About this Abstract |
| Meeting |
MS&T25: Materials Science & Technology
|
| Symposium
|
High-Entropy Materials: Solid Solutions, Intermetallics, Ceramics, Functional Materials and Beyond VI
|
| Presentation Title |
Better, Faster Alloy Simulations: an Improved Hybrid MC/MD Algorithm for LAMMPS |
| Author(s) |
Megan J. McCarthy, Aidan Thompson |
| On-Site Speaker (Planned) |
Megan J. McCarthy |
| Abstract Scope |
Hybrid Monte Carlo/molecular dynamics (MC/MD) methods are a powerful simulation approach widely used to study complex alloys. The availability of hybrid MC/MD in codes such as LAMMPS enables simulation of alloy phenomena outside timescale regimes of MD, such as chemical mixing/ordering and phase segregation. However, the applicability of MC/MD in large-scale simulations is hindered by computational inefficiencies and a lack of generality. Current algorithms in LAMMPS’ MC package suffer from poor scaling due to the nature of underlying energy calculations. Although alternative, faster approaches exist, they are often limited to use with a single category of interatomic potential. In this talk, we will present a new implementation of hybrid MC/MD in LAMMPS that addresses both limitations, offering users a fully general approach with 100-1000x speedup over existing code. We will demonstrate its utility with both classical and machine-learned potentials in simulations of species segregation and phase change in complex alloys. |