| About this Abstract |
| Meeting |
2025 TMS Annual Meeting & Exhibition
|
| Symposium
|
Special Topics in Nuclear Materials: Lessons Learned; Non-Energy Systems; and Coupled Extremes
|
| Presentation Title |
Developing a High-Performance Cluster Dynamics Code: Challenges and Lessons Learned |
| Author(s) |
Sophie Blondel, Philip Fackler, Philip Roth, David Bernholdt, Brian Wirth |
| On-Site Speaker (Planned) |
Sophie Blondel |
| Abstract Scope |
Xolotl is an open source DOE high-performance computing code using spatially dependent advection-reaction-diffusion kinetic rate theory to model the time evolution defects and microstructure in irradiated materials. It was originally developed to support two DOE SciDAC programs (on plasma surface interactions for fusion applications, and nuclear fuel in fission reactors). In the past 10 years, it has grown to model 5 different materials, be coupled to other simulation methods in a variety of ways, and integrate a grouping scheme and Kokkos for computational efficiency. This presentation will go over the challenges faced and the solutions (some tentative) explored along the way. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Modeling and Simulation, Nuclear Materials, |