QC-DMET: a python implementation of density matrix embedding theory for ab initio quantum chemistry

This is a modernized fork of sebwouters/qc-dmet, updated for Python 3.10+ and PySCF 2.0+.

Copyright (C) 2015 Sebastian Wouters sebastianwouters@gmail.com

This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version.

This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.

You should have received a copy of the GNU General Public License along with this program; if not, write to the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.

Building and testing

Environment setup (Python 3.10 + PySCF 2.0+):

> module load miniconda3/24.1.2-py310
> conda create -n qcdmet python=3.10 -y
> conda activate qcdmet
> conda install -y numpy scipy pyscf -c conda-forge

QC-DMET requires numpy, scipy, and pyscf. The ED/FCI solver uses PySCF's built-in pyscf.fci module. Block2 via pip is optional and only required for the DMRG solver. CheMPS2 is supported via the DMRG-CheMPS2 keyword for legacy DMRG.

The path to pyscf can be adjusted in qcdmet_paths.py.

Compile the C++ library libqcdmet.so (requires Intel compilers and MKL):

> module load intel/2021.10.0 intel-oneapi-mkl/2023.2.0
> cd lib
> mkdir build
> cd build
> CXX=icpc CC=icc cmake .. -DMKL=ON
> make
> cd ../..

Start from the files examples/*.py.

Performance testing

1. Find the most costly functions:

> python -m cProfile -o testx.profile testx.py
> python -m pstats testx.profile
>>> sort cumulative
>>> stats

2. Find what makes them most costly:

Place just before the function you want to profile @profile:

@profile
def construct1RDM_loc_response( self, doSCF, umat, list_H1 ):

And then use line_profiler:

> kernprof -lv testx.py