The Community for Technology Leaders
RSS Icon
Subscribe
pp: 1
Shuming Chen , National University of Defense Technology, Chang Sha
Yaohua Wang , National University of Defense Technology, Chang Sha
Sheng Liu , National University of Defense Technology, Chang Sha
Jianghua Wan , National University of Defense Technology, Chang Sha
Haiyan Chen , National University of Defense Technology, Chang Sha
Hengzhu Liu , National University of Defense Technology, Chang Sha
Kai Zhang , National University of Defense Technology, Chang Sha
Xiangyuan Liu , National University of Defense Technology, Chang Sha
Xi Ning , National University of Defense Technology, Chang Sha
ABSTRACT
Vector-SIMD architectures have gained increasing attention due to their high performance in signal processing applications. However, the performance of existing vector-SIMD architectures is still limited due to their inefficiency in the coordinated exploitation of different hardware units. To solve this problem, this paper proposes the FT-Matrix architecture, which improves the coordination of traditional vector-SIMD architectures from three aspects: (1) the cooperation between scalar and SIMD unit is refined with the dynamic coupling execution scheme; (2) the communication among SIMD lanes is enhanced with the matrix-style communication; (3) data sharing among vector memory banks is accomplished by the unaligned vector memory accessing scheme. The evaluation result exhibits an average performance gain of 58.5% against vector-SIMD architectures without the proposed improvements. A four-core chip with each core built upon the FT-Matrix architecture is also under fabrication.
CITATION
Shuming Chen, Yaohua Wang, Sheng Liu, Jianghua Wan, Haiyan Chen, Hengzhu Liu, Kai Zhang, Xiangyuan Liu, Xi Ning, "FT-Matrix: A Coordination-aware Architecture for Signal Processing", IEEE Micro, , no. 1, pp. 1, PrePrints PrePrints, doi:10.1109/MM.2013.129
67 ms
(Ver 2.0)

Marketing Automation Platform Marketing Automation Tool