ERROR: the system is looking for a value outside the grid along the 1 index!

[elena....@gmail.com]

Dear PLUMED community,

First I would like to warn everyone that I'm new to PLUMED. I have performed metadynamics simulations before with NAMD, but not using PLUMED. We decided to switch to GROMACS, so now I am trying to perform a metadynamics simulation with GROMACS 4.6.7 and PLUMED 2.1.1. When I run my simulation without using grid, everything works fine. However once I uncomment the GRID options I receive the following error:

ERROR: the system is looking for a value outside the grid along the 1 index!

I do not really understand where the problem is, since the values that are being explored are well within the GRID_MIN-GRID_MAX range. 

This is my plumed.dat input file:

UNITS  LENGTH=nm  ENERGY=kj/mol  TIME=ps

#

##################################################################

#DEFINE GROUPS AND COMS

##################################################################

ab:       GROUP NDX_FILE=index.ndx NDX_GROUP=AB_Ca

c_bottom: GROUP NDX_FILE=index.ndx NDX_GROUP=C_bottom_Ca

d_bottom: GROUP NDX_FILE=index.ndx NDX_GROUP=D_bottom_Ca

a_i:      GROUP NDX_FILE=index.ndx NDX_GROUP=A_111_112_heavy

b_i:      GROUP NDX_FILE=index.ndx NDX_GROUP=B_111_112_heavy

c_i:      GROUP NDX_FILE=index.ndx NDX_GROUP=C_49_82_heavy

d_i:      GROUP NDX_FILE=index.ndx NDX_GROUP=D_49_82_heavy

#

com_ab:         COM   ATOMS=ab

com_c_bottom:   COM   ATOMS=c_bottom

com_d_bottom:   COM   ATOMS=d_bottom

##################################################################

#DEFINE CVs

##################################################################

ang:    ANGLE ATOMS=com_c_bottom,com_ab,com_d_bottom

cn_a_d: COORDINATION GROUPA=a_i GROUPB=d_i R_0=0.4

cn_b_c: COORDINATION GROUPA=b_i GROUPB=c_i R_0=0.4

##################################################################

#SET WALLS

##################################################################

uwall:  UPPER_WALLS ARG=ang,cn_a_d,cn_b_c AT=2.4,44.0,44.0 KAPPA=150.0,150.0,150.0

lwall:  LOWER_WALLS ARG=ang               AT=1.4           KAPPA=150.0

##################################################################

#METADYNAMICS

##################################################################

METAD ...

LABEL=metad

ARG=ang,cn_a_d,cn_b_c 

PACE=100 

HEIGHT=1.2

SIGMA=0.04,1,1 

#GRID_MIN=1.4,0,0

#GRID_MAX=2.4,44.0,44.0

FILE=HILLS

... METAD

##################################################################

#OUTPUT

##################################################################

PRINT ARG=ang,cn_a_d,cn_b_c,metad.bias STRIDE=50 FILE=COLVAR

##################################################################

This is an excerpt from my COLVAR file without the grid options:

#! FIELDS time ang cn_a_d cn_b_c metad.bias

 0.000000 1.502978 43.750342 31.362944 0.000000

 0.100000 1.502508 41.959095 36.184315 0.000000

 0.200000 1.502326 41.106465 38.363767 0.000000

 0.300000 1.500268 39.061361 37.743294 0.122124

 0.400000 1.498676 36.868988 37.299855 0.000000

 0.500000 1.497909 34.835206 37.018463 0.145790

 0.600000 1.498553 39.112815 34.352828 0.000000

 0.700000 1.499987 40.456059 36.713441 0.278401

 0.800000 1.499625 40.359629 35.582627 0.277794

I could not find anything on this error in google. I probably do not interpret the error message correctly. Any suggestions?

Berst regards,

Elena

[davide branduardi]

Hi Elena, 

by chance, have you edited the input file with a windows application? Or saved with a win-compatible format?

Can you use dos2unix to generate a new file which is 100% guaranteed to be unix?

D

[elena....@gmail.com]

Hi,

no I used emacs on linux :).

I think I solved my problem. I increased KAPPA and EPS of the walls and now everything works fine.

Regards,

Elena

[Rubén Granero]

  Hi there,

I'm having the same problem and I can't find a way to solve it. I'm running PLUMED 2.1.1 with NAMD 2.9 under CentOS 7 (no Windows involved in any step). I'm performing metadynamics, the variables are the x, y, z coordinates of a water molecule. The simulation box contains many molecules, but I'm using walls to confine the variables to a smaller box. I started with a grid a bit larger than the space limited by the walls, but I got the following error:

   ERROR: the system is looking for a value outside the grid along the 2 index!

The position of the water molecule was well inside the walls and the grid when this happened. I increased the KAPPA of the walls a lot just to make sure the water molecule wasn't going out and I ended up making the grid as big as the whole simulation box, but the same problem keeps appearing. Needless to say that the simulation runs smoothly without the grid, but it's going to end up being quite long and I have to look for efficiency.

I copy here my input file:

# UNITS

# Default: energy in kJ/mol, length in nm, time in ps

#

UNITS LENGTH=A TIME=fs

# VARIABLES

# x, y, z coordinates of atom 442

#

pos: POSITION ATOM=442

# WALLS

#

LOWER_WALLS ...

LABEL=lwall

ARG=pos.x,pos.y,pos.z

AT=9.6026,0.9490,20.9610

KAPPA=1000,1000,1000

OFFSET=1,1,1

... LOWER_WALLS

#

UPPER_WALLS ...

LABEL=uwall

ARG=pos.x,pos.y,pos.z

AT=26.2596,37.2020,32.1780

KAPPA=1000,1000,1000

OFFSET=1,1,1

... UPPER_WALLS

# METADYNAMICS SETUP

# Well tempered at 298 K

# Deposit a Gaussian on the three components of the position,

# every 1000 steps,

# with height equal to 0.5 kJ/mol,

# and width equal to 0.4 A

#

METAD ...

LABEL=metad

BIASFACTOR=10

TEMP=298

ARG=pos.x,pos.y,pos.z

PACE=1000

HEIGHT=0.5,0.5,0.5

SIGMA=0.4,0.4,0.4

FILE=HILLS

GRID_MIN=0,0,0

GRID_MAX=36,38,53

GRID_SPACING=0.08,0.08,0.08

... METAD

# MONITORIZATION

#

PRINT STRIDE=1000 ARG=pos.x,pos.y,pos.z,metad.bias FILE=COLVAR

I'm afraid I'm not understanding properly how the grid works and I hope you can give me a hand with this.

   Best regards,

      Rubén