ALGEBRAIC RECURRENCE TRANSFORMATIONS FOR MASSIVE PARALLELISM

被引:0
|
作者
FETTWEIS, GP
THIELE, L
机构
[1] Teknekron Communications Systems, Berkeley
[2] Institute of Microelectronics, University of Saarland, Saarbrücken
关键词
D O I
10.1109/81.269037
中图分类号
TM [电工技术]; TN [电子技术、通信技术];
学科分类号
0808 ; 0809 ;
摘要
If an algorithm needs to be implemented for a high-speed application, a recursive loop is often found to be a bottleneck. Therefore, for some practical examples much work has been done so far to show how to transform single-step recursions into M-step recursions using so-called ''look-ahead computations.'' In this paper, we show that certain algebraic structures are sufficient for the operations of the loop, to allow the application of the well-knwn look-ahead computation. This provides a unifying algebraic basis to derive a new generalization for more complex recursions with more than one operation, which now allows the application of look-ahead computation to a wide area of problems.
引用
收藏
页码:949 / 952
页数:4
相关论文
共 50 条
  • [1] AUTOMATING NON-UNIMODULAR LOOP TRANSFORMATIONS FOR MASSIVE PARALLELISM
    XUE, JL
    PARALLEL COMPUTING, 1994, 20 (05) : 711 - 728
  • [2] Derivation of systolic algorithms for the algebraic path problem by recurrence transformations
    Djamégni, CT
    Quinton, P
    Rajopadhye, S
    Risset, T
    PARALLEL COMPUTING, 2000, 26 (11) : 1429 - 1445
  • [3] Massive Atomics for Massive Parallelism on GPUs
    Egielski, Ian
    Huang, Jesse
    Zhang, Eddy Z.
    ACM SIGPLAN NOTICES, 2014, 49 (11) : 93 - 103
  • [4] Parallelism in AGREE Transformations
    Corradini, Andrea
    Duval, Dominique
    Prost, Frederic
    Ribeiro, Leila
    GRAPH TRANSFORMATION, 2016, 9761 : 37 - 53
  • [5] ON THE HORIZON - MASSIVE PARALLELISM
    RATTNER, J
    ELECTRONIC DESIGN, 1990, 38 (16) : 58 - 59
  • [6] ALGORITHM TRANSFORMATIONS FOR UNLIMITED PARALLELISM
    THIELE, L
    FETTWEIS, G
    AEU-ARCHIV FUR ELEKTRONIK UND UBERTRAGUNGSTECHNIK-INTERNATIONAL JOURNAL OF ELECTRONICS AND COMMUNICATIONS, 1990, 44 (02): : 83 - 91
  • [7] Algorithm transformations for unlimited parallelism
    Thiele, Lothar
    Fettweis, Gerhard
    AEU. Archiv fur Elektronik und Ubertragungstechnik, 1990, 44 (02): : 83 - 91
  • [8] MATRIX CRUNCHING WITH MASSIVE PARALLELISM
    CUSHMAN, B
    VLSI SYSTEMS DESIGN, 1988, 9 (12): : 18 - &
  • [9] MASSIVE PARALLELISM - THE NAME OF THE GAME
    FAGGIN, F
    BYTE, 1992, 17 (02): : 134 - 134
  • [10] MASSIVE TRANSFORMATIONS
    MASSALSKI, TB
    MATERIALS SCIENCE AND ENGINEERING, 1976, 25 (SEP-O): : 119 - 125