Amber

General Description

Assisted Model Building with Energy Refinement (AMBER) is a software suite used for the simulation of biomolecules.

Basic Usage

Pre- and post-processing tools of Amber 18.0 are available after loading its module:

module load amber

A previous version of Amber is also available for use:

module load amber/16.0

A job script example for running parallel version of Amber:

#!/bin/csh
# Example for 24 processors.
#$ -N amber_job_name
#$ -pe smp 24
#$ -M netid@nd.edu
#$ -m abe
# Do not set automatic restart as Amber does not recover well

module load amber/18.0

mpirun -np $NSLOTS $AMBERHOME/bin/sander.MPI  -O -i file.in -o file.out -c file.restrt -p file.prmtop  -r file.restrt -x file.cmd -ref file.restrt

Replace file.{in,out,restrt,prmtop,cmd} name with your files. If these outputs are specified in the mdin file (i.e. file.in in the example) but not on the command line, then files are output with generic names e.g. mdout, mdcrd instead of file.out, file.mdcrd respectively.

A job script example for running MMPBSA:

#!/bin/csh
#$ -M netid@nd.edu    # Email address for job notification
#$ -m abe             # Send mail when job begins, ends and aborts
#$ -N mmpbsa          # Specify job name
#$ -q long
#$ -pe smp 1

module load amber/16.0
MMPBSA -O -i test.in -o FINAL_RESULTS_DEC23.dat -do FINAL_DECOMP_DEC23.dat -sp npc12ncw.top -cp npc12ncnw.top -rp npc1ncnw.top -lp npc2ncnw.top -y npc12ncwmd2.crd.gz npc12ncwmd3.crd.gz

For further information, please consult the Amber documentation: http://ambermd.org/

Amber with GPU Acceleration

More information on Amber acceleration using GPU can be found here. To run the GPU version of Amber, please send a request to CRCSupport@nd.edu. The following is a job script example for running Amber 18.0 accelerated with GPU starting from a previous simulation:

#!/bin/csh
#$ -M netid@nd.edu    # Email address for job notification
#$ -m bea             # Send mail when job begins, ends and aborts
#$ -N Amber-GPU       # Specify job name
#$ -q gpu             # Run on the GPU cluster
#$ -l gpu_card=1      # Run on 1 GPU card
#$ -pe smp 1

module load amber/18.0  # Load Amber accelerated with GPU

# mdin file is included in this job script within a heredoc
#    normally, you'd want to write this as a separate .in file
cat > mdin <<EOF
 short md, nve ensemble
 &cntrl
                  ! This line is a comment.
   irest=1,       ! continue from a previous simulation step
   ntx=5,         ! coordinates AND velocities are read in from the restart file
   ntc=2,         ! use SHAKE to constrain bonds involving hydrogen atoms
   ntf=2,         ! omit force calculations involving hydrogens, which are unnecessary when ntc = 2
   tol=0.0000001, ! SHAKE tolerance
   nstlim=20,     ! number of steps to run simulation for
   ntt=0,         ! constant energy; no temperature scaling
   ntpr=1,        ! frequency of mdout file write in dt
   ntwr=10,       ! frequency of restrt file write in dt
   dt=0.001,      ! timestep of 0.001 femtosecond(s); maximum 0.002 is recommended if SHAKE is in use, 0.001 otherwise
                  ! More information on mdin settings can be found at http://ambermd.org/Manuals.php
 /
 &ewald
  NFFT1=72,NFFT2=64,NFFT3=60,
 /
EOF


pmemd.cuda -O -i mdin -o md.out -c md12.x -r restrt -x mdcrd # 

Further Information

  • Amber16 User's Manual [1]
  • Amber18 User's Manual [2]
  • See the official website: [3]