Delay Update in determinant quantum Monte Carlo

TL;DR

This paper introduces a generalized delay update scheme for determinant quantum Monte Carlo that enhances efficiency and scalability in simulating strongly correlated electron systems.

Contribution

It proposes a novel delay update method applicable to both onsite and extended interactions in DQMC, improving computational efficiency.

Findings

Significantly improves DQMC simulation efficiency.

Enables simulation of larger system sizes.

Applicable to both zero-temperature and finite-temperature DQMC.

Abstract

Determinant quantum Monte Carlo (DQMC) is a widely used unbiased numerical method for simulating strongly correlated electron systems. However, the update process in DQMC is often a bottleneck for its efficiency. To address this issue, we propose a generalized delay update scheme that can handle both onsite and extended interactions. Our delay update scheme can be implemented in both zero-temperature and finite-temperature versions of DQMC. We apply the delay update scheme to various strongly correlated electron models and evaluate its efficiency under different conditions. Our results demonstrate that the proposed delay update scheme significantly improves the efficiency of DQMC simulations, enable it to simulate larger system size.