/ctint-science

Primary LanguageC++Apache License 2.0Apache-2.0

CTIN-SCIENCE

Build Status

Snap Status

Get it from the Snap Store

Notes

This program reproduces the DMFT results that can be found in the following science article: http://science.sciencemag.org/content/early/2018/12/06/science.aat4134.abstract

It is also available on the arxiv:https://arxiv.org/abs/1802.09456 .

Obtaining the full version

This program is a heavily toned down version. Indeed, this program is not optimized for speed, nor is generic or parallel. The main reasons being ease of use, installation and keeping competitive advantages. Please contact the developper to discuss obtaining the fully optimized, generic and parallel version.

Installation

Preferred way: Snap package

The preferred and easiest way to install is to use the snap package. Follow the instructions to install snap capability here: Snap Installation. The snap package can be obtained by clicking the snap store button at the top.

Manual installation

Note : If build problems, please remove the build directory if it exists, then retry :

$ rm -rf build

Dependencies

  1. Armadillo
  2. boost (serialization, filesystem, system)

Pre-Steps

  1. Make sure you have a "bin" directory in your home folder
  2. Append the bin folder to your path. Add the following line to your /.bashrc: export PATH="$PATH:/bin"
  3. $ source ~/.bashrc

Mac (Tested on macOS 10.13.6)

  1. Install the Dependencies (with Homebrew : https://brew.sh/)

    • $ brew install armadillo
    • $ brew install boost

  2. $ mkdir build && cd build && cmake .. && make install

Linux (Ubuntu 16.04 and 18.04)

This installation procedure should work for many recent Linux flavors. For the following we present the instructions specific for Ubuntu or derivatives.

  1. Install the Dependencies $ sudo apt-get install libarmadillo-dev libboost-all-dev cmake liblapack-dev
  2. | $ mkdir build && cd build && cmake .. && make install

Example

  1. Go to the examples/U3_b10 directory:

     $ cd examples/U3_b10
 $ bash run_dmft.sh

  2. The important output is the selfenergy, given at each dmft iteration, by self+{$ITER}.dat .
    The columns are the matsubara frequencies, the real and imaginary parts of the self-energy.
    The parameter files is given by params{$ITER}.json, which is in big part self-explainatory.

Parameters

We use json as the parameter file. Please keep the same structure and the same names (with case) as in the examples.

  • SEED: The seed for the random number generator
  • beta: Inverse temperature
  • mu : Chemical potentiel
  • U : Hubbard Interaction
  • NMAT : The number of matsubara frequencies
  • NTAU : The discretization in imaginary time. Should be at least 1000
  • UPDATESMEAS : The number of updates bewteen each measure
  • THERMALIZATION : The number of UPDATESMEAS for the thermalization, i.e , do UPDATESMEAS*THERMALIZATION updates before measuring
  • TOTALNMEAS : The number of measures
  • CLEANUPDATE : The number of UPDATESMEAS before a cleanupdate
  • delta : Value for the auxiliary Ising Spins
  • HybFile: the name of the hybridization file for the current iteration

Used third-party tools

Contact and help

To get help, please leave an issue or contact me by email at charles-david.hebert@usherbrooke.ca

Authors:

Charles-David Hébert.

Acknowledgments

Maxime charlebois and Patrick Sémon have inspired some portions of the code.