Single Instruction Stream - Multiple Data Stream Machine Interconnection Network Design*
نویسنده
چکیده
An SIMD machine must have an interconnection network to pass data between processing elements. We introduce a model of SIMD machines which allows a formal mathematical analysis and comparison of different interconnection networks. Five interconnection networks that have been proposed in the literature are defined in terms of our model. They include a network similar to the one used in the STARAN, a network similar to the one recommended by Feng to implement data manipulating functions, the Illiac IV network, and the Perfect Shuffle. The networks are evaluated in terms of the upper and lower bounds on the time required for each network to simulate the actions of the others. It is usually impractical to implement all the interconnections that may be needed by the machine to perform a large variety of computations, so the ability of a network to simulate other interconnections is important. The methods used to prove the lower bounds and to construct the simulation algorithms to demonstrate the upper bounds can be generalized and applied to the analysis of other networks.
منابع مشابه
Control of a Partitionable Multimicroprocessor System
A dynamically reconfigurable large-scale multimicroprocessor system capable of operating as one or more independent SIMD (single instruction stream multiple data stream) machines and/or MIMD (multiple instruction stream multiple data stream) machines is described. The system consists of a Parallel Computation Unit, which contains N processors, N memory modules, and an interconnection network; Q...
متن کاملSimd Algorithms to Perform Linear Predictive Coding for Speech Processing Applications
The use of the SIMD (single instruction stream multiple data stream) mode of parallelism to perform the speech analysis task of linear predictive coding is explored. Linear prediction represents one of the major analysis techniques for speech compression, transmission, and recognition applications. Parallel algorithms to perform linear prediction have been developed, and are evaluated in terms ...
متن کاملMultiple Instruction Stream Control for an Associative Model of Parallel Computation
This paper describes a system software design for multiple instruction stream control in a massively parallel associative computing environment. The purpose of providing multiple instruction stream control is to increase throughput and reduce the amount of parallel slackness inherent in single instruction stream parallel programming constructs. The Multiple ASsociative Computing (MASC) model wi...
متن کاملParallel Sequence Alignment in Limited
Sequence comparison with aane gap costs is a problem that is readily parallelizable on simple single-instruction, multiple-data stream (SIMD) parallel processors using only constant space per processing element. Unfortunately, the twin problem of sequence alignment, nding the optimal character-by-character correspondence between two sequences, is more complicated. While the innovative O(n 2)-ti...
متن کاملParallel Sequence Alignment in Limited Space
Sequence comparison with affine gap costs is a problem that is readily parallelizable on simple single-instruction, multiple-data stream (SIMD) parallel processors using only constant space per processing element. Unfortunately, the twin problem of sequence alignment, finding the optimal character-by-character correspondence between two sequences, is more complicated. While the innovative O(n2)...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
عنوان ژورنال:
دوره شماره
صفحات -
تاریخ انتشار 2006