Using hammock graphs to eliminate nonstructured branch statements
The detection of parallelism is based on the dependence analysis of the program. One of the key problems is to partition the program into independent subtasks in the presence of control statements (MIMD). In this paper the nonstructured control statements are converted into structured ones. In this way, a large block of statements is obtained, and the control dependences of the program are easily converted into manageable parallel tasks.
Unable to display preview. Download preview PDF.
- 1.Allen, J.R. 1983. Dependence Analysis for subscript Variables and Its Application to Program Transformations. Ph.D. Dissertation, Department of Mathematical Sciences, Rice University, Houston, Texas(April).Google Scholar
- 2.Allen, J.R., Kennedy, K., Porterfield, C., and Warren, J. 1983. Conversion of Control Dependence to Data Dependence. In Conference Proceedings — The 10th Annual ACM Symposium on Principles of Programming Languages (Austin, Texas, January 24–26), ACM Press, pp. 177–189.Google Scholar
- 3.Banerjee, U., 1988. Dependence Analysis for Supercomputing. The Kluwer international series in engineering and computer science. Parallel processing and fifth generation computing. ISBN 0-89838-289-0. Kluwer Academic Publishers, 1988.Google Scholar
- 4.Erik D'Hollander. Computer Aided Dataflow Analysis for the Conversion of Sequential Programs into Parallel Form. Algorithm and Applications on Vector and Parallel Computers. Edited by H.J.J te Riele, Th.J.Dekker and H.A.van der Vorst. Elsvier Science Publishers B.V.(North-Holland), 1987. pp. 75–102.Google Scholar
- 5.Fubo Zhang and Erik H. D'Hollander. 1993. Using Hammock Graphs to Eliminate Nonstructured Branch Statements. LEM Internal Report DG 93-31.Google Scholar