applied math seminar
Steve Plimpton, Sandia National Laboratories
Monday, January 29, 2018 -
3:30pm to 4:30pm
Title: Challenges for Materials Modeling at the Atomic and Meso Scales Speaker: Steve Plimpton, Sandia National Labs Abstract: Molecular dynamics is a well-established technique for atomistic materials modeling, as is kinetic Monte Carlo (kMC) for modeling materials at the mesoscale. However, both methods suffer from limited accuracy (compared to quantum calculations) and from limited length and time scales (compared to continuum models or experiment). In this talk I'll describe work by myself and by collaborators to address these issues. This includes development of quantum-accurate potentials, parallel implementation of a hyperdynamics method for time-accelerated atomistics, and a constant-time exact kMC algorithm which can be used as part of approximate parallel algorithms to enable large mesoscale simulations. I'll highlight some of the open research issues, both algorithmic and mathematical, associated with the methods. Bio: Steve is a staff member at Sandia National Laboratories, in the Multiscale Science Dept of the Center for Computing Research. He received his Ph.D. from Cornell University in 1989 in Applied & Engineering Physics and has been at Sandia ever since. Most of his research involves development of algorithms and efficient codes which use particles to model materials on large computers. He is the lead developer of the LAMMPS molecular dynamics simulator, and also a co-author of other open-source software packages: SPPARKS (kinetic Monte Carlo), SPARTA (Direct Simulation Monte Carlo), ChemCell (stochastic particle modeling of biological cells), and MR-MPI (MapReduce on top of MPI). See www.sandia.gov/~sjplimp for details. He is a fellow of the American Physical Society and was recently awarded the IEEE Sidney Fernbach award at the SC17 supercomputing conference, for his work on these codes.
Contact Name: Pavel Lushnikov