Implementing a data parallel language on a tightly coupled multiprocessor

Technical Report

Public Deposited

Citeable URL: https://ir.library.oregonstate.edu/concern/technical_reports/6q182t82n

Descriptions

Attribute Name	Values
Creator	Quinn, Michael J. Hatcher, Philip J.
Abstract	Our thesis is that data parallel programs can be translated into programs that execute efficiently on a variety of architectures. Currently we are focusing on the data parallel programming language C, developed by Thinking Machines Corporation for the Connection Machine processor array. Previous papers have described our design of a cross-compiler for hypercube multicomputers. In this paper we propose the design of a cross-compiler that translates C programs into C programs suitable for compilation and execution on a tightly coupled multiprocessor. The C* language is based upon a synchronous model of parallel computation, while the resulting C code executes asynchronously on the multiprocessor. We present an algorithm that, given a list of data dependences between expressions, produces a minimal set of barrier synchronizations necessary to ensure the correctness of the translated program.
Resource Type	Technical Report
Date Issued	1989
Academic Affiliation	Computer Science
Series	Technical report (Oregon State University. Department of Computer Science)
Rights Statement	In Copyright
Publisher	Oregon State University. Department of Computer Science
Peer Reviewed	No
Language	English [eng]

Thumbnail	Title	Date Uploaded	Visibility	Actions
	Quinn_Hatcher_89_60_20.pdf	2022-12-15	Public	Download