Skip to content

Latest commit

 

History

History
137 lines (100 loc) · 4.78 KB

INSTALL.md

File metadata and controls

137 lines (100 loc) · 4.78 KB
Raw

Local installation guide

About

This installation guide covers what are the necessary requirements for executing individual steps of the physics workflow without the need of Docker.

Requirements

To have a local environment able of generating and showering events, there are multiple requirements that we need to have installed and setup in our system.

In terms of software:

In terms of environment:

  • ROOTSYS environment var defined.
  • PATH environment var updated with ROOT binaries path.
  • LD_LIBRARY_PATH environment var defined (Linux).
  • DYLD_LIBRARY_PATH environment var defined (Mac OS).

Installation

This guide covers the installation of ROOT and its environment variables in a distinct way for macOS and linux distributions. Once it is correctly set up, the installation of MadGraph 5, Pythia 8 and Delphes is the same in both architectures.

1. Install ROOT

1.1 Mac OS

Since macOS El Capitan (2015), all macOS devices have a System Integrity Protection (SIP) activated by default. It protects users to modify certain parts of the OS environment. Follow this guide to disable it.

Once it is disabled, install ROOT:

brew install root

The compilation process take around 45 minutes. Once it has finished, brew would have installed ROOT at /usr/local/Cellar/root.

1.2 Linux

For Linux distributions the installation of ROOT is more manual. Follow this guide on how to download and compile ROOT locally.

2. Setup environment

To set up the environment for ROOT, multiple variables need to be defined. These variables will be useful later on in the process as they will tell Madgraph 5 packages where to find ROOT in your local system.

2.1 Mac OS

Modify your shell source file (.bashrc / .zshrc / ...) and add:

export ROOTSYS="/usr/local/Cellar/root"
export PATH="$ROOTSYS/bin:$PATH"
export DYLD_LIBRARY_PATH="$DYLD_LIBRARY_PATH:$ROOTSYS/lib"

2.2 Linux

Modify your shell source file (.bashrc / .zshrc / ...) and add:

export ROOTSYS=<path_to_root_folder>
export PATH="$ROOTSYS/bin:$PATH"
export LD_LIBRARY_PATH="$LD_LIBRARY_PATH:$ROOTSYS/lib"

3. Install MadGraph 5

MadGraph is a software tool to install other packages. It can be downloaded to any path of the local system, although it is recommended to have it in a folder called software at the root folder of this project.

To download MadGraph 5:

mkdir -p software
curl -sSL "https://launchpad.net/mg5amcnlo/lts/2.9.x/+download/MG5_aMC_v2.9.4.tar.gz" | tar -xz -C software

4. Install Pythia and Delphes

Pythia and Delphes are installed using MadGraph 5 as some sort of "package manager". Bear in mind, however, MadGraph asks for an upgrade the first time you launch it.

# For macOS 11.0+ systems
export SYSTEM_VERSION_COMPAT=1

echo "install pythia8" | python3 software/MG5_aMC_v2_9_4/bin/mg5_aMC
echo "install Delphes" | python3 software/MG5_aMC_v2_9_4/bin/mg5_aMC

5. Import generation model

When running MadGraph 5, there are several generative models than could be imported. The one specified in both the signal and background process cards is called Weak boson effective field theory.

This model was originally defined in Python2, so it needs to be converted to Python3. Thankfully, there is a MadGraph option to automatically convert old models to Python3: auto_convert_model.

echo "set auto_convert_model T" | python3 software/MG5_aMC_v2_9_4/bin/mg5_aMC
echo "import model EWdim6-full" | python3 software/MG5_aMC_v2_9_4/bin/mg5_aMC

6. Install Numpy

Finally, as some MadGraph packages use Fortran code internally, an additional binary need to be installed. The binary is used to create Fortran to Python interfaces, it is called f2py, and it is distributed by Numpy.

Given that MadMiner already installs Numpy, there is nothing to do.