HPC Extensions to the OpenKIM Processing Pipeline

TL;DR

This paper discusses enhancements to the OpenKIM processing pipeline, enabling it to efficiently leverage HPC resources for large-scale materials science computations.

Contribution

The paper introduces HPC extensions to the OpenKIM pipeline, improving its scalability and performance for large-scale interatomic model simulations.

Findings

Enhanced pipeline supports large-scale HPC computations

Improved efficiency and scalability of the processing pipeline

Facilitates advanced materials science research

Abstract

The Open Knowledgebase of Interatomic Models (OpenKIM) is an NSF Science Gateway that archives fully functional computer implementations of interatomic models (potentials and force fields) and simulation codes that use them to compute material properties. Interatomic models are coupled with compatible simulation codes and executed in a fully automated manner by the OpenKIM processing pipeline, a cloud-based computation platform. The pipeline as previously introduced in the literature was insufficient to support the large-scale computations that have become necessary within the materials science community. Accordingly, we present extensions made to the pipeline that allow it to utilize High-Performance Computing (HPC) resources in an efficient and performant fashion.