This Article 
   
 Share 
   
 Bibliographic References 
   
 Add to: 
 
Digg
Furl
Spurl
Blink
Simpy
Google
Del.icio.us
Y!MyWeb
 
 Search 
   
Fast Vectorization for Calculating a Moving Sum
November 1995 (vol. 44 no. 11)
pp. 1335-1337

Abstract—A simple vectorized method for calculating a moving sum is developed. Our proposed method is suitable for register-to-register vector computers and entails much less redundant floating-point operations than the vectorized algorithm of Mossberg [3]. We demonstrate the performance of our vectorized algorithm on the CRAY X-MP EA/116se supercomputer.

[1] K. Hwang and F.A. Briggs,Computer Architecture and Parallel Processing.New York: McGraw Hill, 1984.
[2] J.M. Levesque and J.W. Williamson,A Guidebook to Fortran on Supercomputers, Section 2.2.1: “Memory-to-memory vector processors,” pp. 28-35.New York: Academic Press, 1989.
[3] B. Mossberg,“Vectorization of the calculation of a moving sum,” IEEE Trans. Computers, vol. 36, pp. 362-365, Mar. 1987.
[4] “Supercomputer programming (I): Advanced Fortran: Architecture,vectorization, and parallel computing,” Working manual for CRAY X-MP EA/116se, 1991.

Index Terms:
Automatic gain control, CRAY X-MP, Fortran, moving sum, prefix sum, vectorization.
Citation:
Wen-Ming Yan, Kuo-Liang Chung, "Fast Vectorization for Calculating a Moving Sum," IEEE Transactions on Computers, vol. 44, no. 11, pp. 1335-1337, Nov. 1995, doi:10.1109/12.475130
Usage of this product signifies your acceptance of the Terms of Use.