Resource requirements of dataflow programs
David Culler, Arvind Arvind · ACM SIGARCH Computer Architecture News · 1988
Parallel execution of programs requires more resources and more complex resource management than sequential execution. If concurrent tasks can be spawned dynamically, programs may require an inordinate amount of resources when the potential parallelism in the program is much greater than the amount of parallelism the machine can exploit. We describe loop bounding , a technique for dynamically controlling the amount of parallelism exposed in dataflow programs. The effectiveness of the technique in reducing token storage requirements is supported by experimental data in the form of parallelism profiles and waiting-token profiles . Comparisons are made throughout with more conventional approaches to parallel computing.
Cited by 102 works. doi:10.1145/633625.52417
Authors
References
- Register allocation via coloring
- Dependence graphs and compiler optimizations
- Optimizing supercompilers for supercomputers
- Two fundamental issues in multiprocessing
- Executing a program on the MIT tagged-token dataflow architecture
- The U-Interpreter
- A COMPILER FOR THE MIT TAGGED-TOKEN DATAFLOW ARCHITECTURE
- Implementation of multilisp
- COORDINATING LARGE NUMBERS OF PROCESSORS.
- Future scientific programming on parallel machines