Abstract
A previous study of MD algorithms designed for GPU use is extended to cover more recent developments in GPU architecture. Algorithm modifications are described, togther with extensions to more complex systems. New measurements include the effects of increased parallelism on GPU performance, as well as comparisons with multiple-core CPUs using multitasking based on CPU threads and message passing. The results show that the GPU retains a significant performance advantage.
| Original language | English |
|---|---|
| Article number | 012007 |
| Journal | Journal of Physics: Conference Series |
| Volume | 2241 |
| Issue number | 1 |
| DOIs | |
| State | Published - 6 Apr 2022 |
| Event | 33rd Annual CSP Workshop on Recent Developments in Computer Simulation Studies in Condensed Matter Physics, CSP 2020 - Virtual, Online Duration: 17 Feb 2020 → 21 Feb 2020 |
Bibliographical note
Publisher Copyright:© Published under licence by IOP Publishing Ltd.