Efficient systolic structures for LU decomposition and system of linear equations
Systolic architecture is designed to take advantage of VLSI technology to perform parallel implementations of algorithms. A previously proposed systolic structure for LU decomposition [cf. \textit{H. T. Kung} and \textit{C. E. Leiserson} [Sparse Matrix Computations Proc. Symp., Knoxville 1978, 256-282 (1979; Zbl 0404.68037)] has hardware efficiency of 33 \%. This paper describes two more efficient systolic algorithms for this problem. The first assumes that elements of the matrix A are stationary and so must be stored in the structure; this process is 100 \% hardware efficient. The second overlaps the inputting of A with the computations; the cost of this benefit is a reduction to 50 \% hardware efficiency. Figures accompanying the text enable the reader to differentiate between the two methods. A third section describes an integrated systolic structure for solving a system of linear equations which is based on the first of the algorithms.
- A systolic array architecture for linear and inverse matrix systems
- A modular systolic linear array for gaussian elimination
- Part II A systolic array for numerical linear algebra
- On the direct parallel solution of systems of linear equations: New algorithms and systolic structures
- Solution of dense linear systems on an optimal systolic architecture
- A systolic array architecture for linear and inverse matrix systems
- scientific article; zbMATH DE number 4192945 (Why is no real title available?)
- scientific article; zbMATH DE number 3958727 (Why is no real title available?)
- An efficient implementation of LU decomposition in C
- Part II A systolic array for numerical linear algebra
This page was built for publication: Efficient systolic structures for LU decomposition and system of linear equations
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1103323)