Dataflow architectures

By Arvind, David E. Culler. 1986

Subfields: ML Systems & Hardware

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In ML Systems & Hardware

1. Derek Gordon Murray
2. Michael Isard
3. Martín Abadi
4. Quoc V. Le
5. Rajat Monga
6. Jianmin Chen
7. J. R. Paul Hanna
8. Paul Barham

Citations in the corpus

• TensorFlow: A system for large-scale machine learning (2016, positive): “For this purpose, we borrow the Switch and Merge primitives from classic dynamic dataflow architectures [4].”
• Deep Probabilistic Programming (2017, background): “We desire two criteria: (a) integration with computational graphs, an efficient framework where nodes represent operations on data and edges represent data communicated between them (Culler, 1986); and (b) invariance of the representation under the graph, that is, the representation can be reused du”
• Advances in dataflow programming languages (2004, methodology): “The Pure Data.ow Model In the data.ow execution model, a pro­gram is represented by a directed graph [Arvind and Culler 1986; Davis and Keller 1982; Dennis 1974; Dennis and Misunas 1975; Karp and Miller 1966].”
• A computational model for TensorFlow: an introduction (2017, methodology): “In the TensorFlow implementation, they are mapped to dataflow primitives, as in tagged dataflow architectures [6, 7].”
• Organizing Business Knowledge: The MIT Process Handbook (2003, background): “More recently Ashforth and Fried (1988) build on the concept of scripts (Schank and Abelson 1977) to describe ''mindless''routines that can be initiated by a very limited stimulus and run through until completion, unless interrupted.”