Parallel and distributed computation: numerical methods
Parallel and distributed computation: numerical methods
High-Speed Computation
Impact of Physical/Logical Network Topology on Parallel Matrix Computation
International Journal of High Performance Computing Applications
Hi-index | 0.48 |
It has been shown in [Nuclear Science and Engineering 93 (1986) 6799] that the finite difference discretization of Navier-Stoke's equation leads to the solution of N × N system written in the matrix form as My = B, where M is a quasi-tridiagonal having non-zero elements at the top right and bottom left corners. We present an efficient parallel algorithm on a p-processor hypercube implemented in two phases. In phase I a generalization of an algorithm due to Kowalik [High Speed Computation, Springer, New York] is developed which decomposes the above matrix system into smaller quasi-tridiagonal (p + 1) × (p + 1) subsystem, which is then solved in Phase II using an odd-even reduction method.