Parallel Monte Carlo eigenvalue calculations
Shigekazu Matsuura, F.B. Brown, R.N. Blomquist · Transactions of the American Nuclear Society · 1994
In this paper we present theoretical models for the speed-ups attainable from parallel Monte Carlo eigenvalue calculations. The models allow for various hardware features (CPU and communications), different problem scaling strategies, and the iterative nature of the Monte Carlo algorithm. Measured performance results from the IBM-SP1 parallel computer are also presented for realistic Monte Carlo eigenvalue calculations.