Running Many PLAMS Jobs in Parallel

Technical Parallelization Workflow automation

Configure a parallel job runner, read a set of molecules from disk, and dispatch many AMS jobs from one Python workflow.

Downloads: Notebook | Script ?

Requires: AMS2026 or later

Related examples
Related tutorials
Related documentation

Initial Imports

import psutil
from scm.plams import JobRunner, config, Settings, from_smiles, AMSJob

Configure Job Runner

Set the default job runner to run in parallel. Run as many jobs simultaneously as there are cpu on the system. In addition, set the number of cores for each job to 1.

maxjobs = psutil.cpu_count(logical=False)
print("Running up to {} jobs in parallel simultaneously".format(maxjobs))

Running up to 12 jobs in parallel simultaneously

config.default_jobrunner = JobRunner(parallel=True, maxjobs=maxjobs)

config.job.runscript.nproc = 1

Load Molecules

Generate a set of molecules from SMILES strings.

molecule_smiles = {
    "Acetic_acid": "CC(=O)O",
    "Benzene": "c1ccccc1",
    "Butane": "CCCC",
    "Ethane": "CC",
    "Ethanol": "CCO",
    "Formic_acid": "C(=O)O",
    "Methanol": "CO",
    "Water": "O",
}

molecules = {
    name: from_smiles(smiles, forcefield="uff") for name, smiles in molecule_smiles.items()
}

Set Up and Run Jobs

Configure the calculation settings in the Settings object. Run a geometry optimization job for each molecule in parallel.

settings = Settings()
settings.input.ams.Task = "GeometryOptimization"
settings.input.dftb.Model = "GFN1-xTB"

results = []
for name, molecule in sorted(molecules.items()):
    job = AMSJob(molecule=molecule, settings=settings, name=name)
    results.append(job.run())

[10.04|11:35:08] JOB Acetic_acid STARTED
[10.04|11:35:08] JOB Benzene STARTED
[10.04|11:35:08] JOB Butane STARTED
[10.04|11:35:08] JOB Ethane STARTED
[10.04|11:35:08] JOB Ethanol STARTED
[10.04|11:35:08] JOB Formic_acid STARTED
[10.04|11:35:08] JOB Methanol STARTED
[10.04|11:35:08] JOB Water STARTED
[10.04|11:35:08] JOB Acetic_acid RUNNING
[10.04|11:35:08] JOB Benzene RUNNING

Results

Print a table of results only for the successful calculations.

# Only print the results of the succesful caluclations:
for result in [r for r in results if r.ok()]:
    print(
        "Energy for {:<12}: {:>10.3f} kcal/mol".format(
            result.name, result.get_energy(unit="kcal/mol")
        )
    )

[10.04|11:35:08] Waiting for job Acetic_acid to finish
[10.04|11:35:08] JOB Butane RUNNING
[10.04|11:35:08] JOB Ethanol RUNNING
[10.04|11:35:08] JOB Ethane RUNNING
[10.04|11:35:08] JOB Formic_acid RUNNING
[10.04|11:35:08] JOB Water RUNNING
[10.04|11:35:08] JOB Methanol RUNNING
[10.04|11:35:08] JOB Benzene FINISHED
[10.04|11:35:08] JOB Water FINISHED
[10.04|11:35:08] JOB Ethane FINISHED
... output trimmed ....
[10.04|11:35:08] Waiting for job Butane to finish
[10.04|11:35:08] JOB Butane SUCCESSFUL
Energy for Acetic_acid :  -9906.399 kcal/mol
Energy for Benzene     :  -9973.315 kcal/mol
Energy for Butane      :  -8699.182 kcal/mol
Energy for Ethane      :  -4686.354 kcal/mol
Energy for Ethanol     :  -7630.501 kcal/mol
Energy for Formic_acid :  -7898.338 kcal/mol
Energy for Methanol    :  -5621.724 kcal/mol
Energy for Water       :  -3618.401 kcal/mol

See also

Python Script

#!/usr/bin/env python
# coding: utf-8

# ## Initial Imports

import psutil
from scm.plams import JobRunner, config, Settings, from_smiles, AMSJob


# ## Configure Job Runner
#
# Set the default job runner to run in parallel. Run as many jobs simultaneously as there are cpu on the system. In addition, set the number of cores for each job to 1.

maxjobs = psutil.cpu_count(logical=False)
print("Running up to {} jobs in parallel simultaneously".format(maxjobs))


config.default_jobrunner = JobRunner(parallel=True, maxjobs=maxjobs)


config.job.runscript.nproc = 1


# ## Load Molecules
#
# Generate a set of molecules from SMILES strings.

molecule_smiles = {
    "Acetic_acid": "CC(=O)O",
    "Benzene": "c1ccccc1",
    "Butane": "CCCC",
    "Ethane": "CC",
    "Ethanol": "CCO",
    "Formic_acid": "C(=O)O",
    "Methanol": "CO",
    "Water": "O",
}

molecules = {name: from_smiles(smiles, forcefield="uff") for name, smiles in molecule_smiles.items()}


# ## Set Up and Run Jobs
#
# Configure the calculation settings in the `Settings` object. Run a geometry optimization job for each molecule in parallel.

settings = Settings()
settings.input.ams.Task = "GeometryOptimization"
settings.input.dftb.Model = "GFN1-xTB"


results = []
for name, molecule in sorted(molecules.items()):
    job = AMSJob(molecule=molecule, settings=settings, name=name)
    results.append(job.run())


# ## Results
#
# Print a table of results only for the successful calculations.

# Only print the results of the succesful caluclations:
for result in [r for r in results if r.ok()]:
    print("Energy for {:<12}: {:>10.3f} kcal/mol".format(result.name, result.get_energy(unit="kcal/mol")))