Field‐programmable gate array acceleration of the Tersoff potential in LAMMPS

Molecular dynamics simulation is a common method to help humans understand the microscopic world. The traditional general-purpose high-performance computing platforms are hindered by low computational and power efficiency, constraining practical application of large-scale long-time many-body molecular simulations. In order address these problems, novel accelerator for Tersoff potential designed based on field-programmable gate array (FPGA) platforms, which enables acceleration LAMMPS using FPGAs. Firstly, an on-the-fly proposed build neighbor lists reduce storage usage. Besides, multilevel parallelizations implemented enable be flexibly deployed FPGAs different scales achieve good performance. Finally, mathematical models built, determine optimal-performance parameters proposed. Experimental results show that, when tested Xilinx Alveo U200, achieves performance 9.51 ns/day in 55,296-atom system, 2.00 × $$ \times increase compared Intel I7-8700K 1.70 NVIDIA Tesla K40c under same test case. addition, terms efficiency improvements 7.19 I7-8700K, 4.33 2.11 Titan Xp, respectively.