scispace - formally typeset
Search or ask a question

Showing papers by "Thomas Sterling published in 1988"


Patent
12 Aug 1988
TL;DR: In this paper, an associative architecture for a static data flow processing system comprises a functional computation unit in which data processing operations are executed, a data processing execution control structure (template) storage and control unit and communication channels through which the functional computations and the templates communicate with one another.
Abstract: An associative architecture for a static data flow processing system comprises a functional computation unit in which data processing operations are executed, a data processing execution control structure (template) storage and control unit and communication channels through which the functional computation unit and the template storage and control unit communicate with one another. The template storage and control unit controls the supply of data to be processed by the functional computation unit and includes memory for storing a plurality of templates. Each template storage and control unit assembles data processing messages for application to a first of the communication channels for controlling the execution of a data processing operation by the functional computation unit. Each message contains the address of that template to which the result of the data processing operation is returned and stored in a return buffer, an opcode and either the data directly or the address of the template that contains the data to be processed by the functional computation unit. Each template also stores the status of a data processing execution cycle. Each template continuously monitors the communications channels for its address and, upon detecting its address, controllably interfaces prescribed information associated with the execution of a data processing operation with respect to the communication channels.

12 citations





Proceedings ArticleDOI
01 Nov 1988
TL;DR: The associative template mechanism, an application of associative techniques, uses specially configured content-addressable memories to provide efficient flow control for static data-flow program execution, and associative diffusion provides a token-free approach to static Dataflow computation.
Abstract: The static data flow model of computation promises high performance from fine grained parallelism, but conventional token-driven static data flow architectures are inefficient in terms of memory bandwidth and microcycles required per operation. The associative template mechanism, a new application of associative techniques, employs specially configured content-addressable memories to provide efficient flow control for static data flow program execution. It supports static data flow semantics while exhibiting memory bandwidth and microcycle demands comparable to those of conventional uniprocessors. Associative diffusion, a second application of associative methods, provides communication between adjacent nodes of a mesh-connected network of associative-template-based processors. This mechanism achieves nearest-neighbor communication at speeds comparable to intra-node transactions by overlapping domains of associativity across boundaries between neighboring nodes. Together, associative templates and associative diffusion provide a token-free approach to static data flow computation.

1 citations