Installation instructions and pre-workshop exercises

Install on Windows

The different program are foreseen to run on Linux/Mac. If you don't have Linux or Mac, the best solution is to install Linux (I would advice to install Ubuntu) in dual boot and follow the instructions for Linux. This video might help you to do so: ​http://video.google.com/videoplay?docid=-2369893842637434537 virtual machine are fine, but quite often the library readline is not present on the system. This package is not mandatory but allozs mg5 to have history and nice auto-completions. So you might consider to install that library first and the compile python2.6.

Install on Linux

• python 2.6/2.7
  MadGraph 5 requires python 2.6 or python 2.7. In order to check your version of python, you can type: python --version. Some of the repositories are still on python 2.5. So if updating python via your repository manager didn't work. you can download directly python from the following link: ​http://www.python.org/download/ and follow instructions.
• Madgraph5
  You will find madgraph5 package on the following page: ​https://launchpad.net/madgraph5. For this program, you just need to untar it. To check if mg5 is correctly install you directly try to run it by doing: ./bin/mg5 If you don't have a valid python version. It will fail directly.
• MadAnalysis
  This programs is not mandatory but allows to have partonic distributions done automatically. In order to install this program, you need to launch mg5 (./bin/mg5) and then type

     install MadAnalysis
  

Install on MacOs

1. Basic program
   • perl
   • bash
     Those programs are by default present in any Mac distribution, but just in case you should check that they are all installed in your computer (by doing perl --version).
   • gmake
     By default, makefile are not recognize on mac. In order to check, you can do 'make --version'. If it says: -bash: make: command not found then you should install it. The easiest is to install the xcode.
     1. For MacOs 10.5: ​https://connect.apple.com/cgi-bin/WebObjects/MemberSite.woa/wa/getSoftware?bundleID=20414
     2. For MacOs 10.6: ​http://connect.apple.com/cgi-bin/WebObjects/MemberSite.woa/wa/getSoftware?bundleID=20792
     3. For MacOs 10.7: ​http://itunes.apple.com/us/app/xcode/id448457090?mt=12
     In order to be able to download it, you will need to a apple developer account (which is free)
   • gfortran 4.x (To check the version do gfortran do: gfortran --version) This program is often not included in standard mac version. In order to download it, you can download it from:
     1. ​http://prdownloads.sourceforge.net/hpc/gcc-lion.tar.gz?download for MacOs 10.7
     2. ​http://prdownloads.sourceforge.net/hpc/gcc-snwleo-intel-bin.tar.gz?download for MacOs 10.6
     3. ​http://sourceforge.net/projects/hpc/files/hpc/gcc/gcc-leopard-intel-bin.tar.gz/download for MacOs 10.5

2. !MadGraph5
   • python
     1. The default python version of MacOs 10.6 is python 2.6 and is (in 99% of the case) fully working with Madgraph5, so nothing to install for you.
     2. The python version of MacOs 10.7 is python 2.7. And this one works perfectly, so nothing to install for you.
     3. For other version of MacOS, you will need to install python2.7 from the python web site: ​http://www.python.org/download/
   • Madgraph5
     You will find madgraph5 package on the following page: ​https://launchpad.net/madgraph5. For this program, you just need to untar it. To check if mg5 is correctly install you directly try to run it by doing: ./bin/mg5 If you don't have a valid python version. It will fail directly.
   • MadAnalysis
     This programs is not mandatory but allows to have partonic distributions done automatically. In order to install this program, you need to launch mg5 (./bin/mg5) and then type

        install MadAnalysis
     

Instructions for the tutorial exercise

• If you haven't ran the tutorial before the lectures please do it now. Please launch mg5 ./bin/mg5 and type the following command:

     tutorial
  

and follow the tutorial instructions on the screen.

• The exercice of this tutorial is to
  1. Use the model generatedby Feynrules
  2. generate $ p p \to U \bar U $
  3. generate some events
  4. pass the events to Pythia
  5. Restart the chain but this time by including all the decay chain in MG5.

Note that in order to have the correct cross-sections in the second cases or to have the correct branching ratio computed by Pythia in the first case, we need first to compute the width and the branching ratio of the non standard model particles.

• The model
  The simple way to have access to a model in MG5 is to put in the directory: MG5_DIR/models after that you can simply import it in MG5 by doing

     import model MODELNAME
  

  or

     import model MODELNAME --modelname
  

  The option --modelname tells MG5 to use the name of the particles defines in the UFO model, and not the usual MG5 conventions for the particles of the SM and the MSSM. In principle this shouldn't change anything in the model that you load. The first time that you import a new model, it is a good moment to test it! For this
  1. You can run MG5 in debug mode (./bin/mg5 --debug)
  2. You can test the Lorentz invariance/Gauge invariance/... of at least a couple of processes. This is one examples:

        import model MODELNAME
        check p p > a a 
     

• Generation of events (Decay done in pythia)

     import model MODELNAME
     generate p p  > a a
     output
     launch
  

  One of the questions will be to ask you if you want to edit the param_card. At this stage, you can enter the path of the previously created param_card.dat.

• Generation of events (Decay done in MG5) Add the decay in MG5 (such that the full spin correlation is taken into-account) and generate events. In order to learn the syntax of the decay chain. Please type

      help generate
  

  Solution in attachments.