Multi-GPU-based Swendsen-Wang multi-cluster algorithm for the simulation of two-dimensional q-state Potts model
Multi-GPU-based Swendsen-Wang multi-cluster algorithm for the simulation of two-dimensional q-state Potts model
复制标题
DOI：
10.1016/j.cpc.2012.08.006
复制
发表时间：
2012-08
期刊：
影响因子：
--
通讯作者：
Yukihiro Komura;Y. Okabe
中科院分区：
文献类型：
--
作者：
Yukihiro Komura;Y. Okabe
针对二维Q态Potts模型的Swendsen-Wang多集群算法，我们提出了具有通用统一设备架构(CUDA)的多GPU计算。将我们的算法扩展为单GPU计算[Y.Komura，Y.，Okabe，基于GPU的Swendsen-Wang多集群算法，用于模拟二维经典自旋系统，Comput.太棒了。通讯。183(2012)1155-1161]，实现了多个GPU的Swendsen-Wang多集群算法的GPU计算。我们在大型开放科学超级计算机TSUBAME 2.0上实现了我们的代码，并测试了2D Potts模型模拟的性能和可扩展性。在使用256GPU的特斯拉M2050上，对于Q=2的Potts模型(伊辛模型)，在线性系统尺寸L=65536的临界温度下，其性能为每纳秒37.3次自旋翻转。
We present multiple GPU computing with the common unified device architecture (CUDA) for the Swendsen–Wang multi-cluster algorithm of two-dimensional (2D) q-state Potts model. Extending our algorithm for single GPU computing [Y. Komura, Y. Okabe, GPU-based Swendsen–Wang multi-cluster algorithm for the simulation of two-dimensional classical spin systems, Comput. Phys. Comm. 183 (2012) 1155–1161], we realize the GPU computation of the Swendsen–Wang multi-cluster algorithm for multiple GPUs. We implement our code on the large-scale open science supercomputer TSUBAME 2.0, and test the performance and the scalability of the simulation of the 2D Potts model. The performance on Tesla M2050 using 256 GPUs is obtained as 37.3 spin flips per a nano second for the q=2 Potts model (Ising model) at the critical temperature with the linear system size L=65536.