Skip to content

Mathematical & chemical routines e.g. nonlinear optimization etc. with c++ & python frontend

License

Notifications You must be signed in to change notification settings

YifanShenSZ/Fortran-Library

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Fortran-Library

Mathematical & chemical routines e.g. nonlinear optimization etc. with c++ & python interface

Featured utilities

Geometry transformation:

  • Standard geometry (also called standard orientaion)
  • Assimilate a geometry to a reference
  • Cartesian <-> internal coordinate
  • Normal mode and vibrational frequency

Nonlinear optimization: (all parameters conveniently tunable)

  • Unconstrained optimizer: Newton-Raphson, BFGS, limited memory BFGS, conjugate gradient (Dai-Yuan and Polak-Ribiere+), trust region
  • Constrained optimizer: augmented Lagrangian

Integral transform:

  • Fourier transform
  • Fast Fourier transform

Clustering:

  • K-means
  • Gaussian mixture model

Statistics:

  • Variance, R^2
  • Distribution

Chemistry:

  • Gradient of eigenstates
  • Phase fixing
  • Conical intersection adapted coordinate

StringUtility:

General:

  • Random number
  • Sorting
  • Some other basic routines

Mathematics:

  • Selected special functions and combinatorics
  • Quaternion operation
  • Ordinary differential equation
  • 1 dimensional integration

Linear algebra:

  • All kinds of vector & matrix & tensor operation
  • LAPACK wrapper for linear solver, eigensystem, matrix norm

Installation

libFL.a, libFL.so, *.mod, FortranLibrary.hpp, FortranLibrary (python package directory) will be installed to your prefix (default = Fortran-Library)

  1. export LIBRARY_PATH=prefix/lib:$LIBRARY_PATH
  2. export LD_LIBRARY_PATH=prefix/lib:$LD_LIBRARY_PATH
  3. export CPATH=prefix/include:$CPATH
  4. export PYTHONPATH=prefix:$PYTHONPATH
  5. Copy mkl_rci.f90 & mkl_dfti.f90 from MKL installation path to source (for gnu compiler, additionally modify DJACOBI in mkl_rci.f90 to specify the dimension of x, fjac, f explicitly: x(n), fjac(m,n), f(m))
  6. make, make install
  7. (optional) make test, then check the logs

Optional flags to tune make:

  • prefix=
  • compiler= (can be intel or gnu)
  • intelflag=
  • gnuflag=

Usage

  • fortran use FortranLibrary
  • c++ #include <FortranLibrary.hpp>
  • python import FortranLibrary as FL

Source

To see what this library is capable of in detail, you may open certain source file then fold all: routines are categorized and folded into different sections (VS code is recommended: press ctrl+k+0)

Source code level from bottom to top:

  1. StringUtility.f90, General.f90, Mathematics.f90, LinearAlgebra.f90
  2. NonlinearOptimization.f90, IntegralTransform.f90, Clustering.f90, Statistics.f90, Chemistry.f90
  3. GeometryTransformation.f90
  4. FortranLibrary.f90

Dependency

  • MKL
  • mkl_rci.f90 & mkl_dfti.f90 (can be found in MKL installation path)

Publications based on this work

  1. Y. Shen and D. R. Yarkony, J. Phys. Chem. A 2020, 124, 22, 4539–4548 https://doi.org/10.1021/acs.jpca.0c02763
  2. Y. Shen and D. R. Yarkony, J. Phys. Chem. Lett. 2020, 11, 17, 7245–7252 https://doi.org/10.1021/acs.jpclett.0c02199

Reference

  1. J. Nocedal, S. J. Wright, Numerical Optimization 2nd edition (Springer, 2006)
  2. E. B. Wilson, J. C. Decius, P. C. Cross, Molecular viobrations: the theory of infrared and Raman vibrational spectra (Dover, 1980)
  3. D. R. Yarkony, J. Chem. Phys. 112, 2111 (2000)

About

Mathematical & chemical routines e.g. nonlinear optimization etc. with c++ & python frontend

Resources

License

Stars

Watchers

Forks

Packages

No packages published