Abstract
Considering the complex neutron-physical systems with the number of calculation points than 1 million, the calculation time can be an hour or more, and in case of non-stationary process with the small step of time, the waiting can be a month. The aim of the study is to develop effective techniques for the stationary computational modeling using parallelization technology to imply in non-stationary calculations. The parallelization scheme with height decomposition for stationary model was developed. The calculated speedup on the parallel part of the code are in good agreement with the linear dependence. The theoretical possible acceleration was derived using Amdal’s law. Comparing the obtained and theoretical curves demonstrates that the developed parallelization scheme has less acceleration due to losses. The losses and their relative contribution were identified.
Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
