GPU-acceleration of locally mesh allocated two phase flow solver for nuclear reactors

This paper presents a GPU-based Poisson solver on a block-based adaptive mesh refinement (block-AMR) framework. The block-AMR method is essential for high performance GPU computation and efficient description of the nuclear reactor which is composed of complicated structures. In this paper, we successfully implement a conjugate gradient method with a state-of-the-art multi-grid preconditioner (MG-CG) on the block-AMR framework. GPU kernel performance of the MG-CG method was measured on the GPU-based supercomputer TSUBAME3.0. The bandwidth of a vector-vector sum, a matrix-vector product, and a dot product in the CG kernel gave good performance at about 60% of the peak performance. In the MG kernel, the smoothers in a three-stage V-cycle MG method are implemented using a mixed precision red-black SOR (RB-SOR) method, which also gave good performance. For a large-scale Poisson problem with 453.0 x 10⁶ cells, the developed MG-CG method reduced the number of iterations to less than 30% and achieved x 2.5 speedup compared with the original preconditioned CG method. (author)