2. Installation and basic settings¶
- CLUPAN is available only for LINUX, not for Windows and Mac.
- Language : C++
- Licence : GPL v2
- Multi thread calculations using OPENMP
- If you find any bugs, contact Atsuto Seko by e-mail, seko at cms.mtl.kyoto-u.ac.jp.
- If you use CLUPAN for publishing academic papers, cite the paper of [CE1] .
2.1. Download¶
Latest version of CLUPAN can be downloaded from the following web site.
Latest version can also be downloaded using a subversion support in Sourceforge.
svn co https://clupan.svn.sourceforge.net/svnroot/clupan clupan
Source codes can be syncronized with the latest version of CLUPAN by typing
svn update
2.2. Required libraries¶
- BOOST C++ libraries (http://www.boost.org)
- GSL (GNU Scientific Library, http://www.gnu.org/software/gsl)
- Spglib (written by Atsushi Togo, http://sourceforge.net/projects/spglib)
On ubuntu OS, BOOST and GSL can be installed using apt.
aptitude install libboost-dev libgsl0-dev
For multi thread calculations, two kinds of additional libraries are required.
- Intel TBB (http://www.threadingbuildingblocks.org/)
- OPENMP
On ubuntu OS, Intel TBB can be installed using apt.
aptitude install libtbb-dev
2.3. Compilation¶
Extracting clupan.tar.gz file
tar xvfz clupan-0.7.tar.gz
compilation of Spglib
preparation of libsymspg.a (http://sourceforge.net/projects/spglib)
cd clupan-0.7/spglib ./configure make cp ./src/.libs/libsymspg.a .
compilation of CLUPAN
For single thread calculations, use makefile.intel or makefile.g++ as makefile. For multi-thread calculations, use makefile.mt_intel or makefile.mt_g++ as makefile. When compling CLUPAN for multi-thread calculations using g++, use a newer version of g++ than v4.2. When compling CLUPAN using the Intel compiler for multi-thread calculations, type the following commands.
cd clupan-0.7 cp makefile.mt_intel makefile make
2.4. Basic settings¶
In order to change stacksize, type the following command in case of zsh
ulimit -s unlimited
If you use any codes for DFT calculations except for VASP code, you need to modify input files for atomic positions to VASP format.
The number of threads for multi-thread calculations can be changed by setting the environment variable OMP_NUM_THREADS. In case of bash or zsh, type the following command
export OMP_NUM_THREADS=2