A systolic array for SVD updating
نویسندگان
چکیده
In an earlier paper, an approximate SVD updating scheme has been derived as an interlacing of a QR updating on the one hand and a Jacobi-type SVD procedure on the other hand, possibly supplemented with a certain re-orthogonalization scheme. In this paper, this updating algorithm is mapped onto a systolic array with O(n 2 ) parallelism for O(n 2 ) complexity, resulting in an O(n 0 ) throughput. Furthermore it is shown how a square root free implementation is obtained by combining modi ed Givens rotations with approximate SVD schemes.
منابع مشابه
A Systolic Array for Recursive Least Squares Computations - Part II : Mapping directionally weighted RLS on an SVD updating array
In an earlier paper, a systolic algorithm/array was derived for recursive least squares (RLS) estimation, which achieves an O(n0) throughput rate with O(n2) parallelism. The resulting array is specifically tuned towards the RLS problem. Here, a different route is taken, by trying to implement the RLS problem on a systolic array, which is also useful for several other applications, such as, e.g....
متن کاملA systolic algorithm for QSVD updating
In earlier reports, a Jacobi-type algorithm for SVD updating has been developed 8] and implemented on a systolic array 9]. Here, this is extended to a generalized decomposition for a matrix pair, viz. the quotient singular value decomposition (QSVD). Updating problems are considered where new rows are appended to either one or both of the matrices involved. Systolic arrays as well as square roo...
متن کاملSystolic Algorithms for Adaptive Signal Processing
An overview is given of recent work in parallel algorithms development. It is shown how one speci c type of systolic algorithm/array can be used for several `classical' adaptive signal processing tasks, such as recursive least squares parameter estimation, SVD updating, Kalman ltering, beamforming and direction nding, etc.
متن کاملFactored Orthogonal Matrix-vector Multiplication with Applications to Parallel and Adaptive Eigennltering and Svd 1 Factored Orthogonal Matrix -vector Multiplication with Applications to Parallel and Adaptive Eigennltering and Svd
A novel algorithm is presented for adaptive eigennltering and for updating the singular value decomposition (SVD). It is an improvement upon an earlier developed Jacobi-type SVD updating algorithm, where now the exact orthogonality of the matrix of singular vectors/eigenvectors is guaranteed by storing a minimal factorization. This orthogonality property is known to be crucial for the numerical...
متن کاملA Systolic VLSI Architecture for Complex SVD
This thesis presents a systolic algorithm for the SVD of arbitrary complex matrices, based on the cyclic Jacobi method with \parallel ordering". As a basic step in the algorithm, a two-step, two-sided unitary transformation scheme is employed to diagonalize a complex 2 2 matrix. The transformations are tailored to the use of CORDIC (COordinate Rotation Digital Computer) algorithms for high spee...
متن کامل