ComputingTutorials

Introduction

This page is intended for physicists who are just starting to work in the mu2e computing environment. The following is a broad overview of the major components, their terminology, and how you might use them. Following each into paragraph, here are links into more specific tutorials and the rest of the mu2e documentation which is more terse and is intended as a reference once you get through the introductory material.

You probably don't have to work through this entire page and you can stop and any point, please talk to you adviser or mentor to see what's appropriate. The material you are most likely need to use comes first, followed by more in-depth tutorials for people who will be spending years on mu2e and learn to do more complex work.

Prerequisites

Have you completed the Mu2e Day 1 CheckList? If not, please do so.

If you have not already done so, please sign up for these slack channels:

tutorial-questions - ask questions about the tutorials here
tutorial-bug-reports - if you find ambiguities or errors, report them here.
computing_and_software - the annoucements and general discussion list
is_it_me_or_a_bug - ask questions and about anything except the tutorials.

If you find ambiguities or errors in these tutorials, please report them to the tutorial-bug-reports slack channel.

In this tutorial we will assume you are familiar with the topics below. You don't need to master this material before you start but we recommend that you skim it and learn where to find information:

the Mu2e detector, the parts, what they do, and how they work
HEP offline concepts
bash shell commands
- Beginner Cheat Sheet
- Linux FAQ - includes links to suggested references
The basics of C++ code concepts and syntax

When you do these exercises you will start on your local computer and log in to one of the Mu2e interactive machines, which are sometimes called the "central machines". The software you will use is on these machines and you will do your work on these machines. You will be working at a terminal window and typing commands at a prompt. The editors available on these machines are: vi, vim, emacs and nedit. If this way is working is not familiar to you, contact a colleague to help get you started.

A Convention You will See

You will sometimes see an instruction that says something like. "To check that this step worked correctly, do"

> echo $PRODUCTS
/cvmfs/fermilab.opensciencegrid.org/products/common/db

The right arrow character represents the shell prompt and you should type everything to the right of that character at the prompt in your working shell. The expected output of that command is known on the following lines. There may be many lines of output. If the output that you see is that same as shown in the example, then everything is working. If the example output contains a version number or a timestamp, it's OK if your output has different values, unless, of course, the example is there explicitly to check the version number.

Login

We have 6 machines that you can use for the tutorials: mu2egpvm01 ... mu2egpvm06. For purposes of this tutorial these machines are identical and they all mount the same disks, including your home disk. You can start on one machine and restart on another. Pick one and log in. Last I checked Fermilab security policy broke the automatic load balancer so we need to load balance by hand. If your machine is slow for more than a few minutes, try another.

You login to the mu2e interactive machines using your kerberos principal, "your_username@FNAL.GOV". If you need a refresher on kerberos vs SSO, revisit the Day 1 Checklist.

To login, follow the instructions at: LoginTutorial.
If your find ambiguities or errors in the above, let us know on tutorial-questions channel.

Interactive logins

Collaborators can do interactive computing work in several places. Probably the best place to start is the collaboration's interactive linux machines at Fermilab. The disks that the user sees are located on specialized disk server hardware and the same disks are mounted by all the interactive machines. There are six quad-core machines named mu2egpvm01.fnal.gov through mu2egpvm06.fnal.gov, and a code build machine called mu2ebuild01. You will have an account and a home area here (the same home area is on all machines) and some disk space for data. We prefer using the bash shell for all purposes. Collaborators can also compile and run mu2e code on their linux desktops or laptops.

Try:

LoginTutorial checking your access

Become familiar with these references:

Ntuple

The data from the detector, and the reconstructed data, is stored in files in art format. Accessing this data generally requires compiling code and learning a special configuration language, so we will save that for a later tutorial. To simplify, and speed up access to the data, we often run a program to copy a small part of the data in the art file into a convenient format call a root ntuple (pronounced "en-tuple"). This format is easy to browser interactively and make histograms. The ntuple file may contain histograms that were already made, or a list of the tracks in each event along with interesting quantities, such as the number of hits on the track or its reconstructed momentum.

Tutorial:

NtupleTutorial is the tutorial for this section and will guide you through making plots with one of the Mu2e-specific ntuples that are available.

Other useful pages:

ROOT is a very useful resource. In particular: getting started and code reference
Overview of existing mu2e ntuples

Geometry Browser

It is often useful to look at the detector as it is implemented in the simulation either to debug (e.g. double-check the geometry is as you expect) or to get images for presentations.

Tutorial:

Related references:

Code, art and fcl

The main program that is used for simulation, reconstruction and analysis is Mu2e Offline. This is an art-based framework in which the data passes through a series of modules to perform a variety of tasks. Sometimes you will need to build the full Mu2e Offline framework (e.g. if you are a developer) but in other cases you might only need a partial build or use an already existing build.

Tutorial:

The CodeArtFclTutorial
- CodeEnvironment runs you through setting up the environment
- ReleaseList gives the list of releases that are available on cvmfs.

Event Display

Not only is it useful to look at the geometry but it is also useful to look at specific events in the simulation to see what is happening.

Tutorial:

The EventDisplayTutorial provides two example tasks to make you familiar with the REve/Mu2e display
- EventDisplays gives an overview of the current ways to display events.

Art Data Products

All objects (e.g. straw hits, calorimeter clusters) are stored in the art event as art data products. These are accessed and created in Offline modules.

Tutorial:

The ArtDataProductTutorial is still to be written but...
- ReadProducts gives information on reading products; and,
- MakeProducts gives information on creating products.

Checkout and build code

If you need to write your own modules or edit code in Offline itself, then you will need your own build of Offline.

Tutorial:

CheckoutAndBuildCodeTutorial is the build system up to spring 2021
MuseBuildCodeTutorial is the build system after spring 2021
- ReleaseList gives the list of releases that are available on cvmfs.

Modules

There are a few different types of art module that you will encounter. "Analyzers" can only analyze data products that are already in the event; "producers" can create new data products; and, "filters" make a decision as to whether an event passes or fails some criteria.

Tutorial:

The ModulesTutorial is still to be written but...
- Modules discusses module names and labels; and
- FilterModules discusses filter modules.

Geometry and Config

For your study, you might need to edit part of the geometry or change the generated particle that is simulated. This is done with config files

Tutorial:

The GeometryAndConfig tutorial is still to be written but...
- SimpleConfig describes the format of these files.

Staging and Mixing Concepts

For the simulation, we don't just run from protons-on-target (POT) all the way through to hits in the tracker and actually run it in stages. This allows us to re-run specific stages to test new geometries without having to run everything again; and it also allows us to generate large samples of specific processes (e.g. particles emitted after a muon is captured by a nucleus) with better efficiency before mixing them all together into a full microbunch event.

Tutorial:

The StagingAndMixingTutorial is still to be written but...
- Staging describes many of the relevant concepts; and,
- Mixing explains background mixing and how to run this type of job.

Datasets and dCache

dCache is the tape system that we use to store our art files.

Tutorial:

The dCacheTutorial is still to be written but...
- Dcache gives a good introduction

Grids

To run large jobs, we use the grid rather than run on a local machine.

Tutorial:

The GridTutorial is still to be written but...
- Grids has a lot of useful information; and,
- Workflows gives a lot of information on how best to organise your work in Mu2e.

Git commits

Git is a version control system that allows us to coordinate many people developing software at the same time. It is widely used in the software development world so you will be able to find a lot of information online.

Tutorial:

The GitTutorial is still to be written but...
- Git has some Mu2e specific information

Code standards

In order to ensure that our code is stable and doing what we expect, there are various tasks where we enforce standard ways of performing them.

Tutorial:

The CodeStandardsTutorial is still to be written but...
- CodingStandards has a nice summary; and,
- RandomNumbers and RandomNumbersBasic describes the way you should generate random numbers in Offline.