A high-performance and portable implementation of the SISSO method for CPUs and GPUs

TL;DR

This paper presents a GPU-accelerated, portable implementation of the SISSO AI method using Kokkos, enabling efficient high-performance computations across Nvidia and AMD hardware.

Contribution

The authors ported SISSO++ to GPUs with Kokkos, maintaining a single codebase and demonstrating significant speedups on different GPU architectures.

Findings

Achieved hardware portability with Kokkos for SISSO++

Demonstrated substantial speedups on Nvidia and AMD GPUs

Maintained a unified codebase for heterogeneous hardware

Abstract

SISSO (sure-independence screening and sparsifying operator) is an artificial intelligence (AI) method based on symbolic regression and compressed sensing widely used in materials science research. SISSO++ is its C++ implementation that employs MPI and OpenMP for parallelization, rendering it well-suited for high-performance computing (HPC) environments. As heterogeneous hardware becomes mainstream in the HPC and AI fields, we chose to port the SISSO++ code to GPUs using the Kokkos performance-portable library. Kokkos allows us to maintain a single codebase for both Nvidia and AMD GPUs, significantly reducing the maintenance effort. In this work, we summarize the necessary code changes we did to achieve hardware and performance portability. This is accompanied by performance benchmarks on Nvidia and AMD GPUs. We demonstrate the speedups obtained from using GPUs across the three most time-consuming parts of our code.