KIM Application Programmming Interface As a Standard for Molecular Simulations

Nanoscale modeling of materials often involves the use of molecular simulations or multiscale methods. These approaches frequently use empirical (fitted) interatomic potentials to represent the response of the material. Currently, no standardized approach exists for estimating the accuracy of interatomic potentials. In addition, a lack of standardization in programming interfaces for potentials and the lack of a systematic infrastructure for archiving them makes it difficult to use potentials for new applications and to reproduce published results.

The Knowledgebase of Interatomic Models (KIM) (http://openKIM.org) is a 4-year NSF CDI program which attempts to address these limitations.  One difficulty in the use of interatomic models is that simulation codes and interatomic potentials are written in a variety of programming languages. In order to allow the programs and model subroutines to be mutually compatible, it is necessary to define a robust application programming interface (API) designed to accommodate the requirements of molecular simulations. The KIM project has developed such an API in collaboration with key members of the molecular simulation community.  The API is based on the concept of descriptor files, which define all variables and methods needed for input and output for a given application or model. The API addresses issues such as naming conventions for models (unique names), neighbor-list objects and their interface, system of units, unit conversion, and a complete description of variables and methods.

In this talk we will present a brief overview of the KIM framework and then provide details on the KIM API protocol. We invite the molecular simulation community to participate in the KIM project and contribute to the standardization efforts under way.