Adaptive Parallel Monte Carlo Simulations of Low-Pressure Plasma Processing

The adaptive parallel computing model of plasma source for semiconductor processing is developed. By using the Monte Carlo method in the low-pressure plasma regime is characterized by its time-varying problem size. We proposed an adaptive processor allocation methodology for the parallel simulation to dynamically adapt to the workload variation and achieve better performance. Static and dynamic allocation schemes are illustrated to obtain a better understanding of the processor allocation sequences for the adaptive allocation.