High performance computing-I

[6-D-01] Xcompact3D Revisited: A Codebase for Heterogeneous Architectures Based on a Distributed-memory Tridiagonal Matrix Algorithm

*Semih Akkurt¹, Sebastien Lemaire², Paul Bartholomew², Stefano Rolfo³, Cedric Flageul⁴, Filippo Spiga⁵, Charles Moulinec³, Sylvain Laizet¹ (1. Imperial College London, 2. EPCC, University of Edinburgh, 3. STFC Daresbury Laboratory, UKRI, 4. Universite de Poitiers, 5. NVIDIA Corporation)

Keywords:High performance computing, Compact finite difference schemes, GPU