Trillion-atom molecular dynamics simulations with ab initio accuracy

TL;DR

This paper reports a groundbreaking molecular dynamics simulation of 1.62 trillion atoms with ab initio accuracy, achieving unprecedented speed and scalability on supercomputers, bridging microscopic and mesoscale modeling.

Contribution

The work introduces a highly scalable, ultra-large-scale MD simulation framework using NEP, enabling quantum-accurate mesoscale modeling at an unprecedented scale.

Findings

Simulated 1.62 trillion atoms with ab initio accuracy.

Achieved 100x faster simulation speed than previous ML force fields.

Demonstrated 86.9% weak scaling efficiency up to 45,000 GPGPUs.

Abstract

Material properties are fundamentally dictated by multiscale phenomena, which often reach mesoscale in size. The {\mu}m mesoscale is also the size which can be observed directly under an optical microscope, bridging the atomistic microscopic description with the continuous model macroscopic world. In this work, we report an unprecedented molecular dynamics (MD) simulation comprising 1.62 trillion atoms. Utilizing the neuroevolution potential (NEP) framework, we attained ab initio accuracy on China's New-generation Intelligent Supercomputer. Our implementation achieves a time-to-solution (s/step/atom) 100 times faster than previous state-of-the-art machine learning force field simulations, and 1,000 times faster than the Gordon Bell Prize-winning application from six years ago. Furthermore, we demonstrate an 86.9% weak scaling efficiency from a single GPGPU to 45,000 GPGPUs. These results redefine atomistic simulation boundaries, enabling direct mesoscopic modeling with quantum-level precision.