Installation of CyGutz

Prerequisites

CyGutz consists of programs, executables, and scripts, written in Fortran90, c (c++) and Python2.7. Before you start the installation, you must make sure that the following packages are installed in your system.

A very convenient way to install python and the related packages is Anaconda (https://www.continuum.io/downloads).

Build and install

The compilation ans installation consist of three simple steps:

  • Download CyGutz here at https://github.com/yaoyongxin/CyGutz, unzip it and cd to that directory (CyGutz-master). You may simply type:

    $ wget https://codeload.github.com/yaoyongxin/CyGutz/zip/master
    $ unzip master
    $ cd CyGutz-master
    
  • Choose a Makefile.in file in directory template_makefile.in. Check and make sure the path to compilers, their options, and libraries are correctly. For example, you may type:

    $ cp template_makefile.in/Makefile.in.intelmpi_ifort_hal2011 Makefile.in
    $ vi ./Makefile.in
    
  • Run the following command at the top directory to install it:

    $ # Optional, clean any previous installation.
    $ make clean && make clean_bin
    $ ./install.sh
    $ # Get the environment variable ${WIEN_GUTZ_ROOT2}
    $ # for the first-time installation.
    $ source ~/.bashrc
    

Explanations on the scripts and Makefile.in

  • By default, ${WIEN_GUTZ_ROOT2}=${HOME}/WIEN_GUTZ/bin2 will be set in your .bashrc file.
  • Key executables, such as CyGutz, exe_spci, and python scripts are located at ${WIEN_GUTZ_ROOT2}.
  • In Makefile.in file:
    • Use sequential version of mkl, so set MKL_LIB = -mkl=sequential in Makefile.in.
    • The base path of the hdf5 library has to be set correctly. For instance, HDF5_BASE = /home/ykent/OPT/lib/hdf5-1.8.15-intel/