KIM Content Carnival
CALL FOR APPLICATIONS
Come to Maryland this summer during August 18–21, 2014 for an intensive 4-day event to learn about the Knowledgebase of Interatomic Models (KIM) project (https://openkim.org) and contribute content to it. If you are new to KIM, a good place to start is https://openkim.org/doc/overview/getting-started/. At the event:
Date: August 18-21, 2014
Location: University of Maryland, College Park, MD
Requirements: Participation is by invitation only (see “To Apply” below). Applicants should meet the following prerequisites:
To Apply: Applicants will be considered on a first come, first served basis. There are a limited number of places so APPLY EARLY! The application deadline is Friday, August 1, 2014. To apply send an e-mail to tadmor@aem.umn.edu (Prof. Ellad Tadmor, University of Minnesota) with the following information:
This KCC event is organized in collaboration with the National Institute of Standards and Technology (NIST).
The objective of KCC events is to contribute new content to the KIM project. There are four different categories in which you can contribute:
A major element of the KIM project is the development of an application programming interface (API) for atomistic simulations. The KIM API is a standard interface for information exchange between a computer implementation of an interatomic model (a potential) and a simulation code. A model conforming to the KIM API will work seamlessly with any simulation code that supports this API (such as LAMMPS, IMD, DL_POLY, GULP, ASE). Making a model “KIM-compliant” requires changes to the way that information is passed in and out of the model implementation including changes to neighbor list handling. Note that KIM is currently limited to reactive potentials commonly used in materials science and solid-state physics. Bonded force fields such as CHARMM and AMBER are not currently supported.
To apply for Category 1, please provide the following information: describe the potential, which materials it can model and what it was designed to do. What computer language is the code written in? Give a reference(s) if the potential has been published. (Electronic copies of any cited references would be appreciated.) Priority will be given to potentials that can be implemented as KIM “Model Drivers” (i.e. potentials that read in parameters and therefore constitute an entire class of potentials (such as Lennard-Jones) as opposed to a single “black box” potential.)
One of the objectives of the KIM project is the universal adoption of the KIM API by all major atomistic and multiscale simulation codes. Several major simulation codes (“Simulators” in KIM jargon) already support the KIM API (for a list of major software packages that support KIM, see https://openkim.org/projects-using-kim/). KIM-compliant Simulators can be installed in the KIM system as libraries enabling users to develop KIM Tests as input files to these programs (see Category 3). The KIM Team is interested in working with other major code developers to make their programs KIM compliant.
To apply for Category 2, please provide the following information: describe the simulation code including its objectives and capabilities. What computer language is the code written in? Give a reference(s) if an article on the program has been published and/or the code website. (Electronic copies of any cited references would be appreciated.) Priority will be given to simulation codes with large user-bases (as opposed to programs limited to a single research group).
Interatomic models uploaded to KIM are tested against a battery of “KIM Tests” that are contained within the KIM Repository. A KIM Test is a computer program that when coupled with a suitable KIM Model, calculates the prediction of that model for a specific material property. Examples of properties include elastic constants, surface energies, phonon dispersion curves, etc. (Properties are characterized by the KIM Properties Framework described here: https://openkim.org/doc/schema/properties-framework/) A Test can be a stand-alone computer program or an input file to a Simulator library (see Category 2). Currently LAMMPS and ASE are installed as Simulators for tutorials on how to create LAMMPS and ASE based Tests). The power of KIM is that the pool of Tests is user-extendable allowing any researcher to upload Tests for new properties that he/she may be interested in. Newly uploaded Tests are automatically applied to all Models in the KIM Repository. Note that a current limitation of the KIM system is that Tests are run in serial and therefore cannot be overly computationally intensive. The KIM Team is interested in working with researchers to help them create new Tests.
To apply for Category 3, please provide the following information: describe the property or properties for which you would like to create a KIM Test. Why are these of interest? Do you want to write a stand-alone code or use either LAMMPS or ASE? Provide reference(s) for the methodology you plan to use to compute the property. (For example, elastic constants can be computed statically using an analytical expression or numerical differentiation, from vibrational properties of the material, etc.) (Electronic copies of any cited references would be appreciated.) Priority will be given to properties that can be implemented as KIM “Test Drivers” (i.e. Tests that read in parameters and therefore constitute an entire class of Tests, such as a Test Driver for computing the surface energy of an fcc crystal for any desired surface).
The coupling of Tests and Models (as described in Category 3) over time generates a huge amount of data that is stored in the KIM Repository along with experimental and first principles Reference Data that can be uploaded to KIM. The KIM Visualization framework has been developed to allow users to visualize, compare and analyze Test results (i.e. predictions of interatomic models for specific properties) and Reference Data. As in all KIM content, the visualization system is user-extendable. The objective is that over time a host of interesting ways to visualize data will be added by researchers and become part of an extensive toolkit available to the entire community. A visualizer works by performing a query to the KIM Repository using a standard query language and then displaying the information using a set of JavaScript libraries. The KIM Team is interested in working with developers to create new visualizers for important properties.
To apply for Category 4, please provide the following information: describe the property or properties for which you would like to create a KIM Visualizer. What information will be displayed and how will this information help to understand the results? Will the Visualizer have any analysis capabilities (such as curve fitting results)? If relevant, provide reference(s) for any specialized methodologies you plan to use to visualize the property. (For example, “Vitek plots” are specialized diagrams for representing the structure of screw dislocation cores.) (Electronic copies of any cited references would be appreciated.) Priority will be given to innovative Visualizers displaying information in unique and provocative ways (see for example the work of Edward Tufte at https://www.edwardtufte.com/tufte/).