Heterogeneous parallel algorithm design and performance optimization for WENO on the Sunway Taihulight supercomputer
Heterogeneous parallel algorithm design and performance optimization for WENO on the Sunway Taihulight supercomputer
复制标题
神威·太湖之光超级计算机上WENO异构并行算法设计与性能优化
DOI：
10.26599/tst.2018.9010112
复制
发表时间：
2020-02
影响因子：
6.6
通讯作者：
Chen Wenguang
中科院分区：
文献类型：
--
作者：
Huang Jianqiang;Han Wentao;Wang Xiaoying;Chen Wenguang
加权基本无振荡格式（韦诺）是一种双曲守恒律方程的求解方法，适用于求解高密度流体界面不稳定性问题。这些问题具有庞大而复杂的流程结构。为了充分利用高性能计算（HPC）系统的计算能力，有必要开发特定的方法来优化基于特定系统架构的应用程序的性能。Sunway TaihuLight超级计算机目前被评为世界上最快的超级计算机。本文在神威太湖之光上实现了一个高阶韦诺的异构并行算法设计和性能优化。分析了核函数的特点，提出了一种合适的异构并行模型。并在神威太湖之光上实现了该并行算法。通过使用访问优化，数据依赖消除和向量化优化，我们的并行算法可以实现高达172倍的加速在一个单一的节点上，和额外的58倍加速在64个节点上，具有接近线性的可扩展性。
A Weighted Essentially Non-Oscillatory scheme (WENO) is a solution to hyperbolic conservation laws, suitable for solving high-density fluid interface instability with strong intermittency. These problems have a large and complex flow structure. To fully utilize the computing power of High Performance Computing (HPC) systems, it is necessary to develop specific methodologies to optimize the performance of applications based on the particular system's architecture. The Sunway TaihuLight supercomputer is currently ranked as the fastest supercomputer in the world. This article presents a heterogeneous parallel algorithm design and performance optimization of a high-order WENO on Sunway TaihuLight. We analyzed characteristics of kernel functions, and proposed an appropriate heterogeneous parallel model. We also figured out the best division strategy for computing tasks, and implemented the parallel algorithm on Sunway TaihuLight. By using access optimization, data dependency elimination, and vectorization optimization, our parallel algorithm can achieve up to 172X speedup on one single node, and additional 58X speedup on 64 nodes, with nearly linear scalability.
登录
查看更多内容
DOI：
10.1109/sc.2010.2
发表时间：
2010-11
期刊：
2010 ACM/IEEE International Conference for High Performance Computing, Networking, Storage and Analysis
影响因子：
--
作者：
A. Nguyen;N. Satish;J. Chhugani;Changkyu Kim;P. Dubey
通讯作者：
A. Nguyen;N. Satish;J. Chhugani;Changkyu Kim;P. Dubey
DOI：
10.1109/ipdps.2017.53
发表时间：
2017-05
期刊：
2017 IEEE International Parallel and Distributed Processing Symposium (IPDPS)
影响因子：
--
作者：
Heng Lin;Xiongchao Tang;Bowen Yu;Youwei Zhuo;Wenguang Chen;Jidong Zhai;Wanwang Yin;Weimin Zheng
通讯作者：
Heng Lin;Xiongchao Tang;Bowen Yu;Youwei Zhuo;Wenguang Chen;Jidong Zhai;Wanwang Yin;Weimin Zheng
影响因子：
4.6
作者：
F. Grasso;S. Pirozzoli
通讯作者：
F. Grasso;S. Pirozzoli
影响因子：
4.1
作者：
HARTEN, A;ENGQUIST, B;CHAKRAVARTHY, SR
通讯作者：
CHAKRAVARTHY, SR
影响因子：
4.1
作者：
Jiang, GS;Shu, CW
通讯作者：
Shu, CW