The rapidly developed computational capability of the Graphic Processing Unit (GPU) provides an alternative platform for transport transient calculation. In this work, a GPU-based 2D method of characteristic (MOC) transient fixed source problem (TFSP) solver is implemented on lattice physics code ALPHA (Advanced Lattice Physics code based on Heterogeneous Architecture), which employs the full implicit method (FIM) and pin-based coarse mesh finite difference (CMFD) acceleration. C5G7-TD benchmark is calculated to verify the transient solver built-in ALPHA. To exploit the potential capability of GPU, sensitivity studies are performed on convergence criteria of MOC, which demonstrate single float-point precision (FP32) can be applied in the transient simulation. The numerical results of the ALPHA code show a good agreement with the results of MPACT and the ALPHA code with FP32 can provide enough accuracy and good performance on executing time.
Read full abstract