Incremental solver for orbital‐free density functional theory

Abstract

First-principle calculations are still a challenge since they require a great amount of computational time. In this article, we introduce a new algorithm to perform orbital-free density functional theory (OF-DFT) calculations. Our new algorithm focuses computational efforts on important parts of the particle system, which, in the context of adaptively restrained particle simulations (ARPS) allows us to accelerate particle simulations. MethodologyBefore we describe our method, we first introduce the molecular dynamics in which simulations will take place, ARPS, and then recall the main ideas and equations behind the OF-DFT.Adaptively-restrained particle simulations ARPS freeze the slowest particle movements so that the number of moved particles at each time step decreases.Briefly, ARPS manage the behavior of particles by using their kinetic energy K and 2 thresholds: ε r , the fully-restrained threshold, and ε f , the full-dynamics threshold. If K ≤ ε r , the particle is 1 3 . Different implementations of OF-DFT computation have been tested, either in Fourier space [2] or in real space with finite elements [32] or with finite differences. [8] Fastest implementations, like PROFESS, use Fourier space. This is because the computational bottleneck of real-space WWW.C-CHEM.ORG

Orbital-Free Density Functional Theory: An Attractive Electronic Structure Method for Large-Scale First-Principles Simulations

“…Lastly, because FD methods represent the electron density in real space, they are well-suited to aid AIMD accelerators. This was described in detail in ref 526, where a simple FD method was employed to track the electron density and its associated potential with free-space (open) boundary conditions in adaptively restrained particle simulations (ARPSs). An ARPS decomposes the system into slow-and fast-moving particles.…”