SIMONAPreprocessor
Installation Guide

Linux

No matter if you use the Windows or Linux versions of SIMONA you need limited knowledge of working on a shell. If you do not have any experience with Linux whatsoever, it is very recommended that you read a tutorial like this one: Linux Tutorial. Although this tutorial seems very long, you will see that most of the tutorials can be done in a few minutes.

Checking out the Code

To obtain the code you need an account on the Subversion system intfischer. After getting the Account, install subversion and type

1 svn co http://intfischer.ka.fzk.de/svn/poempp/trunk poempp

This will then download the code and put it into the directory poempp.

How to install SIMONA (C++)

  • SIMONA
    SIMONA needs a few programs:
    • The boost libraries
    • doxygen (for source code documentation)
    • scons (After checking out, SIMONA can be build using)
    • Now install SIMONA. Make sure you are in the directory, where SIMONA is located:
      1 scons
      Cleaned using
      1 scons -c
      The main building file is SConstruct - the source dir has another file named SConscript including all the cpp files to be built.
  • OpenCL with NVIDIA You only need to do this if you want to run SIMONA on your nvidia graphics accelerator. If you do not know, if you want it, don't do it.
    • Download and install NVIDIA CUDA Toolkit
    • Download and install mesa dev for your system.
    • set
      1 export CUDA_HOME
    • set
      1 USE_OPENCL=TRUE
    • 1 scons
Todo:
OpenCL with AMD

SIMONA Preprocessor

  • Download and install
    If you are using Linux, install everything with your Installation Manager, unless its written different in the documentation.
    • Python
    • lxml (if lxml does not work for you, try Amara. Please note that out Amara-backend is marked depreacted and will be removed soon.)
    • NumPy
    • SciPy
    • Python-QT4
    • BioPython
      if you install biopython, you can use BioPython in SIMONAgui in stead of POEM PDB engine.
    • If you get errors regarding sip, when you start the preprocessor, install python-sip. Change can be made in Settings, but this change is not recommended.

If you are using Ubuntu or Linux Mint, these packages can be searched with apt-cache search PACKAGENAME and installed with apt-get install PACKAGENAME or with Synaptics using the GUI. If you use SUSE please use Yast

We recommend adding simonagui.py and poempp to your PATH variable, so that you can just type poempp myxml.xml or simonagui.py without the correct path in front. Tutorial to add /home/yourname/bin to your path . Once you added a directory to your path use "ln -s /path/to/mysimonainstall/src/poempp poempp" inside your bin directory.

Installation with windows

It should be emphasized that it is NOT recommended using SIMONA under Windows. It is possible and supported, but the building process is harder as there are no package managers under Windows.
The supported compiler under Windows is MingW in a current version.

  • Download MingW with mingw-get MingW
  • Download manually and install: Python 2.x
  • Download manually and install: SCons
  • Add both Python and the Python/scripts folder to your path variable
  • If you need a commandline SVN client (recommended) download and install SlikSVN
  • From now on, work in your homefolder in msys:
  • Build zlib for example in the folder zlib-1.2.5
    1 cd zlib-1.2.5
    2 ./configure
    3 make -f win32/Makefile.gcc
  • Build <a href="http://www.boost.org/ target="_tab">Boost for example in the folder boost-1.44
    1 bootstap.sh --with-toolset=mingw
    2 edit project_config.xml and change every occurence of mingw with gcc (as of boost 1.48)
    3 b2.exe -sZLIB_SOURCE=../../../../zlib-1.2.5 -sZLIB_INCLUDE=../../../../zlib-1.2.5 -sNO_ZLIB=0 --with-random --with-iostreams --with-filesystem --with-serialization --with-system runtime-link=static release
  • If BOINC is needed, checkout the BOINC source and build it by entering the libdir and typing
    1 make -f Makefile.mingw
  • Afterwards SIMONA should build. Depending on the versions of boost and zlib, paths may need to be changed in the SConstruct

How to install scons on local clusters (C++)

If scons is not installed on a local cluster and you do not know the admin, who could install it for you:

  • 1. download it from: http://www.scons.org/ (chose recent production tarball)
  • 2. unzip it in a directory of your choice (tar xf scons-xxx.tar.gz)
  • 3. run python setup.py install –prefix=$HOME <- Scons will be installed in $HOME/bin
  • 4. make sure $HOME/bin is in your path (as described above).