graph framework: A Domain Specific Compiler for Building Physics Applications

TL;DR

This paper introduces a graph-based compiler framework that translates physics equations into optimized code for various hardware platforms, demonstrated on RF ray tracing in fusion energy.

Contribution

It presents a novel domain-specific compiler framework that abstracts physics from hardware-specific kernels, supporting multiple accelerators including CPUs, Apple GPUs, and NVidia GPUs.

Findings

Successfully compiled physics equations for multiple hardware platforms.

Demonstrated application on RF ray tracing in fusion energy.

Achieved optimized kernel code generation for diverse accelerators.

Abstract

Modern supercomputers are increasingly relying on Graphic Processing Units (GPUs) and other accelerators to achieve exa-scale performance at reasonable energy usage. The challenge of exploiting these accelerators is the incompatibility between different vendors. A scientific code written using CUDA will not operate on a AMD gpu. Frameworks that can abstract the physics from the accelerator kernel code are needed to exploit the current and future hardware. In the world of machine learning, several auto differentiation frameworks have been developed that have the promise of abstracting the math from the compute hardware. However in practice, these framework often lag in supporting non-CUDA platforms. Their reliance on python makes them challenging to embed within non python based applications. In this paper we present the development of a graph computation framework which compiles physics equations to optimized kernel code for the central processing unit (CPUs), Apple GPUs, and NVidia GPUs. The utility of this framework will be demonstrated for a Radio Frequency (RF) ray tracing problems in fusion energy.