The Community for Technology Leaders
2016 International Conference on Parallel Architecture and Compilation Techniques (PACT) (2016)
Haifa, Israel
Sept. 11, 2016 to Sept. 15, 2016
ISBN: 978-1-5090-5308-7
pp: 441-442
Xiaowei Shen , SKL of Computer Architecture, Institute of Computing Technology, Chinese Academy of Sciences, Beijing, China
Xiaochun Ye , SKL of Computer Architecture, Institute of Computing Technology, Chinese Academy of Sciences, Beijing, China
Xu Tan , SKL of Computer Architecture, Institute of Computing Technology, Chinese Academy of Sciences, Beijing, China
Da Wang , SKL of Computer Architecture, Institute of Computing Technology, Chinese Academy of Sciences, Beijing, China
Zhimin Zhang , SKL of Computer Architecture, Institute of Computing Technology, Chinese Academy of Sciences, Beijing, China
Dongrui Fan , SKL of Computer Architecture, Institute of Computing Technology, Chinese Academy of Sciences, Beijing, China
Zhimin Tang , SKL of Computer Architecture, Institute of Computing Technology, Chinese Academy of Sciences, Beijing, China
ABSTRACT
Dataflow computing is proved to be promising in high-performance computing. However, traditional dataflow architectures are general-purpose and not efficient enough when dealing with typical scientific applications due to low utilization of function units. In this paper, we propose an optimization of dataflow architectures for scientific applications. The optimization introduces a request for operands mechanism and a topology-based instruction mapping algorithm to improve the efficiency of dataflow architectures. Experimental results show that the request for operands optimization achieves a 4.6% average performance improvement over the traditional dataflow architectures and the TBIM algorithm achieves a 2.28× and a 1.98× average performance improvement over SPDI and SPS algorithm respectively.
INDEX TERMS
Computer architecture, Optimization, Algorithm design and analysis, Parallel processing, Program processors, Topology, Computers,high-performance computing, Request for operands, topology-based instruction mapping, dataflow computing
CITATION
Xiaowei Shen, Xiaochun Ye, Xu Tan, Da Wang, Zhimin Zhang, Dongrui Fan, Zhimin Tang, "POSTER: An optimization of dataflow architectures for scientific applications", 2016 International Conference on Parallel Architecture and Compilation Techniques (PACT), vol. 00, no. , pp. 441-442, 2016, doi:10.1145/2967938.2974054
97 ms
(Ver 3.3 (11022016))