Enabling Multireference Calculations on Multi-Metallic Systems with Graphic Processing Units

TL;DR

This paper presents a GPU-accelerated implementation of multireference wave function calculations for multimetallic systems, significantly reducing computation time and enabling larger, more complex chemical simulations.

Contribution

It introduces a multi-GPU implementation of LASSCF calculations with density fitting, achieving substantial speedups and demonstrating scalability on different GPU architectures.

Findings

Achieved 5-10x speedup in LASSCF calculations

Successfully utilized up to four NVIDIA A100 GPUs

Demonstrated performance portability across GPU architectures

Abstract

Modeling multimetallic systems efficiently enables faster prediction of desirable chemical properties and design of new materials. This work describes an initial implementation for performing multireference wave function method localized active space self-consistent field (LASSCF) calculations through the use of multiple graphics processing units (GPUs) to accelerate time-to-solution. Density fitting is leveraged to reduce memory requirements, and we demonstrate the ability to fully utilize multi-GPU compute nodes. Performance improvements of 5-10x in total application runtime were observed in LASSCF calculations for multimetallic catalyst systems up to 1200 AOs and an active space of (22e,40o) using up to four NVIDIA A100 GPUs. Written with performance portability in mind, comparable performance is also observed in early runs on the Aurora exascale system using Intel Max Series GPUs.