A concurrent, multigroup, discrete ordinates model of neutron transport

We present an algorithm for the concurrent solution of the linear system arising from a multigroup, discrete ordinates model of neutron transport. The target architectures consist of distributed memory computers ranging from workstation clusters to massively parallel computers. Based on an analysis of the memory requirement and floating point complexity of matrix-vector multiplication in the iterative solution of the linear system, we propose a data layout and communication strategy designed to achieve scalability with respect to all phase space variables. Numerical results are presented to demonstrate the performance of the algorithm on the nCUBE/2.<
>