Venue: 411 West Hall (1085 S. University)
Dr Kent`s research is focusing on predicting and explaining the properties of materials using computer simulation. Over the last two decades, advances in simulation techniques coupled with increasing computer power have led to several methods that are able to predict physical properties of real materials to a useful accuracy. Moreover, these methods use little or no experimental data, making them especially valuable for the study of new materials and devices. Dr. Kent specializes in the application and development of these so-called “first principles” methods.
His research interests are broadly focused on atomistic materials simulation. His ongoing research projects include:
Dr. Kent is the director of the Center for Predictive Simulation of Functional Materials. He also leads the development of the QMCPACK application for exascale computing as part of the Exascale Computing Project. QMCPACK is a high-performance Quantum Monte Carlo code for computing the electronic structure of atoms, molecules and solids, including metals. QMCPACK is open source and available on GitHub.
Dr Kent is a member of the Nanotheory Institute at the Center for Nanophase Materials Sciences (CNMS) and the Computational Chemical and Materials Science group in the Computational Science and Engineering Division. He spent three years at NREL with Alex Zunger after completing his PhD with Richard Needs at the University of Cambridge. For several years he worked with Mark Jarrell at the University of Cincinnati on high-temperature cuprate superconductors. In 2009 he transitioned from JICS/UT Knoxville to ORNL.
Advances in the field of computational materials science have helped to predict, understand, and optimize the properties of many classes of materials. These include new battery electrodes, catalysts, and arguably even higher-temperature superconductors. However, we still lack a widely usable method where all the key uncertainties and approximations in the predictions can be assessed and systematically reduced. This is critical where the approximations in established methods fail, such as in quantum materials, or simply where greater accuracy is desired. In this talk I will first describe our recent advances in Quantum Monte Carlo methods that promise to meet this challenge. Second, I will describe the new algorithms and performance portable software design and development strategies we have adopted to run efficiently on the largest supercomputers powered by GPU accelerators from NVIDIA, AMD and Intel. The lessons learned can be applied in any area of scientific software development.
The MICDE Winter 2023 Seminar Series is open to all. University of Michigan faculty and students interested in predicting and explaining the properties of materials using computer simulation are encouraged to attend.
This seminar is cohosted by the Michigan Institute for Computational Discovery & Engineering (MICDE) and the Department of Physics. Dr. Kent will be hosted by Dr. Emanuel Gull, Associate Professor of Theoretical Condensed Matter Physics.
This is an in-person event.
Graduate Certificate in Computational Discovery and Engineering, and MICDE fellows, please use this form to record your attendance.
Questions? Email MICDEfirstname.lastname@example.org