/NISTfit

A heavily abstracted C++-based kernel for fitting correlations

Primary LanguageC++OtherNOASSERTION

NISTfit

A heavily abstracted C++-based kernel for fitting correlations

Build Status

Why NISTfit?

  • In our work, we develop correlations between input properties and output properties. These correlations are sometimes complex multiparameter equations of state, and other times, simple polynomials. In any case, it was desired to have a generalized framework for carrying out these types of procedures in a generalized, fast(!), and flexible framework. Being open-source and cross-platform, C++ is an ideal language for all three goals.

Information

Usage

python wrapper can be built and installed with:

python setup.py install

Requirements:

  • Cmake
  • C++11 compliant compiler (MSVC 2015+ on windows, most recent versions of g++ or clang will work fine)
  • python (anaconda package is one good option, also includes plotting libraries (matplotlib) needed to run the timing tests)

Notes: to specify that you want the Intel compiler, (on linux) you can do:

CXX=icc python setup.py install

Alternatively, you can checkout and install in one fell swoop with:

pip install git+https://github.com/usnistgov/NISTfit.git

License

  • Public Domain. See LICENSE.txt

Credits