A Precision Emulation Approach to the GPU Acceleration of Ab Initio Electronic Structure Calculations

2026-03-31

AI Summary

This work addresses the challenge of efficiently leveraging modern GPU resources for traditional high-precision FP64 HPC applications, such as ab initio electronic structure calculations. The authors propose a tunable INT8 emulation strategy that enables transparent acceleration of the LSMS application within the MuST suite without modifying the original source code. By integrating cache-coherent unified memory architecture with the SCILIB-Accel automated BLAS offloading tool, the approach transcends the limitations of conventional mixed-precision paradigms. It simultaneously enhances GPU hardware utilization while preserving numerical accuracy and computational performance, thereby demonstrating a viable pathway for harnessing AI-oriented hardware to accelerate scientific computing workloads.

Abstract

This study explores the use of INT8-based emulation for accelerating traditional FP64-based HPC workloads on modern GPU architectures. Through SCILIB-Accel automatic BLAS offload tool for cache-coherent Unified Memory Architecture, we emulate FP64 matrix multiplications in the LSMS CPU application in the MuST suite without code changes. We find that accuracy depends on both arithmetic precision and the properties of the operator, which can be dealt with through tunable precision emulation. Unlike traditional mixed-precision approaches, this method preserves original algorithms while optimizing hardware utilization. We showcase the potential of improving accuracy and performance at the same time. This work highlights the potential of AI-driven hardware to transform HPC, advocating for adaptive precision strategies in future scientific computing.

Problem

• GPU acceleration
• precision emulation
• ab initio electronic structure
• high-performance computing
• mixed-precision

Innovation

• precision emulation
• INT8 acceleration
• GPU computing
• adaptive precision
• HPC