PyDPFT 1.0.6

A package for density potential functional theory

PyDPFT demo

A package for Density Potential Functional Theory introduced by Julian Schwinger and Berge Englert.

A project developed by Ding Ruiqi from National University of Singapore Physics Department for his bachelor thesis

Author: Ding Ruiqi

Date: 6 April 2020

Thesis: https://tesla-cat.github.io/about/Bachelor-Thesis.pdf

Presentation: https://github.com/tesla-cat/PyDPFT/Thesis/PPT

PyPI: https://pypi.org/project/PyDPFT/

Github: https://github.com/tesla-cat/PyDPFT

Run this notebook in Google Colab: https://colab.research.google.com/drive/1bC64gVM2WMk85MH2clTqDnBbaLhLCwTr

Highlights:

• multi-GPU acceleration
• optimized using Tensor operation (instead of for-loop)
• implemented from 1D to 3D with auto detection

import

!pip install PyDPFT
from PyDPFT import PyDPFT
from PyDPFT.plot import plot

Collecting PyDPFT
  Downloading https://files.pythonhosted.org/packages/b9/f4/c9989112b6c1beafbd7c592ebe07864dcae9d8d93f5878af1207187d3331/PyDPFT-0.0.7-py3-none-any.whl
Installing collected packages: PyDPFT
Successfully installed PyDPFT-0.0.7

1D Hartree interaction

config = {
    'space':{'x':[-5,5,100]},
    'loop':{'Imax':1000,'precision':1e-6,'mix':0.05},
    'const':{'epsilon':1e-3},
    'rho':{'N':32},
    'Vint':{'name':'Hartree','coef':1},
}

dpft = PyDPFT(config)
Vext = 1e3*dpft.x ** 2
Vx,Vint,rho,N = dpft(Vext)
plot(dpft,Vx,Vint,rho)

PyDPFT: Written by Ding Ruiqi from NUS for his bachelor thesis
PyDPFT: Detected dim = 1
PyDPFT: Using 1 GPUs !
PyDPFT: Starting the self consistent loop
PyDPFT: Converged after 223 iterations in 2.0902912616729736 seconds!

2D Dipole-dipole interaction in momentum (p) space

config = {
    'space':{'x':[-5,5,50],'y':[-5,5,50]},
    'loop':{'Imax':1000,'precision':1e-6,'mix':0.05},
    'const':{'epsilon':1e-2,'mu':[0.7, 0.7]},
    'rho':{'N':32},
    'Vint':{'name':'Dipole-p','coef':.1},
}

dpft = PyDPFT(config)
Vext = dpft.xx**2 + dpft.yy**2
Vx,Vint,rho,N = dpft(Vext)
plot(dpft,Vx,Vint,rho)

PyDPFT: Written by Ding Ruiqi from NUS for his bachelor thesis
PyDPFT: Detected dim = 2
PyDPFT: Using 1 GPUs !
PyDPFT: Starting the self consistent loop
PyDPFT: Converged after 219 iterations in 11.653722524642944 seconds!

3D Dipole-dipole interaction in position (x) space

config = {
    'space':{'x':[-5,5,20],'y':[-5,5,20],'z':[-5,5,20]},
    'loop':{'Imax':1000,'precision':1e-6,'mix':0.05},
    'const':{'epsilon':1e-2,'mu':[0.7, 0.7, 0]},
    'rho':{'N':32},
    'Vint':{'name':'Dipole-x','coef':5},
}

dpft = PyDPFT(config)
Vext = dpft.xx**2 + dpft.yy**2 + dpft.zz**2
Vx,Vint,rho,N = dpft(Vext)
plot(dpft,Vx,Vint,rho)

PyDPFT: Written by Ding Ruiqi from NUS for his bachelor thesis
PyDPFT: Detected dim = 3
PyDPFT: Using 1 GPUs !
PyDPFT: Starting the self consistent loop
PyDPFT: Converged after 142 iterations in 4.736774921417236 seconds!