High-performance systolic arrays for band matrix multiplication

Yun Yang, Wenqing Zhao, Yasuaki Inoue

    Research output: Chapter in Book/Report/Conference proceedingConference contribution

    8 Citations (Scopus)

    Abstract

    Band matrix multiplication is widely used in DSP systems. However traditional Kung-Leiserson systolic array for band matrix multiplication cannot be realized with high cell-efficiency. In this paper, three high-performance band matrix multiplication systolic arrays (BMMSA) are presented based on the ideas of "Matrix Compression" and "Super Pipelined". These new systolic arrays are realized by compressing the data matrix skillfully and adjusting the operation sequence carefully. The results show that the best systolic array for band matrix multiplication uses almost 100% processing elements(PE) in each step. Also, these modifications increase the operation speed and at best spend only 1/3 processing time to complete the multiplication operation.

    Original languageEnglish
    Title of host publicationProceedings - IEEE International Symposium on Circuits and Systems
    Pages1130-1133
    Number of pages4
    DOIs
    Publication statusPublished - 2005
    EventIEEE International Symposium on Circuits and Systems 2005, ISCAS 2005 - Kobe, Japan
    Duration: 2005 May 232005 May 26

    Other

    OtherIEEE International Symposium on Circuits and Systems 2005, ISCAS 2005
    CountryJapan
    CityKobe
    Period05/5/2305/5/26

    Fingerprint

    Systolic arrays
    Processing

    ASJC Scopus subject areas

    • Electrical and Electronic Engineering

    Cite this

    Yang, Y., Zhao, W., & Inoue, Y. (2005). High-performance systolic arrays for band matrix multiplication. In Proceedings - IEEE International Symposium on Circuits and Systems (pp. 1130-1133). [1464792] https://doi.org/10.1109/ISCAS.2005.1464792

    High-performance systolic arrays for band matrix multiplication. / Yang, Yun; Zhao, Wenqing; Inoue, Yasuaki.

    Proceedings - IEEE International Symposium on Circuits and Systems. 2005. p. 1130-1133 1464792.

    Research output: Chapter in Book/Report/Conference proceedingConference contribution

    Yang, Y, Zhao, W & Inoue, Y 2005, High-performance systolic arrays for band matrix multiplication. in Proceedings - IEEE International Symposium on Circuits and Systems., 1464792, pp. 1130-1133, IEEE International Symposium on Circuits and Systems 2005, ISCAS 2005, Kobe, Japan, 05/5/23. https://doi.org/10.1109/ISCAS.2005.1464792
    Yang Y, Zhao W, Inoue Y. High-performance systolic arrays for band matrix multiplication. In Proceedings - IEEE International Symposium on Circuits and Systems. 2005. p. 1130-1133. 1464792 https://doi.org/10.1109/ISCAS.2005.1464792
    Yang, Yun ; Zhao, Wenqing ; Inoue, Yasuaki. / High-performance systolic arrays for band matrix multiplication. Proceedings - IEEE International Symposium on Circuits and Systems. 2005. pp. 1130-1133
    @inproceedings{84b360ffa0ee49c4af5a8b3dc825af83,
    title = "High-performance systolic arrays for band matrix multiplication",
    abstract = "Band matrix multiplication is widely used in DSP systems. However traditional Kung-Leiserson systolic array for band matrix multiplication cannot be realized with high cell-efficiency. In this paper, three high-performance band matrix multiplication systolic arrays (BMMSA) are presented based on the ideas of {"}Matrix Compression{"} and {"}Super Pipelined{"}. These new systolic arrays are realized by compressing the data matrix skillfully and adjusting the operation sequence carefully. The results show that the best systolic array for band matrix multiplication uses almost 100{\%} processing elements(PE) in each step. Also, these modifications increase the operation speed and at best spend only 1/3 processing time to complete the multiplication operation.",
    author = "Yun Yang and Wenqing Zhao and Yasuaki Inoue",
    year = "2005",
    doi = "10.1109/ISCAS.2005.1464792",
    language = "English",
    pages = "1130--1133",
    booktitle = "Proceedings - IEEE International Symposium on Circuits and Systems",

    }

    TY - GEN

    T1 - High-performance systolic arrays for band matrix multiplication

    AU - Yang, Yun

    AU - Zhao, Wenqing

    AU - Inoue, Yasuaki

    PY - 2005

    Y1 - 2005

    N2 - Band matrix multiplication is widely used in DSP systems. However traditional Kung-Leiserson systolic array for band matrix multiplication cannot be realized with high cell-efficiency. In this paper, three high-performance band matrix multiplication systolic arrays (BMMSA) are presented based on the ideas of "Matrix Compression" and "Super Pipelined". These new systolic arrays are realized by compressing the data matrix skillfully and adjusting the operation sequence carefully. The results show that the best systolic array for band matrix multiplication uses almost 100% processing elements(PE) in each step. Also, these modifications increase the operation speed and at best spend only 1/3 processing time to complete the multiplication operation.

    AB - Band matrix multiplication is widely used in DSP systems. However traditional Kung-Leiserson systolic array for band matrix multiplication cannot be realized with high cell-efficiency. In this paper, three high-performance band matrix multiplication systolic arrays (BMMSA) are presented based on the ideas of "Matrix Compression" and "Super Pipelined". These new systolic arrays are realized by compressing the data matrix skillfully and adjusting the operation sequence carefully. The results show that the best systolic array for band matrix multiplication uses almost 100% processing elements(PE) in each step. Also, these modifications increase the operation speed and at best spend only 1/3 processing time to complete the multiplication operation.

    UR - http://www.scopus.com/inward/record.url?scp=48649091653&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=48649091653&partnerID=8YFLogxK

    U2 - 10.1109/ISCAS.2005.1464792

    DO - 10.1109/ISCAS.2005.1464792

    M3 - Conference contribution

    SP - 1130

    EP - 1133

    BT - Proceedings - IEEE International Symposium on Circuits and Systems

    ER -