Abstract:
The combinatorial geometry Monte Carlo particle transport infrastructure JCOGIN was introduced. The architecture and data structure of JCOGIN were designed. Further, particle parallelism and decomposition parallelism were designed and implemented to support the up to 10 million solids and 10 billion particles. By the testing, JMCT which was developed on JCOGIN reaches the parallel efficiency of 70% on 20 480 cores for outside source problem (P9) with 2G particles. By the decompo-sition parallelism, the pin-by-pin computation of the Dayawan reactor was implemented, and the number of the solids is up to 10 million and the number of particles is up to 10 billion.