|
|
论文:2017,Vol:35,Issue(6):1059-1063 |
|
|
引用本文： |
|
|
罗殊彦, 朱怡安, 王子芃, 王仲轩. 基于SRIO总线的高性能嵌入式并行处理模型研究[J]. 西北工业大学学报 |
|
|
Luo Shuyan, Zhu Yi'an, Wang Zipeng, Wang Zhongxuan. Research on High Performance Embedded Parallel Processing Model Based on SRIO Bus[J]. Northwestern polytechnical university |
|
|
|
|
|
|
|
| 基于SRIO总线的高性能嵌入式并行处理模型研究 |
|
| 罗殊彦1, 朱怡安1, 王子芃2, 王仲轩1 |
|
1. 西北工业大学 计算机学院, 陕西 西安 710072;
2. 西北工业大学软件与微电子学院, 陕西 西安 710072 |
| 摘要: |
| 针对嵌入式系统所处理任务日益复杂，难以满足安全关键任务的执行时间需求问题，提出了一种高性能嵌入式并行处理模型（HPEPPM）。模型可分为3层。底层传输层通过使用SRIO（serial rapid I/O）总线，替代传统的以太网进行板间通信，以提高传输效率，降低通信延迟。中间层为MPI接口层，使用MPI协议进行并行处理，以加快任务执行时间。最上层为应用层，用于执行具体的任务。通过执行HPL测试程序的方式，对模型进行了多个指标的测试，测试结果优于传统模式。 |
| 关键词:
高性能嵌入式系统
并行处理模型
SRIO总线
MPI
HPEPPM
|
|
| Research on High Performance Embedded Parallel Processing Model Based on SRIO Bus |
|
| Luo Shuyan1, Zhu Yi'an1, Wang Zipeng2, Wang Zhongxuan1 |
|
1. School of Computer Science and Engineering, Northwestern Polytechnical University, Xi'an 710072, China;
2. School of Software and Microelectronics of Northwestern Polytechnical University, Northwestern Polytechnical University, Xi'an 710072, China |
| Abstract: |
| The task of embedded system is becoming more and more complex, and it is difficult to meet the requirement of execution time of security critical task, the paper proposes a high performance embedded parallel processing model (HPEPPM) to solve the above problems. The model is divided into three layers, the bottom layer is transmission layer by using Serial Rapid I/O (SRIO) bus to replace the traditional Ethernet communication between boards, improve transmission efficiency and reduce communication delay; the middle layer is MPI interface layer, parallel processing using the MPI protocol, thus speeding up the task execution time; the top layer is the application layer, task specific execution system. In this paper, the model is tested by HPL benchmark, and the results are better than the traditional model. |
| Key words:
high performance embedded system
parallel processing model
SRIO bus
MPI
HPEPPM
big data
cloud computing
computational efficiency
design of experiments
|
|
| 收稿日期: 2016-10-28
修回日期:
|
| DOI: |
| 基金项目: 航空科学基金（20150753010）、民机专项（XJZ-2015-D-76）及陕西省重点研发计划重大重点项目（2016MSZD-G-8-1）资助 |
| 通讯作者:
Email： |
| 作者简介: 罗殊彦(1984-),西北工业大学博士研究生,主要从事计算机性能评价及人工智能研究。
|
|
|
|
|
|
|
|
参考文献: |
|
|
[1] Muraoka D, Yasugi M, Hiraishi T, et al. Evaluation of an MPI-Based Implementation of the Tascell Task-Parallel Language on Massively Parallel Systems[C]//International Conference on Parallel Processing Workshops, 2016:161-170
[2] Si M, A, A J, Balaji P, et al. MT-MPI:Multithreaded MPI for Many-Core Environments[C]//ACM International Conference on Supercomputing, 2014:125-134
[3] Wu X, Taylor V. Performance Characteristics of Hybrid MPI/OpenMP Scientific Applications on a Large-Scale Multithreaded BlueGene/Q Supercomputer[C]//ACIS International Conference on Software Engineering, Artificial Intelligence, 2013:303-309
[4] Aubrey-Jones T, Fischer B. Synthesizing MPI Implementations from Functional Data-Parallel Programs[J]. International Journal of Parallel Programming, 2016, 44(3):1-22
[5] Matveeva N, Suvorova E, Sheynin Y. QoS Mechanisms in Space Fibre and RapidIO:SpaceWire Networks and Protocols, Long Paper[C]//International SpaceWire Conference, 2016
[6] Baymani S, Alexopoulos K, Valat S. Exploring RapidIO Technology within a DAQ System Event Building Network[C]//IEEE-NPSS Real Time Conference, 2016:1-4
[7] Klilou A, Belkouch S, Elmaizi A, et al. Case Studies of Data Traffic Management on a High-Performance Computing System Based on Multi-DSPs and Serial RapidIO Interconnect[C]//IEEE International Conference on Information Technology for Organizations Development, 2016:1-6
[8] Shu L, Hao J, Song Y, et al. Optimal Many-to-Many Personalized Concurrent Communication in RapidIO-Based Fat-Trees[C]//IEEE/ACIS International Conference on Software Engineering, Artificial Intelligence, Networking and Parallel/Distributed Computing, 2016:343-350
[9] Goodwin S A, Weed R A, Sankar L N, et al. Toward Cost-Effective Aero-elastic Analysis on Advanced Parallel Computing Systems[J]. Journal of Aircraft, 2015, 36(36):710-715
[10] Prokopec A. Data Structures and Algorithms for Data-Parallel Computing in a Managed Runtime[D]. Switzerland, Ecole Polytechnique Fédérale de Lausanne, 2014 |
|
|
|
|
|
|
|
