KaiEDJ

KaiEDJ: Integrated DMFT-ED and MFT Framework for Correlated Magnetic Systems

[Contact]

• Hyeong Jun Lee -- hjuntaf (at) gmail.com
• Taekjung Kim -- tj.kim (at) kaist.ac.kr
• Hongkee yoon -- hongkeeyoon (at) kangwon.ac.kr

The KaiEDJ is maintained by KAIST-ELST Group (https://github.com/KAIST-ELST).

[Main Features]

• Dynamical mean-field theory (DMFT) calculation
  • Exact diagonalization (ED) solver
  • Support an interface to external solver packages
  • DFT+DMFT interface - compatible with multiple density functional theory (DFT) codes
• Magnetic force theorem (MFT) or magnetic force response theory calculation with correlated electronic structure.
• Spin-wave dispersion analysis

[Citation]

(Submitted) KaiEDJ: A program conducting dynamical mean-field theory and magnetic force theory calculation for correlated magnetic materials, Hyeong Jun Lee, Taek Jung Kim, Hongkee Yoon, Myung Joon Han

0. Preparation

• Download KaiEDJ.

  git clone https://github.com/KAIST-ELST/KaiEDJ.git
  

• Julia should be installed. (Download Julia here.) Julia doesn't require any configurations or compile-processes in the beginning. You will just download it and set its bin/ path to your PATH to use julia excutable.

  $ export PATH=<your-julia-full-path>/bin/:$PATH    # a shell-command example.
  export PATH=<your-julia-full-path>/bin/:$PATH      # a shell-script example such as in .bashrc or .bash_profile
  

  Version-test successfully done for

  • julia >= 1.8.5, <= 1.11.3

• Install a set of required packages supported by Julia by running setup.jl provided in our package.

  $ julia setup.jl
  

  This will install the following packages. Formatting LinearAlgebra SparseArrays DelimitedFiles FastGaussQuadrature BenchmarkTools Optimization OptimizationOptimJL Arpack ThreadedSparseCSR SparseMatricesCSR Plots TickTock TimerOutputs FFTW JSON Dierckx ImageFiltering DFTforge

• Export your PROJECT_PATH_KAIEDJ to your project-environment directory. (setup.jl will automatically generate the project-environment directory of envs/KEDJ.)

  $ export PROJECT_PATH_KAIEDJ=<your-KaiEDJ-full-path>/envs/KEDJ    # shell-command
  export PROJECT_PATH_KAIEDJ=<your-KaiEDJ-full-path>/envs/KEDJ      # shell-script such as .bashrc or .bash_profile
  

1. Files

• run.sh is a single running script for all processes including the installation and all of the example codes.
• setup.jl will install the required packages, activating a Julia-project-environment TOML directory.
• kaiedj.jl is a main script for running KaiEDJ within Julia.
• eddmft.jl is a script for running ED-DMFT within Julia.
• dmftmft.jl is a script for running DMFT+MFT within Julia.
• examples/ contains example codes of La2CuO4 for ED-DMFT, DMFT+MFT, Spinwave, and of Fe for (QMC-DMFT,) DMFT+MFT, Spinwave calculations.
• src/ contains source codes.
• envs/ is a temporary directory for Julia-project-environment TOML files. setup.jl will use this location and generate envs/KED/*.toml files.
• scripts/ contains scripts for data-converting and patches.
• hotfix/ contains a hotfix files.

2. Usage

Preparing the TOML script, e.g. example.TOML, you can run KaiEDJ as follow.

$ julia kaiedj.jl -T example.TOML

In order to activate the Julia-project-environment, you can add a flag as julia --project=<julia-project-environment path> ....

For tutorials, you can download in github.com/hjuntaf/KaiED_tutorials.git and run as follows.

$ julia tutorials/01_manybody_binary_basis.jl

Every tutorial code here contains include("../src/mybase.jl") with relative path to include mybase.jl and use KaiED. You will need to modify this path (../src/) when you want to run in a different directory/location.

Jupyter notebook

We also provide some tutorials within jupyter notebook (.ipynb). You can use the jupyter notebooks after an installation of IJulia as follows.

Installation

   _       _ _(_)_     |  Documentation: https://docs.julialang.org
  (_)     | (_) (_)    |
   _ _   _| |_  __ _   |  Type "?" for help, "]?" for Pkg help.
  | | | | | | |/ _` |  |
  | | |_| | | | (_| |  |  Version 1.9.3 (2023-08-24)
 _/ |\__'_|_|_|\__'_|  |  Official https://julialang.org/ release
|__/                   |

julia> using Pkg

julia> Pkg.add("IJulia")

or enter the Pkg REPL by pressing ] and install it :

(@v1.9) pkg> add IJulia

Run

$ jupyter notebook

3. Tutorials

We provide some tutorial codes in formats of .jl (julia source code) and .ipynb (jupyter notebook). You can see and download them in the following repo.: github.com/hjuntaf/KaiED_tutorials

Following basic concepts are included.

• Quantum many-body basis state manipulations
• Quantum many-body wavefunction manipulations
• Green functions
• Impurity models
• Tight-binding models
• Exact diagonalization method
• Dynamical mean-field theory

4. (Optional) External Packages

• Our interface supports DMFT calculations with external DMFT solver packages.
• Currently, it is compatible with two continuous-time quantum Monte Carlo (CT-QMC) solvers.

  • CT-QMC of COMSUITE package (https://github.com/comscope/comsuite)

  • CT-QMC of EDMFTF (http://hauleweb.rutgers.edu/tutorials/)

    CT-QMC of EDMFTF package installation

    Below is a very brief introduction to the installation of EDMFTF package. Therefore, it may not work in your computer environment. For a more successful installation, please visit the EDMFTF package site directly (http://hauleweb.rutgers.edu/tutorials/).

    Download and extract compressed EDMFTF package

    $ wget http://hauleweb.rutgers.edu/downloads/EDMFTF.tgz
    $ tar -zxvf EDMFTF.tgz
    

    Modify file configure.py to suit your computer's compiler environment

    $ vim EDMFTF-*/configure.py
    

    ,and run

    $ python setup.py
    

    If at least the following files are created in /bini folder, it is possible to run Jx_DMFT.

    ../bini/ctqmc
    ../bini/atom_d.py
    ../bini/maxent_run.py
    

    Finally, add the following lines to $ vi ~/.bashrc

    export WIEN_DMFT_ROOT=EDMFTF-installed folder/bini
    export PYTHONPATH=$PYTHONPATH:$WIEN_DMFT_ROOT
    export SCRATCH="."
    export PATH=$WIEN_DMFT_ROOT:$PATH
    

    CT-QMC(ComCTQMC) of COMSUITE package installation

    Below is a very brief introduction to the installation of the COMSUITE package. Therefore, it may not work in your computer environment. For a more successful installation, please visit the COMSUITE package site directly (https://github.com/comscope/comsuite).

    Download zip file from https://github.com/comscope/comsuite and unzip the compressed folder

    $ unzip comsuite-master.zip 
    

    Modify file arch.mk to suit your computer's compiler environment

    $ vim comsuite-master/arch.mk
    

    and, add the following two lines to $ vi ~/.bashrc

    export COMSUITE_BIN= comsuite-installed folder/comsuite-master/bin  
    export PATH=$COMSUITE_BIN:$PATH
    

    Finally,

    $ make all
    

    If the following files are created in /bin folder, you are ready to run Jx_DMFT.

    ./comsuite-master/bin/CTQMC  
    ./comsuite-master/bin/EVALSIM