Scattering arrays for matrix computations
Several new mesh connected multiprocessor architectures are adapted to execute highly parallel algorithms for matrix algebra and signal processing, such as triangular- and eigen-decomposition, inversion and low-rank updating of general matrices, as well as Toeplitz and Hankel related matrices. These algorithms are based on scattering theory concepts and information preserving transformations hence they exhibit local communication, and simple control and memory management, ideal for VLSI implementation. The architectures are based on two-dimensional scattering arrays that can be folded into linear arrays either through time-sharing or due to simple computation wavefronts, or due to special structures of the matrices involved, such as Toeplitz. 21 references.
- Research Organization:
- Stanford Univ., CA
- OSTI ID:
- 6412333
- Journal Information:
- Proc. SPIE Int. Soc. Opt. Eng.; (United States), Vol. 298
- Country of Publication:
- United States
- Language:
- English
Similar Records
Parallel sparse matrix computations: Wavefront minimization of sparse matrices. Final report for the period ending June 14, 1998
Efficient matrix multiplications on a concurrent data-loading array processor