A task scheduling algorithm for the parallel expression evaluation in a reconfigurable fully digit on-line network
In this paper, we present a task scheduling algorithm which accounts for digit-level pipelines of on-line arithmetic units when a limited number of heterogeneous on-line arithmetic units can be connected totally with each other and the network is reconfigurable during the execution. In on-line arithmetic, an arithmetic unit can be reused after the completion of its computing process while its result is available digit by digit during the computation. Thus, a new criterion, called the maximal delay is introduced to take into account additional precedence constraints based on the on-line delay.
Unable to display preview. Download preview PDF.
- 1.M. D. Ercegovac, “On-line arithmetic: an overview,” SPIE Vol. 495 Real Time Signal Processing VII, pp. 86–92, 1984.Google Scholar
- 2.M. Irwin and R. Owens, “Fully digit on-line networks,” IEEE Trans. Comput., vol. c-32, no. 4, pp. 402–406, 1983.Google Scholar
- 3.S. P. Johansen, “A universal online bit-serial cell for parallel expression evaluation,” in Proc. of 2nd IEEE Symp. on Parallel and Dist Processing, (Dallas, Texas, USA), pp. 335–338, Sept. 1990.Google Scholar
- 4.P. K. G. Tu and M. D. Ercegovac, “Application of on-line arithmetic algorithms to the svd computation: preliminary results,” in Proc. 10th Symp. on Comput. Arith., pp. 246–255, 1991.Google Scholar
- 5.J. Duprat, M. Fiallos, J. M. Muller, and H. J. Yeh, “Delays of on-line floating-point operators in borrow save notation,” in Algorithms and parallel VLSI architectures II, pp. 273–278, North Holland, 1992.Google Scholar
- 6.D. Bernstein, M. Rodeh, and I. Gertner, “Approximation algorithms for scheduling arithmetic expressions on pipelined machines,” J. of Algorithms, vol. 10, pp. 120–139, 1989.Google Scholar
- 7.J. Hwang, Y. Chow, F. D. Anger, and C. Lee, “Scheduling precedence graphs in systems with interprocessor communication times,” SIAM J. Comput., vol. 18, no. 2, pp. 244–257, 1989.Google Scholar
- 8.G. C. Sih and E. A. Lee, “Dynamic-level scheduling for heterogeneous processor networks,” in Proc. of 2nd IEEE Symp. on Parallel and Dist. Processing, (Dallas, Texas, USA), pp. 42–49, Dec. 1990.Google Scholar