Partitioning parallel programs for macro-dataflow
Vivek Sarkar, John LeRoy Hennessy · 1986
Partitioning techniques are necessary to execute functional programs at a coarse granularity. Fine granularity execution is inefficient on general purpose multiprocessors. There is a trade off between parallelism and the overhead of exploiting parallelism. A compile time partitioning approach to achieve this trade-off is presented.
Cited by 107 works. doi:10.1145/319838.319863
Authors
References
- Computers and Intractability: A Guide to the Theory of NP-Completeness
- The Manchester prototype dataflow computer
- Data Flow Program Graphs
- Compile-time partitioning and scheduling of parallel programs
- Serial combinators: "optimal" grains of parallelism
- Automatic partitioning and scheduling on a network of personal computers
- Computer data structures