Turing: CPMD

Introduction

The CPMD (Car-Parrinello Molecular Dynamics) software is installed at IDRIS on the Turing machine.

You may consult the program manual on the CPMD official site. A version of the manual is also available here in PDF.

If your CPMD results are used in a publication, please include the following reference: CPMD, http://www.cpmd.org/, Copyright IBM Corp 1990-2015, Copyright MPI für Festkörperforschung Stuttgart 1997-2001.

Availability

The following versions are available:

  • 3.13.2
  • 3.15.3
  • 3.17.1 (version par défaut)
  • 4.1

The parallel OpenMP/MPI hybrid versions are installed.

Launching script

Here is an example of a launching script for a calculation executed in the WORKDIR :

job.ll
# @ job_name         = CPMD
# @ output           = $(job_name).$(jobid)
# @ error            = $(job_name).$(jobid)
# @ job_type         = BLUEGENE
# @ bg_size          = 64
# @ wall_clock_limit = 20:00:00
# @ queue
 
### Module initialisation ###
module load cpmd
 
### Command echoes ###
set -x
 
### Run the calculation ###
runjob --np 1024 --ranks-per-node=16 --envs "OMP_NUM_THREADS=4" : $CPMD_EXEDIR/cpmd.x fichier.input > fichier.out

Here is an example of a launching script for a calculation executed in the TMPDIR :

job.ll
# @ job_name         = CPMD
# @ output           = $(job_name).$(jobid)
# @ error            = $(job_name).$(jobid)
# @ job_type         = BLUEGENE
# @ bg_size          = 64
# @ wall_clock_limit = 20:00:00
# @ queue
 
### Module initialisation ###
module load cpmd
 
### Command echoes ###
set -x
 
### Copy to the TMPDIR ###
cp ./* $TMPDIR
 
### Run the calculation ###
cd $TMPDIR
runjob --np 1024 --ranks-per-node=16 --envs "OMP_NUM_THREADS=4" : $CPMD_EXEDIR/cpmd.x fichier.input > fichier.out
 
### Copy to the submission folder ###
cd -
cp $TMPDIR/* .

You may refer to the documentation about the module command.

You may also refer to our more detailed advice about submitting a job in queue parallelised with MPI and OpenMP on the Turing machine.

Specific software parameters

In the 3.17.1 version, the TASKGROUPS parameter is called CG_GROUPS. This parameter manages the sharing of MPI processes when there is a high number of them. However, it is generally not useful except for a number of processes superior to about 256.