On mapping algorithms to linear and fault-tolerant systolic arrays
The authors develop a simple mapping technique to design systolic arrays with limited I/O capability. Using this, improved systolic algorithms are derived for some matrix computations, on linearly connected arrays of processing elements (PEs) with constant I/O bandwidth. The important features of these designs are modularity with constant hardware in each PE, few control lines, simple data input/ouput format, and improved delay time. They extend the technique to design an optimal time systolic algorithm for n x n matrix multiplication. In this model, the propagation delay is assumed to be proportional to wire length. Fault reconfiguration is achieved by using buffers to bypass faulty PE's, which does not affect the clock rate of the system. The unidirectional flow of control and data in our design assures correctness of the algorithm in the presence of faulty PE's. This design can be implemented on reconfigurable fault-tolerant VLSI arrays using the Diogenes methodology. The authors compare their designs to those in the literature and are shown to be superior with respect to I/O format, control, and delay from input to output.
- Research Organization:
- Dept. of Electrical Engineering-Systems, Univ. of Southern California, Los Angeles, CA (US)
- OSTI ID:
- 6243175
- Journal Information:
- IEEE Trans. Comput.; (United States), Journal Name: IEEE Trans. Comput.; (United States) Vol. 38:3; ISSN ITCOB
- Country of Publication:
- United States
- Language:
- English
Similar Records
On the design of fault-tolerant two-dimensional systolic arrays for yield enhancement
Design of easily testable and reconfigurable systolic arrays
Related Subjects
990210* -- Supercomputers-- (1987-1989)
990230 -- Mathematics & Mathematical Models-- (1987-1989)
ALGORITHMS
ARRAY PROCESSORS
COMPUTER CALCULATIONS
COMPUTER CODES
COMPUTERS
DESIGN
DIGITAL COMPUTERS
ELEMENTS
FAULT TOLERANT COMPUTERS
LINEAR PROGRAMMING
MAPPING
MATHEMATICAL LOGIC
MATHEMATICAL MODELS
OPTIMIZATION
PROGRAMMING
SEMIMETALS
SILICON