Available support resources

Last update: 15 Nov 2023 [History] [Edit]

This tutorial provides an introduction to many aspects of an ATLAS analysis. However, it cannot cover all possible scenarios you may encounter in the course of your analysis work. There are numerous resources that you can use to seek help for any computing difficulties you have. This page summarizes the main resources available.

Who to Contact

The best group to contact depends on your issue. Below are several different categories of help you may need.

During the Tutorial

During the tutorial, the tutors are your first point of contact for help. If you are participating in an in-person tutorial, talk to the tutors in the room. If you are doing a remote/asynchronous tutorial, you can ask for help on the Discord server if one is provided, otherwise you can send an email to the tutors at atlas-sw-tutors-organisers@cern.ch. For any questions/issues not directly related to the tutorial material, please keep reading to find the appropriate channel to get help.

CERN Service Portal

Experts who are able to help with all questions and technical issues related to CERN computing resources. The service desk is the primary point of contact for most issues.

ATLAS Twiki Pages

The majority of ATLAS internal information is documented on twiki pages. Within the twiki pages, you will find details about technical implementations, analysis and combined performance recommendations, and contact information for conveners and other experts. This tutorial makes references and provides links to many twiki pages in relevant sections, but a good place to start is the main ATLAS page:

Some of the main pages that might be useful to you are:

Mattermost

Mattermost is a team communication service very similar in functionality to Slack. This is supported at CERN and can be accessed by the Mattermost App or through the CERN Mattermost page. You should have access to the service by default if you have a CERN account.

CERN Mattermost has several “teams” you can join, each of which has multiple channels. A good place for software questions is the atlassoftware team.

Analysis Model Group

The Analysis Model Group (AMG) is responsible for developing and overseeing analysis frameworks and tools. This tutorial is under the purview of AMG. The main AMG twiki page and three useful mailing lists are:

Physics Modeling Group

The Physics Modeling Group (PMG) is responsible for the development and validation of Monte Carlo (MC) simulation of physics processes. The simulation of physics processes that are specific to an analysis are the responsibility of the analyzers or the physics group, but PMG is responsible for giving the final approval for all production requests. Contact information is available on the PMG twiki for experts in all aspects of MC generation and a central mailing list is available for general questions.

Statistics

The Statistics Forum is responsible for all recommendations and support related to statistical analysis. They provide support for numerous commonly used statistical tools and are available to help with any analysis-specific questions you may have. Contact information for experts for specific tools are available on the twiki page and there is a central statistics help mailing list for any questions.

PanDA

The Production and Distributed Analysis (PanDA) system is used to process computing jobs on the grid. It is covered in detail in this tutorial. Central PanDA documentation is available.

CERN Batch Service

CERN provides access to the HTCondor batch system (replacing the previous LXBatch system). The system’s usage will be presented in the tutorial. More details can be found on the following page:

What To Include

When contacting the mailing lists please be as detailed as possible. When asking for software help, you will never be accused of providing too much detail.

Include things such as:

  • A minimal example that reproduces the problem
  • Log files
  • Links to grid jobs
  • Exact command(s) used
  • Your source code (via GitLab)
  • Input files

If log files or input files are large, share them via CERNBox.