Physics Computing Track Overview


PC:  Physics Computing


Rudi Frühwirth, Andreas Pfeiffer


The track will cover two important aspects of computing in High Energy Physics: how to store the information generated by the various processing stages during an experiment (aka "Object Persistency") and the simulation of the experimental setup to optimize the detector, to test and improve the reconstruction software, and to gain a detailed understanding  of the data.

After an introduction explaining the various steps in the data reduction and analysis chain of an experiment, the first series of lectures will introduce the concepts of Object Persistency.  Here we will cover issues like Meta-data handling with distributed databases as well as the use of Grid services and local XML based catalogs. Following this the lecture then will concentrate on the POOL object store and it's dictionary. In the exercises the students will learn how to use the POOL object store and a catalog in making a simple data model persistent.

In the second series of lectures in this track, the emphasis will be on simulation of the experiments using Geant4. After an introduction in the why, what and how of simulation in HEP, some basic concepts of detector description and detector simulation are introduced. This is followed by a more detailed explanation on  how an experiment uses the Geant4 toolkit. In the exercises for this series of lectures, the students will learn how to do a simulation of an experimental testbeam setup, with emphasis on how to set up  the geometry, activate the required physics processes, extract hit information, and store the hits (using POOL) and optionally to read them back to do some analysis.



General Introduction to Physics Computing Series

R. Frühwirth




The lectures give an overview of the software components required for the processing of the experimental data, from the source - the detector - to the final input to the physics analysis. The emphasis is on the concepts, not on the implementation details. The key concept is data reduction, both in terms of rate and in terms of information density. The flow of the real data is the main topic, but the need for and the production of simulated data is discussed as well. Some basic requirements to detector description databases are examined.

Databases and object persistency Series

D. Düllmann  M.Girone






LCG-POOL project

Starting from 2007 the LHC experiments at CERN will generate a vast amount of data. Some 100 Peta-Bytes of event, calibration and analysis data will be stored and analysed in a distributed grid environment. Over the last years several R&D projects have studied the use of object oriented technology as key element of the LHC software systems and object persistency as a simple but powerful paradigm for data storage has emerged as baseline approach for most of the LHC experiments. The POOL project as part of the LHC computing grid (LCG) provides a common framework to support this paradigm. At the same time POOL insulates the experiment software against potential technology changes over time and allows to combine technologies like an object store (ROOT I/O) for event data with transactionally safe data base storage for meta data which needs concurrent access and requires frequent updates.
This series of lectures will start with a brief overview of the main features of object persistency systems pointing out their advantages compared to more traditional storage technologies. In particular the simplicity of transparent object navigation and the strict separation between logical and physical data model will be demonstrated in more detail.
The second part of the lectures will focus on the architecture of the POOL system explaining the key design goals of the system such as
- end user interface independence of any back end technology details
- strict component model approach loading implementation components as required at runtime
- seamless integration into a large variety of computing environments, ranging from isolated laptops to fully grid connected systems
We will conclude the lectures with a report on the practical integration and production experience gained in several large-scale data challenges using POOL in the LHC computing grid and show real life data management configurations involving several sites, explaining how consistent data management can be realised using POOL and lower level grid services.
The lecture will be accompanied by a series of hands-on exercises ranging from the definition of a persistent object model, a simple population of a distributed store up to a batch analysis of physics data exploiting advanced query techniques.

Experiment Simulation Series

M. Liendl







"Experiment Simulation" Lectures
This series of four lectures gives an introduction to the domain of simulating high energy physics experiments with practical examples taken from the CERN/LHC based CMS experiment. The lectures focus on event simulation based on the GEANT4 toolkit.  GEANT4 is a toolkit for simulating the passage of particles through matter. It includes a complete range of functionality including tracking, geometry, physics models, and hits. The physics processes offered cover a comprehensive range, including electromagnetic, hadronic and optical processes, a large set of long-lived particles, materials and elements, over a wide energy range starting, in some cases, from 250 eV and extending in others to the TeV energy range. It has been designed and constructed to expose the physics models utilised, to handle complex geometries, and to enable its easy adaptation for optimal use in different sets of applications.
Basic simulation requirements are explained, such as experimental setup in terms of geometry, materials, and external electromagnetic fields, physics processes principles, selection and configuration of physics processes, run and event concepts, and extraction of hit information. It is shown how this requirements are met by applying GEANT4. Cross references to the simulation package of the CMS detector give illustrative insights on how these tasks are realized in a complex, state of the art high energy physics experiment. The programming language used throughout the lectures is C++.

Physics in Geant 4  Series

A. Ribon




This optional series of two lectures gives a more detailed description  of the design of the basic physics model utilised by Geant4, together with an overview of the electromagnetic and hadronic physics  processes. Some examples of how these physics models are validated with test-beam data are also discussed 





Databases and object persistency  Series

D. Düllmann  M.Girone




Exercises on Databases and object persistency

Experiment Simulation Series

M. Liendl








Exercises on Experiment Simulation
The aim of the exercises is to simulate a simple testbeam setup in order to store hit information persistently, using POOL. Starting from a provided skeleton implementation, students have to complete the geometrical description of the testbeam setup, activate and configure appropriate physics processes, extract useful hit information, execute several simulation runs and store the results using POOL. If there is time left, the stored results of several exercise groups are combined in a simple data analysis task. The programming language used in the exercises is C++





Total hours




 Italics: Optional lectures 

Last edited: 08-Sep-04 .