GPU Acceleration of High-Speed Collision Molecular Dynamics Simulation
Dept. of Comput., Nat. Univ. of Defense Technol., Changsha, China
Ninth IEEE International Conference on Computer and Information Technology, 2009. CIT ’09, vol. 2, p.254-259
@conference{yang2009gpu,
title={GPU Acceleration of High-Speed Collision Molecular Dynamics Simulation},
author={Yang, C. and Wu, Q. and Chen, J. and Ge, Z.},
booktitle={IEEE Ninth International Conference on Computer and Information Technology},
pages={254–259},
year={2009},
organization={IEEE}
}
We discuss an implementation and optimization of GPU-accelerated Molecular Dynamics (MD) simulation of high-speed collision molecular model in NVIDIA CUDA language. A series of optimization methods are presented: spatial decomposition, use of shared memory and use of blockcell-link structure. These optimization methods effectively improve the performance by reducing data transfer time between CPU and GPU and reducing memory access time on GPU. We test our GPU-accelerated MD algorithm on a modern GPU, NVIDIA Tesla C870. The performance of our code implemented on the GPU with AMD Athlon64 4400+ is compared to the CPU-only version. GPU-accelerated MD simulation can achieve speedup of 38.37 times compared to the sequential version running on single core of the CPU. The peak performance of C870 reaches 15 GFLOPS.
March 25, 2011 by hgpu
