Quantum ESPRESSO: Difference between revisions

From RCSWiki
Jump to navigation Jump to search
Line 160: Line 160:


Now you can submit the job to the SLURM scheduler and check the job:
Now you can submit the job to the SLURM scheduler and check the job:
<pre>
<source lang=bash>
$ sbatch pw-scf.slurm
$ sbatch pw-scf.slurm
Submitted batch job 29235262
Submitted batch job 29235262
Line 168: Line 168:
JOBID    USER      STATE  PARTITION TIME_LIMIT  TIME  NODES  TASKS  CPUS  MIN_MEMORY TRES_PER_NREASON  NODELIST
JOBID    USER      STATE  PARTITION TIME_LIMIT  TIME  NODES  TASKS  CPUS  MIN_MEMORY TRES_PER_NREASON  NODELIST
29235262  username  PENDING ....
29235262  username  PENDING ....
</pre>
</source>
The computation is very short, about 2 seconds. You will not be able to see it in the queue if it starts quickly.
The computation is very short, about 2 seconds. You will not be able to see it in the queue if it starts quickly.
If the cluster is busy, it may take some time until the job runs, until then the job will be in the '''PENDING''' state.
If the cluster is busy, it may take some time until the job runs, until then the job will be in the '''PENDING''' state.

Revision as of 20:36, 24 April 2024

General

Installation: https://www.quantum-espresso.org/Doc/user_guide/node7.html



Quantum Espresso is an integrated suite of Open-Source computer codes for electronic-structure calculations and materials modeling at the nanoscale. It is based on density-functional theory, plane waves, and pseudopotentials.

Quantum ESPRESSO has evolved into a distribution of independent and inter-operable codes in the spirit of an open-source project. The Quantum ESPRESSO distribution consists of a “historical” core set of components, and a set of plug-ins that perform more advanced tasks, plus a number of third-party packages designed to be inter-operable with the core components.


The projector augmented wave method (PAW) is a technique used in ab initio electronic structure calculations. It is a generalization of the pseudopotential and linear augmented-plane-wave methods, and allows for density functional theory calculations to be performed with greater computational efficiency.

See Wikipedia: https://en.wikipedia.org/wiki/Projector_augmented_wave_method


Since 26 April 2016 EPW is distributed as part of the Quantum ESPRESSO suite.

EPW is the short name for "Electron-phonon Wannier". EPW is an open-source F90/MPI code which calculates properties related to the electron-phonon interaction using Density-Functional Perturbation Theory and Maximally Localized Wannier Functions.


  • PSLIBRARY -- a library for generating ultra-soft pseudo-potentials.
https://dalcorso.github.io/pslibrary/pslibrary_help.html


QE on ARC

To see available versions of QE on ARC, use the module command:

$ module avail espresso
-------------------------- /global/software/Modules/4.6.0/modulefiles -----------------------------
espresso/6.3-gnu  espresso/7.2  

The versions:

  • espresso/7.2 was released on 31 March 2023.
Built with GCC and gfortran 8.5.0
OpenMPI 4.1.1
LibXC v6.2.2
OpenBLAS v0.3.23
No GPU support.
No HDF5 support.


  • espresso/6.3-gnu is an older version of QE and is provided for compatibility purpose.


Test case

Input file

QE PW input file for some Diamond crystal, pw-scf.in:

 &control
    calculation     = 'scf'
    prefix          = 'diam'
    restart_mode    = 'from_scratch'
    wf_collect      = .false.
    pseudo_dir      = 'pseudo'
    outdir          = './out'
    tprnfor         = .true.
    tstress         = .true.
 /
 &system
    ibrav           = 2
    celldm(1)       = 6.64245
    nat             = 2
    ntyp            = 1
    ecutwfc         = 60
    occupations     = 'smearing'
    smearing        = 'mp'
    degauss         = 0.02
    nbnd            = 4
 /
 &electrons
    diagonalization = 'david'
    mixing_beta     = 0.7
    conv_thr        = 1.0d-10
 /
ATOMIC_SPECIES
  C  12.01078  C_3.98148.UPF
ATOMIC_POSITIONS alat
  C   0.00  0.00  0.00
  C   0.25  0.25  0.25
K_POINTS automatic
8 8 8 1 1 1

We will run this calculation on a compute node from the default list using 4 MPI processes.

Job script

The job script for this computation, pw-scf.slurm:

#!/bin/bash 
# ===========================================================================
#SBATCH --job-name=qe-pw-test

#SBATCH --nodes=1
#SBATCH --ntasks=4
#SBATCH --cpus-per-task=1
#SBATCH --mem=16gb
#SBATCH --time=0-01:00:00
# ===========================================================================
module load espresso/7.2

# Create a symbolic link to the pseudopotentials
ln -s $ESPRESSO_PSEUDO .

mpiexec pw.x -npool $SLURM_NTASKS < scf.in 

# ===========================================================================

This job script requests 1 hour or run time on a compute node in one of the default partitions. It also requests 4 MPI processes and 16 Gb of RAM.

Running the test

You have to create a directory to contain all the files for this test case, and "change" to that directory:

$ cd 
$ mkdir -p my-jobs/qe-tests/diamond-scf
$ cd my-jobs/qe-tests/diamond-scf

Then you have to create the pw-scf.in input file, as well as the job script, pw-scf.slurm. You can use the nano text editor for this and copy/paste the text for the files from this page.

$ nano pw-scf.in
....
$ nano pw-scf.slurm
....

Once you create both the files, you can check if they are in the current directory:

$ pwd
/home/username/my-jobs/qe-tests/diamond-scf

$ ls -l
-rw-r----- 1 username username   751 Apr 24 14:06 pw-scf.in
-rwxr-x--x 1 username username   488 Apr 24 14:15 pw-scf.slurm

Now you can submit the job to the SLURM scheduler and check the job:

$ sbatch pw-scf.slurm
Submitted batch job 29235262

# Check if it has started yet.
$ squeue-long -j 29235262
JOBID     USER      STATE   PARTITION TIME_LIMIT  TIME  NODES  TASKS  CPUS  MIN_MEMORY TRES_PER_NREASON   NODELIST
29235262  username  PENDING ....

The computation is very short, about 2 seconds. You will not be able to see it in the queue if it starts quickly. If the cluster is busy, it may take some time until the job runs, until then the job will be in the PENDING state.

Once the job is completed, the output is saved by SLURM to the slurm-29235262.out output file.

So, we can check the results:

$ ls -l

drwxr-xr-x 3 username username  4096 Apr 24 14:16 out
lrwxrwxrwx 1 username username    30 Apr 24 14:16 pseudo -> /global/software/qe/7.2/pseudo
-rw-r----- 1 username username   751 Apr 24 14:06 pw-scf.in
-rwxr-x--x 1 username username   488 Apr 24 14:15 pw-scf.slurm
-rw-r--r-- 1 username username 21727 Apr 24 14:16 slurm-29235262.out

# Check the output directory
$ ls out
diam.save  diam.xml

# Check the output file (the end of it only).
$ tail slurm-29235262.out 
     Parallel routines

     PWSCF        :      1.09s CPU      1.22s WALL


   This run was terminated on:  14:16:23  24Apr2024            

=------------------------------------------------------------------------------=
   JOB DONE.
=------------------------------------------------------------------------------=

Success!

Links

ARC Software