採択課題 【詳細】
| jh260047 | High-resolution and cell-resolved simulation of cardiac electrophysiology |
|---|---|
| 課題代表者 | 中島研吾(東京大学情報基盤センター) Kengo Nakajima (Information Technology Center, The University of Tokyo) |
| 概要 | Cell-resolved modeling of cardiac electrophysiology has lately received increasing interest in the research domain of computational cardiology. In particular, the so-called EMI (extracellular-membrane-intracellular) model allows investigating individual cells and the coupling in-between. This stands in strong contrast to the conventional monodomain and bidomain models that are based on an averaged approach, where the individual cells are not present. However, simulations of the EMI model require computations that are several orders of magnitude larger than those needed by conventional models, thus relying on an efficient use of modern supercomputers especially those based on GPUs. As a continuation of the ongoing jh250028 project, we will continue our work on multiphysics-adapted mesh partitioning, advanced code optimization and GPU-accelerated linear algebra software, for the purpose of enabling cell-resolved simulations of cardiac physiology. Such huge-scale simulations will deepen the scientific understanding of vital electrical activities in the heart. The significance is due to the possibility of studying subcellular heterogeneity and non-uniform details along the cell membranes, which was impossible with conventional models. With respect to hardware, our aim is to secure the efficiency of GPU computing, especially on the latest GH200 superchips (available on the Miyabi system). A successful outcome of this project will be unprecedented, huge-scale cell-resolved simulations, providing new knowledge in the domains of both computational physiology and HPC. |
| 関連Webページ | |
| 報告書等 | 研究紹介ポスター ／ 最終報告書 |
無断転載禁止