Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

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

Hang Liu, Junjie Li, Yinzhi Wang, Niraj K. Nepal, Yang Wang

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.

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

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.

Paper Structure

This paper contains 9 sections, 1 figure, 1 table.

Table of Contents

  1. Introduction
  2. Related Work
  3. Previous Work on Automatic BLAS Offload
  4. GEMM Emulation on Integer Matrix Multiplication Unit
  5. Experiment
  6. Implementation
  7. Experiments with MuST
  8. Discussion
  9. Conclusion

Figures (1)

  • Figure 1: Impact of mixed-precision emulation on the LSMS method using cuda13 and GEMMul8. The $G(z)$ counts all energy points and atomic sites, requiring invertion of a $33,750 \times 33,750$ double complex matrix, representing a significant computational challenge for standard FP64 solvers.