CPLEX
CPLEX is software for optimization, developed by IBM. It is available for academic users through IBM's academic program.
Download
To use CPLEX on Compute Canada clusters, you must first register with the IBM academic program and then download a personal version of the software.
Installation
The file is an executable archive which will perform the installation after you have answered a few questions. To execute the archive, you need to type the command bash ./cplex_studioXYZ.linux-x86.bin.
To access the software, you can create a personal module. Modules are normally created and placed in a directory hierarchy. In order for your modules to be found, you need to modify the configuration file $HOME/.bashrc so that it points to the root of this hierarchy, by adding the following line:
[name@server ~]$ export MODULEPATH=$HOME/modulefiles:$MODULEPATH
Next, you need to create a directory structure in which to put your new cplex module:
[name@server ~]$ mkdir -p $HOME/modulefiles/mycplex
In this directory, you should create a file (e.g. $HOME/modulefiles/mycplex/12.8.0) with the version number corresponding to the version you downloaded (e.g. 12.8.0) and with the following content:
#%Module1.0####
##
## cplex
##
proc ModulesHelp { } {
global cplexversion
puts stderr "\tIBM ILOG cplex "
puts stderr "\tThis module provides configuration for cplex, concert, cpoptimizer and opl"
module-whatis "IBM ILOG cplex (cplex, concert, cpoptimizer, opl). This version doesn't ask for a licence file."
# for Tcl script use only
set cplexversion XYZ
set studio_root <root path of cplex>
set cplexroot $studio_root/cplex
set concertroot $studio_root/concert
set oplroot $studio_root/opl
set cpoptimizerroot $studio_root/cpoptimizer
set cplexbin x86-64_linux
set cplexlib $cplexbin/static_pic
set concertbin x86-64_linux
set concertlib $concertbin/static_pic
set oplbin x86-64_linux
set opllib $oplbin/static_pic
set cpoptimizerbin x86-64_linux
set cpoptimizerlib $cpoptimizerbin/static_pic
prepend-path PATH $cplexroot/bin/$cplexbin
prepend-path PATH $oplroot/bin/$oplbin
prepend-path PATH $cpoptimizerroot/bin/$cpoptimizerbin
prepend-path CPATH $cplexroot/include
prepend-path CPATH $concertroot/include
prepend-path CPATH $oplroot/include
prepend-path CPATH $cpoptimizerroot/include
prepend-path -d " " CPATH_EXPANDED -I$cplexroot/include
prepend-path -d " " CPATH_EXPANDED -I$concertroot/include
prepend-path -d " " CPATH_EXPANDED -I$oplroot/include
prepend-path -d " " CPATH_EXPANDED -I$cpoptimizerroot/include
prepend-path LIBRARY_PATH $cplexroot/lib/$cplexlib
prepend-path LIBRARY_PATH $concertroot/lib/$concertlib
prepend-path LIBRARY_PATH $oplroot/lib/$opllib
prepend-path LIBRARY_PATH $oplroot/bin/x86-64_linux/
prepend-path LIBRARY_PATH $cpoptimizerroot/lib/$cpoptimizerlib
prepend-path -d " " LIBRARY_PATH_EXPANDED -L$cplexroot/lib/$cplexlib
prepend-path -d " " LIBRARY_PATH_EXPANDED -L$concertroot/lib/$concertlib
prepend-path -d " " LIBRARY_PATH_EXPANDED -L$oplroot/lib/$opllib
prepend-path -d " " LIBRARY_PATH_EXPANDED -L$oplroot/bin/x86-64_linux/
prepend-path -d " " LIBRARY_PATH_EXPANDED -L$cpoptimizerroot/lib/$cpoptimizerlib
prepend-path LD_LIBRARY_PATH $cplexroot/bin/$cplexbin
prepend-path LD_LIBRARY_PATH $oplroot/bin/$oplbin
prepend-path CLASSPATH $cplexroot/lib/cplex.jar
prepend-path MATLABPATH $cplexroot/matlab
Adjust as necessary the lines corresponding to the variables cplexversion and studio_root so that they have the right values: the version number downloaded and the path to the software, i.e. the one you chose during the installation.
Java
If you use Java, you will have some further steps to carry out. Firstly, in your .bashrc file, you can add the line
[nom@serveur ~]$ export CLASSPATH=.
which will allow your code to be found during execution.
Next, you should modify the dynamic library of cplex. Look for this library in the directory hierarchy of the installation directory, make a copy of it and then execute the command
[name@server ~]$ setrpaths.sh --path libcplex1280.so
It's possible that during your compilation, you receive an error message because of a lack of memory. In this case, you should request a compute node using an interactive job to do the compilation. For example,
[name@server ~]$ salloc --time=1:0:0 --ntasks=1 --mem-per-cpu=8G