Dataflow Architectures.
Abstract
Dataflow graphs are described as a machine language for parallel machines. Static and dynamic dataflow architectures are presented as two implementations of the abstract dataflow model. Static dataflow allows at most one token per arc in dataflow graphs and thus only approximates the abstract model where unbounded token storage per arch is assumed. Dynamic architectures tag each token and keep them in a common pool of storage, thus permitting a better approximation of the abstract model. The relative merits of the two approaches are discussed. Functional data structures and I-structures are presented as two views of data structures which are both compatible with the dataflow model. These views are contrasted and compared in regard to efficiency and exploitation of potential parallelism in programs. A discussion of major dataflow projects and a prognosis for data flow architectures are also presented. Keywords: Mulitprocessors, Parallel computation, Parallel computers. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Feb 12, 1986
- Accession Number
- ADA166235
Entities
People
- David E. Culler
Organizations
- Massachusetts Institute of Technology