BATSRUS GPU: Faster-than-Real-Time Magnetospheric Simulations with a Block-Adaptive Grid Code

TL;DR

This paper presents a GPU-accelerated version of the BATSRUS magnetospheric simulation code, achieving faster-than-real-time performance and high scalability on modern supercomputers.

Contribution

The authors successfully ported BATSRUS to GPUs using OpenACC, developed a new message passing algorithm for multi-GPU support, and demonstrated significant performance improvements.

Findings

GPU implementation achieves 3.6x faster than real time on a single A100 GPU.

Strong scaling up to 256 GPUs with 50-60% efficiency.

Performance on 4 GPUs is 6.9 times faster than real time.

Abstract

BATSRUS, our state-of-the-art extended magnetohydrodynamic code, is the most used and one of the most resource-consuming models in the Space Weather Modeling Framework. It has always been our objective to improve its efficiency and speed with emerging techniques, such as GPU acceleration. To utilize the GPU nodes on modern supercomputers, we port BATSRUS to GPUs with the OpenACC API. Porting the code to a single GPU requires rewriting and optimizing the most used functionalities of the original code into a new solver, which accounts for around 1% of the entire program in length. To port it to multiple GPUs, we implement a new message passing algorithm to support its unique block-adaptive grid feature. We conduct weak scaling tests on as many as 256 GPUs and find good performance. The program has 50-60% parallel efficiency on up to 256 GPUs, and up to 95% efficiency within a single node (4 GPUs). Running large problems on more than one node has reduced efficiency due to hardware bottlenecks. We also demonstrate our ability to run representative magnetospheric simulations on GPUs. The performance for a single A100 GPU is about the same as 270 AMD "Rome" CPU cores, and it runs 3.6 times faster than real time. The simulation can run 6.9 times faster than real time on four A100 GPUs.