Detailed modeling and massive tallying of nuclear reactors lead to memory overload for a single core processor.It could not be calculated by Monte Carlo particle transport with particle parallelism only.Domain decomposition is one of solutions.Domain decomposition needs to interchange particles between processors, so that inherit technique of pseudo-random number could not make identical results between serial and parallel.Two techniques of pseudo-random number are described to obtain identical results on different numbers of domains in a Monte Carlo particle simulation code.

Time consuming analysis of grid search in JMCT, a continuous-energy Monte Carlo neutron transport code, is done. Region projecting acceleration is performed on grids of main energy, fission yield neutron and unresolved resonance data. Quality of project grid is shown with arithmetic progression and geometric progression. At last, optimized grid search is test on several k_∞ problems. The models are fuel lattice cell at begin of cycle and middle of cycle. It shows that optimized grid search improves speed of JMCT remarkably in middle of cycle problem.

A parallel software module of FMM (fast multipole method) for three-dimensional Laplace kernel functions, JASMIN-3DLapFMM, is designed and implemented. The module is based on two phases parallel strategy of both processes and threads. A parallel software module is successfully used to solve the far field potential of electrostatic fields. With a fixed problem size of single processor,almost linear weak parallel scalability is obtained for a grand scale problem with 10¹⁰ particles on more than ten thousand processors. With a fixed size of total problem and 1024 processors, about three times speedup is obtained on four threads.