Evaluation of working set algorithms for data-flow machines
This paper extends the algorithms for implementing the working set concept for data-flow machines proposed in  in order to cope with dynamic data-flow programs which include branches, loops, and procedure calls. The algorithms for computing the D-, E-, and L-levels are revised and the performance for static data-flow programs is reevaluated. Algorithms to manage branches, loops, and function calls are developed. The effectiveness of these algorithms is shown through computer simulation.
Unable to display preview. Download preview PDF.
- W. B. Ackerman, "A Structure Processing Facility for Data Flow Computers," Proc. 8th International Conference on Parallel Processing, 1978.Google Scholar
- W. B. Ackerman, "A Structure Controller for Data Flow Computers," CGS Memo 156, MIT, Jan 1978.Google Scholar
- Arvind and R. Thomas, "I-structures: An efficient data type for functional languages," Univ. of California Irvine Technical Report, 1980.Google Scholar
- Arvind and V. Katathail, "A Multiple processor Data-Flow Machine that Supports Generalized Procedures," Proc. 8th Annual Symposium on Computer Architecture, Mineapolis, Minnesota, 1981, 291–302.Google Scholar
- H. Asakura, H. Sunahara and M. Tokoro, "A Note on the Structure Memory for the Data-flow Machines with the Working Set Concept," Proc. of the 28th National Conference of Information Processing Society of Japan, Tokyo Japan, March 1984, 4F-10 (in Japanese).Google Scholar
- J. B. Dennis and D. P. Misunas, "A preliminary Architecture for a Basic Data Flow Processor," Proc. 2nd Annual Symp osium on Computer Architecture, Jan 1975, 126–132.Google Scholar
- M. Tokoro, J. R. Jagannathan and H. Sunahara, "On the Working Set Concept for Data-Flow Machine," Proc. of the 10th Annual Symposium on Computer Architecture, Jun. 1983, 90–97.Google Scholar