Preliminary Investigation of Accelerating Molecular Dynamics Simulation on Godson-T Many-Core Processor

Molecular dynamics (MD) simulation is widely used in computational science, however, its irregular memory-access pattern imposes great difficulty on performance optimization. This paper presents a joint application/ architecture study to accelerate MD on an emerging unconventional computing platform-Godson-T many-core architecture. We propose three incremental optimizations: (1) a divide-and-conquer algorithm adaptive to on-chip memory; (2) a novel data-layout to re-organize linked-list cell data structures to improve data locality; (3) an on-chip locality-aware parallel algorithm to enhance data reuse. Experiments on an event-driven, cycle-accurate Godson-T simulator achieve excellent speedup of 62 on 64 cores.