Accelerating Ab Initio Nuclear Physics Calculations with GPUs

TL;DR

This paper demonstrates how GPU acceleration significantly speeds up ab initio nuclear structure calculations, particularly in matrix construction, by modifying existing software to leverage GPU computing resources.

Contribution

The paper introduces GPU acceleration techniques for the MFDn nuclear structure eigensolver, achieving notable speedups in matrix construction and overall computation.

Findings

Speedup of 2.2x - 2.7x in matrix construction

Speedup of 1.2x - 1.4x for entire calculations

Effective GPU integration in nuclear physics software

Abstract

This paper describes some applications of GPU acceleration in ab initio nuclear structure calculations. Specifically, we discuss GPU acceleration of the software package MFDn, a parallel nuclear structure eigensolver. We modify the matrix construction stage to run partly on the GPU. On the Titan supercomputer at the Oak Ridge Leadership Computing Facility, this produces a speedup of approximately 2.2x - 2.7x for the matrix construction stage and 1.2x - 1.4x for the entire run.