/16reps/replica4
Command line:
  gmx_mpi grompp -c ../../../conf.gro -p ../../../topol.top

Setting the LD random seed to 2012205054
Generated 2145 of the 2145 non-bonded parameter combinations
Generating 1-4 interactions: fudge = 0.5
Generated 2145 of the 2145 1-4 parameter combinations
Excluding 3 bonded neighbours molecule type 'Protein'
Excluding 2 bonded neighbours molecule type 'SOL'
Number of degrees of freedom in T-Coupling group Protein is 53.95
Number of degrees of freedom in T-Coupling group SOL is 3135.05
Estimate for the relative computational load of the PME mesh part: 0.19

GROMACS reminds you: "Gabba Gabba Hey!" (The Ramones)

turning H bonds into constraints...
turning H bonds into constraints...
Analysing residue names:
There are:     3    Protein residues
There are:   523      Water residues
Analysing Protein...
Determining Verlet buffer for a tolerance of 0.005 kJ/mol/ps at 389.684 K
Calculated rlist for 1x1 atom pair-list as 1.045 nm, buffer size 0.045 nm
Set rlist, assuming 4x4 atom pair-list, to 1.003 nm, buffer size 0.003 nm
Note that mdrun will redetermine rlist based on the actual pair-list setup
Calculating fourier grid dimensions for X Y Z
Using a fourier grid of 24x24x24, spacing 0.118 0.118 0.118
This run will generate roughly 12 Mb of data

! grep -A 4 average probabilities 1b/32reps/replica0/md.log

grep: probabilities: No such file or directory
1b/32reps/replica0/md.log:Repl  average probabilities:
1b/32reps/replica0/md.log-Repl     0    1    2    3    4    5    6    7    8    9   10   11   12   13   14   15   16   17   18   19   20   21   22   23   24   25   26   27   28   29   30   31
1b/32reps/replica0/md.log-Repl      .25  .24  .24  .26  .27  .27  .27  .30  .29  .31  .31  .31  .33  .34  .36  .34  .36  .37  .38  .36  .39  .38  .40  .43  .44  .42  .44  .44  .45  .46  .45
1b/32reps/replica0/md.log-Repl  number of exchanges:
1b/32reps/replica0/md.log-Repl     0    1    2    3    4    5    6    7    8    9   10   11   12   13   14   15   16   17   18   19   20   21   22   23   24   25   26   27   28   29   30   31
--
1b/32reps/replica0/md.log:Repl  average number of exchanges:
1b/32reps/replica0/md.log-Repl     0    1    2    3    4    5    6    7    8    9   10   11   12   13   14   15   16   17   18   19   20   21   22   23   24   25   26   27   28   29   30   31
1b/32reps/replica0/md.log-Repl      .25  .24  .22  .25  .27  .28  .28  .31  .28  .30  .31  .32  .34  .35  .37  .34  .36  .37  .38  .37  .40  .38  .41  .41  .45  .41  .43  .44  .44  .48  .45
1b/32reps/replica0/md.log-
1b/32reps/replica0/md.log-

Acceptances are very reasonable. They are not exactly uniform (22% - 45%), and slighly below the requested value (30%). You could try e.g. with 48 replicas to see if you can arrive to 30%. Or you could try to redistribute replicas better (more replicas at low temperature and less at high temperature) to make acceptance more homogenous.

colvar=[]

for i in range(len(Ts)):
    with cd("1b/32reps/replica"+str(i)):
        with open("plumed.dat","w") as f:
            print("""
MOLINFO STRUCTURE=../../../ala.pdb 
phi: TORSION ATOMS=@phi-2
psi: TORSION ATOMS=@psi-2
PRINT ARG=phi,psi FILE=COLVAR 
                  """,file=f)
        subprocess.run("plumed driver --ixtc traj_comp.xtc --plumed plumed.dat",shell=True)
            
        colvar.append(plumed.read_as_pandas("COLVAR"))

PLUMED: PLUMED is starting
PLUMED: Version: 2.8.0 (git: ddb59d16b) compiled on May 17 2022 at 19:07:56
PLUMED: Please cite these papers when using PLUMED [1][2]
PLUMED: For further information see the PLUMED web page at http://www.plumed.org
PLUMED: Root: /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed
PLUMED: For installed feature, see /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed/src/config/config.txt
PLUMED: Molecular dynamics engine: driver
PLUMED: Precision of reals: 8
PLUMED: Running over 1 node
PLUMED: Number of threads: 1
PLUMED: Cache line size: 512
PLUMED: Number of atoms: 1591
PLUMED: File suffix: 
PLUMED: FILE: plumed.dat
PLUMED: Action MOLINFO
PLUMED:   with label @0
PLUMED:   pdb file named ../../../ala.pdb contains 1 chains 
PLUMED:   chain named   contains residues 1 to 526 and atoms 1 to 1591 
PLUMED: Action TORSION
PLUMED:   with label phi
PLUMED:   between atoms 5 7 9 15
PLUMED:   using periodic boundary conditions
PLUMED: Action TORSION
PLUMED:   with label psi
PLUMED:   between atoms 7 9 15 17
PLUMED:   using periodic boundary conditions
PLUMED: Action PRINT
PLUMED:   with label @3
PLUMED:   with stride 1
PLUMED:   with arguments phi psi
PLUMED:   on file COLVAR
PLUMED:   with format  %f
PLUMED: END FILE: plumed.dat
PLUMED: Timestep: 1.000000
PLUMED: KbT has not been set by the MD engine
PLUMED: It should be set by hand where needed
PLUMED: Relevant bibliography:
PLUMED:   [1] The PLUMED consortium, Nat. Methods 16, 670 (2019)
PLUMED:   [2] Tribello, Bonomi, Branduardi, Camilloni, and Bussi, Comput. Phys. Commun. 185, 604 (2014)
PLUMED: Please read and cite where appropriate!
PLUMED: Finished setup
PLUMED:                                               Cycles        Total      Average      Minimum      Maximum
PLUMED:                                                    1     0.090283     0.090283     0.090283     0.090283
PLUMED: 1 Prepare dependencies                          1001     0.001998     0.000002     0.000001     0.000013
PLUMED: 2 Sharing data                                  1001     0.009986     0.000010     0.000009     0.000266
PLUMED: 3 Waiting for data                              1001     0.003193     0.000003     0.000003     0.000020
PLUMED: 4 Calculating (forward loop)                    1001     0.010955     0.000011     0.000010     0.000042
PLUMED: 5 Applying (backward loop)                      1001     0.014539     0.000015     0.000014     0.000026
PLUMED: 6 Update                                        1001     0.009344     0.000009     0.000008     0.000057

+++ Loading the PLUMED kernel runtime +++
+++ PLUMED_KERNEL="/net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/libplumedKernel.so" +++

PLUMED: PLUMED is starting
PLUMED: Version: 2.8.0 (git: ddb59d16b) compiled on May 17 2022 at 19:07:56
PLUMED: Please cite these papers when using PLUMED [1][2]
PLUMED: For further information see the PLUMED web page at http://www.plumed.org
PLUMED: Root: /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed
PLUMED: For installed feature, see /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed/src/config/config.txt
PLUMED: Molecular dynamics engine: driver
PLUMED: Precision of reals: 8
PLUMED: Running over 1 node
PLUMED: Number of threads: 1
PLUMED: Cache line size: 512
PLUMED: Number of atoms: 1591
PLUMED: File suffix: 
PLUMED: FILE: plumed.dat
PLUMED: Action MOLINFO
PLUMED:   with label @0
PLUMED:   pdb file named ../../../ala.pdb contains 1 chains 
PLUMED:   chain named   contains residues 1 to 526 and atoms 1 to 1591 
PLUMED: Action TORSION
PLUMED:   with label phi
PLUMED:   between atoms 5 7 9 15
PLUMED:   using periodic boundary conditions
PLUMED: Action TORSION
PLUMED:   with label psi
PLUMED:   between atoms 7 9 15 17
PLUMED:   using periodic boundary conditions
PLUMED: Action PRINT
PLUMED:   with label @3
PLUMED:   with stride 1
PLUMED:   with arguments phi psi
PLUMED:   on file COLVAR
PLUMED:   with format  %f
PLUMED: END FILE: plumed.dat
PLUMED: Timestep: 1.000000
PLUMED: KbT has not been set by the MD engine
PLUMED: It should be set by hand where needed
PLUMED: Relevant bibliography:
PLUMED:   [1] The PLUMED consortium, Nat. Methods 16, 670 (2019)
PLUMED:   [2] Tribello, Bonomi, Branduardi, Camilloni, and Bussi, Comput. Phys. Commun. 185, 604 (2014)
PLUMED: Please read and cite where appropriate!
PLUMED: Finished setup
PLUMED:                                               Cycles        Total      Average      Minimum      Maximum
PLUMED:                                                    1     0.091145     0.091145     0.091145     0.091145
PLUMED: 1 Prepare dependencies                          1001     0.002002     0.000002     0.000002     0.000013
PLUMED: 2 Sharing data                                  1001     0.010173     0.000010     0.000009     0.000264
PLUMED: 3 Waiting for data                              1001     0.003270     0.000003     0.000003     0.000020
PLUMED: 4 Calculating (forward loop)                    1001     0.011068     0.000011     0.000011     0.000042
PLUMED: 5 Applying (backward loop)                      1001     0.015032     0.000015     0.000015     0.000028
PLUMED: 6 Update                                        1001     0.009298     0.000009     0.000008     0.000060

+++ Loading the PLUMED kernel runtime +++
+++ PLUMED_KERNEL="/net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/libplumedKernel.so" +++

PLUMED: PLUMED is starting
PLUMED: Version: 2.8.0 (git: ddb59d16b) compiled on May 17 2022 at 19:07:56
PLUMED: Please cite these papers when using PLUMED [1][2]
PLUMED: For further information see the PLUMED web page at http://www.plumed.org
PLUMED: Root: /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed
PLUMED: For installed feature, see /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed/src/config/config.txt
PLUMED: Molecular dynamics engine: driver
PLUMED: Precision of reals: 8
PLUMED: Running over 1 node
PLUMED: Number of threads: 1
PLUMED: Cache line size: 512
PLUMED: Number of atoms: 1591
PLUMED: File suffix: 
PLUMED: FILE: plumed.dat
PLUMED: Action MOLINFO
PLUMED:   with label @0
PLUMED:   pdb file named ../../../ala.pdb contains 1 chains 
PLUMED:   chain named   contains residues 1 to 526 and atoms 1 to 1591 
PLUMED: Action TORSION
PLUMED:   with label phi
PLUMED:   between atoms 5 7 9 15
PLUMED:   using periodic boundary conditions
PLUMED: Action TORSION
PLUMED:   with label psi
PLUMED:   between atoms 7 9 15 17
PLUMED:   using periodic boundary conditions
PLUMED: Action PRINT
PLUMED:   with label @3
PLUMED:   with stride 1
PLUMED:   with arguments phi psi
PLUMED:   on file COLVAR
PLUMED:   with format  %f
PLUMED: END FILE: plumed.dat
PLUMED: Timestep: 1.000000
PLUMED: KbT has not been set by the MD engine
PLUMED: It should be set by hand where needed
PLUMED: Relevant bibliography:
PLUMED:   [1] The PLUMED consortium, Nat. Methods 16, 670 (2019)
PLUMED:   [2] Tribello, Bonomi, Branduardi, Camilloni, and Bussi, Comput. Phys. Commun. 185, 604 (2014)
PLUMED: Please read and cite where appropriate!
PLUMED: Finished setup
PLUMED:                                               Cycles        Total      Average      Minimum      Maximum
PLUMED:                                                    1     0.090897     0.090897     0.090897     0.090897
PLUMED: 1 Prepare dependencies                          1001     0.002092     0.000002     0.000002     0.000011
PLUMED: 2 Sharing data                                  1001     0.010071     0.000010     0.000009     0.000255
PLUMED: 3 Waiting for data                              1001     0.003381     0.000003     0.000003     0.000021
PLUMED: 4 Calculating (forward loop)                    1001     0.011126     0.000011     0.000010     0.000062
PLUMED: 5 Applying (backward loop)                      1001     0.014544     0.000015     0.000014     0.000029
PLUMED: 6 Update                                        1001     0.009246     0.000009     0.000008     0.000057
PLUMED: PLUMED is starting
PLUMED: Version: 2.8.0 (git: ddb59d16b) compiled on May 17 2022 at 19:07:56
PLUMED: Please cite these papers when using PLUMED [1][2]
PLUMED: For further information see the PLUMED web page at http://www.plumed.org
PLUMED: Root: /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed
PLUMED: For installed feature, see /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed/src/config/config.txt
PLUMED: Molecular dynamics engine: driver
PLUMED: Precision of reals: 8
PLUMED: Running over 1 node
PLUMED: Number of threads: 1
PLUMED: Cache line size: 512
PLUMED: Number of atoms: 1591
PLUMED: File suffix: 
PLUMED: FILE: plumed.dat
PLUMED: Action MOLINFO
PLUMED:   with label @0
PLUMED:   pdb file named ../../../ala.pdb contains 1 chains 
PLUMED:   chain named   contains residues 1 to 526 and atoms 1 to 1591 
PLUMED: Action TORSION
PLUMED:   with label phi
PLUMED:   between atoms 5 7 9 15
PLUMED:   using periodic boundary conditions
PLUMED: Action TORSION
PLUMED:   with label psi
PLUMED:   between atoms 7 9 15 17
PLUMED:   using periodic boundary conditions
PLUMED: Action PRINT
PLUMED:   with label @3
PLUMED:   with stride 1
PLUMED:   with arguments phi psi
PLUMED:   on file COLVAR
PLUMED:   with format  %f
PLUMED: END FILE: plumed.dat
PLUMED: Timestep: 1.000000
PLUMED: KbT has not been set by the MD engine
PLUMED: It should be set by hand where needed
PLUMED: Relevant bibliography:
PLUMED:   [1] The PLUMED consortium, Nat. Methods 16, 670 (2019)
PLUMED:   [2] Tribello, Bonomi, Branduardi, Camilloni, and Bussi, Comput. Phys. Commun. 185, 604 (2014)
PLUMED: Please read and cite where appropriate!
PLUMED: Finished setup

+++ Loading the PLUMED kernel runtime +++
+++ PLUMED_KERNEL="/net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/libplumedKernel.so" +++

PLUMED:                                               Cycles        Total      Average      Minimum      Maximum
PLUMED:                                                    1     0.090466     0.090466     0.090466     0.090466
PLUMED: 1 Prepare dependencies                          1001     0.001991     0.000002     0.000001     0.000011
PLUMED: 2 Sharing data                                  1001     0.010066     0.000010     0.000009     0.000255
PLUMED: 3 Waiting for data                              1001     0.003405     0.000003     0.000003     0.000021
PLUMED: 4 Calculating (forward loop)                    1001     0.011021     0.000011     0.000010     0.000042
PLUMED: 5 Applying (backward loop)                      1001     0.014642     0.000015     0.000014     0.000033
PLUMED: 6 Update                                        1001     0.009120     0.000009     0.000008     0.000056

+++ Loading the PLUMED kernel runtime +++
+++ PLUMED_KERNEL="/net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/libplumedKernel.so" +++

PLUMED: PLUMED is starting
PLUMED: Version: 2.8.0 (git: ddb59d16b) compiled on May 17 2022 at 19:07:56
PLUMED: Please cite these papers when using PLUMED [1][2]
PLUMED: For further information see the PLUMED web page at http://www.plumed.org
PLUMED: Root: /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed
PLUMED: For installed feature, see /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed/src/config/config.txt
PLUMED: Molecular dynamics engine: driver
PLUMED: Precision of reals: 8
PLUMED: Running over 1 node
PLUMED: Number of threads: 1
PLUMED: Cache line size: 512
PLUMED: Number of atoms: 1591
PLUMED: File suffix: 
PLUMED: FILE: plumed.dat
PLUMED: Action MOLINFO
PLUMED:   with label @0
PLUMED:   pdb file named ../../../ala.pdb contains 1 chains 
PLUMED:   chain named   contains residues 1 to 526 and atoms 1 to 1591 
PLUMED: Action TORSION
PLUMED:   with label phi
PLUMED:   between atoms 5 7 9 15
PLUMED:   using periodic boundary conditions
PLUMED: Action TORSION
PLUMED:   with label psi
PLUMED:   between atoms 7 9 15 17
PLUMED:   using periodic boundary conditions
PLUMED: Action PRINT
PLUMED:   with label @3
PLUMED:   with stride 1
PLUMED:   with arguments phi psi
PLUMED:   on file COLVAR
PLUMED:   with format  %f
PLUMED: END FILE: plumed.dat
PLUMED: Timestep: 1.000000
PLUMED: KbT has not been set by the MD engine
PLUMED: It should be set by hand where needed
PLUMED: Relevant bibliography:
PLUMED:   [1] The PLUMED consortium, Nat. Methods 16, 670 (2019)
PLUMED:   [2] Tribello, Bonomi, Branduardi, Camilloni, and Bussi, Comput. Phys. Commun. 185, 604 (2014)
PLUMED: Please read and cite where appropriate!
PLUMED: Finished setup
PLUMED:                                               Cycles        Total      Average      Minimum      Maximum
PLUMED:                                                    1     0.094982     0.094982     0.094982     0.094982
PLUMED: 1 Prepare dependencies                          1001     0.002018     0.000002     0.000002     0.000012
PLUMED: 2 Sharing data                                  1001     0.011020     0.000011     0.000009     0.000254
PLUMED: 3 Waiting for data                              1001     0.003375     0.000003     0.000003     0.000030
PLUMED: 4 Calculating (forward loop)                    1001     0.012080     0.000012     0.000011     0.000043
PLUMED: 5 Applying (backward loop)                      1001     0.016923     0.000017     0.000014     0.000033
PLUMED: 6 Update                                        1001     0.008946     0.000009     0.000008     0.000058

+++ Loading the PLUMED kernel runtime +++
+++ PLUMED_KERNEL="/net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/libplumedKernel.so" +++

for i in range(len(Ts)):
    plt.plot(colvar[i].time,colvar[i].phi,".")

plt.plot(colvar[i].time,colvar[31].phi,".")

[<matplotlib.lines.Line2D at 0x7efc04d130d0>]

pop_pt=[]
for i in range(len(Ts)):
    pop_pt.append(
        np.sum(np.logical_and(colvar[i].phi<2.0,colvar[i].phi>0))/
        np.sum(np.logical_or(colvar[i].phi>2.0,colvar[i].phi<0))
    )

plt.plot(Ts,pop_pt,"x-")
plt.xlabel("T")
plt.ylabel("relative population")
plt.show()

Exercise 2a: Setting up scaled Hamiltonians¶

We now learn how to use the plumed partial_tempering tool to prepared scaled Hamiltonians. Notice that you could do the same using some other tool or editing by hand the top file.

The first thing that we have to do is to construct a self-contained top file

with cd("2a"):
    run_mpi("gmx_mpi grompp -p ../topol.top -f ../grompp.mdp -c ../conf.gro -pp processed.top")

               :-) GROMACS - gmx grompp, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx grompp, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/2a
Command line:
  gmx_mpi grompp -p ../topol.top -f ../grompp.mdp -c ../conf.gro -pp processed.top

Setting the LD random seed to 2143280311

Back Off! I just backed up processed.top to ./#processed.top.1#
Generated 2145 of the 2145 non-bonded parameter combinations
Generating 1-4 interactions: fudge = 0.5
Generated 2145 of the 2145 1-4 parameter combinations
Excluding 3 bonded neighbours molecule type 'Protein'
Excluding 2 bonded neighbours molecule type 'SOL'

turning H bonds into constraints...
turning H bonds into constraints...
Analysing residue names:
There are:     3    Protein residues
There are:   523      Water residues
Analysing Protein...
Determining Verlet buffer for a tolerance of 0.005 kJ/mol/ps at 300 K

Number of degrees of freedom in T-Coupling group Protein is 53.95
Number of degrees of freedom in T-Coupling group SOL is 3135.05
Estimate for the relative computational load of the PME mesh part: 0.19

Back Off! I just backed up topol.tpr to ./#topol.tpr.1#

GROMACS reminds you: "This Puke Stinks Like Beer" (LIVE)

Calculated rlist for 1x1 atom pair-list as 1.036 nm, buffer size 0.036 nm
Set rlist, assuming 4x4 atom pair-list, to 1.000 nm, buffer size 0.000 nm
Note that mdrun will redetermine rlist based on the actual pair-list setup
Calculating fourier grid dimensions for X Y Z
Using a fourier grid of 24x24x24, spacing 0.118 0.118 0.118
This run will generate roughly 12 Mb of data

Then we can edit it. Usually this can be done with a text editor. However, if you want to batch process many systems, or do it in a notebook as here, you can use the following script:

with cd("2a/"):
    with open("processed.top") as f:
        with open("processed_selected.top","w") as ff:
            in_molecules=False
            in_atoms=False
            ala_done=False
            for l in f:
                if(in_atoms) and not re.match(" *;.*",l) and not re.match(r"\[.*",l) and len(l.split())>2:
                    ll=l.split()
                    ll[1]=ll[1]+"_"
                    ll=" ".join(ll)
                    print(ll,file=ff)
                else:
                    print(l,end="",file=ff)
                if not ala_done and re.match(r"\[ *moleculetype *\]",l):
                    in_molecules=True
                    ala_done=True # track when alanine has been done so that we only modify that
                elif in_molecules and re.match(r"\[ *atoms *\]",l):
                    in_atoms=True
                elif re.match(r"\[.*",l):
                    in_molecules=False
                    in_atoms=False
    # then we generate two scaled topologies with factors 1.0 and 0.5
    run_mpi("plumed partial_tempering 1.0 < processed_selected.top > scaled1.0.top")
    run_mpi("plumed partial_tempering 0.5 < processed_selected.top > scaled0.5.top")

We can see the differences between the two scaled topologies:

New atom types have been added (e.g., C_)
Charges have been modified
All torsions have been put "inline". No generic HC CT C O torsions, but specific 1 2 5 6 torsions. This allows to scale only portions of a molecule.

!diff 2a/scaled1.0.top 2a/scaled0.5.top

40c40
< Br_ Br 79.90 0.0000 A 0.395559 1.33888  ; scaled
---
> Br_ Br 79.90 0.0000 A 0.395559 0.66944  ; scaled
42c42
< C_ C 12.01 0.0000 A 0.339967 0.359824  ; scaled
---
> C_ C 12.01 0.0000 A 0.339967 0.179912  ; scaled
44c44
< CA_ CA 12.01 0.0000 A 0.339967 0.359824  ; scaled
---
> CA_ CA 12.01 0.0000 A 0.339967 0.179912  ; scaled
46c46
< CB_ CB 12.01 0.0000 A 0.339967 0.359824  ; scaled
---
> CB_ CB 12.01 0.0000 A 0.339967 0.179912  ; scaled
48c48
< CC_ CC 12.01 0.0000 A 0.339967 0.359824  ; scaled
---
> CC_ CC 12.01 0.0000 A 0.339967 0.179912  ; scaled
50c50
< CK_ CK 12.01 0.0000 A 0.339967 0.359824  ; scaled
---
> CK_ CK 12.01 0.0000 A 0.339967 0.179912  ; scaled
52c52
< CM_ CM 12.01 0.0000 A 0.339967 0.359824  ; scaled
---
> CM_ CM 12.01 0.0000 A 0.339967 0.179912  ; scaled
54c54
< CN_ CN 12.01 0.0000 A 0.339967 0.359824  ; scaled
---
> CN_ CN 12.01 0.0000 A 0.339967 0.179912  ; scaled
56c56
< CQ_ CQ 12.01 0.0000 A 0.339967 0.359824  ; scaled
---
> CQ_ CQ 12.01 0.0000 A 0.339967 0.179912  ; scaled
58c58
< CR_ CR 12.01 0.0000 A 0.339967 0.359824  ; scaled
---
> CR_ CR 12.01 0.0000 A 0.339967 0.179912  ; scaled
60c60
< CT_ CT 12.01 0.0000 A 0.339967 0.45773  ; scaled
---
> CT_ CT 12.01 0.0000 A 0.339967 0.228865  ; scaled
62c62
< CV_ CV 12.01 0.0000 A 0.339967 0.359824  ; scaled
---
> CV_ CV 12.01 0.0000 A 0.339967 0.179912  ; scaled
64c64
< CW_ CW 12.01 0.0000 A 0.339967 0.359824  ; scaled
---
> CW_ CW 12.01 0.0000 A 0.339967 0.179912  ; scaled
66c66
< C*_ C* 12.01 0.0000 A 0.339967 0.359824  ; scaled
---
> C*_ C* 12.01 0.0000 A 0.339967 0.179912  ; scaled
68c68
< C0_ C0 40.08 0.0000 A 0.30524 1.92376  ; scaled
---
> C0_ C0 40.08 0.0000 A 0.30524 0.96188  ; scaled
70c70
< F_ F 19.00 0.0000 A 0.311815 0.255224  ; scaled
---
> F_ F 19.00 0.0000 A 0.311815 0.127612  ; scaled
72c72
< H_ H 1.008 0.0000 A 0.106908 0.0656888  ; scaled
---
> H_ H 1.008 0.0000 A 0.106908 0.0328444  ; scaled
74c74
< HC_ HC 1.008 0.0000 A 0.264953 0.0656888  ; scaled
---
> HC_ HC 1.008 0.0000 A 0.264953 0.0328444  ; scaled
76c76
< H1_ H1 1.008 0.0000 A 0.247135 0.0656888  ; scaled
---
> H1_ H1 1.008 0.0000 A 0.247135 0.0328444  ; scaled
78c78
< H2_ H2 1.008 0.0000 A 0.229317 0.0656888  ; scaled
---
> H2_ H2 1.008 0.0000 A 0.229317 0.0328444  ; scaled
80c80
< H3_ H3 1.008 0.0000 A 0.211499 0.0656888  ; scaled
---
> H3_ H3 1.008 0.0000 A 0.211499 0.0328444  ; scaled
82c82
< HA_ HA 1.008 0.0000 A 0.259964 0.06276  ; scaled
---
> HA_ HA 1.008 0.0000 A 0.259964 0.03138  ; scaled
84c84
< H4_ H4 1.008 0.0000 A 0.251055 0.06276  ; scaled
---
> H4_ H4 1.008 0.0000 A 0.251055 0.03138  ; scaled
86c86
< H5_ H5 1.008 0.0000 A 0.242146 0.06276  ; scaled
---
> H5_ H5 1.008 0.0000 A 0.242146 0.03138  ; scaled
90c90
< HS_ HS 1.008 0.0000 A 0.106908 0.0656888  ; scaled
---
> HS_ HS 1.008 0.0000 A 0.106908 0.0328444  ; scaled
94c94
< HP_ HP 1.008 0.0000 A 0.195998 0.0656888  ; scaled
---
> HP_ HP 1.008 0.0000 A 0.195998 0.0328444  ; scaled
96c96
< I_ I 126.9 0.0000 A 0.418722 1.6736  ; scaled
---
> I_ I 126.9 0.0000 A 0.418722 0.8368  ; scaled
98c98
< Cl_ Cl 35.45 0.0000 A 0.440104 0.4184  ; scaled
---
> Cl_ Cl 35.45 0.0000 A 0.440104 0.2092  ; scaled
100c100
< Na_ Na 22.99 0.0000 A 0.33284 0.0115897  ; scaled
---
> Na_ Na 22.99 0.0000 A 0.33284 0.00579485  ; scaled
102c102
< IB_ IB 131.0 0.0000 A 0.890899 0.4184  ; scaled
---
> IB_ IB 131.0 0.0000 A 0.890899 0.2092  ; scaled
104c104
< MG_ MG 24.305 0.0000 A 0.141225 3.74342  ; scaled
---
> MG_ MG 24.305 0.0000 A 0.141225 1.87171  ; scaled
106c106
< N_ N 14.01 0.0000 A 0.325 0.71128  ; scaled
---
> N_ N 14.01 0.0000 A 0.325 0.35564  ; scaled
108c108
< NA_ NA 14.01 0.0000 A 0.325 0.71128  ; scaled
---
> NA_ NA 14.01 0.0000 A 0.325 0.35564  ; scaled
110c110
< NB_ NB 14.01 0.0000 A 0.325 0.71128  ; scaled
---
> NB_ NB 14.01 0.0000 A 0.325 0.35564  ; scaled
112c112
< NC_ NC 14.01 0.0000 A 0.325 0.71128  ; scaled
---
> NC_ NC 14.01 0.0000 A 0.325 0.35564  ; scaled
114c114
< N2_ N2 14.01 0.0000 A 0.325 0.71128  ; scaled
---
> N2_ N2 14.01 0.0000 A 0.325 0.35564  ; scaled
116c116
< N3_ N3 14.01 0.0000 A 0.325 0.71128  ; scaled
---
> N3_ N3 14.01 0.0000 A 0.325 0.35564  ; scaled
118c118
< N*_ N* 14.01 0.0000 A 0.325 0.71128  ; scaled
---
> N*_ N* 14.01 0.0000 A 0.325 0.35564  ; scaled
120c120
< O_ O 16.00 0.0000 A 0.295992 0.87864  ; scaled
---
> O_ O 16.00 0.0000 A 0.295992 0.43932  ; scaled
122c122
< OW_ OW 16.00 0.0000 A 0.315061 0.636386  ; scaled
---
> OW_ OW 16.00 0.0000 A 0.315061 0.318193  ; scaled
124c124
< OH_ OH 16.00 0.0000 A 0.306647 0.880314  ; scaled
---
> OH_ OH 16.00 0.0000 A 0.306647 0.440157  ; scaled
126c126
< OS_ OS 16.00 0.0000 A 0.300001 0.71128  ; scaled
---
> OS_ OS 16.00 0.0000 A 0.300001 0.35564  ; scaled
128c128
< O2_ O2 16.00 0.0000 A 0.295992 0.87864  ; scaled
---
> O2_ O2 16.00 0.0000 A 0.295992 0.43932  ; scaled
130c130
< P_ P 30.97 0.0000 A 0.374177 0.8368  ; scaled
---
> P_ P 30.97 0.0000 A 0.374177 0.4184  ; scaled
132c132
< S_ S 32.06 0.0000 A 0.356359 1.046  ; scaled
---
> S_ S 32.06 0.0000 A 0.356359 0.523  ; scaled
134c134
< SH_ SH 32.06 0.0000 A 0.356359 1.046  ; scaled
---
> SH_ SH 32.06 0.0000 A 0.356359 0.523  ; scaled
136c136
< CU_ CU 63.55 0.0000 A 0.339967 0.359824  ; scaled
---
> CU_ CU 63.55 0.0000 A 0.339967 0.179912  ; scaled
138c138
< K_ K 39.10 0.0000 A 0.473602 0.00137235  ; scaled
---
> K_ K 39.10 0.0000 A 0.473602 0.000686175  ; scaled
140c140
< Rb_ Rb 85.47 0.0000 A 0.526699 0.00071128  ; scaled
---
> Rb_ Rb 85.47 0.0000 A 0.526699 0.00035564  ; scaled
142c142
< Cs_ Cs 132.91 0.0000 A 0.60492 0.00033723  ; scaled
---
> Cs_ Cs 132.91 0.0000 A 0.60492 0.000168615  ; scaled
145c145
< OW_spc_ OW_spc 15.9994 0.0000 A 0.316557 0.650629  ; scaled
---
> OW_spc_ OW_spc 15.9994 0.0000 A 0.316557 0.325315  ; scaled
149c149
< Li_ Li 6.94 0.0000 A 0.20259 0.0765672  ; scaled
---
> Li_ Li 6.94 0.0000 A 0.20259 0.0382836  ; scaled
151c151
< Zn_ Zn 65.4 0.0000 A 0.195998 0.0523  ; scaled
---
> Zn_ Zn 65.4 0.0000 A 0.195998 0.02615  ; scaled
156c156
< OW_tip4pew_ OW_tip4pew 16.00 0.0000 A 0.316435 0.680946  ; scaled
---
> OW_tip4pew_ OW_tip4pew 16.00 0.0000 A 0.316435 0.340473  ; scaled
161c161
< OW_tip4p_ OW_tip4p 16.00 0.0000 A 0.315365 0.64852  ; scaled
---
> OW_tip4p_ OW_tip4p 16.00 0.0000 A 0.315365 0.32426  ; scaled
166c166
< OW_tip5p_ OW_tip5p 16.00 0.0000 A 0.312 0.66944  ; scaled
---
> OW_tip5p_ OW_tip5p 16.00 0.0000 A 0.312 0.33472  ; scaled
736,741c736,741
< 1 HC_ 1 ACE HH31 1 0.1123 1.008 
< 2 CT_ 1 ACE CH3 2 -0.3662 12.01 
< 3 HC_ 1 ACE HH32 3 0.1123 1.008 
< 4 HC_ 1 ACE HH33 4 0.1123 1.008 
< 5 C_ 1 ACE C 5 0.5972 12.01 
< 6 O_ 1 ACE O 6 -0.5679 16 ; qtot 0
---
> 1 HC_ 1 ACE HH31 1 0.0794081 1.008 
> 2 CT_ 1 ACE CH3 2 -0.258943 12.01 
> 3 HC_ 1 ACE HH32 3 0.0794081 1.008 
> 4 HC_ 1 ACE HH33 4 0.0794081 1.008 
> 5 C_ 1 ACE C 5 0.422284 12.01 
> 6 O_ 1 ACE O 6 -0.401566 16 ; qtot 0
743,752c743,752
< 7 N_ 2 ALA N 7 -0.4157 14.01 
< 8 H_ 2 ALA H 8 0.2719 1.008 
< 9 CT_ 2 ALA CA 9 0.0337 12.01 
< 10 H1_ 2 ALA HA 10 0.0823 1.008 
< 11 CT_ 2 ALA CB 11 -0.1825 12.01 
< 12 HC_ 2 ALA HB1 12 0.0603 1.008 
< 13 HC_ 2 ALA HB2 13 0.0603 1.008 
< 14 HC_ 2 ALA HB3 14 0.0603 1.008 
< 15 C_ 2 ALA C 15 0.5973 12.01 
< 16 O_ 2 ALA O 16 -0.5679 16 ; qtot 0
---
> 7 N_ 2 ALA N 7 -0.293944 14.01 
> 8 H_ 2 ALA H 8 0.192262 1.008 
> 9 CT_ 2 ALA CA 9 0.0238295 12.01 
> 10 H1_ 2 ALA HA 10 0.0581949 1.008 
> 11 CT_ 2 ALA CB 11 -0.129047 12.01 
> 12 HC_ 2 ALA HB1 12 0.0426385 1.008 
> 13 HC_ 2 ALA HB2 13 0.0426385 1.008 
> 14 HC_ 2 ALA HB3 14 0.0426385 1.008 
> 15 C_ 2 ALA C 15 0.422355 12.01 
> 16 O_ 2 ALA O 16 -0.401566 16 ; qtot 0
754,759c754,759
< 17 N_ 3 NME N 17 -0.4157 14.01 
< 18 H_ 3 NME H 18 0.2719 1.008 
< 19 CT_ 3 NME CH3 19 -0.149 12.01 
< 20 H1_ 3 NME HH31 20 0.0976 1.008 
< 21 H1_ 3 NME HH32 21 0.0976 1.008 
< 22 H1_ 3 NME HH33 22 0.0976 1.008 ; qtot 0
---
> 17 N_ 3 NME N 17 -0.293944 14.01 
> 18 H_ 3 NME H 18 0.192262 1.008 
> 19 CT_ 3 NME CH3 19 -0.105359 12.01 
> 20 H1_ 3 NME HH31 20 0.0690136 1.008 
> 21 H1_ 3 NME HH32 21 0.0690136 1.008 
> 22 H1_ 3 NME HH33 22 0.0690136 1.008 ; qtot 0
871,872c871,872
< 1 2 5 6 9 0.0 3.3472 1
< 1 2 5 6 9 180.0 0.33472 3
---
> 1 2 5 6 9 0.0 1.6736 1
> 1 2 5 6 9 180.0 0.16736 3
876,877c876,877
< 3 2 5 6 9 0.0 3.3472 1
< 3 2 5 6 9 180.0 0.33472 3
---
> 3 2 5 6 9 0.0 1.6736 1
> 3 2 5 6 9 180.0 0.16736 3
881,882c881,882
< 4 2 5 6 9 0.0 3.3472 1
< 4 2 5 6 9 180.0 0.33472 3
---
> 4 2 5 6 9 0.0 1.6736 1
> 4 2 5 6 9 180.0 0.16736 3
886c886
< 2 5 7 8 9 180.0 10.46 2
---
> 2 5 7 8 9 180.0 5.23 2
888c888
< 2 5 7 9 9 180.0 10.46 2
---
> 2 5 7 9 9 180.0 5.23 2
890,891c890,891
< 6 5 7 8 9 180.0 10.46 2
< 6 5 7 8 9 0.0 8.368 1
---
> 6 5 7 8 9 180.0 5.23 2
> 6 5 7 8 9 0.0 4.184 1
893c893
< 6 5 7 9 9 180.0 10.46 2
---
> 6 5 7 9 9 180.0 5.23 2
897,899c897,899
< 5 7 9 11 9 0.0 8.368 1
< 5 7 9 11 9 0.0 8.368 2
< 5 7 9 11 9 0.0 1.6736 3
---
> 5 7 9 11 9 0.0 4.184 1
> 5 7 9 11 9 0.0 4.184 2
> 5 7 9 11 9 0.0 0.8368 3
901,902c901,902
< 5 7 9 15 9 0.0 1.12968 2
< 5 7 9 15 9 0.0 1.75728 3
---
> 5 7 9 15 9 0.0 0.56484 2
> 5 7 9 15 9 0.0 0.87864 3
910c910
< 7 9 11 12 9 0.0 0.65084 3
---
> 7 9 11 12 9 0.0 0.32542 3
912c912
< 7 9 11 13 9 0.0 0.65084 3
---
> 7 9 11 13 9 0.0 0.32542 3
914c914
< 7 9 11 14 9 0.0 0.65084 3
---
> 7 9 11 14 9 0.0 0.32542 3
916c916
< 10 9 11 12 9 0.0 0.65084 3
---
> 10 9 11 12 9 0.0 0.32542 3
918c918
< 10 9 11 13 9 0.0 0.65084 3
---
> 10 9 11 13 9 0.0 0.32542 3
920c920
< 10 9 11 14 9 0.0 0.65084 3
---
> 10 9 11 14 9 0.0 0.32542 3
922c922
< 15 9 11 12 9 0.0 0.65084 3
---
> 15 9 11 12 9 0.0 0.32542 3
924c924
< 15 9 11 13 9 0.0 0.65084 3
---
> 15 9 11 13 9 0.0 0.32542 3
926c926
< 15 9 11 14 9 0.0 0.65084 3
---
> 15 9 11 14 9 0.0 0.32542 3
930,932c930,932
< 7 9 15 17 9 180.0 1.8828 1
< 7 9 15 17 9 180.0 6.61072 2
< 7 9 15 17 9 180.0 2.3012 3
---
> 7 9 15 17 9 180.0 0.9414 1
> 7 9 15 17 9 180.0 3.30536 2
> 7 9 15 17 9 180.0 1.1506 3
934,935c934,935
< 10 9 15 16 9 0.0 3.3472 1
< 10 9 15 16 9 180.0 0.33472 3
---
> 10 9 15 16 9 0.0 1.6736 1
> 10 9 15 16 9 180.0 0.16736 3
941,943c941,943
< 11 9 15 17 9 0.0 0.8368 1
< 11 9 15 17 9 0.0 0.8368 2
< 11 9 15 17 9 0.0 1.6736 3
---
> 11 9 15 17 9 0.0 0.4184 1
> 11 9 15 17 9 0.0 0.4184 2
> 11 9 15 17 9 0.0 0.8368 3
945c945
< 9 15 17 18 9 180.0 10.46 2
---
> 9 15 17 18 9 180.0 5.23 2
947c947
< 9 15 17 19 9 180.0 10.46 2
---
> 9 15 17 19 9 180.0 5.23 2
949,950c949,950
< 16 15 17 18 9 180.0 10.46 2
< 16 15 17 18 9 0.0 8.368 1
---
> 16 15 17 18 9 180.0 5.23 2
> 16 15 17 18 9 0.0 4.184 1
952c952
< 16 15 17 19 9 180.0 10.46 2
---
> 16 15 17 19 9 180.0 5.23 2
969c969
< 2 7 5 6 4 180.00 43.932 2
---
> 2 7 5 6 4 180.00 21.966 2
971c971
< 5 9 7 8 4 180.00 4.6024 2
---
> 5 9 7 8 4 180.00 2.3012 2
973c973
< 9 17 15 16 4 180.00 43.932 2
---
> 9 17 15 16 4 180.00 21.966 2
975c975
< 15 19 17 18 4 180.00 4.6024 2
---
> 15 19 17 18 4 180.00 2.3012 2

Exercise 2b: Sanity check on generated topologies¶

We now do some sanity check on the resulting topologies. This is mostly to make sure that the plumed partial_tempering script edited them correctly.

To this aim we generate a short reference trajectory first (it's not important which force field we use at this step).

with cd("2b"):
    run_mpi("gmx_mpi grompp -f ../grompp.mdp -c ../conf.gro -p ../topol.top")
    run_mpi("gmx_mpi mdrun -nsteps 5000")

               :-) GROMACS - gmx grompp, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx grompp, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/2b
Command line:
  gmx_mpi grompp -f ../grompp.mdp -c ../conf.gro -p ../topol.top

Setting the LD random seed to -80025769
Generated 2145 of the 2145 non-bonded parameter combinations
Generating 1-4 interactions: fudge = 0.5
Generated 2145 of the 2145 1-4 parameter combinations
Excluding 3 bonded neighbours molecule type 'Protein'
Excluding 2 bonded neighbours molecule type 'SOL'

turning H bonds into constraints...
turning H bonds into constraints...
Analysing residue names:
There are:     3    Protein residues
There are:   523      Water residues
Analysing Protein...

Number of degrees of freedom in T-Coupling group Protein is 53.95
Number of degrees of freedom in T-Coupling group SOL is 3135.05
Estimate for the relative computational load of the PME mesh part: 0.19

Determining Verlet buffer for a tolerance of 0.005 kJ/mol/ps at 300 K
Calculated rlist for 1x1 atom pair-list as 1.036 nm, buffer size 0.036 nm
Set rlist, assuming 4x4 atom pair-list, to 1.000 nm, buffer size 0.000 nm
Note that mdrun will redetermine rlist based on the actual pair-list setup
Calculating fourier grid dimensions for X Y Z
Using a fourier grid of 24x24x24, spacing 0.118 0.118 0.118
This run will generate roughly 12 Mb of data

Back Off! I just backed up topol.tpr to ./#topol.tpr.1#

GROMACS reminds you: "It all works because Avogadro's number is closer to infinity than to 10." (Ralph Baierlein)

                :-) GROMACS - gmx mdrun, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx mdrun, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/2b
Command line:
  gmx_mpi mdrun -nsteps 5000


Back Off! I just backed up md.log to ./#md.log.1#
Compiled SIMD: SSE2, but for this host/run AVX_256 might be better (see log).
The current CPU can measure timings more accurately than the code in
gmx mdrun was configured to use. This might affect your simulation
speed as accurate timings are needed for load-balancing.
Please consider rebuilding gmx mdrun with the GMX_USE_RDTSCP=ON CMake option.
Reading file topol.tpr, VERSION 2020.6-plumed-2.8.0 (single precision)

Overriding nsteps with value passed on the command line: 5000 steps, 10 ps
Changing nstlist from 10 to 50, rlist from 1 to 1.111


Using 1 MPI process

Non-default thread affinity set, disabling internal thread affinity

Using 24 OpenMP threads 


Back Off! I just backed up traj_comp.xtc to ./#traj_comp.xtc.1#

Back Off! I just backed up ener.edr to ./#ener.edr.1#
starting mdrun 'alanine dipeptide in vacuum in water'
5000 steps,     10.0 ps.

Writing final coordinates.

Back Off! I just backed up confout.gro to ./#confout.gro.1#

NOTE: Detected invalid cycle counts, probably because threads moved between CPU cores that do not have synchronized cycle counters. Will not print the cycle accounting.

               Core t (s)   Wall t (s)        (%)
       Time:     1958.671       81.611     2400.0
                 (ns/day)    (hour/ns)
Performance:       10.589        2.267

GROMACS reminds you: "If I Wanted You to Understand This, I Would Explain it Better" (J. Cruijff)

We then "rerun" its energy with different topologies. Specifically, we use the original tpr file and a new one obtained from topol_scaled1.0.top, which has been obtained with scaling factor 1.0. The two calculations should return identical energies

with cd("2b"):
    run_mpi("cp ../2a/processed_selected.top processed_selected.top")
    run_mpi("plumed partial_tempering 1.0 < processed_selected.top > topol_scaled1.0.top")
    run_mpi("gmx_mpi grompp -f ../grompp.mdp -c ../conf.gro -p topol_scaled1.0.top -o topol_scaled1.0.tpr")
    with cd("rerun0"):
        run_mpi("gmx_mpi mdrun -rerun ../traj_comp.xtc -s ../topol.tpr")
        run_mpi("echo 1 2 3 4 5 6 7 8 9 10 11 | gmx_mpi energy -xvg no")
    with cd("rerun1"):
        run_mpi("gmx_mpi mdrun -rerun ../traj_comp.xtc -s ../topol_scaled1.0.tpr")
        run_mpi("echo 1 2 3 4 5 6 7 8 9 10 11 | gmx_mpi energy -xvg no")

               :-) GROMACS - gmx grompp, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx grompp, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/2b
Command line:
  gmx_mpi grompp -f ../grompp.mdp -c ../conf.gro -p topol_scaled1.0.top -o topol_scaled1.0.tpr

Setting the LD random seed to 2079717087
Generated 8515 of the 8515 non-bonded parameter combinations
Generating 1-4 interactions: fudge = 0.5
Generated 8515 of the 8515 1-4 parameter combinations
Excluding 3 bonded neighbours molecule type 'Protein'
Excluding 2 bonded neighbours molecule type 'SOL'

turning H bonds into constraints...
turning H bonds into constraints...
Analysing residue names:
There are:     3    Protein residues
There are:   523      Water residues
Analysing Protein...
Determining Verlet buffer for a tolerance of 0.005 kJ/mol/ps at 300 K

Number of degrees of freedom in T-Coupling group Protein is 53.95
Number of degrees of freedom in T-Coupling group SOL is 3135.05

Calculated rlist for 1x1 atom pair-list as 1.036 nm, buffer size 0.036 nm
Set rlist, assuming 4x4 atom pair-list, to 1.000 nm, buffer size 0.000 nm
Note that mdrun will redetermine rlist based on the actual pair-list setup
Calculating fourier grid dimensions for X Y Z
Using a fourier grid of 24x24x24, spacing 0.118 0.118 0.118
This run will generate roughly 12 Mb of data

Estimate for the relative computational load of the PME mesh part: 0.19

Back Off! I just backed up topol_scaled1.0.tpr to ./#topol_scaled1.0.tpr.1#

GROMACS reminds you: "There are only two hard things in computer science - cache invalidation, naming things and off-by-one errors." (Anonymous)

                :-) GROMACS - gmx mdrun, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx mdrun, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/2b/rerun0
Command line:
  gmx_mpi mdrun -rerun ../traj_comp.xtc -s ../topol.tpr


Back Off! I just backed up md.log to ./#md.log.1#
Compiled SIMD: SSE2, but for this host/run AVX_256 might be better (see log).
The current CPU can measure timings more accurately than the code in
gmx mdrun was configured to use. This might affect your simulation
speed as accurate timings are needed for load-balancing.
Please consider rebuilding gmx mdrun with the GMX_USE_RDTSCP=ON CMake option.
Reading file ../topol.tpr, VERSION 2020.6-plumed-2.8.0 (single precision)
Changing nstlist from 10 to 50, rlist from 1 to 1.111


Using 1 MPI process

Non-default thread affinity set, disabling internal thread affinity

Using 24 OpenMP threads 


Back Off! I just backed up ener.edr to ./#ener.edr.1#
starting md rerun 'alanine dipeptide in vacuum in water', reading coordinates from input trajectory '../traj_comp.xtc'

Last frame         10 time   10.000   

NOTE: 33 % of the run time was spent in pair search,
      you might want to increase nstlist (this has no effect on accuracy)

               Core t (s)   Wall t (s)        (%)
       Time:        7.318        0.305     2399.0
                 (ns/day)    (hour/ns)
Performance:     2833.047        0.008

GROMACS reminds you: "#define QUESTION ((bb),| !(bb))" (William Shakespeare)

               :-) GROMACS - gmx energy, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx energy, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/2b/rerun0
Command line:
  gmx_mpi energy -xvg no

Opened ener.edr as single precision energy file

Select the terms you want from the following list by
selecting either (part of) the name or the number or a combination.
End your selection with an empty line or a zero.
-------------------------------------------------------------------
  1  Bond             2  Angle            3  Proper-Dih.      4  Improper-Dih. 
  5  LJ-14            6  Coulomb-14       7  LJ-(SR)          8  Disper.-corr. 
  9  Coulomb-(SR)    10  Coul.-recip.    11  Potential     


Back Off! I just backed up energy.xvg to ./#energy.xvg.1#
Last energy frame read 10 time   10.000           

GROMACS reminds you: "#define QUESTION ((bb),| !(bb))" (William Shakespeare)

Statistics over 5001 steps [ 0.0000 through 10.0000 ps ], 11 data sets
All statistics are over 11 points (frames)

Energy                      Average   Err.Est.       RMSD  Tot-Drift
-------------------------------------------------------------------------------
Bond                        12.8501       0.93    5.54295    1.76035  (kJ/mol)
Angle                       31.4149        3.3    9.29962   -14.1766  (kJ/mol)
Proper Dih.                 45.7409        1.7    4.41161   -2.62365  (kJ/mol)
Improper Dih.               2.41675       0.48    1.38529  -0.191024  (kJ/mol)
LJ-14                       13.6088        1.6    4.45048    3.70408  (kJ/mol)
Coulomb-14                  195.498        2.7     6.9361   -11.1147  (kJ/mol)
LJ (SR)                     3321.67         54    169.128    -173.23  (kJ/mol)
Disper. corr.              -188.172          0          0          0  (kJ/mol)
Coulomb (SR)               -24668.2         78    291.718    198.491  (kJ/mol)
Coul. recip.                135.806        3.5    12.0772  0.0296166  (kJ/mol)
Potential                  -21097.4         45    203.431    2.64933  (kJ/mol)

                :-) GROMACS - gmx mdrun, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx mdrun, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/2b/rerun1
Command line:
  gmx_mpi mdrun -rerun ../traj_comp.xtc -s ../topol_scaled1.0.tpr


Back Off! I just backed up md.log to ./#md.log.1#
Compiled SIMD: SSE2, but for this host/run AVX_256 might be better (see log).
The current CPU can measure timings more accurately than the code in
gmx mdrun was configured to use. This might affect your simulation
speed as accurate timings are needed for load-balancing.
Please consider rebuilding gmx mdrun with the GMX_USE_RDTSCP=ON CMake option.
Reading file ../topol_scaled1.0.tpr, VERSION 2020.6-plumed-2.8.0 (single precision)
Changing nstlist from 10 to 50, rlist from 1 to 1.111


Using 1 MPI process

Non-default thread affinity set, disabling internal thread affinity

Using 24 OpenMP threads 


Back Off! I just backed up ener.edr to ./#ener.edr.1#
starting md rerun 'alanine dipeptide in vacuum in water', reading coordinates from input trajectory '../traj_comp.xtc'

Last frame         10 time   10.000   

NOTE: 36 % of the run time was spent in pair search,
      you might want to increase nstlist (this has no effect on accuracy)

               Core t (s)   Wall t (s)        (%)
       Time:        6.654        0.277     2398.9
                 (ns/day)    (hour/ns)
Performance:     3115.312        0.008

GROMACS reminds you: "Take Your Medications and Preparations and Ram It Up Your Snout" (F. Zappa)

Statistics over 5001 steps [ 0.0000 through 10.0000 ps ], 11 data sets
All statistics are over 11 points (frames)

Energy                      Average   Err.Est.       RMSD  Tot-Drift
-------------------------------------------------------------------------------
Bond                        12.8501       0.93    5.54295    1.76035  (kJ/mol)
Angle                       31.4149        3.3    9.29962   -14.1766  (kJ/mol)
Proper Dih.                 45.7409        1.7    4.41161   -2.62365  (kJ/mol)
Improper Dih.               2.41675       0.48    1.38529  -0.191024  (kJ/mol)
LJ-14                       13.6088        1.6    4.45048    3.70408  (kJ/mol)
Coulomb-14                  195.498        2.7     6.9361   -11.1147  (kJ/mol)
LJ (SR)                     3321.67         54    169.128    -173.23  (kJ/mol)
Disper. corr.              -188.172          0          0          0  (kJ/mol)
Coulomb (SR)               -24668.2         78    291.718    198.491  (kJ/mol)
Coul. recip.                135.806        3.5    12.0772  0.0296645  (kJ/mol)
Potential                  -21097.4         45    203.431    2.64933  (kJ/mol)

               :-) GROMACS - gmx energy, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx energy, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/2b/rerun1
Command line:
  gmx_mpi energy -xvg no

Opened ener.edr as single precision energy file

Select the terms you want from the following list by
selecting either (part of) the name or the number or a combination.
End your selection with an empty line or a zero.
-------------------------------------------------------------------
  1  Bond             2  Angle            3  Proper-Dih.      4  Improper-Dih. 
  5  LJ-14            6  Coulomb-14       7  LJ-(SR)          8  Disper.-corr. 
  9  Coulomb-(SR)    10  Coul.-recip.    11  Potential     


Back Off! I just backed up energy.xvg to ./#energy.xvg.1#

Reading energy frame      0 time    0.000         
Reading energy frame      1 time    1.000         
Reading energy frame      2 time    2.000         
Reading energy frame      3 time    3.000         
Reading energy frame      4 time    4.000         
Reading energy frame      5 time    5.000         
Reading energy frame      6 time    6.000         
Reading energy frame      7 time    7.000         
Reading energy frame      8 time    8.000         
Reading energy frame      9 time    9.000         
Reading energy frame     10 time   10.000         
Last energy frame read 10 time   10.000         

GROMACS reminds you: "To survive science you have to become science." (Gerrit Groenhof)

ene=pandas.read_csv("2b/rerun0/energy.xvg",header=None,delim_whitespace=True)
ene_scaled=pandas.read_csv("2b/rerun1/energy.xvg",header=None,delim_whitespace=True)
difference=np.array(ene)-np.array(ene_scaled)
for i in range(difference.shape[1]):
    plt.plot(difference[:,i],label="column {}".format(i))
plt.legend()
plt.show()

Only column 10 (long range Ewald) has some minor difference (1e-5), due to some (unknown?) numerical reason. This is not going to be a problem, since this error is several orders of magnitude lower than kBT

Next we generate a top file where all atoms were selected for scaling. From the statistical mechanics point of view, this is identical to increasing the temperature (indeed, you can show that by properly modifying the time step you should obtain identical trajectories). Here it is interesting because we can check if the energy terms do what we expect.

with cd("2b/"):
    with open("../2a/processed.top") as f:
        with open("processed_selected_whole.top","w") as ff:
            in_molecules=False
            in_atoms=False
            ala_done=False
            for l in f:
                ala_done=False # also modify water
                if(in_atoms) and not re.match(" *;.*",l) and not re.match(r"\[.*",l) and len(l.split())>2:
                    ll=l.split()
                    ll[1]=ll[1]+"_"
                    ll=" ".join(ll)
                    print(ll,file=ff)
                else:
                    print(l,end="",file=ff)
                if not ala_done and re.match(r"\[ *moleculetype *\]",l):
                    in_molecules=True
                    ala_done=True
                elif in_molecules and re.match(r"\[ *atoms *\]",l):
                    in_atoms=True
                elif re.match(r"\[.*",l):
                    in_molecules=False
                    in_atoms=False
    run_mpi("plumed partial_tempering 0.1 < processed_selected_whole.top > topol_scaled_whole0.1.top")
    run_mpi("gmx_mpi grompp -f ../grompp.mdp -c ../conf.gro -p topol_scaled_whole0.1.top -o topol_scaled_whole0.1.tpr")
    with cd("rerun_whole"):
        run_mpi("gmx_mpi mdrun -rerun ../traj_comp.xtc -s ../topol_scaled_whole0.1.tpr")
        run_mpi("echo 1 2 3 4 5 6 7 8 9 10 11 | gmx_mpi energy -xvg no")

               :-) GROMACS - gmx grompp, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx grompp, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/2b
Command line:
  gmx_mpi grompp -f ../grompp.mdp -c ../conf.gro -p topol_scaled_whole0.1.top -o topol_scaled_whole0.1.tpr

Setting the LD random seed to -805658753
Generated 8515 of the 8515 non-bonded parameter combinations
Generating 1-4 interactions: fudge = 0.5
Generated 8515 of the 8515 1-4 parameter combinations
Excluding 3 bonded neighbours molecule type 'Protein'
Excluding 2 bonded neighbours molecule type 'SOL'

turning H bonds into constraints...
turning H bonds into constraints...
Analysing residue names:
There are:     3    Protein residues
There are:   523      Water residues
Analysing Protein...
Determining Verlet buffer for a tolerance of 0.005 kJ/mol/ps at 300 K

Number of degrees of freedom in T-Coupling group Protein is 53.95
Number of degrees of freedom in T-Coupling group SOL is 3135.05
Estimate for the relative computational load of the PME mesh part: 0.19

Back Off! I just backed up topol_scaled_whole0.1.tpr to ./#topol_scaled_whole0.1.tpr.1#

Calculated rlist for 1x1 atom pair-list as 1.000 nm, buffer size 0.000 nm
Set rlist, assuming 4x4 atom pair-list, to 1.000 nm, buffer size 0.000 nm
Note that mdrun will redetermine rlist based on the actual pair-list setup
Calculating fourier grid dimensions for X Y Z
Using a fourier grid of 24x24x24, spacing 0.118 0.118 0.118
This run will generate roughly 12 Mb of data

GROMACS reminds you: "... an excellent man, almost worthy of such a wife ..." (Jane Eyre in Jane Eyre by Charlotte Bronte)

                :-) GROMACS - gmx mdrun, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx mdrun, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/2b/rerun_whole
Command line:
  gmx_mpi mdrun -rerun ../traj_comp.xtc -s ../topol_scaled_whole0.1.tpr


Back Off! I just backed up md.log to ./#md.log.1#
Compiled SIMD: SSE2, but for this host/run AVX_256 might be better (see log).
The current CPU can measure timings more accurately than the code in
gmx mdrun was configured to use. This might affect your simulation
speed as accurate timings are needed for load-balancing.
Please consider rebuilding gmx mdrun with the GMX_USE_RDTSCP=ON CMake option.
Reading file ../topol_scaled_whole0.1.tpr, VERSION 2020.6-plumed-2.8.0 (single precision)
Changing nstlist from 10 to 100, rlist from 1 to 1


Using 1 MPI process

Non-default thread affinity set, disabling internal thread affinity

Using 24 OpenMP threads 


Back Off! I just backed up ener.edr to ./#ener.edr.1#
starting md rerun 'alanine dipeptide in vacuum in water', reading coordinates from input trajectory '../traj_comp.xtc'

Last frame         10 time   10.000   

NOTE: 36 % of the run time was spent in pair search,
      you might want to increase nstlist (this has no effect on accuracy)

               Core t (s)   Wall t (s)        (%)
       Time:        5.228        0.218     2399.0
                 (ns/day)    (hour/ns)
Performance:     3965.059        0.006

GROMACS reminds you: "I Live the Life They Wish They Did" (Tricky)

Statistics over 5001 steps [ 0.0000 through 10.0000 ps ], 11 data sets
All statistics are over 11 points (frames)

Energy                      Average   Err.Est.       RMSD  Tot-Drift
-------------------------------------------------------------------------------
Bond                        12.8501       0.93    5.54295    1.76035  (kJ/mol)
Angle                       31.4149        3.3    9.29962   -14.1766  (kJ/mol)
Proper Dih.                 4.57409       0.17   0.441161  -0.262365  (kJ/mol)
Improper Dih.              0.241675      0.048   0.138529 -0.0191024  (kJ/mol)
LJ-14                       1.36088       0.16   0.445048   0.370408  (kJ/mol)
Coulomb-14                  19.5498       0.27   0.693611   -1.11147  (kJ/mol)
LJ (SR)                     332.167        5.4    16.9128   -17.3229  (kJ/mol)
Disper. corr.              -18.8172          0          0          0  (kJ/mol)
Coulomb (SR)               -2466.83        7.8    29.1721    19.8498  (kJ/mol)
Coul. recip.                13.5806       0.35    1.20772  0.0029531  (kJ/mol)
Potential                   -2069.9        6.2    26.4716   -10.9089  (kJ/mol)

               :-) GROMACS - gmx energy, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx energy, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/2b/rerun_whole
Command line:
  gmx_mpi energy -xvg no

Opened ener.edr as single precision energy file

Select the terms you want from the following list by
selecting either (part of) the name or the number or a combination.
End your selection with an empty line or a zero.
-------------------------------------------------------------------
  1  Bond             2  Angle            3  Proper-Dih.      4  Improper-Dih. 
  5  LJ-14            6  Coulomb-14       7  LJ-(SR)          8  Disper.-corr. 
  9  Coulomb-(SR)    10  Coul.-recip.    11  Potential     


Back Off! I just backed up energy.xvg to ./#energy.xvg.1#

Reading energy frame      0 time    0.000         
Reading energy frame      1 time    1.000         
Reading energy frame      2 time    2.000         
Reading energy frame      3 time    3.000         
Reading energy frame      4 time    4.000         
Reading energy frame      5 time    5.000         
Reading energy frame      6 time    6.000         
Reading energy frame      7 time    7.000         
Reading energy frame      8 time    8.000         
Reading energy frame      9 time    9.000         
Reading energy frame     10 time   10.000         
Last energy frame read 10 time   10.000         

GROMACS reminds you: "I Live the Life They Wish They Did" (Tricky)

ene_whole_scaled=pandas.read_csv("2b/rerun_whole/energy.xvg",header=None,delim_whitespace=True)

for i in range(1,12):
    print(ene[i]/ene_whole_scaled[i])

0     1.0
1     1.0
2     1.0
3     1.0
4     1.0
5     1.0
6     1.0
7     1.0
8     1.0
9     1.0
10    1.0
Name: 1, dtype: float64
0     1.0
1     1.0
2     1.0
3     1.0
4     1.0
5     1.0
6     1.0
7     1.0
8     1.0
9     1.0
10    1.0
Name: 2, dtype: float64
0     10.000000
1     10.000000
2     10.000000
3      9.999999
4     10.000001
5      9.999998
6      9.999999
7     10.000001
8     10.000001
9     10.000001
10    10.000000
Name: 3, dtype: float64
0     10.000004
1     10.000042
2     10.000014
3     10.000012
4      9.999997
5      9.999916
6     10.000000
7     10.000008
8      9.999977
9      9.999989
10     9.999997
Name: 4, dtype: float64
0      9.999994
1      9.999998
2     10.000001
3     10.000004
4     10.000002
5      9.999996
6     10.000000
7      9.999999
8      9.999999
9     10.000001
10    10.000000
Name: 5, dtype: float64
0     10.000010
1     10.000011
2     10.000012
3     10.000014
4     10.000009
5     10.000010
6     10.000008
7     10.000012
8     10.000014
9     10.000012
10    10.000011
Name: 6, dtype: float64
0     10.000000
1     10.000003
2     10.000003
3     10.000002
4     10.000003
5     10.000002
6      9.999999
7      9.999999
8     10.000002
9     10.000000
10    10.000004
Name: 7, dtype: float64
0     10.000001
1     10.000001
2     10.000001
3     10.000001
4     10.000001
5     10.000001
6     10.000001
7     10.000001
8     10.000001
9     10.000001
10    10.000001
Name: 8, dtype: float64
0     9.999997
1     9.999991
2     9.999991
3     9.999992
4     9.999994
5     9.999996
6     9.999996
7     9.999995
8     9.999992
9     9.999996
10    9.999996
Name: 9, dtype: float64
0      9.999997
1     10.000009
2      9.999993
3     10.000005
4     10.000014
5     10.000005
6      9.999998
7     10.000008
8     10.000012
9      9.999997
10    10.000013
Name: 10, dtype: float64
0     10.246152
1     10.284829
2     10.157962
3     10.135779
4     10.171296
5     10.190454
6     10.229319
7     10.196831
8     10.196758
9     10.109526
10    10.203299
Name: 11, dtype: float64

All energy terms have been scaled down by a factor 10, with the exception of bonds and bends that were not touched.

Notice that last column is total potential energy, so it contains both unscaled and scaled terms.

Exercise 2c: Sanity check on replica-exchange implementation¶

Once we know the topologies are fine, we should double check if the gromacs+plumed implementation is correctly calculating the acceptance.

We run a replica exchange simulation with 2 replicas with identical Hamiltonians. Acceptance shoud be 1.

with cd("2c/"):
    with open("plumed.dat","w") as f:
        print("",file=f) # empty
    run_mpi("cp ../2b/topol_scaled1.0.top .")
    with cd ("rep0"):
        run_mpi("gmx_mpi grompp -c ../../conf.gro -f ../../grompp.mdp -p ../../topol.top")
    with cd ("rep1"):
        run_mpi("gmx_mpi grompp -c ../../conf.gro -f ../../grompp.mdp -p ../topol_scaled1.0.top")
    run_mpi("gmx_mpi mdrun -nsteps 1000 -multidir rep? -replex 200 -hrex -plumed ../plumed.dat",2)

               :-) GROMACS - gmx grompp, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx grompp, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/2c/rep0
Command line:
  gmx_mpi grompp -c ../../conf.gro -f ../../grompp.mdp -p ../../topol.top

Setting the LD random seed to 1069470717
Generated 2145 of the 2145 non-bonded parameter combinations
Generating 1-4 interactions: fudge = 0.5
Generated 2145 of the 2145 1-4 parameter combinations
Excluding 3 bonded neighbours molecule type 'Protein'
Excluding 2 bonded neighbours molecule type 'SOL'

turning H bonds into constraints...
turning H bonds into constraints...
Analysing residue names:
There are:     3    Protein residues
There are:   523      Water residues
Analysing Protein...
Determining Verlet buffer for a tolerance of 0.005 kJ/mol/ps at 300 K

Number of degrees of freedom in T-Coupling group Protein is 53.95
Number of degrees of freedom in T-Coupling group SOL is 3135.05

Calculated rlist for 1x1 atom pair-list as 1.036 nm, buffer size 0.036 nm
Set rlist, assuming 4x4 atom pair-list, to 1.000 nm, buffer size 0.000 nm
Note that mdrun will redetermine rlist based on the actual pair-list setup
Calculating fourier grid dimensions for X Y Z
Using a fourier grid of 24x24x24, spacing 0.118 0.118 0.118
This run will generate roughly 12 Mb of data

Estimate for the relative computational load of the PME mesh part: 0.19

Back Off! I just backed up topol.tpr to ./#topol.tpr.1#

GROMACS reminds you: "I have had my results for a long time, but I do not yet know how I am to arrive at them." (Carl Friedrich Gauss)

               :-) GROMACS - gmx grompp, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx grompp, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/2c/rep1
Command line:
  gmx_mpi grompp -c ../../conf.gro -f ../../grompp.mdp -p ../topol_scaled1.0.top

Setting the LD random seed to -1511515254
Generated 8515 of the 8515 non-bonded parameter combinations
Generating 1-4 interactions: fudge = 0.5
Generated 8515 of the 8515 1-4 parameter combinations
Excluding 3 bonded neighbours molecule type 'Protein'
Excluding 2 bonded neighbours molecule type 'SOL'

turning H bonds into constraints...
turning H bonds into constraints...
Analysing residue names:
There are:     3    Protein residues
There are:   523      Water residues
Analysing Protein...
Determining Verlet buffer for a tolerance of 0.005 kJ/mol/ps at 300 K

Number of degrees of freedom in T-Coupling group Protein is 53.95
Number of degrees of freedom in T-Coupling group SOL is 3135.05
Estimate for the relative computational load of the PME mesh part: 0.19

Back Off! I just backed up topol.tpr to ./#topol.tpr.1#

Calculated rlist for 1x1 atom pair-list as 1.036 nm, buffer size 0.036 nm
Set rlist, assuming 4x4 atom pair-list, to 1.000 nm, buffer size 0.000 nm
Note that mdrun will redetermine rlist based on the actual pair-list setup
Calculating fourier grid dimensions for X Y Z
Using a fourier grid of 24x24x24, spacing 0.118 0.118 0.118
This run will generate roughly 12 Mb of data

GROMACS reminds you: "Those people who think they know everything are a great annoyance to those of us who do." (Isaac Asimov)

                :-) GROMACS - gmx mdrun, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx mdrun, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/2c
Command line:
  gmx_mpi mdrun -nsteps 1000 -multidir rep0 rep1 -replex 200 -hrex -plumed ../plumed.dat

+++ Loading the PLUMED kernel runtime +++
+++ PLUMED_KERNEL="/net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/libplumedKernel.so" +++
+++ Loading the PLUMED kernel runtime +++
+++ PLUMED_KERNEL="/net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/libplumedKernel.so" +++
+++ Loading the PLUMED kernel runtime +++
+++ PLUMED_KERNEL="/net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/libplumedKernel.so" +++
+++ Loading the PLUMED kernel runtime +++
+++ PLUMED_KERNEL="/net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/libplumedKernel.so" +++

Back Off! I just backed up md.log to ./#md.log.1#

Back Off! I just backed up md.log to ./#md.log.1#
Compiled SIMD: SSE2, but for this host/run AVX_256 might be better (see log).
The current CPU can measure timings more accurately than the code in
gmx mdrun was configured to use. This might affect your simulation
speed as accurate timings are needed for load-balancing.
Please consider rebuilding gmx mdrun with the GMX_USE_RDTSCP=ON CMake option.
The current CPU can measure timings more accurately than the code in
gmx mdrun was configured to use. This might affect your simulation
speed as accurate timings are needed for load-balancing.
Please consider rebuilding gmx mdrun with the GMX_USE_RDTSCP=ON CMake option.
Reading file topol.tpr, VERSION 2020.6-plumed-2.8.0 (single precision)
Reading file topol.tpr, VERSION 2020.6-plumed-2.8.0 (single precision)

Overriding nsteps with value passed on the command line: 1000 steps, 2 ps

Overriding nsteps with value passed on the command line: 1000 steps, 2 ps
Changing nstlist from 10 to 50, rlist from 1 to 1.111

Changing nstlist from 10 to 50, rlist from 1 to 1.111


This is simulation 0 out of 2 running as a composite GROMACS
multi-simulation job. Setup for this simulation:

Using 1 MPI process

This is simulation 1 out of 2 running as a composite GROMACS
multi-simulation job. Setup for this simulation:

Using 1 MPI process

Non-default thread affinity set, disabling internal thread affinity

Non-default thread affinity set, disabling internal thread affinity

Using 12 OpenMP threads 


Using 12 OpenMP threads 


Back Off! I just backed up traj_comp.xtc to ./#traj_comp.xtc.1#

Back Off! I just backed up ener.edr to ./#ener.edr.1#

Back Off! I just backed up traj_comp.xtc to ./#traj_comp.xtc.1#

Back Off! I just backed up ener.edr to ./#ener.edr.1#
starting mdrun 'alanine dipeptide in vacuum in water'
1000 steps,      2.0 ps.
starting mdrun 'alanine dipeptide in vacuum in water'
1000 steps,      2.0 ps.

Writing final coordinates.

Writing final coordinates.

Back Off! I just backed up confout.gro to ./#confout.gro.1#

Back Off! I just backed up confout.gro to ./#confout.gro.1#

NOTE: Detected invalid cycle counts, probably because threads moved between CPU cores that do not have synchronized cycle counters. Will not print the cycle accounting.

GROMACS reminds you: "Your assumptions are your windows on the world. Scrub them off every once in a while, or the light won't come in." (Isaac Asimov)

!grep Repl 2c/rep0/md.log

Initializing Replica Exchange
Repl  There are 2 replicas:
Replica-exchange molecular dynamics method for protein folding
Replica exchange in temperature
Replica exchange interval: 200
Replica random seed: -4301313
Replica exchange information below: ex and x = exchange, pr = probability
Replica exchange at step 200 time 0.40000
Repl 0 <-> 1  dE_term = -0.000e+00 (kT)
Repl ex  0 x  1
Repl pr   1.0
Replica exchange at step 400 time 0.80000
Repl ex  0    1
Repl pr      
Replica exchange at step 600 time 1.20000
Repl 0 <-> 1  dE_term = -0.000e+00 (kT)
Repl ex  0 x  1
Repl pr   1.0
Replica exchange at step 800 time 1.60000
Repl ex  0    1
Repl pr      
Replica exchange statistics
Repl  4 attempts, 2 odd, 2 even
Repl  average probabilities:
Repl     0    1
Repl      1.0
Repl  number of exchanges:
Repl     0    1
Repl        2
Repl  average number of exchanges:
Repl     0    1
Repl      1.0
Repl        Empirical Transition Matrix
Repl       1       2
Repl  0.5000  0.5000  0
Repl  0.5000  0.5000  1

! grep "^dplumed" 2c/rep0/md.log

dplumed =  0.000e+00  dE_Term =  0.000e+00 (kT)
dplumed =  0.000e+00  dE_Term =  0.000e+00 (kT)

Exercise 2d: Find minimum scaling factor needed to cross the barrier¶

lambdas=[1.0,0.9,0.8,0.7,0.6,0.5,0.4,0.3,0.2,0.1]

#this box takes a while to run
for l in lambdas:
    with cd("2d/l"+str(l)):
        run_mpi("cp ../../2a/processed_selected.top processed_selected.top")
        run_mpi("plumed partial_tempering {} < processed_selected.top > topol.top".format(l))
        run_mpi("gmx_mpi grompp -c ../../conf.gro -f ../../grompp.mdp")
        run_mpi("gmx_mpi mdrun -nsteps 500000 -pin on -ntomp 12")

               :-) GROMACS - gmx grompp, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx grompp, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/2d/l1.0
Command line:
  gmx_mpi grompp -c ../../conf.gro -f ../../grompp.mdp

Setting the LD random seed to -579600386
Generated 8515 of the 8515 non-bonded parameter combinations
Generating 1-4 interactions: fudge = 0.5
Generated 8515 of the 8515 1-4 parameter combinations
Excluding 3 bonded neighbours molecule type 'Protein'
Excluding 2 bonded neighbours molecule type 'SOL'
Number of degrees of freedom in T-Coupling group Protein is 53.95
Number of degrees of freedom in T-Coupling group SOL is 3135.05
Estimate for the relative computational load of the PME mesh part: 0.19

GROMACS reminds you: "I have no responsibility to live up to what others expect of me. That's their mistake, not my failing." (Richard Feynman)

turning H bonds into constraints...
turning H bonds into constraints...
Analysing residue names:
There are:     3    Protein residues
There are:   523      Water residues
Analysing Protein...
Determining Verlet buffer for a tolerance of 0.005 kJ/mol/ps at 300 K
Calculated rlist for 1x1 atom pair-list as 1.036 nm, buffer size 0.036 nm
Set rlist, assuming 4x4 atom pair-list, to 1.000 nm, buffer size 0.000 nm
Note that mdrun will redetermine rlist based on the actual pair-list setup
Calculating fourier grid dimensions for X Y Z
Using a fourier grid of 24x24x24, spacing 0.118 0.118 0.118
This run will generate roughly 12 Mb of data

                :-) GROMACS - gmx mdrun, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx mdrun, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/2d/l1.0
Command line:
  gmx_mpi mdrun -nsteps 500000 -pin on -ntomp 12

Compiled SIMD: SSE2, but for this host/run AVX_256 might be better (see log).
The current CPU can measure timings more accurately than the code in
gmx mdrun was configured to use. This might affect your simulation
speed as accurate timings are needed for load-balancing.
Please consider rebuilding gmx mdrun with the GMX_USE_RDTSCP=ON CMake option.
Reading file topol.tpr, VERSION 2020.6-plumed-2.8.0 (single precision)

Overriding nsteps with value passed on the command line: 500000 steps, 1e+03 ps
Changing nstlist from 10 to 50, rlist from 1 to 1.111


Using 1 MPI process
Using 12 OpenMP threads 


Overriding thread affinity set outside gmx mdrun
starting mdrun 'alanine dipeptide in vacuum in water'
500000 steps,   1000.0 ps.

Writing final coordinates.

               Core t (s)   Wall t (s)        (%)
       Time:     6591.107      549.259     1200.0
                 (ns/day)    (hour/ns)
Performance:      157.303        0.153

GROMACS reminds you: "The sign of wisdom is to have more questions than answers." (Abhijit Naskar)

               :-) GROMACS - gmx grompp, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx grompp, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/2d/l0.9
Command line:
  gmx_mpi grompp -c ../../conf.gro -f ../../grompp.mdp

Setting the LD random seed to -1229314
Generated 8515 of the 8515 non-bonded parameter combinations
Generating 1-4 interactions: fudge = 0.5
Generated 8515 of the 8515 1-4 parameter combinations
Excluding 3 bonded neighbours molecule type 'Protein'
Excluding 2 bonded neighbours molecule type 'SOL'
Number of degrees of freedom in T-Coupling group Protein is 53.95
Number of degrees of freedom in T-Coupling group SOL is 3135.05
Estimate for the relative computational load of the PME mesh part: 0.19

GROMACS reminds you: "Praise those of your critics for whom nothing is up to standard." (Dag Hammarskjold)

turning H bonds into constraints...
turning H bonds into constraints...
Analysing residue names:
There are:     3    Protein residues
There are:   523      Water residues
Analysing Protein...
Determining Verlet buffer for a tolerance of 0.005 kJ/mol/ps at 300 K
Calculated rlist for 1x1 atom pair-list as 1.036 nm, buffer size 0.036 nm
Set rlist, assuming 4x4 atom pair-list, to 1.000 nm, buffer size 0.000 nm
Note that mdrun will redetermine rlist based on the actual pair-list setup
Calculating fourier grid dimensions for X Y Z
Using a fourier grid of 24x24x24, spacing 0.118 0.118 0.118
This run will generate roughly 12 Mb of data

                :-) GROMACS - gmx mdrun, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx mdrun, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/2d/l0.9
Command line:
  gmx_mpi mdrun -nsteps 500000 -pin on -ntomp 12

Compiled SIMD: SSE2, but for this host/run AVX_256 might be better (see log).
The current CPU can measure timings more accurately than the code in
gmx mdrun was configured to use. This might affect your simulation
speed as accurate timings are needed for load-balancing.
Please consider rebuilding gmx mdrun with the GMX_USE_RDTSCP=ON CMake option.
Reading file topol.tpr, VERSION 2020.6-plumed-2.8.0 (single precision)

Overriding nsteps with value passed on the command line: 500000 steps, 1e+03 ps
Changing nstlist from 10 to 50, rlist from 1 to 1.111


Using 1 MPI process
Using 12 OpenMP threads 


Overriding thread affinity set outside gmx mdrun
starting mdrun 'alanine dipeptide in vacuum in water'
500000 steps,   1000.0 ps.

Writing final coordinates.

               Core t (s)   Wall t (s)        (%)
       Time:     6552.731      546.061     1200.0
                 (ns/day)    (hour/ns)
Performance:      158.224        0.152

GROMACS reminds you: "The scientist is not the person who always gives the right answers, he is the one who asks the right questions." (Claude Levi-Strauss)

               :-) GROMACS - gmx grompp, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx grompp, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/2d/l0.8
Command line:
  gmx_mpi grompp -c ../../conf.gro -f ../../grompp.mdp

Setting the LD random seed to 1876946775
Generated 8515 of the 8515 non-bonded parameter combinations
Generating 1-4 interactions: fudge = 0.5
Generated 8515 of the 8515 1-4 parameter combinations
Excluding 3 bonded neighbours molecule type 'Protein'
Excluding 2 bonded neighbours molecule type 'SOL'
Number of degrees of freedom in T-Coupling group Protein is 53.95
Number of degrees of freedom in T-Coupling group SOL is 3135.05
Estimate for the relative computational load of the PME mesh part: 0.19

GROMACS reminds you: "I am at two with nature." (Woody Allen)

turning H bonds into constraints...
turning H bonds into constraints...
Analysing residue names:
There are:     3    Protein residues
There are:   523      Water residues
Analysing Protein...
Determining Verlet buffer for a tolerance of 0.005 kJ/mol/ps at 300 K
Calculated rlist for 1x1 atom pair-list as 1.036 nm, buffer size 0.036 nm
Set rlist, assuming 4x4 atom pair-list, to 1.000 nm, buffer size 0.000 nm
Note that mdrun will redetermine rlist based on the actual pair-list setup
Calculating fourier grid dimensions for X Y Z
Using a fourier grid of 24x24x24, spacing 0.118 0.118 0.118
This run will generate roughly 12 Mb of data

                :-) GROMACS - gmx mdrun, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx mdrun, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/2d/l0.8
Command line:
  gmx_mpi mdrun -nsteps 500000 -pin on -ntomp 12

Compiled SIMD: SSE2, but for this host/run AVX_256 might be better (see log).
The current CPU can measure timings more accurately than the code in
gmx mdrun was configured to use. This might affect your simulation
speed as accurate timings are needed for load-balancing.
Please consider rebuilding gmx mdrun with the GMX_USE_RDTSCP=ON CMake option.
Reading file topol.tpr, VERSION 2020.6-plumed-2.8.0 (single precision)

Overriding nsteps with value passed on the command line: 500000 steps, 1e+03 ps
Changing nstlist from 10 to 50, rlist from 1 to 1.111


Using 1 MPI process
Using 12 OpenMP threads 


Overriding thread affinity set outside gmx mdrun
starting mdrun 'alanine dipeptide in vacuum in water'
500000 steps,   1000.0 ps.

Writing final coordinates.

               Core t (s)   Wall t (s)        (%)
       Time:     6531.176      544.265     1200.0
                 (ns/day)    (hour/ns)
Performance:      158.747        0.151

GROMACS reminds you: "Same sex marriage is not a gay privilege, it's equal rights. Privilege would be something like gay people not paying taxes. Like churches don't." (Ricky Gervais)

               :-) GROMACS - gmx grompp, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx grompp, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/2d/l0.7
Command line:
  gmx_mpi grompp -c ../../conf.gro -f ../../grompp.mdp

Setting the LD random seed to 1073568699
Generated 8515 of the 8515 non-bonded parameter combinations
Generating 1-4 interactions: fudge = 0.5
Generated 8515 of the 8515 1-4 parameter combinations
Excluding 3 bonded neighbours molecule type 'Protein'
Excluding 2 bonded neighbours molecule type 'SOL'
Number of degrees of freedom in T-Coupling group Protein is 53.95
Number of degrees of freedom in T-Coupling group SOL is 3135.05
Estimate for the relative computational load of the PME mesh part: 0.19

GROMACS reminds you: "Njuta men inte frossa, springa men inte fly" (Paganus)

turning H bonds into constraints...
turning H bonds into constraints...
Analysing residue names:
There are:     3    Protein residues
There are:   523      Water residues
Analysing Protein...
Determining Verlet buffer for a tolerance of 0.005 kJ/mol/ps at 300 K
Calculated rlist for 1x1 atom pair-list as 1.036 nm, buffer size 0.036 nm
Set rlist, assuming 4x4 atom pair-list, to 1.000 nm, buffer size 0.000 nm
Note that mdrun will redetermine rlist based on the actual pair-list setup
Calculating fourier grid dimensions for X Y Z
Using a fourier grid of 24x24x24, spacing 0.118 0.118 0.118
This run will generate roughly 12 Mb of data

                :-) GROMACS - gmx mdrun, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx mdrun, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/2d/l0.7
Command line:
  gmx_mpi mdrun -nsteps 500000 -pin on -ntomp 12

Compiled SIMD: SSE2, but for this host/run AVX_256 might be better (see log).
The current CPU can measure timings more accurately than the code in
gmx mdrun was configured to use. This might affect your simulation
speed as accurate timings are needed for load-balancing.
Please consider rebuilding gmx mdrun with the GMX_USE_RDTSCP=ON CMake option.
Reading file topol.tpr, VERSION 2020.6-plumed-2.8.0 (single precision)

Overriding nsteps with value passed on the command line: 500000 steps, 1e+03 ps
Changing nstlist from 10 to 50, rlist from 1 to 1.111


Using 1 MPI process
Using 12 OpenMP threads 


Overriding thread affinity set outside gmx mdrun
starting mdrun 'alanine dipeptide in vacuum in water'
500000 steps,   1000.0 ps.

Writing final coordinates.

               Core t (s)   Wall t (s)        (%)
       Time:     6559.981      546.665     1200.0
                 (ns/day)    (hour/ns)
Performance:      158.050        0.152

GROMACS reminds you: "It seemed a good idea at first" (Gerrit Groenhof)

               :-) GROMACS - gmx grompp, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx grompp, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/2d/l0.6
Command line:
  gmx_mpi grompp -c ../../conf.gro -f ../../grompp.mdp

Setting the LD random seed to -570726071
Generated 8515 of the 8515 non-bonded parameter combinations
Generating 1-4 interactions: fudge = 0.5
Generated 8515 of the 8515 1-4 parameter combinations
Excluding 3 bonded neighbours molecule type 'Protein'
Excluding 2 bonded neighbours molecule type 'SOL'
Number of degrees of freedom in T-Coupling group Protein is 53.95
Number of degrees of freedom in T-Coupling group SOL is 3135.05
Estimate for the relative computational load of the PME mesh part: 0.19

GROMACS reminds you: "It seemed a good idea at first" (Gerrit Groenhof)

turning H bonds into constraints...
turning H bonds into constraints...
Analysing residue names:
There are:     3    Protein residues
There are:   523      Water residues
Analysing Protein...
Determining Verlet buffer for a tolerance of 0.005 kJ/mol/ps at 300 K
Calculated rlist for 1x1 atom pair-list as 1.036 nm, buffer size 0.036 nm
Set rlist, assuming 4x4 atom pair-list, to 1.000 nm, buffer size 0.000 nm
Note that mdrun will redetermine rlist based on the actual pair-list setup
Calculating fourier grid dimensions for X Y Z
Using a fourier grid of 24x24x24, spacing 0.118 0.118 0.118
This run will generate roughly 12 Mb of data

colvar=[]

for l in lambdas:
    with cd("2d/l"+str(l)):
        with open("plumed.dat","w") as f:
                  print("""
MOLINFO STRUCTURE=../../ala.pdb 
phi: TORSION ATOMS=@phi-2
psi: TORSION ATOMS=@psi-2
PRINT ARG=phi,psi FILE=COLVAR 
                  """, file=f)
        subprocess.run("plumed driver --ixtc traj_comp.xtc --plumed plumed.dat",shell=True)
        colvar.append(plumed.read_as_pandas("COLVAR"))

PLUMED: PLUMED is starting
PLUMED: Version: 2.8.0 (git: ddb59d16b) compiled on May 17 2022 at 19:07:56
PLUMED: Please cite these papers when using PLUMED [1][2]
PLUMED: For further information see the PLUMED web page at http://www.plumed.org
PLUMED: Root: /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed
PLUMED: For installed feature, see /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed/src/config/config.txt
PLUMED: Molecular dynamics engine: driver
PLUMED: Precision of reals: 8
PLUMED: Running over 1 node
PLUMED: Number of threads: 1
PLUMED: Cache line size: 512
PLUMED: Number of atoms: 1591
PLUMED: File suffix: 
PLUMED: FILE: plumed.dat
PLUMED: Action MOLINFO
PLUMED:   with label @0
PLUMED:   pdb file named ../../ala.pdb contains 1 chains 
PLUMED:   chain named   contains residues 1 to 526 and atoms 1 to 1591 
PLUMED: Action TORSION
PLUMED:   with label phi
PLUMED:   between atoms 5 7 9 15
PLUMED:   using periodic boundary conditions
PLUMED: Action TORSION
PLUMED:   with label psi
PLUMED:   between atoms 7 9 15 17
PLUMED:   using periodic boundary conditions
PLUMED: Action PRINT
PLUMED:   with label @3
PLUMED:   with stride 1
PLUMED:   with arguments phi psi
PLUMED:   on file COLVAR
PLUMED:   with format  %f
PLUMED: END FILE: plumed.dat
PLUMED: Timestep: 1.000000
PLUMED: KbT has not been set by the MD engine
PLUMED: It should be set by hand where needed
PLUMED: Relevant bibliography:
PLUMED:   [1] The PLUMED consortium, Nat. Methods 16, 670 (2019)
PLUMED:   [2] Tribello, Bonomi, Branduardi, Camilloni, and Bussi, Comput. Phys. Commun. 185, 604 (2014)
PLUMED: Please read and cite where appropriate!
PLUMED: Finished setup
PLUMED:                                               Cycles        Total      Average      Minimum      Maximum
PLUMED:                                                    1     0.091916     0.091916     0.091916     0.091916
PLUMED: 1 Prepare dependencies                          1001     0.001996     0.000002     0.000001     0.000012
PLUMED: 2 Sharing data                                  1001     0.010454     0.000010     0.000009     0.000259
PLUMED: 3 Waiting for data                              1001     0.003329     0.000003     0.000003     0.000021
PLUMED: 4 Calculating (forward loop)                    1001     0.011175     0.000011     0.000010     0.000043
PLUMED: 5 Applying (backward loop)                      1001     0.014713     0.000015     0.000014     0.000025
PLUMED: 6 Update                                        1001     0.009302     0.000009     0.000008     0.000062

+++ Loading the PLUMED kernel runtime +++
+++ PLUMED_KERNEL="/net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/libplumedKernel.so" +++

PLUMED: PLUMED is starting
PLUMED: Version: 2.8.0 (git: ddb59d16b) compiled on May 17 2022 at 19:07:56
PLUMED: Please cite these papers when using PLUMED [1][2]
PLUMED: For further information see the PLUMED web page at http://www.plumed.org
PLUMED: Root: /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed
PLUMED: For installed feature, see /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed/src/config/config.txt
PLUMED: Molecular dynamics engine: driver
PLUMED: Precision of reals: 8
PLUMED: Running over 1 node
PLUMED: Number of threads: 1
PLUMED: Cache line size: 512
PLUMED: Number of atoms: 1591
PLUMED: File suffix: 
PLUMED: FILE: plumed.dat
PLUMED: Action MOLINFO
PLUMED:   with label @0
PLUMED:   pdb file named ../../ala.pdb contains 1 chains 
PLUMED:   chain named   contains residues 1 to 526 and atoms 1 to 1591 
PLUMED: Action TORSION
PLUMED:   with label phi
PLUMED:   between atoms 5 7 9 15
PLUMED:   using periodic boundary conditions
PLUMED: Action TORSION
PLUMED:   with label psi
PLUMED:   between atoms 7 9 15 17
PLUMED:   using periodic boundary conditions
PLUMED: Action PRINT
PLUMED:   with label @3
PLUMED:   with stride 1
PLUMED:   with arguments phi psi
PLUMED:   on file COLVAR
PLUMED:   with format  %f
PLUMED: END FILE: plumed.dat
PLUMED: Timestep: 1.000000
PLUMED: KbT has not been set by the MD engine
PLUMED: It should be set by hand where needed
PLUMED: Relevant bibliography:
PLUMED:   [1] The PLUMED consortium, Nat. Methods 16, 670 (2019)
PLUMED:   [2] Tribello, Bonomi, Branduardi, Camilloni, and Bussi, Comput. Phys. Commun. 185, 604 (2014)
PLUMED: Please read and cite where appropriate!
PLUMED: Finished setup
PLUMED:                                               Cycles        Total      Average      Minimum      Maximum
PLUMED:                                                    1     0.091494     0.091494     0.091494     0.091494
PLUMED: 1 Prepare dependencies                          1001     0.002040     0.000002     0.000001     0.000007
PLUMED: 2 Sharing data                                  1001     0.010125     0.000010     0.000009     0.000261
PLUMED: 3 Waiting for data                              1001     0.003387     0.000003     0.000003     0.000021
PLUMED: 4 Calculating (forward loop)                    1001     0.011161     0.000011     0.000010     0.000042
PLUMED: 5 Applying (backward loop)                      1001     0.014639     0.000015     0.000014     0.000026
PLUMED: 6 Update                                        1001     0.009336     0.000009     0.000008     0.000058

+++ Loading the PLUMED kernel runtime +++
+++ PLUMED_KERNEL="/net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/libplumedKernel.so" +++

PLUMED: PLUMED is starting
PLUMED: Version: 2.8.0 (git: ddb59d16b) compiled on May 17 2022 at 19:07:56
PLUMED: Please cite these papers when using PLUMED [1][2]
PLUMED: For further information see the PLUMED web page at http://www.plumed.org
PLUMED: Root: /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed
PLUMED: For installed feature, see /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed/src/config/config.txt
PLUMED: Molecular dynamics engine: driver
PLUMED: Precision of reals: 8
PLUMED: Running over 1 node
PLUMED: Number of threads: 1
PLUMED: Cache line size: 512
PLUMED: Number of atoms: 1591
PLUMED: File suffix: 
PLUMED: FILE: plumed.dat
PLUMED: Action MOLINFO
PLUMED:   with label @0
PLUMED:   pdb file named ../../ala.pdb contains 1 chains 
PLUMED:   chain named   contains residues 1 to 526 and atoms 1 to 1591 
PLUMED: Action TORSION
PLUMED:   with label phi
PLUMED:   between atoms 5 7 9 15
PLUMED:   using periodic boundary conditions
PLUMED: Action TORSION
PLUMED:   with label psi
PLUMED:   between atoms 7 9 15 17
PLUMED:   using periodic boundary conditions
PLUMED: Action PRINT
PLUMED:   with label @3
PLUMED:   with stride 1
PLUMED:   with arguments phi psi
PLUMED:   on file COLVAR
PLUMED:   with format  %f
PLUMED: END FILE: plumed.dat
PLUMED: Timestep: 1.000000
PLUMED: KbT has not been set by the MD engine
PLUMED: It should be set by hand where needed
PLUMED: Relevant bibliography:
PLUMED:   [1] The PLUMED consortium, Nat. Methods 16, 670 (2019)
PLUMED:   [2] Tribello, Bonomi, Branduardi, Camilloni, and Bussi, Comput. Phys. Commun. 185, 604 (2014)
PLUMED: Please read and cite where appropriate!
PLUMED: Finished setup
PLUMED:                                               Cycles        Total      Average      Minimum      Maximum
PLUMED:                                                    1     0.091360     0.091360     0.091360     0.091360
PLUMED: 1 Prepare dependencies                          1001     0.002026     0.000002     0.000001     0.000011
PLUMED: 2 Sharing data                                  1001     0.010217     0.000010     0.000009     0.000287
PLUMED: 3 Waiting for data                              1001     0.003349     0.000003     0.000003     0.000021
PLUMED: 4 Calculating (forward loop)                    1001     0.011293     0.000011     0.000011     0.000043
PLUMED: 5 Applying (backward loop)                      1001     0.014634     0.000015     0.000014     0.000030
PLUMED: 6 Update                                        1001     0.009294     0.000009     0.000008     0.000058

+++ Loading the PLUMED kernel runtime +++
+++ PLUMED_KERNEL="/net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/libplumedKernel.so" +++

PLUMED: PLUMED is starting
PLUMED: Version: 2.8.0 (git: ddb59d16b) compiled on May 17 2022 at 19:07:56
PLUMED: Please cite these papers when using PLUMED [1][2]
PLUMED: For further information see the PLUMED web page at http://www.plumed.org
PLUMED: Root: /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed
PLUMED: For installed feature, see /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed/src/config/config.txt
PLUMED: Molecular dynamics engine: driver
PLUMED: Precision of reals: 8
PLUMED: Running over 1 node
PLUMED: Number of threads: 1
PLUMED: Cache line size: 512
PLUMED: Number of atoms: 1591
PLUMED: File suffix: 
PLUMED: FILE: plumed.dat
PLUMED: Action MOLINFO
PLUMED:   with label @0
PLUMED:   pdb file named ../../ala.pdb contains 1 chains 
PLUMED:   chain named   contains residues 1 to 526 and atoms 1 to 1591 
PLUMED: Action TORSION
PLUMED:   with label phi
PLUMED:   between atoms 5 7 9 15
PLUMED:   using periodic boundary conditions
PLUMED: Action TORSION
PLUMED:   with label psi
PLUMED:   between atoms 7 9 15 17
PLUMED:   using periodic boundary conditions
PLUMED: Action PRINT
PLUMED:   with label @3
PLUMED:   with stride 1
PLUMED:   with arguments phi psi
PLUMED:   on file COLVAR
PLUMED:   with format  %f
PLUMED: END FILE: plumed.dat
PLUMED: Timestep: 1.000000
PLUMED: KbT has not been set by the MD engine
PLUMED: It should be set by hand where needed
PLUMED: Relevant bibliography:
PLUMED:   [1] The PLUMED consortium, Nat. Methods 16, 670 (2019)
PLUMED:   [2] Tribello, Bonomi, Branduardi, Camilloni, and Bussi, Comput. Phys. Commun. 185, 604 (2014)
PLUMED: Please read and cite where appropriate!
PLUMED: Finished setup
PLUMED:                                               Cycles        Total      Average      Minimum      Maximum
PLUMED:                                                    1     0.093822     0.093822     0.093822     0.093822
PLUMED: 1 Prepare dependencies                          1001     0.002083     0.000002     0.000001     0.000012
PLUMED: 2 Sharing data                                  1001     0.010190     0.000010     0.000009     0.000244
PLUMED: 3 Waiting for data                              1001     0.003433     0.000003     0.000003     0.000020
PLUMED: 4 Calculating (forward loop)                    1001     0.011125     0.000011     0.000010     0.000043
PLUMED: 5 Applying (backward loop)                      1001     0.014630     0.000015     0.000014     0.000026
PLUMED: 6 Update                                        1001     0.009266     0.000009     0.000008     0.000056

+++ Loading the PLUMED kernel runtime +++
+++ PLUMED_KERNEL="/net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/libplumedKernel.so" +++

PLUMED: PLUMED is starting
PLUMED: Version: 2.8.0 (git: ddb59d16b) compiled on May 17 2022 at 19:07:56
PLUMED: Please cite these papers when using PLUMED [1][2]
PLUMED: For further information see the PLUMED web page at http://www.plumed.org
PLUMED: Root: /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed
PLUMED: For installed feature, see /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed/src/config/config.txt
PLUMED: Molecular dynamics engine: driver
PLUMED: Precision of reals: 8
PLUMED: Running over 1 node
PLUMED: Number of threads: 1
PLUMED: Cache line size: 512
PLUMED: Number of atoms: 1591
PLUMED: File suffix: 
PLUMED: FILE: plumed.dat
PLUMED: Action MOLINFO
PLUMED:   with label @0
PLUMED:   pdb file named ../../ala.pdb contains 1 chains 
PLUMED:   chain named   contains residues 1 to 526 and atoms 1 to 1591 
PLUMED: Action TORSION
PLUMED:   with label phi
PLUMED:   between atoms 5 7 9 15
PLUMED:   using periodic boundary conditions
PLUMED: Action TORSION
PLUMED:   with label psi
PLUMED:   between atoms 7 9 15 17
PLUMED:   using periodic boundary conditions
PLUMED: Action PRINT
PLUMED:   with label @3
PLUMED:   with stride 1
PLUMED:   with arguments phi psi
PLUMED:   on file COLVAR
PLUMED:   with format  %f
PLUMED: END FILE: plumed.dat
PLUMED: Timestep: 1.000000
PLUMED: KbT has not been set by the MD engine
PLUMED: It should be set by hand where needed
PLUMED: Relevant bibliography:
PLUMED:   [1] The PLUMED consortium, Nat. Methods 16, 670 (2019)
PLUMED:   [2] Tribello, Bonomi, Branduardi, Camilloni, and Bussi, Comput. Phys. Commun. 185, 604 (2014)
PLUMED: Please read and cite where appropriate!
PLUMED: Finished setup
PLUMED:                                               Cycles        Total      Average      Minimum      Maximum
PLUMED:                                                    1     0.092173     0.092173     0.092173     0.092173
PLUMED: 1 Prepare dependencies                          1001     0.002011     0.000002     0.000001     0.000010
PLUMED: 2 Sharing data                                  1001     0.009920     0.000010     0.000009     0.000241
PLUMED: 3 Waiting for data                              1001     0.003448     0.000003     0.000003     0.000025
PLUMED: 4 Calculating (forward loop)                    1001     0.011430     0.000011     0.000011     0.000048
PLUMED: 5 Applying (backward loop)                      1001     0.014946     0.000015     0.000014     0.000032
PLUMED: 6 Update                                        1001     0.009377     0.000009     0.000008     0.000069

+++ Loading the PLUMED kernel runtime +++
+++ PLUMED_KERNEL="/net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/libplumedKernel.so" +++

for i in range(len(lambdas)):
    plt.title("lambda={}".format(lambdas[i]))
    plt.plot(colvar[i].time,colvar[i].phi,".")
    plt.show()

Looks like 0.15 might be a reasonable value. Notice that it is not too dissimilar from 300/800=0.375

In any case, we expect the need to go to a lower scaling value in this case because solvent is not scaled at all. I will proceed with 0.15 as a reasonable choice

300/800

0.375

Exercise 3: Run Hamiltonian replica exchange simulations¶

Now that we chose the minumum lambda we can setup a tentative hrex simulation. I am trying first a linearly spacing lambda. By experience, linear spacing works better for lambda than logarithmic spacing (though this could be very system dependent!). In any case, we will see later how to optimize the lambda ladder if needed.

lambda_min=0.15
lambdas=np.linspace(1.0,lambda_min,6)

#this box takes a while to run
with cd("3/6reps"):
    with open("plumed.dat","w") as f:
        print("",file=f) # empty
    for i in range(len(lambdas)):
        with cd("replica{}".format(i)):
            run_mpi("cp ../../../2a/processed_selected.top processed_selected.top")
            run_mpi("plumed partial_tempering {} < processed_selected.top > topol.top".format(lambdas[i]))
            run_mpi("gmx_mpi grompp -c ../../../conf.gro -f ../../../grompp.mdp")
    run_mpi("gmx_mpi mdrun -multidir replica? -replex 200 -nsteps 500000 -ntomp 4 -hrex -plumed ../plumed.dat",len(lambdas))

               :-) GROMACS - gmx grompp, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx grompp, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/3/6reps/replica0
Command line:
  gmx_mpi grompp -c ../../../conf.gro -f ../../../grompp.mdp

Setting the LD random seed to -612644117
Generated 8515 of the 8515 non-bonded parameter combinations
Generating 1-4 interactions: fudge = 0.5
Generated 8515 of the 8515 1-4 parameter combinations
Excluding 3 bonded neighbours molecule type 'Protein'
Excluding 2 bonded neighbours molecule type 'SOL'
Number of degrees of freedom in T-Coupling group Protein is 53.95
Number of degrees of freedom in T-Coupling group SOL is 3135.05
Estimate for the relative computational load of the PME mesh part: 0.19

GROMACS reminds you: "She Needs Cash to Buy Aspirine For Her Pain" (LIVE)

turning H bonds into constraints...
turning H bonds into constraints...
Analysing residue names:
There are:     3    Protein residues
There are:   523      Water residues
Analysing Protein...
Determining Verlet buffer for a tolerance of 0.005 kJ/mol/ps at 300 K
Calculated rlist for 1x1 atom pair-list as 1.036 nm, buffer size 0.036 nm
Set rlist, assuming 4x4 atom pair-list, to 1.000 nm, buffer size 0.000 nm
Note that mdrun will redetermine rlist based on the actual pair-list setup
Calculating fourier grid dimensions for X Y Z
Using a fourier grid of 24x24x24, spacing 0.118 0.118 0.118
This run will generate roughly 12 Mb of data

               :-) GROMACS - gmx grompp, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx grompp, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/3/6reps/replica1
Command line:
  gmx_mpi grompp -c ../../../conf.gro -f ../../../grompp.mdp

Setting the LD random seed to -1646347489
Generated 8515 of the 8515 non-bonded parameter combinations
Generating 1-4 interactions: fudge = 0.5
Generated 8515 of the 8515 1-4 parameter combinations
Excluding 3 bonded neighbours molecule type 'Protein'
Excluding 2 bonded neighbours molecule type 'SOL'
Number of degrees of freedom in T-Coupling group Protein is 53.95
Number of degrees of freedom in T-Coupling group SOL is 3135.05
Estimate for the relative computational load of the PME mesh part: 0.19

GROMACS reminds you: "Pump Up the Volume Along With the Tempo" (Jazzy Jeff)

turning H bonds into constraints...
turning H bonds into constraints...
Analysing residue names:
There are:     3    Protein residues
There are:   523      Water residues
Analysing Protein...
Determining Verlet buffer for a tolerance of 0.005 kJ/mol/ps at 300 K
Calculated rlist for 1x1 atom pair-list as 1.036 nm, buffer size 0.036 nm
Set rlist, assuming 4x4 atom pair-list, to 1.000 nm, buffer size 0.000 nm
Note that mdrun will redetermine rlist based on the actual pair-list setup
Calculating fourier grid dimensions for X Y Z
Using a fourier grid of 24x24x24, spacing 0.118 0.118 0.118
This run will generate roughly 12 Mb of data

               :-) GROMACS - gmx grompp, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx grompp, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/3/6reps/replica2
Command line:
  gmx_mpi grompp -c ../../../conf.gro -f ../../../grompp.mdp

Setting the LD random seed to -101321185
Generated 8515 of the 8515 non-bonded parameter combinations
Generating 1-4 interactions: fudge = 0.5
Generated 8515 of the 8515 1-4 parameter combinations
Excluding 3 bonded neighbours molecule type 'Protein'
Excluding 2 bonded neighbours molecule type 'SOL'
Number of degrees of freedom in T-Coupling group Protein is 53.95
Number of degrees of freedom in T-Coupling group SOL is 3135.05
Estimate for the relative computational load of the PME mesh part: 0.19

GROMACS reminds you: "Pump Up the Volume Along With the Tempo" (Jazzy Jeff)

turning H bonds into constraints...
turning H bonds into constraints...
Analysing residue names:
There are:     3    Protein residues
There are:   523      Water residues
Analysing Protein...
Determining Verlet buffer for a tolerance of 0.005 kJ/mol/ps at 300 K
Calculated rlist for 1x1 atom pair-list as 1.036 nm, buffer size 0.036 nm
Set rlist, assuming 4x4 atom pair-list, to 1.000 nm, buffer size 0.000 nm
Note that mdrun will redetermine rlist based on the actual pair-list setup
Calculating fourier grid dimensions for X Y Z
Using a fourier grid of 24x24x24, spacing 0.118 0.118 0.118
This run will generate roughly 12 Mb of data

               :-) GROMACS - gmx grompp, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx grompp, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/3/6reps/replica3
Command line:
  gmx_mpi grompp -c ../../../conf.gro -f ../../../grompp.mdp

Setting the LD random seed to 1454894975
Generated 8515 of the 8515 non-bonded parameter combinations
Generating 1-4 interactions: fudge = 0.5
Generated 8515 of the 8515 1-4 parameter combinations
Excluding 3 bonded neighbours molecule type 'Protein'
Excluding 2 bonded neighbours molecule type 'SOL'
Number of degrees of freedom in T-Coupling group Protein is 53.95
Number of degrees of freedom in T-Coupling group SOL is 3135.05
Estimate for the relative computational load of the PME mesh part: 0.19

GROMACS reminds you: "We Have No Money" (E. Clementi)

turning H bonds into constraints...
turning H bonds into constraints...
Analysing residue names:
There are:     3    Protein residues
There are:   523      Water residues
Analysing Protein...
Determining Verlet buffer for a tolerance of 0.005 kJ/mol/ps at 300 K
Calculated rlist for 1x1 atom pair-list as 1.036 nm, buffer size 0.036 nm
Set rlist, assuming 4x4 atom pair-list, to 1.000 nm, buffer size 0.000 nm
Note that mdrun will redetermine rlist based on the actual pair-list setup
Calculating fourier grid dimensions for X Y Z
Using a fourier grid of 24x24x24, spacing 0.118 0.118 0.118
This run will generate roughly 12 Mb of data

               :-) GROMACS - gmx grompp, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx grompp, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/3/6reps/replica4
Command line:
  gmx_mpi grompp -c ../../../conf.gro -f ../../../grompp.mdp

Setting the LD random seed to 987199374
Generated 8515 of the 8515 non-bonded parameter combinations
Generating 1-4 interactions: fudge = 0.5
Generated 8515 of the 8515 1-4 parameter combinations
Excluding 3 bonded neighbours molecule type 'Protein'
Excluding 2 bonded neighbours molecule type 'SOL'
Number of degrees of freedom in T-Coupling group Protein is 53.95
Number of degrees of freedom in T-Coupling group SOL is 3135.05
Estimate for the relative computational load of the PME mesh part: 0.19

GROMACS reminds you: "It is now quite lawful for a Catholic woman to avoid pregnancy by a resort to mathematics, though she is still forbidden to resort to physics and chemistry." (Henry Louis Mencken)

turning H bonds into constraints...
turning H bonds into constraints...
Analysing residue names:
There are:     3    Protein residues
There are:   523      Water residues
Analysing Protein...
Determining Verlet buffer for a tolerance of 0.005 kJ/mol/ps at 300 K
Calculated rlist for 1x1 atom pair-list as 1.036 nm, buffer size 0.036 nm
Set rlist, assuming 4x4 atom pair-list, to 1.000 nm, buffer size 0.000 nm
Note that mdrun will redetermine rlist based on the actual pair-list setup
Calculating fourier grid dimensions for X Y Z
Using a fourier grid of 24x24x24, spacing 0.118 0.118 0.118
This run will generate roughly 12 Mb of data

! grep -A 4 average probabilities 3/6reps/replica0/md.log

grep: probabilities: No such file or directory
3/6reps/replica0/md.log:Repl  average probabilities:
3/6reps/replica0/md.log-Repl     0    1    2    3    4    5
3/6reps/replica0/md.log-Repl      .55  .47  .41  .30  .12
3/6reps/replica0/md.log-Repl  number of exchanges:
3/6reps/replica0/md.log-Repl     0    1    2    3    4    5
--
3/6reps/replica0/md.log:Repl  average number of exchanges:
3/6reps/replica0/md.log-Repl     0    1    2    3    4    5
3/6reps/replica0/md.log-Repl      .57  .46  .42  .30  .12
3/6reps/replica0/md.log-
3/6reps/replica0/md.log-

Accepntace is fine, except for in the last pair of replicas. We will fix this later.

We then read the resulting trajectories.

colvar=[]

for i in range(len(lambdas)):
    with cd("3/6reps/replica"+str(i)):
        with open("plumed.dat","w") as f:
            print("""
MOLINFO STRUCTURE=../../../ala.pdb 
phi: TORSION ATOMS=@phi-2
psi: TORSION ATOMS=@psi-2
PRINT ARG=phi,psi FILE=COLVAR 
                  """,file=f)
        subprocess.run("plumed driver --ixtc traj_comp.xtc --plumed plumed.dat",shell=True)
            
        colvar.append(plumed.read_as_pandas("COLVAR"))

PLUMED: PLUMED is starting
PLUMED: Version: 2.8.0 (git: ddb59d16b) compiled on May 17 2022 at 19:07:56
PLUMED: Please cite these papers when using PLUMED [1][2]
PLUMED: For further information see the PLUMED web page at http://www.plumed.org
PLUMED: Root: /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed
PLUMED: For installed feature, see /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed/src/config/config.txt
PLUMED: Molecular dynamics engine: driver
PLUMED: Precision of reals: 8
PLUMED: Running over 1 node
PLUMED: Number of threads: 1
PLUMED: Cache line size: 512
PLUMED: Number of atoms: 1591
PLUMED: File suffix: 
PLUMED: FILE: plumed.dat
PLUMED: Action MOLINFO
PLUMED:   with label @0
PLUMED:   pdb file named ../../../ala.pdb contains 1 chains 
PLUMED:   chain named   contains residues 1 to 526 and atoms 1 to 1591 
PLUMED: Action TORSION
PLUMED:   with label phi
PLUMED:   between atoms 5 7 9 15
PLUMED:   using periodic boundary conditions
PLUMED: Action TORSION
PLUMED:   with label psi
PLUMED:   between atoms 7 9 15 17
PLUMED:   using periodic boundary conditions
PLUMED: Action PRINT
PLUMED:   with label @3
PLUMED:   with stride 1
PLUMED:   with arguments phi psi
PLUMED:   on file COLVAR
PLUMED:   with format  %f
PLUMED: END FILE: plumed.dat
PLUMED: Timestep: 1.000000
PLUMED: KbT has not been set by the MD engine
PLUMED: It should be set by hand where needed
PLUMED: Relevant bibliography:
PLUMED:   [1] The PLUMED consortium, Nat. Methods 16, 670 (2019)
PLUMED:   [2] Tribello, Bonomi, Branduardi, Camilloni, and Bussi, Comput. Phys. Commun. 185, 604 (2014)
PLUMED: Please read and cite where appropriate!
PLUMED: Finished setup
PLUMED:                                               Cycles        Total      Average      Minimum      Maximum
PLUMED:                                                    1     0.091165     0.091165     0.091165     0.091165
PLUMED: 1 Prepare dependencies                          1001     0.002075     0.000002     0.000002     0.000013
PLUMED: 2 Sharing data                                  1001     0.009754     0.000010     0.000009     0.000250
PLUMED: 3 Waiting for data                              1001     0.003255     0.000003     0.000003     0.000020
PLUMED: 4 Calculating (forward loop)                    1001     0.011189     0.000011     0.000010     0.000042
PLUMED: 5 Applying (backward loop)                      1001     0.014746     0.000015     0.000014     0.000026
PLUMED: 6 Update                                        1001     0.009348     0.000009     0.000008     0.000055

+++ Loading the PLUMED kernel runtime +++
+++ PLUMED_KERNEL="/net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/libplumedKernel.so" +++

PLUMED: PLUMED is starting
PLUMED: Version: 2.8.0 (git: ddb59d16b) compiled on May 17 2022 at 19:07:56
PLUMED: Please cite these papers when using PLUMED [1][2]
PLUMED: For further information see the PLUMED web page at http://www.plumed.org
PLUMED: Root: /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed
PLUMED: For installed feature, see /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed/src/config/config.txt
PLUMED: Molecular dynamics engine: driver
PLUMED: Precision of reals: 8
PLUMED: Running over 1 node
PLUMED: Number of threads: 1
PLUMED: Cache line size: 512
PLUMED: Number of atoms: 1591
PLUMED: File suffix: 
PLUMED: FILE: plumed.dat
PLUMED: Action MOLINFO
PLUMED:   with label @0
PLUMED:   pdb file named ../../../ala.pdb contains 1 chains 
PLUMED:   chain named   contains residues 1 to 526 and atoms 1 to 1591 
PLUMED: Action TORSION
PLUMED:   with label phi
PLUMED:   between atoms 5 7 9 15
PLUMED:   using periodic boundary conditions
PLUMED: Action TORSION
PLUMED:   with label psi
PLUMED:   between atoms 7 9 15 17
PLUMED:   using periodic boundary conditions
PLUMED: Action PRINT
PLUMED:   with label @3
PLUMED:   with stride 1
PLUMED:   with arguments phi psi
PLUMED:   on file COLVAR
PLUMED:   with format  %f
PLUMED: END FILE: plumed.dat
PLUMED: Timestep: 1.000000
PLUMED: KbT has not been set by the MD engine
PLUMED: It should be set by hand where needed
PLUMED: Relevant bibliography:
PLUMED:   [1] The PLUMED consortium, Nat. Methods 16, 670 (2019)
PLUMED:   [2] Tribello, Bonomi, Branduardi, Camilloni, and Bussi, Comput. Phys. Commun. 185, 604 (2014)
PLUMED: Please read and cite where appropriate!
PLUMED: Finished setup
PLUMED:                                               Cycles        Total      Average      Minimum      Maximum
PLUMED:                                                    1     0.091243     0.091243     0.091243     0.091243
PLUMED: 1 Prepare dependencies                          1001     0.001989     0.000002     0.000002     0.000013
PLUMED: 2 Sharing data                                  1001     0.010143     0.000010     0.000009     0.000251
PLUMED: 3 Waiting for data                              1001     0.003354     0.000003     0.000002     0.000020
PLUMED: 4 Calculating (forward loop)                    1001     0.011083     0.000011     0.000010     0.000042
PLUMED: 5 Applying (backward loop)                      1001     0.014545     0.000015     0.000014     0.000026
PLUMED: 6 Update                                        1001     0.009419     0.000009     0.000008     0.000060

+++ Loading the PLUMED kernel runtime +++
+++ PLUMED_KERNEL="/net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/libplumedKernel.so" +++

PLUMED: PLUMED is starting
PLUMED: Version: 2.8.0 (git: ddb59d16b) compiled on May 17 2022 at 19:07:56
PLUMED: Please cite these papers when using PLUMED [1][2]
PLUMED: For further information see the PLUMED web page at http://www.plumed.org
PLUMED: Root: /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed
PLUMED: For installed feature, see /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed/src/config/config.txt
PLUMED: Molecular dynamics engine: driver
PLUMED: Precision of reals: 8
PLUMED: Running over 1 node
PLUMED: Number of threads: 1
PLUMED: Cache line size: 512
PLUMED: Number of atoms: 1591
PLUMED: File suffix: 
PLUMED: FILE: plumed.dat
PLUMED: Action MOLINFO
PLUMED:   with label @0
PLUMED:   pdb file named ../../../ala.pdb contains 1 chains 
PLUMED:   chain named   contains residues 1 to 526 and atoms 1 to 1591 
PLUMED: Action TORSION
PLUMED:   with label phi
PLUMED:   between atoms 5 7 9 15
PLUMED:   using periodic boundary conditions
PLUMED: Action TORSION
PLUMED:   with label psi
PLUMED:   between atoms 7 9 15 17
PLUMED:   using periodic boundary conditions
PLUMED: Action PRINT
PLUMED:   with label @3
PLUMED:   with stride 1
PLUMED:   with arguments phi psi
PLUMED:   on file COLVAR
PLUMED:   with format  %f
PLUMED: END FILE: plumed.dat
PLUMED: Timestep: 1.000000
PLUMED: KbT has not been set by the MD engine
PLUMED: It should be set by hand where needed
PLUMED: Relevant bibliography:
PLUMED:   [1] The PLUMED consortium, Nat. Methods 16, 670 (2019)
PLUMED:   [2] Tribello, Bonomi, Branduardi, Camilloni, and Bussi, Comput. Phys. Commun. 185, 604 (2014)
PLUMED: Please read and cite where appropriate!
PLUMED: Finished setup
PLUMED:                                               Cycles        Total      Average      Minimum      Maximum
PLUMED:                                                    1     0.091534     0.091534     0.091534     0.091534
PLUMED: 1 Prepare dependencies                          1001     0.002018     0.000002     0.000001     0.000012
PLUMED: 2 Sharing data                                  1001     0.010278     0.000010     0.000009     0.000256
PLUMED: 3 Waiting for data                              1001     0.003398     0.000003     0.000003     0.000021
PLUMED: 4 Calculating (forward loop)                    1001     0.011329     0.000011     0.000011     0.000043
PLUMED: 5 Applying (backward loop)                      1001     0.014719     0.000015     0.000014     0.000031
PLUMED: 6 Update                                        1001     0.009225     0.000009     0.000008     0.000063

+++ Loading the PLUMED kernel runtime +++
+++ PLUMED_KERNEL="/net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/libplumedKernel.so" +++

PLUMED: PLUMED is starting
PLUMED: Version: 2.8.0 (git: ddb59d16b) compiled on May 17 2022 at 19:07:56
PLUMED: Please cite these papers when using PLUMED [1][2]
PLUMED: For further information see the PLUMED web page at http://www.plumed.org
PLUMED: Root: /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed
PLUMED: For installed feature, see /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed/src/config/config.txt
PLUMED: Molecular dynamics engine: driver
PLUMED: Precision of reals: 8
PLUMED: Running over 1 node
PLUMED: Number of threads: 1
PLUMED: Cache line size: 512
PLUMED: Number of atoms: 1591
PLUMED: File suffix: 
PLUMED: FILE: plumed.dat
PLUMED: Action MOLINFO
PLUMED:   with label @0
PLUMED:   pdb file named ../../../ala.pdb contains 1 chains 
PLUMED:   chain named   contains residues 1 to 526 and atoms 1 to 1591 
PLUMED: Action TORSION
PLUMED:   with label phi
PLUMED:   between atoms 5 7 9 15
PLUMED:   using periodic boundary conditions
PLUMED: Action TORSION
PLUMED:   with label psi
PLUMED:   between atoms 7 9 15 17
PLUMED:   using periodic boundary conditions
PLUMED: Action PRINT
PLUMED:   with label @3
PLUMED:   with stride 1
PLUMED:   with arguments phi psi
PLUMED:   on file COLVAR
PLUMED:   with format  %f
PLUMED: END FILE: plumed.dat
PLUMED: Timestep: 1.000000
PLUMED: KbT has not been set by the MD engine
PLUMED: It should be set by hand where needed
PLUMED: Relevant bibliography:
PLUMED:   [1] The PLUMED consortium, Nat. Methods 16, 670 (2019)
PLUMED:   [2] Tribello, Bonomi, Branduardi, Camilloni, and Bussi, Comput. Phys. Commun. 185, 604 (2014)
PLUMED: Please read and cite where appropriate!
PLUMED: Finished setup
PLUMED:                                               Cycles        Total      Average      Minimum      Maximum
PLUMED:                                                    1     0.093062     0.093062     0.093062     0.093062
PLUMED: 1 Prepare dependencies                          1001     0.002012     0.000002     0.000001     0.000011
PLUMED: 2 Sharing data                                  1001     0.010312     0.000010     0.000009     0.000263
PLUMED: 3 Waiting for data                              1001     0.003453     0.000003     0.000003     0.000021
PLUMED: 4 Calculating (forward loop)                    1001     0.011439     0.000011     0.000010     0.000043
PLUMED: 5 Applying (backward loop)                      1001     0.014898     0.000015     0.000014     0.000030
PLUMED: 6 Update                                        1001     0.009661     0.000010     0.000009     0.000060
PLUMED: PLUMED is starting
PLUMED: Version: 2.8.0 (git: ddb59d16b) compiled on May 17 2022 at 19:07:56
PLUMED: Please cite these papers when using PLUMED [1][2]
PLUMED: For further information see the PLUMED web page at http://www.plumed.org
PLUMED: Root: /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed
PLUMED: For installed feature, see /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed/src/config/config.txt
PLUMED: Molecular dynamics engine: driver
PLUMED: Precision of reals: 8
PLUMED: Running over 1 node
PLUMED: Number of threads: 1
PLUMED: Cache line size: 512
PLUMED: Number of atoms: 1591
PLUMED: File suffix: 
PLUMED: FILE: plumed.dat
PLUMED: Action MOLINFO
PLUMED:   with label @0
PLUMED:   pdb file named ../../../ala.pdb contains 1 chains 
PLUMED:   chain named   contains residues 1 to 526 and atoms 1 to 1591 
PLUMED: Action TORSION
PLUMED:   with label phi
PLUMED:   between atoms 5 7 9 15
PLUMED:   using periodic boundary conditions
PLUMED: Action TORSION
PLUMED:   with label psi
PLUMED:   between atoms 7 9 15 17
PLUMED:   using periodic boundary conditions
PLUMED: Action PRINT
PLUMED:   with label @3
PLUMED:   with stride 1
PLUMED:   with arguments phi psi
PLUMED:   on file COLVAR
PLUMED:   with format  %f
PLUMED: END FILE: plumed.dat
PLUMED: Timestep: 1.000000
PLUMED: KbT has not been set by the MD engine
PLUMED: It should be set by hand where needed
PLUMED: Relevant bibliography:
PLUMED:   [1] The PLUMED consortium, Nat. Methods 16, 670 (2019)
PLUMED:   [2] Tribello, Bonomi, Branduardi, Camilloni, and Bussi, Comput. Phys. Commun. 185, 604 (2014)
PLUMED: Please read and cite where appropriate!
PLUMED: Finished setup

+++ Loading the PLUMED kernel runtime +++
+++ PLUMED_KERNEL="/net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/libplumedKernel.so" +++

PLUMED:                                               Cycles        Total      Average      Minimum      Maximum
PLUMED:                                                    1     0.092980     0.092980     0.092980     0.092980
PLUMED: 1 Prepare dependencies                          1001     0.002059     0.000002     0.000002     0.000013
PLUMED: 2 Sharing data                                  1001     0.010187     0.000010     0.000009     0.000251
PLUMED: 3 Waiting for data                              1001     0.003470     0.000003     0.000003     0.000020
PLUMED: 4 Calculating (forward loop)                    1001     0.011443     0.000011     0.000011     0.000044
PLUMED: 5 Applying (backward loop)                      1001     0.015032     0.000015     0.000014     0.000079
PLUMED: 6 Update                                        1001     0.009593     0.000010     0.000008     0.000054
PLUMED: PLUMED is starting
PLUMED: Version: 2.8.0 (git: ddb59d16b) compiled on May 17 2022 at 19:07:56
PLUMED: Please cite these papers when using PLUMED [1][2]
PLUMED: For further information see the PLUMED web page at http://www.plumed.org
PLUMED: Root: /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed
PLUMED: For installed feature, see /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed/src/config/config.txt
PLUMED: Molecular dynamics engine: driver
PLUMED: Precision of reals: 8
PLUMED: Running over 1 node
PLUMED: Number of threads: 1
PLUMED: Cache line size: 512
PLUMED: Number of atoms: 1591
PLUMED: File suffix: 
PLUMED: FILE: plumed.dat
PLUMED: Action MOLINFO
PLUMED:   with label @0
PLUMED:   pdb file named ../../../ala.pdb contains 1 chains 
PLUMED:   chain named   contains residues 1 to 526 and atoms 1 to 1591 
PLUMED: Action TORSION
PLUMED:   with label phi
PLUMED:   between atoms 5 7 9 15
PLUMED:   using periodic boundary conditions
PLUMED: Action TORSION
PLUMED:   with label psi
PLUMED:   between atoms 7 9 15 17
PLUMED:   using periodic boundary conditions
PLUMED: Action PRINT
PLUMED:   with label @3
PLUMED:   with stride 1
PLUMED:   with arguments phi psi
PLUMED:   on file COLVAR
PLUMED:   with format  %f
PLUMED: END FILE: plumed.dat
PLUMED: Timestep: 1.000000
PLUMED: KbT has not been set by the MD engine
PLUMED: It should be set by hand where needed
PLUMED: Relevant bibliography:
PLUMED:   [1] The PLUMED consortium, Nat. Methods 16, 670 (2019)
PLUMED:   [2] Tribello, Bonomi, Branduardi, Camilloni, and Bussi, Comput. Phys. Commun. 185, 604 (2014)
PLUMED: Please read and cite where appropriate!
PLUMED: Finished setup

+++ Loading the PLUMED kernel runtime +++
+++ PLUMED_KERNEL="/net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/libplumedKernel.so" +++

It looks like there are transitions between the two minima (though probably not so many independent ones):

for i in range(len(lambdas)):
    plt.plot(colvar[i].time,colvar[i].phi,".")

for i in range(len(lambdas)):
    plt.plot(colvar[i].time,colvar[i].psi,".")

We then compute the lambda dependent temperature and show it as if it was a "temperature dependence"

pop_hrex=[]
for i in range(len(lambdas)):
    pop_hrex.append(
        np.sum(np.logical_and(colvar[i].phi<2.0,colvar[i].phi>0))/
        np.sum(np.logical_or(colvar[i].phi>2.0,colvar[i].phi<0))
    )

plt.plot(Ts,pop_pt,"x-",label="pt")
plt.plot(Ts[0]/lambdas,pop_hrex,"x-",label="hrex")
plt.xlabel("T or effective T")
plt.ylabel("relative population")
plt.show()

Clearly populations are noisy. You might rerun these simulations longer to have better estimates. In any case, there is clearly some correlation between the results.

Exercise 4: Analyze Hamiltonian replica exchange simulations with WHAM¶

We are now ready to use a full binless WHAM analysis to recover the maximum amount of statistics from our hrex simulation.

# this box takes a while to run
with cd("3/6reps/"):
    run_mpi("gmx_mpi trjcat -cat -f replica?/traj_comp.xtc -o traj.xtc")
    for i in range(len(lambdas)):
        run_mpi("gmx_mpi mdrun -rerun traj.xtc -s replica{}/topol.tpr -e energy{}.edr".format(i,i))
        subprocess.run("echo 11 | gmx_mpi energy -f energy{}.edr -xvg no -o energy{}.xvg".format(i,i),shell=True)

               :-) GROMACS - gmx trjcat, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx trjcat, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/3/6reps
Command line:
  gmx_mpi trjcat -cat -f replica0/traj_comp.xtc replica1/traj_comp.xtc replica2/traj_comp.xtc replica3/traj_comp.xtc replica4/traj_comp.xtc replica5/traj_comp.xtc -o traj.xtc

Reading frame       1 time    1.000   

Summary of files and start times used:

          File                Start time       Time step
---------------------------------------------------------
   replica0/traj_comp.xtc        0.000 ps        1.000 ps
   replica1/traj_comp.xtc        0.000 ps        1.000 ps WARNING: same Start time as previous
   replica2/traj_comp.xtc        0.000 ps        1.000 ps WARNING: same Start time as previous
   replica3/traj_comp.xtc        0.000 ps        1.000 ps WARNING: same Start time as previous
   replica4/traj_comp.xtc        0.000 ps        1.000 ps WARNING: same Start time as previous
   replica5/traj_comp.xtc        0.000 ps        1.000 ps WARNING: same Start time as previous

Reading frame       0 time    0.000   
Continue writing frames from replica0/traj_comp.xtc t=0 ps, frame=0      
Reading frame     600 time  600.000    ->  frame    600 time  600.000 ps

Note that major changes are planned in future for trjcat, to improve usability and utility.
lasttime 0

Last frame       1000 time 1000.000   
Reading frame       0 time    0.000   
Continue writing frames from replica1/traj_comp.xtc t=0 ps, frame=1001      
 ->  frame   1320 time  319.000 ps

lasttime 1000

Last frame       1000 time 1000.000   
Reading frame       0 time    0.000   
Continue writing frames from replica2/traj_comp.xtc t=0 ps, frame=2002      
 ->  frame   2440 time  438.000 ps

lasttime 1000

Last frame       1000 time 1000.000   
Reading frame       0 time    0.000   
Continue writing frames from replica3/traj_comp.xtc t=0 ps, frame=3003      
Reading frame     150 time  150.000    ->  frame   3160 time  157.000 ps

lasttime 1000

Last frame       1000 time 1000.000   
Reading frame       0 time    0.000   
Continue writing frames from replica4/traj_comp.xtc t=0 ps, frame=4004      
 ->  frame   4640 time  636.000 ps

lasttime 1000

import wham
print(wham.__file__) # make sure you are using the wham script provided with this masterclass
energies=[]
with cd("3/6reps"):
    for i in range(len(lambdas)):
        energies.append(np.loadtxt("energy{}.xvg".format(i),usecols=1))
energies=np.array(energies).T

/scratch/bussi/masterclass-22-10-solution/wham.py

energies.shape

(6006, 6)

# energies[i,j] is the energy of frame i according to the j-th Hamiltonian
kBT=0.00831446261815324*300
w=wham.wham(energies,T=kBT)
# w["logW"] are the Boltzmann factors
# notice that we did not specify yet to which ensemble we are reweighting
# this can be done with the following line:
logW=w["logW"]-energies[:,0]/kBT
# logW are the logaritm weights to obtain properties corresponding to replica 0
logW-=np.max(logW) # avoid numerical errors in exp
weights=np.exp(logW)
weights/=np.sum(weights) # normalize weights
# these weights can be used to compute weighted averages

Now we plot the weights corresponding to the first replica. These weights can be obtained to compute averages in the reference ensemble

plt.plot(weights)
plt.xlabel("frame numbers")
plt.ylabel("weight")
plt.show()

Clearly the weights decrease for frames corresponding to higher replicas.

Interestingly, the weight of a conformation depends only on the conformation, and not on the number of the frames. As a consequence, the very first frame of each replica, which was equilibrated using the reference potential energy, has a high weight. Probably these frames should be discarded, as system is not equilibrated yet. Anyway, they likely contribute very little to the weights.

The cumulative sum shows clear "kinks" when we pass from one of the replicas to the next one, showing that clearly the average weight depends on the replica index.

plt.plot(np.cumsum(weights))
plt.ylabel("cumulative sum of weights")
plt.ylabel("frame index")
plt.show()

c=[]
for i in range(len(colvar)):
    c.append(np.array(colvar[i].phi))
colvar_cat=np.hstack(c)
colvar_cat.shape

(6006,)

np.array(colvar[i].phi)
print(
    np.sum(weights*np.logical_and(colvar_cat<2.0,colvar_cat>0))/
    np.sum(weights*np.logical_or(colvar_cat>2.0,colvar_cat<0))
)

0.043619042468797053

weight_of_lambda=[]
for i in range(len(lambdas)): 
    de=(-energies[:,i]+energies[:,0])/kBT
    weight_of_lambda.append(weights*np.exp(de-np.max(de))/np.sum(weights*np.exp(de-np.max(de))))
    
    
pop_wham=[]
for i in range(len(lambdas)):    
    pop_wham.append(
        np.sum(weight_of_lambda[i]*np.logical_and(colvar_cat<2.0,colvar_cat>0))/
        np.sum(weight_of_lambda[i]*np.logical_or(colvar_cat>2.0,colvar_cat<0))
    )

plt.plot(Ts,pop_pt,label="pt")
plt.plot(Ts[0]/lambdas,pop_hrex,label="hrex, one replica")
plt.plot(Ts[0]/lambdas,pop_wham,label="wham")
plt.legend()
plt.xlabel("T or effective T")
plt.ylabel("relative population")
plt.show()

Notice that we can also plot the weights that allow to compute averages at replicas different from the first one.

From the plot below, you can easily see that when you compute weights corresponding to the higher replicas, the frames close to the end of the concatenated trajectory become more important.

for i in range(len(weight_of_lambda)):
    plt.title("weights for replica {}".format(i))
    plt.plot(weight_of_lambda[i])
    plt.show()

Exercise 5: Optimize the lambda ladder¶

We next try to optimize the lambda ladder. To do so, we will run a minimization procedure. In this minimization we should be able to compute averages corresponding to new values of lambda (including values not simulated yet!) quickly. To do so, we realize that the energy of a frame is a quadratic function of the sqrt(lambda).

# this is what we see plotting energy as a function of lambda for a random frame
plt.plot(lambdas,energies[100])
plt.xlabel("lambda")
plt.ylabel("energy")
plt.show()

One might think that energy should decrease monotonously with lambda based on the plot above. However, we can just search from frames where this is not true.

check=[]
for i in range(energies.shape[0]):
    ok=False
    for j in range(energies.shape[1]-1):
        if(energies[i,j]>energies[i,j+1]):
            ok=True
    if ok:
        check.append(i)
check[0]
plt.plot(lambdas,energies[check[0],:])
# this is just one for example
plt.xlabel("lambda")
plt.ylabel("energy")
plt.show()

Then, for each frame, we fit the coefficients of the function that allows us to recompute energies for arbitrary values of lambda

coefficients=[]
for i in range(energies.shape[0]):
    c=np.polyfit(np.sqrt(lambdas),energies[i],2)
    coefficients.append(c)
coefficients=np.array(coefficients)
residual=np.dot(coefficients,
                np.array((lambdas,np.sqrt(lambdas),np.ones(len(lambdas)))))-energies

# This is the residual, very small
plt.plot(residual.T.flatten(),".")
plt.xlabel("frame index")
plt.ylabel("residual")
plt.show()

Now we can write a function that predicts the acceptance between two arbitrary values of lambda

def predict_acceptance(l1,l2):
    # energies corresponding to l1 and l2, using the interpolated function
    e1=coefficients[:,0]*l1+coefficients[:,1]*np.sqrt(l1)+coefficients[:,2]
    e2=coefficients[:,0]*l2+coefficients[:,1]*np.sqrt(l2)+coefficients[:,2]
    
    # weights corresponding to l1 and l2
    logW1=w["logW"]-e1/kBT
    logW2=w["logW"]-e2/kBT
    logW1-=np.max(logW1)
    logW2-=np.max(logW2)
    w1=np.exp(logW1)
    w2=np.exp(logW2)
    w1/=np.sum(w1)
    w2/=np.sum(w2)
    
    # then we have to compute the average of the acceptance, which means the average of:
    # np.minimum(w1[j]*w2[i]/(w1[i]*w2[j]),1.0)
    # (compare with Metropolis rule)
    
    # We can do it in two ways. One way is to randomly sample a number (10000, below)
    # of frames from the two ensembles and using them to compute the average acceptance.
    
    #c1=np.random.choice(len(w1),10000,p=w1)
    #c2=np.random.choice(len(w2),10000,p=w2)
    #a=np.minimum(w1[c2]*w2[c1]/(w1[c1]*w2[c2]),1.0)
    #return np.average(a)
    
    # the problem of this approach is that it is random, thus not good to construct a function to be minimized
    
    # alternatively we consider all pairs of frame, with a skip (5) here to just run faster
    w1=w1[::5]
    w2=w2[::5]
    a=[]
    for i in range(len(w1)):
        a.append(np.average(np.minimum(w1[:]*w2[i]/(w1[i]*w2[:]),1.0),weights=w2))
    return np.average(a,weights=w1)

We can now try to "predict" the acceptance corresponding to the simulated lambdas:

for i in range(len(lambdas)-1):
    print(predict_acceptance(lambdas[i],lambdas[i+1]))

0.5340728519771433
0.4940495074929892
0.39317545103849977
0.2990660445826373
0.11859391076651352

!grep -A 3 "Repl  average probabilities" 3/6reps/replica0/md.log

Repl  average probabilities:
Repl     0    1    2    3    4    5
Repl      .55  .47  .41  .30  .12
Repl  number of exchanges:

Results are very close to the observed averages!

For instance, these would be the acceptance rates predicted for a geometric (rather than linear) arrangement of lambdas

geometric_lambdas=np.geomspace(1.0,lambdas[-1],len(lambdas))
for i in range(len(lambdas)-1):
    print(predict_acceptance(geometric_lambdas[i],geometric_lambdas[i+1]))

0.2308635220970505
0.2783660937613743
0.3601825408394625
0.405309116856007
0.46225107980567376

The acceptance rates are more even. However, they kind of suffer the opposite problem, namely, acceptance is lower at low replicas. The ladder giving homogeneous rates is somewhere in between linear and geometric.

# lambda is the array with the replicas we used for out simulation (a starting point(
# first and last replicas are fixed (1.0 and lambda_min)
# we thus optimize all elements excluding the first and the last (lambda[1:-1])

lambda_min=lambdas[-1]
 
# this function compute the acceptances given the three intermediate values of
# lambda
def predict_all_acceptances(x):
    x=np.array(x)
    acc=[]
    # lambdas should not be negative
    if np.any(x<=0.0):
        return np.zeros(len(x)+1)
    # nor larger than 1
    if np.any(x>=1.0):
        return np.zeros(len(x)+1)
    # first replica has lambda=1.0
    acc.append(predict_acceptance(1.0,x[0]))
    for i in range(len(x)-1):
        acc.append(predict_acceptance(x[i],x[i+1]))
    # last replica has lambda=lambda_min
    acc.append(predict_acceptance(x[-1],lambda_min))
    return acc
 
# this is function to be minimized
def func(x):
    acc=predict_all_acceptances(x)
    print(np.prod(acc)**(1/len(acc)))
    return -np.prod(acc) # negative, since we minimize the result
 
from scipy.optimize import minimize
res=minimize(func,lambdas[1:-1]) # starting values
optimal_lambdas=np.hstack((1.0,res.x,lambda_min))

print("optimal lambdas: ",optimal_lambdas)
print(predict_all_acceptances(res.x))

0.3259547940141271
0.325954793893094
0.3259547933599866
0.32595479341658784
0.32595479063001087
0.32872898680400764
0.3287289866628208
0.3287289861609084
0.32872898589937016
0.32872898428899244
0.3324338906848852
0.332433890492614
0.3324338900161506
0.332433889264336
0.33243389043460636
0.3335674496997589
0.3335674494601702
0.3335674489523692
0.33356744824582124
0.3335674507617446
0.33492136496655117
0.3349213646673383
0.3349213641049616
0.3349213638673992
0.33492136635635655
0.3370537872278969
0.33705378681141973
0.33705378622693233
0.3370537869108689
0.33705378833631455
0.3387366060583005
0.33873660542528716
0.3387366051464819
0.33873660690403357
0.338736606139612
0.3392440704975935
0.33924406983730954
0.3392440698256017
0.3392440712567459
0.3392440704650873
0.3400495109208426
0.34004951026656627
0.3400495107221155
0.3400495113067745
0.34004951084923557
0.34059499359211687
0.3405949930240917
0.3405949940213559
0.3405949934500063
0.34059499350976274
0.3407499482921569
0.3407499478259862
0.3407499487076727
0.3407499480902852
0.3407499482273833
0.34099700348003786
0.34099700321399296
0.34099700376025366
0.3409970033143528
0.3409970034358405
0.3411313735410154
0.34113137355160816
0.3411313735570486
0.3411313735160984
0.34113137353392803
0.3411321736743728
0.3411321736759918
0.3411321736790199
0.3411321736687054
0.3411321736721058
0.34113221701542207
0.34113221701546226
0.341132217015375
0.3411322170154294
0.3411322170153383
optimal lambdas:  [1.         0.7490915  0.54002926 0.37503371 0.24605241 0.15      ]
[0.34973133878840834, 0.3417108301273635, 0.34368442978425084, 0.3414464459182263, 0.3294108774587619]

Predicted acceptance are all ~ 0.33.

Below we compare the possible ladders (linear, geometric, and optimal)

plt.plot(lambdas,label="linear")
plt.plot(optimal_lambdas,label="optimal")
plt.plot(np.geomspace(1.0,lambdas[-1],len(lambdas)),label="geometric")
plt.legend()
plt.xlabel("replica index")
plt.xlabel("lambda")
plt.show()

Exercise 6: Combine with metadynamics on psi¶

We next run a metadynamics + solute tempering simulation on psi.

Notice that psi is a bad variable for metadynamics on alanine dipeptide.

The setup is basically identical to the one in masterclass 21.5

# hardcoded here if I don't want to rerun previous boxes
optimal_lambdas=np.array([1.,0.7490915,0.54002926,0.37503371,0.24605241,0.15])

# this takes a while
with cd("5/6reps"):
    with open("plumed.dat","w") as f:
        print("""
# vim:ft=plumed
MOLINFO STRUCTURE=../../../ala.pdb
phi: TORSION ATOMS=@phi-2 
psi: TORSION ATOMS=@psi-2 
# You can use the same parameters that you used in masterclass 21.4
m: METAD ...
   ARG=psi
   SIGMA=0.3 
   HEIGHT=1.2
   BIASFACTOR=8 
   PACE=500 
   GRID_MIN=-pi 
   GRID_MAX=pi
   FILE=../HILLS_PTMETAD
...
PRINT ARG=phi,psi FILE=../colvar_ptmetad.dat STRIDE=100
        """,file=f) # empty
    for i in range(len(optimal_lambdas)):
        with cd("replica{}".format(i)):
            run_mpi("cp ../../../2a/processed_selected.top processed_selected.top")
            run_mpi("plumed partial_tempering {} < processed_selected.top > topol.top".format(optimal_lambdas[i]))
            run_mpi("gmx_mpi grompp -c ../../../conf.gro -f ../../../grompp.mdp")
    run_mpi("gmx_mpi mdrun -multidir replica? -replex 200 -nsteps 500000 -ntomp 4 -hrex -plumed ../plumed.dat",len(lambdas))

               :-) GROMACS - gmx grompp, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx grompp, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/5/6reps/replica0
Command line:
  gmx_mpi grompp -c ../../../conf.gro -f ../../../grompp.mdp

Setting the LD random seed to -270665771
Generated 8515 of the 8515 non-bonded parameter combinations
Generating 1-4 interactions: fudge = 0.5
Generated 8515 of the 8515 1-4 parameter combinations
Excluding 3 bonded neighbours molecule type 'Protein'
Excluding 2 bonded neighbours molecule type 'SOL'
Number of degrees of freedom in T-Coupling group Protein is 53.95
Number of degrees of freedom in T-Coupling group SOL is 3135.05
Estimate for the relative computational load of the PME mesh part: 0.19

GROMACS reminds you: "History has expired" (PubMed Central)

turning H bonds into constraints...
turning H bonds into constraints...
Analysing residue names:
There are:     3    Protein residues
There are:   523      Water residues
Analysing Protein...
Determining Verlet buffer for a tolerance of 0.005 kJ/mol/ps at 300 K
Calculated rlist for 1x1 atom pair-list as 1.036 nm, buffer size 0.036 nm
Set rlist, assuming 4x4 atom pair-list, to 1.000 nm, buffer size 0.000 nm
Note that mdrun will redetermine rlist based on the actual pair-list setup
Calculating fourier grid dimensions for X Y Z
Using a fourier grid of 24x24x24, spacing 0.118 0.118 0.118
This run will generate roughly 12 Mb of data

               :-) GROMACS - gmx grompp, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx grompp, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/5/6reps/replica1
Command line:
  gmx_mpi grompp -c ../../../conf.gro -f ../../../grompp.mdp

Setting the LD random seed to -69311494
Generated 8515 of the 8515 non-bonded parameter combinations
Generating 1-4 interactions: fudge = 0.5
Generated 8515 of the 8515 1-4 parameter combinations
Excluding 3 bonded neighbours molecule type 'Protein'
Excluding 2 bonded neighbours molecule type 'SOL'
Number of degrees of freedom in T-Coupling group Protein is 53.95
Number of degrees of freedom in T-Coupling group SOL is 3135.05
Estimate for the relative computational load of the PME mesh part: 0.19

GROMACS reminds you: "I think it would be a good idea." (Mahatma Gandhi, when asked what he thought of Western civilization)

turning H bonds into constraints...
turning H bonds into constraints...
Analysing residue names:
There are:     3    Protein residues
There are:   523      Water residues
Analysing Protein...
Determining Verlet buffer for a tolerance of 0.005 kJ/mol/ps at 300 K
Calculated rlist for 1x1 atom pair-list as 1.036 nm, buffer size 0.036 nm
Set rlist, assuming 4x4 atom pair-list, to 1.000 nm, buffer size 0.000 nm
Note that mdrun will redetermine rlist based on the actual pair-list setup
Calculating fourier grid dimensions for X Y Z
Using a fourier grid of 24x24x24, spacing 0.118 0.118 0.118
This run will generate roughly 12 Mb of data

               :-) GROMACS - gmx grompp, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx grompp, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/5/6reps/replica2
Command line:
  gmx_mpi grompp -c ../../../conf.gro -f ../../../grompp.mdp

Setting the LD random seed to -68178437
Generated 8515 of the 8515 non-bonded parameter combinations
Generating 1-4 interactions: fudge = 0.5
Generated 8515 of the 8515 1-4 parameter combinations
Excluding 3 bonded neighbours molecule type 'Protein'
Excluding 2 bonded neighbours molecule type 'SOL'
Number of degrees of freedom in T-Coupling group Protein is 53.95
Number of degrees of freedom in T-Coupling group SOL is 3135.05
Estimate for the relative computational load of the PME mesh part: 0.19

GROMACS reminds you: "They don't have half hours in the north" (Carl Caleman)

turning H bonds into constraints...
turning H bonds into constraints...
Analysing residue names:
There are:     3    Protein residues
There are:   523      Water residues
Analysing Protein...
Determining Verlet buffer for a tolerance of 0.005 kJ/mol/ps at 300 K
Calculated rlist for 1x1 atom pair-list as 1.036 nm, buffer size 0.036 nm
Set rlist, assuming 4x4 atom pair-list, to 1.000 nm, buffer size 0.000 nm
Note that mdrun will redetermine rlist based on the actual pair-list setup
Calculating fourier grid dimensions for X Y Z
Using a fourier grid of 24x24x24, spacing 0.118 0.118 0.118
This run will generate roughly 12 Mb of data

               :-) GROMACS - gmx grompp, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx grompp, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/5/6reps/replica3
Command line:
  gmx_mpi grompp -c ../../../conf.gro -f ../../../grompp.mdp

Setting the LD random seed to -70288261
Generated 8515 of the 8515 non-bonded parameter combinations
Generating 1-4 interactions: fudge = 0.5
Generated 8515 of the 8515 1-4 parameter combinations
Excluding 3 bonded neighbours molecule type 'Protein'
Excluding 2 bonded neighbours molecule type 'SOL'
Number of degrees of freedom in T-Coupling group Protein is 53.95
Number of degrees of freedom in T-Coupling group SOL is 3135.05
Estimate for the relative computational load of the PME mesh part: 0.19

GROMACS reminds you: "They don't have half hours in the north" (Carl Caleman)

turning H bonds into constraints...
turning H bonds into constraints...
Analysing residue names:
There are:     3    Protein residues
There are:   523      Water residues
Analysing Protein...
Determining Verlet buffer for a tolerance of 0.005 kJ/mol/ps at 300 K
Calculated rlist for 1x1 atom pair-list as 1.036 nm, buffer size 0.036 nm
Set rlist, assuming 4x4 atom pair-list, to 1.000 nm, buffer size 0.000 nm
Note that mdrun will redetermine rlist based on the actual pair-list setup
Calculating fourier grid dimensions for X Y Z
Using a fourier grid of 24x24x24, spacing 0.118 0.118 0.118
This run will generate roughly 12 Mb of data

               :-) GROMACS - gmx grompp, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx grompp, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/5/6reps/replica4
Command line:
  gmx_mpi grompp -c ../../../conf.gro -f ../../../grompp.mdp

Setting the LD random seed to -1143472131
Generated 8515 of the 8515 non-bonded parameter combinations
Generating 1-4 interactions: fudge = 0.5
Generated 8515 of the 8515 1-4 parameter combinations
Excluding 3 bonded neighbours molecule type 'Protein'
Excluding 2 bonded neighbours molecule type 'SOL'
Number of degrees of freedom in T-Coupling group Protein is 53.95
Number of degrees of freedom in T-Coupling group SOL is 3135.05
Estimate for the relative computational load of the PME mesh part: 0.19

GROMACS reminds you: "Wild Pointers Couldn't Drag Me Away" (K.A. Feenstra)

turning H bonds into constraints...
turning H bonds into constraints...
Analysing residue names:
There are:     3    Protein residues
There are:   523      Water residues
Analysing Protein...
Determining Verlet buffer for a tolerance of 0.005 kJ/mol/ps at 300 K
Calculated rlist for 1x1 atom pair-list as 1.036 nm, buffer size 0.036 nm
Set rlist, assuming 4x4 atom pair-list, to 1.000 nm, buffer size 0.000 nm
Note that mdrun will redetermine rlist based on the actual pair-list setup
Calculating fourier grid dimensions for X Y Z
Using a fourier grid of 24x24x24, spacing 0.118 0.118 0.118
This run will generate roughly 12 Mb of data

Nicely, the acceptance are in agreement with our prediction

!grep -A 3 "Repl  average probabilities" 5/6reps/replica0/md.log

Repl  average probabilities:
Repl     0    1    2    3    4    5
Repl      .34  .34  .32  .35  .33
Repl  number of exchanges:

colvar=[]

for i in range(len(optimal_lambdas)):
    with cd("5/6reps/replica"+str(i)):
        with open("plumed_plot.dat","w") as f:
            print("""
MOLINFO STRUCTURE=../../../ala.pdb 
phi: TORSION ATOMS=@phi-2
psi: TORSION ATOMS=@psi-2
PRINT ARG=phi,psi FILE=COLVAR 
                  """,file=f)
        subprocess.run("plumed driver --ixtc traj_comp.xtc --plumed plumed_plot.dat",shell=True)
            
        colvar.append(plumed.read_as_pandas("COLVAR"))

PLUMED: PLUMED is starting
PLUMED: Version: 2.8.0 (git: ddb59d16b) compiled on May 17 2022 at 19:07:56
PLUMED: Please cite these papers when using PLUMED [1][2]
PLUMED: For further information see the PLUMED web page at http://www.plumed.org
PLUMED: Root: /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed
PLUMED: For installed feature, see /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed/src/config/config.txt
PLUMED: Molecular dynamics engine: driver
PLUMED: Precision of reals: 8
PLUMED: Running over 1 node
PLUMED: Number of threads: 1
PLUMED: Cache line size: 512
PLUMED: Number of atoms: 1591
PLUMED: File suffix: 
PLUMED: FILE: plumed_plot.dat
PLUMED: Action MOLINFO
PLUMED:   with label @0
PLUMED:   pdb file named ../../../ala.pdb contains 1 chains 
PLUMED:   chain named   contains residues 1 to 526 and atoms 1 to 1591 
PLUMED: Action TORSION
PLUMED:   with label phi
PLUMED:   between atoms 5 7 9 15
PLUMED:   using periodic boundary conditions
PLUMED: Action TORSION
PLUMED:   with label psi
PLUMED:   between atoms 7 9 15 17
PLUMED:   using periodic boundary conditions
PLUMED: Action PRINT
PLUMED:   with label @3
PLUMED:   with stride 1
PLUMED:   with arguments phi psi
PLUMED:   on file COLVAR
PLUMED:   with format  %f
PLUMED: END FILE: plumed_plot.dat
PLUMED: Timestep: 1.000000
PLUMED: KbT has not been set by the MD engine
PLUMED: It should be set by hand where needed
PLUMED: Relevant bibliography:
PLUMED:   [1] The PLUMED consortium, Nat. Methods 16, 670 (2019)
PLUMED:   [2] Tribello, Bonomi, Branduardi, Camilloni, and Bussi, Comput. Phys. Commun. 185, 604 (2014)
PLUMED: Please read and cite where appropriate!
PLUMED: Finished setup
PLUMED:                                               Cycles        Total      Average      Minimum      Maximum
PLUMED:                                                    1     0.092033     0.092033     0.092033     0.092033
PLUMED: 1 Prepare dependencies                          1001     0.002104     0.000002     0.000002     0.000015
PLUMED: 2 Sharing data                                  1001     0.010230     0.000010     0.000009     0.000250
PLUMED: 3 Waiting for data                              1001     0.003381     0.000003     0.000003     0.000020
PLUMED: 4 Calculating (forward loop)                    1001     0.011169     0.000011     0.000010     0.000042
PLUMED: 5 Applying (backward loop)                      1001     0.014775     0.000015     0.000014     0.000035
PLUMED: 6 Update                                        1001     0.009514     0.000010     0.000009     0.000056

+++ Loading the PLUMED kernel runtime +++
+++ PLUMED_KERNEL="/net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/libplumedKernel.so" +++

PLUMED: PLUMED is starting
PLUMED: Version: 2.8.0 (git: ddb59d16b) compiled on May 17 2022 at 19:07:56
PLUMED: Please cite these papers when using PLUMED [1][2]
PLUMED: For further information see the PLUMED web page at http://www.plumed.org
PLUMED: Root: /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed
PLUMED: For installed feature, see /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed/src/config/config.txt
PLUMED: Molecular dynamics engine: driver
PLUMED: Precision of reals: 8
PLUMED: Running over 1 node
PLUMED: Number of threads: 1
PLUMED: Cache line size: 512
PLUMED: Number of atoms: 1591
PLUMED: File suffix: 
PLUMED: FILE: plumed_plot.dat
PLUMED: Action MOLINFO
PLUMED:   with label @0
PLUMED:   pdb file named ../../../ala.pdb contains 1 chains 
PLUMED:   chain named   contains residues 1 to 526 and atoms 1 to 1591 
PLUMED: Action TORSION
PLUMED:   with label phi
PLUMED:   between atoms 5 7 9 15
PLUMED:   using periodic boundary conditions
PLUMED: Action TORSION
PLUMED:   with label psi
PLUMED:   between atoms 7 9 15 17
PLUMED:   using periodic boundary conditions
PLUMED: Action PRINT
PLUMED:   with label @3
PLUMED:   with stride 1
PLUMED:   with arguments phi psi
PLUMED:   on file COLVAR
PLUMED:   with format  %f
PLUMED: END FILE: plumed_plot.dat
PLUMED: Timestep: 1.000000
PLUMED: KbT has not been set by the MD engine
PLUMED: It should be set by hand where needed
PLUMED: Relevant bibliography:
PLUMED:   [1] The PLUMED consortium, Nat. Methods 16, 670 (2019)
PLUMED:   [2] Tribello, Bonomi, Branduardi, Camilloni, and Bussi, Comput. Phys. Commun. 185, 604 (2014)
PLUMED: Please read and cite where appropriate!
PLUMED: Finished setup
PLUMED:                                               Cycles        Total      Average      Minimum      Maximum
PLUMED:                                                    1     0.092327     0.092327     0.092327     0.092327
PLUMED: 1 Prepare dependencies                          1001     0.002092     0.000002     0.000002     0.000013
PLUMED: 2 Sharing data                                  1001     0.010137     0.000010     0.000009     0.000268
PLUMED: 3 Waiting for data                              1001     0.003397     0.000003     0.000003     0.000021
PLUMED: 4 Calculating (forward loop)                    1001     0.011290     0.000011     0.000010     0.000054
PLUMED: 5 Applying (backward loop)                      1001     0.014742     0.000015     0.000014     0.000030
PLUMED: 6 Update                                        1001     0.009562     0.000010     0.000008     0.000057
PLUMED: PLUMED is starting
PLUMED: Version: 2.8.0 (git: ddb59d16b) compiled on May 17 2022 at 19:07:56
PLUMED: Please cite these papers when using PLUMED [1][2]
PLUMED: For further information see the PLUMED web page at http://www.plumed.org
PLUMED: Root: /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed
PLUMED: For installed feature, see /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed/src/config/config.txt
PLUMED: Molecular dynamics engine: driver
PLUMED: Precision of reals: 8
PLUMED: Running over 1 node
PLUMED: Number of threads: 1
PLUMED: Cache line size: 512
PLUMED: Number of atoms: 1591
PLUMED: File suffix: 
PLUMED: FILE: plumed_plot.dat
PLUMED: Action MOLINFO
PLUMED:   with label @0
PLUMED:   pdb file named ../../../ala.pdb contains 1 chains 
PLUMED:   chain named   contains residues 1 to 526 and atoms 1 to 1591 
PLUMED: Action TORSION
PLUMED:   with label phi
PLUMED:   between atoms 5 7 9 15
PLUMED:   using periodic boundary conditions
PLUMED: Action TORSION
PLUMED:   with label psi
PLUMED:   between atoms 7 9 15 17
PLUMED:   using periodic boundary conditions
PLUMED: Action PRINT
PLUMED:   with label @3
PLUMED:   with stride 1
PLUMED:   with arguments phi psi
PLUMED:   on file COLVAR
PLUMED:   with format  %f
PLUMED: END FILE: plumed_plot.dat
PLUMED: Timestep: 1.000000
PLUMED: KbT has not been set by the MD engine
PLUMED: It should be set by hand where needed
PLUMED: Relevant bibliography:
PLUMED:   [1] The PLUMED consortium, Nat. Methods 16, 670 (2019)
PLUMED:   [2] Tribello, Bonomi, Branduardi, Camilloni, and Bussi, Comput. Phys. Commun. 185, 604 (2014)
PLUMED: Please read and cite where appropriate!
PLUMED: Finished setup

+++ Loading the PLUMED kernel runtime +++
+++ PLUMED_KERNEL="/net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/libplumedKernel.so" +++

PLUMED:                                               Cycles        Total      Average      Minimum      Maximum
PLUMED:                                                    1     0.091335     0.091335     0.091335     0.091335
PLUMED: 1 Prepare dependencies                          1001     0.002040     0.000002     0.000002     0.000013
PLUMED: 2 Sharing data                                  1001     0.010122     0.000010     0.000009     0.000279
PLUMED: 3 Waiting for data                              1001     0.003369     0.000003     0.000003     0.000020
PLUMED: 4 Calculating (forward loop)                    1001     0.011145     0.000011     0.000010     0.000043
PLUMED: 5 Applying (backward loop)                      1001     0.014699     0.000015     0.000014     0.000026
PLUMED: 6 Update                                        1001     0.009292     0.000009     0.000008     0.000056

+++ Loading the PLUMED kernel runtime +++
+++ PLUMED_KERNEL="/net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/libplumedKernel.so" +++

PLUMED: PLUMED is starting
PLUMED: Version: 2.8.0 (git: ddb59d16b) compiled on May 17 2022 at 19:07:56
PLUMED: Please cite these papers when using PLUMED [1][2]
PLUMED: For further information see the PLUMED web page at http://www.plumed.org
PLUMED: Root: /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed
PLUMED: For installed feature, see /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed/src/config/config.txt
PLUMED: Molecular dynamics engine: driver
PLUMED: Precision of reals: 8
PLUMED: Running over 1 node
PLUMED: Number of threads: 1
PLUMED: Cache line size: 512
PLUMED: Number of atoms: 1591
PLUMED: File suffix: 
PLUMED: FILE: plumed_plot.dat
PLUMED: Action MOLINFO
PLUMED:   with label @0
PLUMED:   pdb file named ../../../ala.pdb contains 1 chains 
PLUMED:   chain named   contains residues 1 to 526 and atoms 1 to 1591 
PLUMED: Action TORSION
PLUMED:   with label phi
PLUMED:   between atoms 5 7 9 15
PLUMED:   using periodic boundary conditions
PLUMED: Action TORSION
PLUMED:   with label psi
PLUMED:   between atoms 7 9 15 17
PLUMED:   using periodic boundary conditions
PLUMED: Action PRINT
PLUMED:   with label @3
PLUMED:   with stride 1
PLUMED:   with arguments phi psi
PLUMED:   on file COLVAR
PLUMED:   with format  %f
PLUMED: END FILE: plumed_plot.dat
PLUMED: Timestep: 1.000000
PLUMED: KbT has not been set by the MD engine
PLUMED: It should be set by hand where needed
PLUMED: Relevant bibliography:
PLUMED:   [1] The PLUMED consortium, Nat. Methods 16, 670 (2019)
PLUMED:   [2] Tribello, Bonomi, Branduardi, Camilloni, and Bussi, Comput. Phys. Commun. 185, 604 (2014)
PLUMED: Please read and cite where appropriate!
PLUMED: Finished setup
PLUMED:                                               Cycles        Total      Average      Minimum      Maximum
PLUMED:                                                    1     0.091119     0.091119     0.091119     0.091119
PLUMED: 1 Prepare dependencies                          1001     0.001986     0.000002     0.000002     0.000012
PLUMED: 2 Sharing data                                  1001     0.009956     0.000010     0.000009     0.000251
PLUMED: 3 Waiting for data                              1001     0.003412     0.000003     0.000003     0.000020
PLUMED: 4 Calculating (forward loop)                    1001     0.011114     0.000011     0.000010     0.000043
PLUMED: 5 Applying (backward loop)                      1001     0.014684     0.000015     0.000014     0.000030
PLUMED: 6 Update                                        1001     0.009175     0.000009     0.000008     0.000057

+++ Loading the PLUMED kernel runtime +++
+++ PLUMED_KERNEL="/net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/libplumedKernel.so" +++

PLUMED: PLUMED is starting
PLUMED: Version: 2.8.0 (git: ddb59d16b) compiled on May 17 2022 at 19:07:56
PLUMED: Please cite these papers when using PLUMED [1][2]
PLUMED: For further information see the PLUMED web page at http://www.plumed.org
PLUMED: Root: /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed
PLUMED: For installed feature, see /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed/src/config/config.txt
PLUMED: Molecular dynamics engine: driver
PLUMED: Precision of reals: 8
PLUMED: Running over 1 node
PLUMED: Number of threads: 1
PLUMED: Cache line size: 512
PLUMED: Number of atoms: 1591
PLUMED: File suffix: 
PLUMED: FILE: plumed_plot.dat
PLUMED: Action MOLINFO
PLUMED:   with label @0
PLUMED:   pdb file named ../../../ala.pdb contains 1 chains 
PLUMED:   chain named   contains residues 1 to 526 and atoms 1 to 1591 
PLUMED: Action TORSION
PLUMED:   with label phi
PLUMED:   between atoms 5 7 9 15
PLUMED:   using periodic boundary conditions
PLUMED: Action TORSION
PLUMED:   with label psi
PLUMED:   between atoms 7 9 15 17
PLUMED:   using periodic boundary conditions
PLUMED: Action PRINT
PLUMED:   with label @3
PLUMED:   with stride 1
PLUMED:   with arguments phi psi
PLUMED:   on file COLVAR
PLUMED:   with format  %f
PLUMED: END FILE: plumed_plot.dat
PLUMED: Timestep: 1.000000
PLUMED: KbT has not been set by the MD engine
PLUMED: It should be set by hand where needed
PLUMED: Relevant bibliography:
PLUMED:   [1] The PLUMED consortium, Nat. Methods 16, 670 (2019)
PLUMED:   [2] Tribello, Bonomi, Branduardi, Camilloni, and Bussi, Comput. Phys. Commun. 185, 604 (2014)
PLUMED: Please read and cite where appropriate!
PLUMED: Finished setup
PLUMED:                                               Cycles        Total      Average      Minimum      Maximum
PLUMED:                                                    1     0.091733     0.091733     0.091733     0.091733
PLUMED: 1 Prepare dependencies                          1001     0.002053     0.000002     0.000001     0.000014
PLUMED: 2 Sharing data                                  1001     0.010138     0.000010     0.000009     0.000272
PLUMED: 3 Waiting for data                              1001     0.003430     0.000003     0.000003     0.000020
PLUMED: 4 Calculating (forward loop)                    1001     0.011154     0.000011     0.000010     0.000043
PLUMED: 5 Applying (backward loop)                      1001     0.014941     0.000015     0.000014     0.000026
PLUMED: 6 Update                                        1001     0.009502     0.000009     0.000009     0.000067

+++ Loading the PLUMED kernel runtime +++
+++ PLUMED_KERNEL="/net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/libplumedKernel.so" +++

for i in range(len(optimal_lambdas)):
    plt.plot(colvar[i].time,colvar[i].phi,".")

for i in range(len(lambdas)):
    plt.plot(colvar[i].time,colvar[i].psi,".")

Then we concatenate the trajectories to do a WHAM analysis. We also use mdrun -rerun to compute energies corresponding to all replicas. In theory, we could do this for 3 replicas only and then interpolate the other values using the trick that we have used in the previous exercise. For simplicitly, we here just rerun with all topologies.

# this takes a while
with cd("5/6reps/"):
    run_mpi("gmx_mpi trjcat -cat -f replica?/traj_comp.xtc -o traj.xtc")
    for i in range(len(optimal_lambdas)):
        run_mpi("gmx_mpi mdrun -rerun traj.xtc -s replica{}/topol.tpr -e energy{}.edr".format(i,i))
        subprocess.run("echo 11 | gmx_mpi energy -f energy{}.edr -xvg no -o energy{}.xvg".format(i,i),shell=True)

               :-) GROMACS - gmx trjcat, 2020.6-plumed-2.8.0 (-:

                            GROMACS is written by:
     Emile Apol      Rossen Apostolov      Paul Bauer     Herman J.C. Berendsen
    Par Bjelkmar      Christian Blau   Viacheslav Bolnykh     Kevin Boyd    
 Aldert van Buuren   Rudi van Drunen     Anton Feenstra       Alan Gray     
  Gerrit Groenhof     Anca Hamuraru    Vincent Hindriksen  M. Eric Irrgang  
  Aleksei Iupinov   Christoph Junghans     Joe Jordan     Dimitrios Karkoulis
    Peter Kasson        Jiri Kraus      Carsten Kutzner      Per Larsson    
  Justin A. Lemkul    Viveca Lindahl    Magnus Lundborg     Erik Marklund   
    Pascal Merz     Pieter Meulenhoff    Teemu Murtola       Szilard Pall   
    Sander Pronk      Roland Schulz      Michael Shirts    Alexey Shvetsov  
   Alfons Sijbers     Peter Tieleman      Jon Vincent      Teemu Virolainen 
 Christian Wennberg    Maarten Wolf      Artem Zhmurov   
                           and the project leaders:
        Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2019, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS:      gmx trjcat, version 2020.6-plumed-2.8.0
Executable:   /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/bin/gmx_mpi
Data prefix:  /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022
Working dir:  /scratch/bussi/masterclass-22-10-solution/5/6reps
Command line:
  gmx_mpi trjcat -cat -f replica0/traj_comp.xtc replica1/traj_comp.xtc replica2/traj_comp.xtc replica3/traj_comp.xtc replica4/traj_comp.xtc replica5/traj_comp.xtc -o traj.xtc

Reading frame       1 time    1.000   

Summary of files and start times used:

          File                Start time       Time step
---------------------------------------------------------
   replica0/traj_comp.xtc        0.000 ps        1.000 ps
   replica1/traj_comp.xtc        0.000 ps        1.000 ps WARNING: same Start time as previous
   replica2/traj_comp.xtc        0.000 ps        1.000 ps WARNING: same Start time as previous
   replica3/traj_comp.xtc        0.000 ps        1.000 ps WARNING: same Start time as previous
   replica4/traj_comp.xtc        0.000 ps        1.000 ps WARNING: same Start time as previous
   replica5/traj_comp.xtc        0.000 ps        1.000 ps WARNING: same Start time as previous

Reading frame       0 time    0.000   
Continue writing frames from replica0/traj_comp.xtc t=0 ps, frame=0      
Reading frame     600 time  600.000    ->  frame    600 time  600.000 ps

Note that major changes are planned in future for trjcat, to improve usability and utility.
lasttime 0

Last frame       1000 time 1000.000   
Reading frame       0 time    0.000   
Continue writing frames from replica1/traj_comp.xtc t=0 ps, frame=1001      
Reading frame     300 time  300.000    ->  frame   1310 time  309.000 ps

lasttime 1000

Last frame       1000 time 1000.000   
Reading frame       0 time    0.000   
Continue writing frames from replica2/traj_comp.xtc t=0 ps, frame=2002      
Reading frame      20 time   20.000

lasttime 1000

Last frame       1000 time 1000.000   
Reading frame       0 time    0.000   
Continue writing frames from replica3/traj_comp.xtc t=0 ps, frame=3003      
Reading frame     130 time  130.000

lasttime 1000

Last frame       1000 time 1000.000   
Reading frame       0 time    0.000   
Continue writing frames from replica4/traj_comp.xtc t=0 ps, frame=4004      
 ->  frame   4550 time  546.000 ps

lasttime 1000

Additionally, we have to compute the value of the bias potential according to all replicas

# this also takes a while
with cd("5/6reps/"):
    with open("plumed_ptmetad_process.dat","w") as f:
        print("""
# vim:ft=plumed
MOLINFO STRUCTURE=../../ala.pdb 
phi: TORSION ATOMS=@phi-2 
psi: TORSION ATOMS=@psi-2 
m: METAD ...
   ARG=psi
   SIGMA=0.3 
   HEIGHT=0.0
   BIASFACTOR=8 
   PACE=5000000 
   GRID_MIN=-pi 
   GRID_MAX=pi
   FILE=HILLS_PTMETAD
   RESTART=YES
   TEMP=300
...
PRINT ARG=phi,psi,m.bias FILE=colvar_ptmetad_concat.dat
""".format(i,i),file=f)
    run_mpi("plumed driver --multi {} --plumed plumed_ptmetad_process.dat --ixtc traj.xtc".format(len(optimal_lambdas)),len(optimal_lambdas))

PLUMED: PLUMED is starting
PLUMED: Version: 2.8.0 (git: ddb59d16b) compiled on May 17 2022 at 19:07:56
PLUMED: Please cite these papers when using PLUMED [1][2]
PLUMED: For further information see the PLUMED web page at http://www.plumed.org
PLUMED: PLUMED is starting
PLUMED: Version: 2.8.0 (git: ddb59d16b) compiled on May 17 2022 at 19:07:56
PLUMED: Please cite these papers when using PLUMED [1][2]
PLUMED: For further information see the PLUMED web page at http://www.plumed.org
PLUMED: Root: /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed
PLUMED: PLUMED is starting
PLUMED: Version: 2.8.0 (git: ddb59d16b) compiled on May 17 2022 at 19:07:56
PLUMED: Please cite these papers when using PLUMED [1][2]
PLUMED: For further information see the PLUMED web page at http://www.plumed.org
PLUMED: Root: /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed
PLUMED: For installed feature, see /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed/src/config/config.txt
PLUMED: PLUMED is starting
PLUMED: Version: 2.8.0 (git: ddb59d16b) compiled on May 17 2022 at 19:07:56
PLUMED: PLUMED is starting
PLUMED: Version: 2.8.0 (git: ddb59d16b) compiled on May 17 2022 at 19:07:56
PLUMED: Root: /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed
PLUMED: For installed feature, see /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed/src/config/config.txt
PLUMED: Molecular dynamics engine: driver
PLUMED: Precision of reals: 8
PLUMED: Running over 1 node
PLUMED: Number of threads: 1
PLUMED: PLUMED is starting
PLUMED: Version: 2.8.0 (git: ddb59d16b) compiled on May 17 2022 at 19:07:56
PLUMED: Please cite these papers when using PLUMED [1][2]
PLUMED: For further information see the PLUMED web page at http://www.plumed.org
PLUMED: For installed feature, see /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed/src/config/config.txt
PLUMED: Molecular dynamics engine: driver
PLUMED: Precision of reals: 8
PLUMED: Running over 1 node
PLUMED: Number of threads: 1
PLUMED: Cache line size: 512
PLUMED: Number of atoms: 1591
PLUMED: Molecular dynamics engine: driver
PLUMED: Precision of reals: 8
PLUMED: Running over 1 node
PLUMED: Number of threads: 1
PLUMED: Cache line size: 512
PLUMED: Number of atoms: 1591
PLUMED: GROMACS-like replica exchange is on
PLUMED: File suffix: .5
PLUMED: FILE: plumed_ptmetad_process.dat
PLUMED: GROMACS-like replica exchange is on
PLUMED: File suffix: .3
PLUMED: FILE: plumed_ptmetad_process.dat
PLUMED: Please cite these papers when using PLUMED [1][2]
PLUMED: For further information see the PLUMED web page at http://www.plumed.org
PLUMED: Root: /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed
PLUMED: For installed feature, see /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed/src/config/config.txt
PLUMED: Molecular dynamics engine: driver
PLUMED: Precision of reals: 8
PLUMED: Running over 1 node
PLUMED: Number of threads: 1
PLUMED: Cache line size: 512
PLUMED: Number of atoms: 1591
PLUMED: Please cite these papers when using PLUMED [1][2]
PLUMED: For further information see the PLUMED web page at http://www.plumed.org
PLUMED: Root: /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed
PLUMED: For installed feature, see /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed/src/config/config.txt
PLUMED: Molecular dynamics engine: driver
PLUMED: Precision of reals: 8
PLUMED: Running over 1 node
PLUMED: Number of threads: 1
PLUMED: Cache line size: 512
PLUMED: Cache line size: 512
PLUMED: Number of atoms: 1591
PLUMED: GROMACS-like replica exchange is on
PLUMED: File suffix: .1
PLUMED: FILE: plumed_ptmetad_process.dat
PLUMED: Root: /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed
PLUMED: For installed feature, see /net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/plumed/src/config/config.txt
PLUMED: Molecular dynamics engine: driver
PLUMED: Precision of reals: 8
PLUMED: Running over 1 node
PLUMED: Number of threads: 1
PLUMED: Cache line size: 512
PLUMED: GROMACS-like replica exchange is on
PLUMED: File suffix: .0
PLUMED: FILE: plumed_ptmetad_process.dat
PLUMED: Number of atoms: 1591
PLUMED: GROMACS-like replica exchange is on
PLUMED: File suffix: .2
PLUMED: FILE: plumed_ptmetad_process.dat
PLUMED: Number of atoms: 1591
PLUMED: GROMACS-like replica exchange is on
PLUMED: File suffix: .4
PLUMED: FILE: plumed_ptmetad_process.dat
PLUMED: Action MOLINFO
PLUMED:   with label @0
PLUMED: Action MOLINFO
PLUMED:   with label @0
PLUMED: Action MOLINFO
PLUMED:   with label @0
PLUMED: Action MOLINFO
PLUMED:   with label @0
PLUMED: Action MOLINFO
PLUMED:   with label @0
PLUMED: Action MOLINFO
PLUMED:   with label @0
PLUMED:   pdb file named ../../ala.pdb contains 1 chains 
PLUMED:   chain named   contains residues 1 to 526 and atoms 1 to 1591 
PLUMED:   pdb file named ../../ala.pdb contains 1 chains 
PLUMED:   pdb file named ../../ala.pdb contains 1 chains 
PLUMED: Action TORSION
PLUMED:   with label phi
PLUMED:   chain named   contains residues 1 to 526 and atoms 1 to 1591 
PLUMED:   chain named   contains residues 1 to 526 and atoms 1 to 1591 
PLUMED:   between atoms 5 7 9 15
PLUMED:   using periodic boundary conditions
PLUMED: Action TORSION
PLUMED:   with label phi
PLUMED: Action TORSION
PLUMED:   with label phi
PLUMED:   between atoms 5 7 9 15
PLUMED:   using periodic boundary conditions
PLUMED: Action TORSION
PLUMED:   with label psi
PLUMED:   between atoms 5 7 9 15
PLUMED:   using periodic boundary conditions
PLUMED:   between atoms 7 9 15 17
PLUMED:   using periodic boundary conditions
PLUMED:   pdb file named ../../ala.pdb contains 1 chains 
PLUMED: Action TORSION
PLUMED:   with label psi
PLUMED:   between atoms 7 9 15 17
PLUMED: Action TORSION
PLUMED:   with label psi
PLUMED:   between atoms 7 9 15 17
PLUMED:   using periodic boundary conditions
PLUMED:   chain named   contains residues 1 to 526 and atoms 1 to 1591 
PLUMED:   using periodic boundary conditions
PLUMED:   pdb file named ../../ala.pdb contains 1 chains 
PLUMED: Action TORSION
PLUMED:   with label phi
PLUMED:   chain named   contains residues 1 to 526 and atoms 1 to 1591 
PLUMED:   between atoms 5 7 9 15
PLUMED:   using periodic boundary conditions
PLUMED:   pdb file named ../../ala.pdb contains 1 chains 
PLUMED: Action TORSION
PLUMED:   with label phi
PLUMED:   chain named   contains residues 1 to 526 and atoms 1 to 1591 
PLUMED: Action METAD
PLUMED:   with label m
PLUMED: Action TORSION
PLUMED:   with label psi
PLUMED:   with arguments psi
PLUMED:   added component to this action:  m.bias 
PLUMED: Action METAD
PLUMED:   with label m
PLUMED:   between atoms 5 7 9 15
PLUMED:   using periodic boundary conditions
PLUMED: Action METAD
PLUMED:   with label m
PLUMED:   between atoms 7 9 15 17
PLUMED:   using periodic boundary conditions
PLUMED:   with arguments psi
PLUMED:   added component to this action:  m.bias 
PLUMED: Action TORSION
PLUMED:   with label phi
PLUMED: Action TORSION
PLUMED:   with label psi
PLUMED:   with arguments psi
PLUMED:   added component to this action:  m.bias 
PLUMED:   Binsize not specified, 1/5 of sigma will be be used
PLUMED:   between atoms 5 7 9 15
PLUMED:   using periodic boundary conditions
PLUMED:   between atoms 7 9 15 17
PLUMED:   using periodic boundary conditions
PLUMED:   Binsize not specified, 1/5 of sigma will be be used
PLUMED:   Binsize not specified, 1/5 of sigma will be be used
PLUMED:   Gaussian width  0.300000  Gaussian height 0.000000
PLUMED:   Gaussian deposition pace 5000000
PLUMED: Action TORSION
PLUMED:   with label psi
PLUMED:   Gaussian file HILLS_PTMETAD
PLUMED:   Well-Tempered Bias Factor 8.000000
PLUMED:   Hills relaxation time (tau) inf
PLUMED:   KbT 2.494339
PLUMED:   Grid min -pi
PLUMED:   Gaussian width  0.300000  Gaussian height 0.000000
PLUMED:   Gaussian deposition pace 5000000
PLUMED:   Gaussian file HILLS_PTMETAD
PLUMED:   Well-Tempered Bias Factor 8.000000
PLUMED:   Hills relaxation time (tau) inf
PLUMED:   Gaussian width  0.300000  Gaussian height 0.000000
PLUMED:   Gaussian deposition pace 5000000
PLUMED:   Gaussian file HILLS_PTMETAD
PLUMED:   Grid max pi
PLUMED:   Grid bin 105
PLUMED:   Grid uses spline interpolation
PLUMED:   Well-Tempered Bias Factor 8.000000
PLUMED:   Hills relaxation time (tau) inf
PLUMED:   KbT 2.494339
PLUMED:   Grid min -pi
PLUMED:   Grid max pi
PLUMED:   KbT 2.494339
PLUMED:   Grid min -pi
PLUMED:   Grid max pi
PLUMED:   Grid bin 105
PLUMED:   Grid uses spline interpolation
PLUMED:   between atoms 7 9 15 17
PLUMED:   using periodic boundary conditions
PLUMED: Action METAD
PLUMED:   Grid bin 105
PLUMED:   Grid uses spline interpolation
PLUMED:   with label m
PLUMED:   with arguments psi
PLUMED:   added component to this action:  m.bias 
PLUMED: Action METAD
PLUMED:   with label m
PLUMED:   with arguments psi
PLUMED:   added component to this action:  m.bias 
PLUMED:   Binsize not specified, 1/5 of sigma will be be used
PLUMED: Action METAD
PLUMED:   with label m
PLUMED:   Gaussian width  0.300000  Gaussian height 0.000000
PLUMED:   Gaussian deposition pace 5000000
PLUMED:   Gaussian file HILLS_PTMETAD
PLUMED:   Well-Tempered Bias Factor 8.000000
PLUMED:   Hills relaxation time (tau) inf
PLUMED:   Binsize not specified, 1/5 of sigma will be be used
PLUMED:   with arguments psi
PLUMED:   added component to this action:  m.bias 
PLUMED:   KbT 2.494339
PLUMED:   Grid min -pi
PLUMED:   Grid max pi
PLUMED:   Grid bin 105
PLUMED:   Gaussian width  0.300000  Gaussian height 0.000000
PLUMED:   Grid uses spline interpolation
PLUMED:   Gaussian deposition pace 5000000
PLUMED:   Gaussian file HILLS_PTMETAD
PLUMED:   Well-Tempered Bias Factor 8.000000
PLUMED:   Hills relaxation time (tau) inf
PLUMED:   KbT 2.494339
PLUMED:   Binsize not specified, 1/5 of sigma will be be used
PLUMED:   Grid min -pi
PLUMED:   Grid max pi
PLUMED:   Grid bin 105
PLUMED:   Grid uses spline interpolation
PLUMED:   Gaussian width  0.300000  Gaussian height 0.000000
PLUMED:   Gaussian deposition pace 5000000
PLUMED:   Gaussian file HILLS_PTMETAD
PLUMED:   Well-Tempered Bias Factor 8.000000
PLUMED:   Hills relaxation time (tau) inf
PLUMED:   KbT 2.494339
PLUMED:   Grid min -pi
PLUMED:   Grid max pi
PLUMED:   Grid bin 105
PLUMED:   Grid uses spline interpolation
PLUMED:   Restarting from HILLS_PTMETAD:      1000 Gaussians read
PLUMED:   Bibliography [3][4]
PLUMED: Action PRINT
PLUMED:   with label @4
PLUMED:   Restarting from HILLS_PTMETAD:      1000 Gaussians read
PLUMED:   with stride 1
PLUMED:   with arguments phi psi m.bias
PLUMED:   Restarting from HILLS_PTMETAD:      1000 Gaussians read
PLUMED:   Bibliography [3][4]
PLUMED: Action PRINT
PLUMED:   with label @4
PLUMED:   Restarting from HILLS_PTMETAD:      1000 Gaussians read
PLUMED:   on file colvar_ptmetad_concat.dat
PLUMED:   with format  %f
PLUMED:   with stride 1
PLUMED:   with arguments phi psi m.bias
PLUMED:   Bibliography [3][4]
PLUMED: END FILE: plumed_ptmetad_process.dat
PLUMED:   Bibliography [3][4]
PLUMED: Timestep: 1.000000
PLUMED: KbT has not been set by the MD engine
PLUMED: It should be set by hand where needed
PLUMED:   Restarting from HILLS_PTMETAD:      1000 Gaussians read
PLUMED: Relevant bibliography:
PLUMED: Action PRINT
PLUMED:   with label @4
PLUMED:   [1] The PLUMED consortium, Nat. Methods 16, 670 (2019)
PLUMED:   [2] Tribello, Bonomi, Branduardi, Camilloni, and Bussi, Comput. Phys. Commun. 185, 604 (2014)
PLUMED:   [3] Laio and Parrinello, PNAS 99, 12562 (2002)
PLUMED:   [4] Barducci, Bussi, and Parrinello, Phys. Rev. Lett. 100, 020603 (2008)
PLUMED: Please read and cite where appropriate!
PLUMED:   with stride 1
PLUMED:   with arguments phi psi m.bias
PLUMED:   Bibliography [3][4]
PLUMED:   on file colvar_ptmetad_concat.dat
PLUMED:   with format  %f
PLUMED: END FILE: plumed_ptmetad_process.dat
PLUMED: Action PRINT
PLUMED:   with label @4
PLUMED:   with stride 1
PLUMED: Finished setup
PLUMED:   with arguments phi psi m.bias
PLUMED: Timestep: 1.000000
PLUMED: KbT has not been set by the MD engine
PLUMED: It should be set by hand where needed
PLUMED: Relevant bibliography:
PLUMED:   [1] The PLUMED consortium, Nat. Methods 16, 670 (2019)
PLUMED:   [2] Tribello, Bonomi, Branduardi, Camilloni, and Bussi, Comput. Phys. Commun. 185, 604 (2014)
PLUMED:   [3] Laio and Parrinello, PNAS 99, 12562 (2002)
PLUMED:   [4] Barducci, Bussi, and Parrinello, Phys. Rev. Lett. 100, 020603 (2008)
PLUMED:   on file colvar_ptmetad_concat.dat
PLUMED:   with format  %f
PLUMED: Action PRINT
PLUMED:   with label @4
PLUMED: Please read and cite where appropriate!
PLUMED: Finished setup
PLUMED:   with stride 1
PLUMED: END FILE: plumed_ptmetad_process.dat
PLUMED:   with arguments phi psi m.bias
PLUMED: Timestep: 1.000000
PLUMED: KbT has not been set by the MD engine
PLUMED:   on file colvar_ptmetad_concat.dat
PLUMED:   with format  %f
PLUMED: It should be set by hand where needed
PLUMED: Relevant bibliography:
PLUMED:   [1] The PLUMED consortium, Nat. Methods 16, 670 (2019)
PLUMED:   [2] Tribello, Bonomi, Branduardi, Camilloni, and Bussi, Comput. Phys. Commun. 185, 604 (2014)
PLUMED:   [3] Laio and Parrinello, PNAS 99, 12562 (2002)
PLUMED:   [4] Barducci, Bussi, and Parrinello, Phys. Rev. Lett. 100, 020603 (2008)
PLUMED: Please read and cite where appropriate!
PLUMED:   on file colvar_ptmetad_concat.dat
PLUMED:   with format  %f
PLUMED: END FILE: plumed_ptmetad_process.dat
PLUMED: Timestep: 1.000000
PLUMED: KbT has not been set by the MD engine
PLUMED: It should be set by hand where needed
PLUMED: Finished setup
PLUMED: Relevant bibliography:
PLUMED:   [1] The PLUMED consortium, Nat. Methods 16, 670 (2019)
PLUMED:   [2] Tribello, Bonomi, Branduardi, Camilloni, and Bussi, Comput. Phys. Commun. 185, 604 (2014)
PLUMED:   [3] Laio and Parrinello, PNAS 99, 12562 (2002)
PLUMED:   [4] Barducci, Bussi, and Parrinello, Phys. Rev. Lett. 100, 020603 (2008)
PLUMED: END FILE: plumed_ptmetad_process.dat
PLUMED: Timestep: 1.000000
PLUMED: KbT has not been set by the MD engine
PLUMED:   Restarting from HILLS_PTMETAD:      1000 Gaussians read
PLUMED: Please read and cite where appropriate!
PLUMED: Finished setup
PLUMED: It should be set by hand where needed
PLUMED: Relevant bibliography:
PLUMED:   [1] The PLUMED consortium, Nat. Methods 16, 670 (2019)
PLUMED:   [2] Tribello, Bonomi, Branduardi, Camilloni, and Bussi, Comput. Phys. Commun. 185, 604 (2014)
PLUMED:   [3] Laio and Parrinello, PNAS 99, 12562 (2002)
PLUMED:   [4] Barducci, Bussi, and Parrinello, Phys. Rev. Lett. 100, 020603 (2008)
PLUMED: Please read and cite where appropriate!
PLUMED: Finished setup
PLUMED:   Bibliography [3][4]
PLUMED: Action PRINT
PLUMED:   with label @4
PLUMED:   with stride 1
PLUMED:   with arguments phi psi m.bias
PLUMED:   on file colvar_ptmetad_concat.dat
PLUMED:   with format  %f
PLUMED: END FILE: plumed_ptmetad_process.dat
PLUMED: Timestep: 1.000000
PLUMED: KbT has not been set by the MD engine
PLUMED: It should be set by hand where needed
PLUMED: Relevant bibliography:
PLUMED:   [1] The PLUMED consortium, Nat. Methods 16, 670 (2019)
PLUMED:   [2] Tribello, Bonomi, Branduardi, Camilloni, and Bussi, Comput. Phys. Commun. 185, 604 (2014)
PLUMED:   [3] Laio and Parrinello, PNAS 99, 12562 (2002)
PLUMED:   [4] Barducci, Bussi, and Parrinello, Phys. Rev. Lett. 100, 020603 (2008)
PLUMED: Please read and cite where appropriate!
PLUMED: Finished setup
PLUMED:                                               Cycles        Total      Average      Minimum      Maximum
PLUMED:                                                    1     0.582387     0.582387     0.582387     0.582387
PLUMED: 1 Prepare dependencies                          6006     0.012921     0.000002     0.000001     0.000014
PLUMED: 2 Sharing data                                  6006     0.060595     0.000010     0.000009     0.000261
PLUMED: 3 Waiting for data                              6006     0.021028     0.000004     0.000003     0.000022
PLUMED: 4 Calculating (forward loop)                    6006     0.095677     0.000016     0.000014     0.000062
PLUMED: 5 Applying (backward loop)                      6006     0.094292     0.000016     0.000014     0.000033
PLUMED: 6 Update                                        6006     0.069700     0.000012     0.000010     0.000082
PLUMED:                                               Cycles        Total      Average      Minimum      Maximum
PLUMED:                                                    1     0.578578     0.578578     0.578578     0.578578
PLUMED: 1 Prepare dependencies                          6006     0.012613     0.000002     0.000001     0.000039
PLUMED: 2 Sharing data                                  6006     0.060105     0.000010     0.000009     0.000277
PLUMED: 3 Waiting for data                              6006     0.021026     0.000004     0.000003     0.000047
PLUMED: 4 Calculating (forward loop)                    6006     0.094993     0.000016     0.000014     0.000066
PLUMED: 5 Applying (backward loop)                      6006     0.094447     0.000016     0.000014     0.000084
PLUMED: 6 Update                                        6006     0.069739     0.000012     0.000010     0.000082
PLUMED:                                               Cycles        Total      Average      Minimum      Maximum
PLUMED:                                                    1     0.587870     0.587870     0.587870     0.587870
PLUMED: 1 Prepare dependencies                          6006     0.013446     0.000002     0.000001     0.000013
PLUMED: 2 Sharing data                                  6006     0.061476     0.000010     0.000009     0.000310
PLUMED: 3 Waiting for data                              6006     0.021331     0.000004     0.000003     0.000018
PLUMED: 4 Calculating (forward loop)                    6006     0.096095     0.000016     0.000014     0.000051
PLUMED: 5 Applying (backward loop)                      6006     0.095719     0.000016     0.000015     0.000040
PLUMED: 6 Update                                        6006     0.069868     0.000012     0.000010     0.000075
PLUMED:                                               Cycles        Total      Average      Minimum      Maximum
PLUMED:                                                    1     0.589687     0.589687     0.589687     0.589687
PLUMED: 1 Prepare dependencies                          6006     0.013095     0.000002     0.000001     0.000016
PLUMED: 2 Sharing data                                  6006     0.061798     0.000010     0.000009     0.000252
PLUMED: 3 Waiting for data                              6006     0.021391     0.000004     0.000003     0.000032
PLUMED: 4 Calculating (forward loop)                    6006     0.097182     0.000016     0.000014     0.000051
PLUMED: 5 Applying (backward loop)                      6006     0.096035     0.000016     0.000015     0.000033
PLUMED: 6 Update                                        6006     0.070426     0.000012     0.000010     0.000075
PLUMED:                                               Cycles        Total      Average      Minimum      Maximum
PLUMED:                                                    1     0.590121     0.590121     0.590121     0.590121
PLUMED: 1 Prepare dependencies                          6006     0.013244     0.000002     0.000001     0.000017
PLUMED: 2 Sharing data                                  6006     0.061635     0.000010     0.000009     0.000255
PLUMED: 3 Waiting for data                              6006     0.021584     0.000004     0.000003     0.000347
PLUMED: 4 Calculating (forward loop)                    6006     0.098626     0.000016     0.000015     0.000159
PLUMED: 5 Applying (backward loop)                      6006     0.096428     0.000016     0.000014     0.000185
PLUMED: 6 Update                                        6006     0.070997     0.000012     0.000010     0.000078
PLUMED:                                               Cycles        Total      Average      Minimum      Maximum
PLUMED:                                                    1     0.590560     0.590560     0.590560     0.590560
PLUMED: 1 Prepare dependencies                          6006     0.013603     0.000002     0.000001     0.000039
PLUMED: 2 Sharing data                                  6006     0.061908     0.000010     0.000009     0.000257
PLUMED: 3 Waiting for data                              6006     0.021286     0.000004     0.000003     0.000046
PLUMED: 4 Calculating (forward loop)                    6006     0.098118     0.000016     0.000014     0.000069
PLUMED: 5 Applying (backward loop)                      6006     0.095375     0.000016     0.000014     0.000063
PLUMED: 6 Update                                        6006     0.071330     0.000012     0.000010     0.000078

Then we read all the calculated energies

energies=[]
bias=[]
with cd("5/6reps"):
    for i in range(len(optimal_lambdas)):
        energies.append(np.loadtxt("energy{}.xvg".format(i),usecols=1))
        bias.append(plumed.read_as_pandas("colvar_ptmetad_concat.{}.dat".format(i))["m.bias"])
energies=np.array(energies).T
bias=np.array(bias).T
energies.shape,bias.shape

+++ Loading the PLUMED kernel runtime +++
+++ PLUMED_KERNEL="/net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/libplumedKernel.so" +++
+++ Loading the PLUMED kernel runtime +++
+++ PLUMED_KERNEL="/net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/libplumedKernel.so" +++
+++ Loading the PLUMED kernel runtime +++
+++ PLUMED_KERNEL="/net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/libplumedKernel.so" +++
+++ Loading the PLUMED kernel runtime +++
+++ PLUMED_KERNEL="/net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/libplumedKernel.so" +++
+++ Loading the PLUMED kernel runtime +++
+++ PLUMED_KERNEL="/net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/libplumedKernel.so" +++
+++ Loading the PLUMED kernel runtime +++
+++ PLUMED_KERNEL="/net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/libplumedKernel.so" +++

((6006, 6), (6006, 6))

import wham
kBT=0.00831446261815324*300
w=wham.wham(energies+bias,T=kBT)
# w["logW"] are the Boltzmann factors
# notice that we did not specify yet to which ensemble we are reweighting
# this can be done with the following line:
logW=w["logW"]-energies[:,0]/kBT
# logW are the logaritm weights to obtain properties corresponding to replica 0
logW-=np.max(logW) # avoid numerical errors in exp
weights=np.exp(logW)
weights/=np.sum(weights) # normalize weights
# these weights can be used to compute weighted averages

with cd("5/6reps"):
    colvar_cat=plumed.read_as_pandas("colvar_ptmetad_concat.0.dat")["phi"] # any replica would be fine

+++ Loading the PLUMED kernel runtime +++
+++ PLUMED_KERNEL="/net/sbp/sbpstore1/bussi/miniconda3/envs/plumed-masterclass-2022/lib/libplumedKernel.so" +++

print(
    np.sum(weights*np.logical_and(colvar_cat<2.0,colvar_cat>0))/
    np.sum(weights*np.logical_or(colvar_cat>2.0,colvar_cat<0))
)

0.02000781840905731

weight_of_lambda=[]
for i in range(len(optimal_lambdas)): 
    de=(-energies[:,i]+energies[:,0])/kBT
    weight_of_lambda.append(weights*np.exp(de-np.max(de))/np.sum(weights*np.exp(de-np.max(de))))
    
    
pop_metad=[]
for i in range(len(optimal_lambdas)):    
    pop_metad.append(
        np.sum(weight_of_lambda[i]*np.logical_and(colvar_cat<2.0,colvar_cat>0))/
        np.sum(weight_of_lambda[i]*np.logical_or(colvar_cat>2.0,colvar_cat<0))
    )

plt.plot(Ts,pop_pt,label="pt")
plt.plot(Ts[0]/lambdas,pop_hrex,label="hrex")
plt.plot(Ts[0]/lambdas,pop_wham,label="wham")
plt.plot(Ts[0]/optimal_lambdas,pop_metad,label="metad+wham")
plt.show()

Weird, there is some difference. I now test the result using individual replicas

pop_metad_single=[]
for i in range(6):
    colvar1=colvar_cat[1001*i:1001*(i+1)]
    weights1=np.exp((bias[1001*i:1001*(i+1),i]-np.max(bias[1001*i:1001*(i+1),i]))/kBT)
    weights1=np.exp(weights1)
    weights1/=np.sum(weights1)
    pop_metad_single.append(
        np.sum(weights1*np.logical_and(colvar1<2.0,colvar1>0))/
        np.sum(weights1*np.logical_or(colvar1>2.0,colvar1<0))
    )

plt.plot(Ts,pop_pt,label="pt")
plt.plot(Ts[0]/lambdas,pop_hrex,label="hrex")
plt.plot(Ts[0]/lambdas,pop_wham,label="wham")
plt.plot(Ts[0]/optimal_lambdas,pop_metad,label="metad+wham")
plt.plot(Ts[0]/optimal_lambdas,pop_metad_single,label="metad")
plt.legend()
plt.show()

Likely there is some convergence problem at small lambda!

Compare with results from a longer simulation