Innovative Multigrid Methods
中島研吾(東京大学)

In the present work, we are developing robust and efficient GMG and AMG methods, where we are focusing on development of algorithms for (1) efficient and robust smoother, (2) parallel global reordering/aggregation methods for robustness, (3) utilization of near-kernel vectors for robustness, and (4) hierarchical methods for scalability. Moreover, we develop new algorithms for Parallel-in-Space/Time (PinST).

研究紹介ポスター／最終報告書