Applications/Nwchem

From HPC
Revision as of 15:22, 3 February 2017 by Pysdlb (talk | contribs)

Jump to: navigation , search

Application Details

  • Description : NWChem provides computational chemistry tools that are scalable both in their ability to treat large scientific computational chemistry problems efficiently, and in their use of available parallel computing resources from high-performance parallel supercomputers to conventional workstation clusters.
  • Versions : x86/6.6 and cuda/6.6
  • Module names : nwchem-x86/6.6 and nwchem-cuda/6.6
  • License: Open Source
  • Further information: http://www.nwchem-sw.org/ nwchem


Toolboxes

No toolboxes are listed here.

Modules Available

  • module load nwchem-x86/6.6
  • module load nwchem-cuda/6.6

Usage Examples

Batch Submission

This first example is based on the compute queue:


#!/bin/bash
#SBATCH -J nwch-cpu
#SBATCH -N 100
#SBATCH --ntasks-per-node=28
#SBATCH -o %N.%j.%a.out
#SBATCH -e %N.%j.%a.err
#SBATCH -p compute
#SBATCH --exclusive

echo $SLURM_JOB_NODELIST

module purge
module load intel/mkl/64/11.3.2
module load intel/mpi/64/5.1.3.181
module load intel/compiler/64/2016

export I_MPI_FABRICS=shm:tmi
export I_MPI_FALLBACK=no

module load nwchem-x86/6.6
export NWCHEM_ROOT=$HOME/NWCHEM

module list
mpirun --version

# calculating the number of processes
NP=$(( $SLURM_JOB_NUM_NODES * $SLURM_NTASKS_PER_NODE ))
echo $NP "processes"
echo $SLURM_NTASKS

NWCHEM=/trinity/clustervision/CentOS/7/apps/nwchem-6.6/x86/bin/LINUX64/nwchem

mpirun -genvall  -np $NP $NWCHEM py-c1-vdz.inp > py-c1-vdz-2800_CV.out


[username@login01 ~]$ sbatch nwchem-x86.job
Submitted batch job 289522

Secondly, the script is based on the CUDA version of nwchem:


#!/bin/bash
#SBATCH -J nwchem-gpu
#SBATCH -N 1
#SBATCH --ntasks-per-node 24
#SBATCH -o %N.%j.%a.out
#SBATCH -e %N.%j.%a.err
#SBATCH -p gpu
#SBATCH --gres=gpu:1
#SBATCH --exclusive

module purge

module load nwchem-cuda/6.6
module load cuda/7.5.18
module load intel/mkl/64/11.3.2
module load intel/mpi/64/5.1.3.181

module list
nvidia-smi -a

mpirun --version

# calculating the number of processes
NP=$(( $SLURM_JOB_NUM_NODES * $SLURM_NTASKS_PER_NODE ))
echo $NP "processes"

NWCHEM=/trinity/clustervision/CentOS/7/apps/nwchem-6.6/cuda/bin/LINUX64/nwchem

mpirun -np $NP $NWCHEM py-c1-vdz.inp > py-c1-vdz-24c_0GPU.out

[username@login01 ~]$ sbatch nwchem-cuda.job
Submitted batch job 289522