GPU Acceleration of the Discontinuous Galerkin Shallow Water Equation Model with OpenACC
TL;DR
This paper presents porting of a discontinuous Galerkin shallow water equation model (DG-SWEM version 3.0.2), a discontinuous Galerkin solver for hurricane storm surge, to NVIDIA graphics processing units (GPUs) using OpenACC and unified memory.
Abstract
This paper presents porting of a discontinuous Galerkin shallow water equation model (DG-SWEM version 3.0.2), a discontinuous Galerkin solver for hurricane storm surge, to NVIDIA graphics processing units (GPUs). Time-explicit discontinuous Galerkin methods contain a large number of degrees of freedom but have been shown to exhibit a large amount of data parallelism due to the loose coupling between elements and thus are naturally mapped to the central processing unit (CPU) architecture. A previous framework in porting DG-SWEM to GPUs required converting subroutines from Fortran to C + + to be used with CUDA C + + . By using OpenACC and unified memory, we simplified the porting process and maintained a single codebase for both CPU and GPU versions. We tested the code using a large Hurricane Harvey scenario on NVIDIA’s Grace Hopper chip and compared the GPU code’s performance on multiple H200 nodes to the CPU code on Grace and AMD Milan CPU nodes.