• Search Research Projects
  • Search Researchers
  • How to Use
  1. Back to project page

2019 Fiscal Year Final Research Report

Development of the hybrid density functional theory on the real-space grid method and its implementation on massively parallel computers

Research Project

  • PDF
Project/Area Number 17K05138
Research Category

Grant-in-Aid for Scientific Research (C)

Allocation TypeMulti-year Fund
Section一般
Research Field Computational science
Research InstitutionTohoku University

Principal Investigator

Takahashi Hideaki  東北大学, 理学研究科, 准教授 (10291436)

Project Period (FY) 2017-04-01 – 2020-03-31
Keywords超並列計算 / 実空間グリッド法 / 密度汎関数法 / Hartree-Fock交換エネルギー / FFT / CG
Outline of Final Research Achievements

In this subject we developed two types of implementation of the Hartree-Fock(HF) exchange energy using the real-space grid approach to achieve high efficiency in the parallel execution of the hybrid exchange functional in the density functional theory. We first reduce the problem of calculating the HF exchange energy to the solution of the Poisson equation presented on the discrete real-space grids. Then, the Poisson equations for the electrostatic potentials of the products of the orbital pairs were solved iteratively through the conjugate gradient (CG) method where the operation of Laplacian was parallelized by domain decomposition scheme. In the other approach, we adapted the three-dimensional fast Fourier transform (FFT), referred to as PFFT, to solve the Poisson equations.
It was shown in the benchmark tests that the parallel execution with the FFT approach is faster than that using the CG method because a larger bandwidth can be made available in the parallel execution of FFT.

Free Research Field

量子化学

Academic Significance and Societal Importance of the Research Achievements

密度汎関数法(DFT)は化学や物理の理論の分野で必須のツールであり、これを並列計算によって高速化することは、DFT法の応用範囲を拡張する上で重要である。DFT計算の精度を向上させる為には、Hartree-Fockの交換エネルギーを部分的に含む、ハイブリッド汎関数の使用が標準となっている。実空間グリッド基底はDFTの並列計算を高速化する上で極めて有効であることが実証されているが、HF交換ポテンシャルのような非局所な演算子の実行には不利である。本研究の遂行によって実空間グリッド基底によるハイブリッド汎関数の実行が高速化され、ひいては、機能性分子の合成や反応のメカニズムの解明に資する。

URL: 

Published: 2021-02-19  

Information User Guide FAQ News Terms of Use Attribution of KAKENHI

Powered by NII kakenhi