Grid Technology with Dynamic Load Balancing for Monte Carlo Simulations

Authors:
Y. P. Galyuk;Vladimir Memnonov;S. E. Zhuravleva;V. I. Zolotarev
Affiliations:
-;-;-;-
Venue:
PARA '02 Proceedings of the 6th International Conference on Applied Parallel Computing Advanced Scientific Computing
Year:
2002

Citing 1
Cited 2

Parallel Implementation of a Corrected DSMC Method

PaCT '01 Proceedings of the 6th International Conference on Parallel Computing Technologies

Parallel implementation of the DSMC method coupled with a continuum solution: simulation of a lubrication problem in magnetic disc storage

ICCS'03 Proceedings of the 2003 international conference on Computational science: PartII
Monte Carlo simulation on heterogeneous distributed systems: A computing framework with parallel merging and checkpointing strategies

Future Generation Computer Systems

Quantified Score

Hi-index	0.00

Visualization

Abstract

With the help of Monte Carlo simulation we study flows in very narrow channels in transitional regime. The standard difficulty of large statistical scattering in Monte Carlo simulation results, which for these problems is especially severe because of the small values of mean gas velocities in comparison with the thermal molecular speeds, has been overcomed by enlarging the samples through Internet connection of several clusters into a metacomputer. Special dynamic load balancing technique under distributed memory conditions was developed. It is described in detail and quantitative estimation of its performance is given. It is shown in particular that under condition of only 1 Mb/s Internet channel it is possible to obtain efficiency of processor's utilisation as high as 93% - 94%. This is just 2% - 3% lower than for pure separate computation on the same cluster. Some numerical simulation results are outlined.