CORSIKA Cosmic Ray Simulation Workshop Karlsruhe

Hotel ACHAT PLAZA Karlsruhe

Hotel ACHAT PLAZA Karlsruhe

Mendelssohnplatz, 76131 Karlsruhe
Ralf Ulrich (KIT)

General and open workshop to discuss technology and science of air shower simulations, and meeting of the CORSIKA 8 Project.

  • Abdul Rehman
  • Alan Coleman
  • Alexander Sandrock
  • Anna Schäfer
  • Bernardino Spisso
  • Blahoslav Pastirčák
  • Darko Veberic
  • Dieter Heck
  • Dmitriy Kostunin
  • Dominik Baack
  • Doris Wochele
  • Eva Santos
  • felix riehn
  • Frank Schröder
  • G. Schatz
  • Gabor Veres
  • Hans Dembinski
  • Ioana Alexandra Caracas
  • Jan Ebr
  • Jean-Marco Alameddine
  • Jean-Noel Capdevielle
  • Johannes Blümer
  • Johannes Knapp
  • Juergen Oehlschlaeger
  • Konrad Bernlöhr
  • Luisa Arrabito
  • Lukas Nellen
  • Marcel Koepke
  • Mario Pecimotika
  • Markus Roth
  • Maximilian Reininghaus
  • Michael Unger
  • Oliver Suranyi
  • Peter Grieder
  • Piotr Kalaczyński
  • Ralf Ulrich
  • Ralph Engel
  • Roxanne Turcotte
  • Sabine Bucher
  • Sergey Ostapchenko
  • Simona Maria Stellacci
  • Stephan Meighen-Berger
  • Suren Chilingaryan
  • Takashi Sako
  • Tanguy Pierog
  • Tim Huege
  • Yating Chai
    • 09:30 10:30
      Arrival, Registration, Coffee 1h
      Speaker: Organizers
    • 10:30 11:00
      Break 30m
    • 11:00 11:10
      Welcome 10m
      Speaker: Ralf Ulrich (KIT)
    • 11:10 12:25
      Status and progress of air shower simulations

      The current state-of-the-art in air shower simulations, with a focus on the status and progress of CORSIKA 8.

      • 11:10
        Introduction 15m
        Speaker: Ralph Engel
      • 11:25
        The CORSIKA 8 Project 20m
        Speaker: Ralf Ulrich (KIT)
      • 11:45
        Status of capabilities of CORSIKA 8, first results and comparisons 20m
        Speaker: Maximilian Reininghaus (KIT / IKP)
      • 12:05
        The lepton propagator PROPOSAL and the air shower simulation code CORSIKA 20m

        This talk describes the lepton propagator PROPOSAL, its current capabilities and its possible use in future versions of CORSIKA, and necessary extensions to function as a replacement for the EGS4 electromagnetic shower simulation code.

        Speaker: Alexander Sandrock (Technische Universität Dortmund)
    • 12:30 14:00
      Lunch 1h 30m
    • 14:00 15:00
      Applications and requirements of CORSIKA

      Contributions from experimental collaborations, phenomenologist or theorists about important application of CORSIKA, and future wishes for new functionality.

      • 14:00
        Measured and simulated muon flux for the first KM3NeT Detection Units 20m

        The KM3NeT Collaboration has successfully deployed its first detection units in the Mediterranean Sea in 2016 (ARCA) and 2017 (ORCA). The sample of data collected between December 2016 and December 2017 has been used to measure the atmospheric muon flux  at two different depths under the sea level:  3.5 km with ARCA and 2.5 km with ORCA. The atmospheric muon flux represents an abundant signal for a neutrino telescope and can be used to test the reliability of the Monte Carlo simulation chain. In this case, the measurements are compared to Monte Carlo simulations based on MUPAGE and CORSIKA codes. MUPAGE is fast, parametric simulation code developed for the ANTARES experiment that generates muon events, induced by cosmic rays impinging the Earth atmosphere, at different undersea depths. The atmospheric muon distributions at the sea level obtained with MUPAGE/CORSIKA are propagated to the detector with the MUSIC code, a 3D program that takes into account all muon energy loss processes. Finally, all muons arriving at the detector are transported through the active volume, Cherenkov light is generated and the same triggering algorithms used during data taking are applied. The main features of the simulation and reconstruction chain are discussed and presented.

        Speaker: Piotr Kalaczyński (NCBJ Warsaw)
      • 14:20
        Corsika simulations of background in Baikal experiment 15m

        The Baikal experiment was proposed to register cosmological neutrinos and map the high-energy neutrino sky in the Southern Hemisphere. It will use a km3-scale high-energy neutrino telescope in Lake Baikal. The selection of events from neutrino induced upward going muons, suggests a fairly reliable estimation of the expected background from atmospheric muons. The procedure for simulating background events from atmospheric muons in an array is performed in several step. The CORSIKA7.64 was used to simulate the flux of atmospheric muons at sea level with appropriate chemical composition of the primary cosmic rays. The muon propagation through water and rock to the array level was then simulated with the MUM code. As the last step of simulation chain simulation of the detector response to the Cherenkov radiation of muons by taking into account the features of array measuring systems was performed. The main features of the CORSIKA Monte Carlo code and following steps of the full simulation chain will be briefly reported.

        Speaker: Blahoslav Pastircak (IEP SAS Kosice)
      • 14:35
        Proton-Oxygen collisions at the LHC for air shower research 25m

        New measurements at the LHC are needed to resolve the Muon Puzzle in cosmic-ray induced air showers and to reduce the systematic uncertainties of air shower simulations overall. In particular, the very forward production of light hadrons has to be studied with LHCb and other forward detectors. The LHC has accelerated proton and lead beams so far (and a bit of Xenon). The hadronic interaction models used to simulate air showers are primarily tuned to p-p data, while the most common collision in an air shower is between a high energy hadron and a oxygen or nitrogen nucleus. The predictions for these interactions vary between models by 50 % in multiplicity. This variance needs to be reduced to 10 % and the fraction of energy which ends up in neutral pions need to be clarified. This can be achieved by studying proton-oxygen collisions at the LHC, which have been proposed for 2023. We review the science case for measuring proton-oxygen collisions and report on the status of the proposal.

        Speaker: Hans Dembinski (Max Planck Institute for Nuclear Physics, Heidelberg)
    • 15:00 15:30
      Coffee 30m
    • 15:30 17:05
      Status and progress of air shower simulations

      The current state-of-the-art in air shower simulations, with a focus on the status and progress of CORSIKA 8.

      • 15:30
        The COSMOS air shower simulation program 20m
        Speaker: Takashi Sako (ICRR, University of Tokyo)
      • 15:50
        Hands-On: Simulate your first CORSIKA 8 air shower 55m

        Tutorial part (DIY) and introduction, overview, results

        Speaker: Ralf Ulrich (KIT)
      • 16:45
        Introduction to template meta-programming in C++ 20m

        Template meta-programming (TMP) is a fancy and mysterious sounding topic. CORSIKA 8 is using TMP a lot, so in this talk I will explain when it is useful and demonstrate a few key techniques and best practices, based on my experience as the author of Boost.Histogram, which is another TMP-heavy C++ library.

        Speaker: Hans Dembinski (Max Planck Institute for Nuclear Physics, Heidelberg)
    • 09:00 12:45
      Celebration: 30years of CORSIKA

      The first and original version of the program "COsmic Ray SImulation for KAscade" was released in 1989 - 30 years in the past. We want to celebrate the extraordinary development CORSIKA has taken until today.

      • 09:00
        Introduction 5m
        Speaker: Johannes Blümer (KIT)
      • 09:10
        Original motivation for CORSIKA 20m
        Speaker: Gerd Schatz
      • 09:30
        From Quarks to Air Showers: A Historic Review of the Development of the Simulation Program System ASICO, Results and Impact, and its Evolution to CORSIKA 30m

        I set the scene with a brief presentation of the status of our knowledge of high energy cosmic rays and particle physics around 1960. This is followed by an overview of the quark hunt in the early 1960s, when it was believed that free quarks exist, and the search that was undertaken at the Niels Bohr Institute in Copenhagen in 1965, and elsewhere. Instead of quarks, the unexpected copious production of $N\overline{N}$ in air showers was discovered, which prompted me to develop what later on became the air shower simulation program system ASICO. I discuss briefly the contemporary shower and high energy hadronic interaction models, the sources of information to construct such models, the architecture of the initial cascade simulation process, and the first results. This is followed by an outline of the unique Echo Lake experiment, a landmark in cosmic ray physics, its results and the problem it caused. Subsequently, a series of significant discoveries and new theoretical developments of the epoch are summarized, which culminated in the coming into operation of the CERN ISR. This collider with its approximately 1800\,GeV laboratory equivalent energy yielded an enormous wealth of new information and insight into a hitherto "uncharted" energy domain. Its results confirmed several cosmic ray discoveries and solved the Echo Lake problem. The presentation closes with a potpourri of results of ASICO after over twenty years of development, and its transition to CORSIKA.

        Speaker: Prof. Peter Grieder (University of Bern, Switzerland)
      • 10:05
        Collisions from 40 years ago… 20m
        Speaker: Capdevielle Jean-Noel
      • 10:25
        Coffee 30m
      • 11:25
        Developments of CORSIKA 20m
        Speaker: Dieter Heck
      • 11:50
        CORSIKA and Imaging Air-Cherenkov Telescopes 15m
        Speaker: Konrad Bernlöhr
      • 12:15
        Workshop Photo 30m
        Speaker: Everybody
    • 12:45 14:15
      Lunch 1h 30m
    • 14:15 17:40
      Applications and requirements of CORSIKA

      Contributions from experimental collaborations, phenomenologist or theorists about important application of CORSIKA, and future wishes for new functionality.

      • 14:15
        Modeling of high energy cosmic ray interactions: selected topics 40m
        Speaker: Sergey Ostapchenko
      • 14:55
        SMASH - a novel transport model to simulate low-energy hadronic interactions 30m
        Speaker: Anna Schäfer
      • 15:30
        Coffee 30m
      • 16:00
        Opportunities with Calorimeters at LHC: How to solve the muon puzzle? 20m

        Zero degree calorimeters at LHC like LHCf or in CMS provide the opportunity to identify pion-exchange reactions at very high energies p+p --> p+pi+n where the neutron escapes at zero degree and the pion and the proton subsequently collide in the detector. This provides a unique opportunity to learn about pion collisions, which are one of the largest uncertainties in air showers.

        Speaker: Tanguy Pierog (IKP, KIT)
      • 16:20
        Zero-degree neutron measurements with CMS 20m
        Speaker: Oliver Suranyi (Budapest)
      • 16:40
        Results from the LHCf calorimeter 20m
        Speaker: Takashi Sako (ICRR, University of Tokyo)
      • 17:00
        Time of Flight detector for cosmic muons? 15m
        Speaker: Gabor Veres
      • 17:15
        GHOST - a new generator for nuclear collisions 20m
        Speaker: Jean-Noel Capdevielle
    • 09:00 10:30
      Applications and requirements of CORSIKA

      Contributions from experimental collaborations, phenomenologist or theorists about important application of CORSIKA, and future wishes for new functionality.

      • 09:00
        Radio detection of extensive air showers 15m
        Speaker: Tim Huege (KIT)
      • 09:15
        Use of CORSIKA for the IceTop Radio Enhancement 20m

        The detector is currently in an R&D phase for an upgrade that will include adding an array of scintillator panels and radio antennas, to the existing array of ice Cherenkov detectors. Understanding the potential physics output of such an array requires a large library of CoREAS simulations. I will present how we are currently using CORSIKA to develop the radio component of the IceTop upgrade as well as how our needs could be better met with improvements to the simulation software.

        Speaker: Alan Coleman (University of Delaware)
      • 09:35
        Handling the metadata for large amount of CORSIKA simulations 20m

        There are many reasons for producing millions of equitype CORSIKA simulations, e.g. to cover necessary parameter space for the model or to fit experimental data. Moreover, sometimes it is necessary to perform fast pre-simulations, e.g. with CONEX or other computation-reduced options. We present a SIMulation Manager (SIMM) framework for steering, running and evaluating status of CORSIKA simulations. The main features of SIMM is the interface for CORSIKA configuration, scheduling the simulations for running on distributed resources, integration with Auger Offline software and flexible plugin-based framework for custom simulation sequences (i.e. CONEX-CORSIKA-CoREAS with evaluation at each step). The software is implemented on Python with SQLAlchemy framework, which allows one to use any of SQL engines for metadata storage. We are successfully using SIMM more than five years for the handling simulations for Tunka-Rex and TAIGA experiments. The ideas developed in the frame of SIMM can be further used in future Astroparticle Data Life Cycle frameworks.

        Speaker: Dmitriy Kostunin (DESY)
      • 09:55
        CORELib: COsmic Ray Event Library 20m

        CORELib is a collection of simulated showers induced by primary cosmic rays and a common framework to produce and manage such large production. Indeed, CORELib provides high statics background simulation for various astroparticles and astronomy experiments. The simulations are obtained employing widely used CORSIKA simulation framework and the GRID infrastructure as computational resources; in the current productions the observation level is placed at sea level, while the Standard European Atmospheric Model is used. The parameters of simulations are available and easily accessible to the users. Two different primary energy spectra are considered: one with -2 spectral index and one with flat spectrum. Four different high energy models are evaluated: QGSJET01 with CHARM option, QGSJET01 with TAULEP option, QGSJETII with TAULEP option and EPOS with TAULEP option. The common low energy model chosen is GHEISHA. Protons and Heavy Nuclei (He, C, N, O, Fe) induced showers are simulated. Proton induced showers two production are available: with and without Cherenkov radiation. All the information about every single shower is stored in SQLite database queryable by the user. In order to simplify the access at the library, a dedicate tool named CORANT is used to translate from binary to ASCII the CORSIKA output files. The information about showers is placed in separated files, one for each particles class (EM, Hadrons, Muons, Neutrinos ). The whole production of CORELib will be stored at the CNAF, the Information Technology national center of INFN (Italian Institute for Nuclear Physics) . It can be downloaded through gridFTP with X509 certificate.

        Speaker: Simona Maria Stellacci
    • 10:30 11:00
      Coffee 30m
    • 11:00 12:35
      Status and progress of air shower simulations

      The current state-of-the-art in air shower simulations, with a focus on the status and progress of CORSIKA 8.

      • 11:00
        Particle propagation for CORSIKA in PROPOSAL 20m

        This talk intends to give technical insights into the propagation process of PROPOSAL, a tool to propagate leptons through media. Furthermore, we are going to discuss the requirements and challenges regarding possible use cases for PROPOSAL in CORSIKA in the future.

        Speaker: Jean-Marco Alameddine
      • 11:20
        EmCa - Electromagnetic Cascades Simulation Package 20m

        Emca is a new python simulation package, based on MCEq, for the calculation of electromagnetic particle fluxes in various materials. The calculations are based on a cascade equation approach, allowing for an iterative calculation of showers. This allows for fast and efficient calculations of cascades, allowing easier testing and prototyping compared to a Monte Carlo approach. The methodology of EmCa and comparisons between it and CONEX will be shown.

        Speaker: Stephan Meighen-Berger
      • 11:40
        Hands-On: Physics processes 50m

        Tutorial part for DIY, introduction and concepts

        Speaker: felix riehn (LIP, Lisbon)
    • 12:35 14:00
      Lunch 1h 25m
    • 14:00 17:00
      Status and progress of air shower simulations

      The current state-of-the-art in air shower simulations, with a focus on the status and progress of CORSIKA 8.

      • 14:00
        Python in High-Energy Physics 30m

        In high-energy physics (HEP), C++ is still the dominant language, but Python is continuously growing and should overtake C++ in the near future. In the industry, Python is already the dominant language for data science and driving the leading frameworks for machine learning. In this talk, I will show these and other interesting facts and explain how a "slow" interpreted language like Python is able to beat a "fast" language like C++, in a field where code execution speed actually matters. I will present the Scikit-HEP project that aims to provide key functionality in Python for HEP analyses that's currently missing, like suitable histograms. Finally, I will argue why the upcoming ROOT 7, the first backward incompatible change in ROOT's history and a bold step forward, will not win the users back.

        Speaker: Hans Dembinski (Max Planck Institute for Nuclear Physics, Heidelberg)
      • 14:30
        Optimization and Vectorization 20m
        Speaker: Luisa Arrabito
      • 14:50
        Atmosphere implementation and approximations 20m
        Speaker: Konrad Bernlöhr
      • 15:10
        Simulation of radio emission from particle showers in CORSIKA 7 and CORSIKA 8 15m
        Speaker: Tim Huege (KIT)
      • 15:25
        Coffee 30m
    • 18:00 21:00
      Workshop Dinner 3h Schützenhaus am Turmberg

      Schützenhaus am Turmberg

      Schützenhaus auf dem Turmberg Jean-Ritzert-Strasse 8 | 76227 Karlsruhe

      Telefon (0721) 94318122

      see also:

    • 09:00 09:40
      Applications and requirements of CORSIKA

      Contributions from experimental collaborations, phenomenologist or theorists about important application of CORSIKA, and future wishes for new functionality.

      • 09:00
        Simulations for the atmosphere with clouds 20m

        The Cherenkov Telescope Array (CTA) is the next generation ground-based array of IACTs, the world’s largest and most sensitive high-energy gamma-ray observatory. The full setup will consist of more than 100 telescopes located in the northern (CTA-N) and southern hemispheres (CTA-S). The Observatorio del Roque de Los Muchachos (ORM) on the LaPalma island, Spain is selected as a site for the northern array (CTA-N).

        In order to determine the influence of clouds on the atmospheric transparency and performance of CTA-N, transmittance simulations and simulations of instrument response in such an atmosphere are being performed.

        Speaker: Mario Pecimotika (University of Rijeka, Department of Physics, Croatia)
      • 09:20
        Testing modifications to hadronic intractions in air showers 20m
        Speaker: Jan Ebr
    • 09:40 10:20
      Status and progress of air shower simulations

      The current state-of-the-art in air shower simulations, with a focus on the status and progress of CORSIKA 8.

    • 10:20 10:50
      Coffee 30m
    • 10:50 12:20
      Status and progress of air shower simulations

      The current state-of-the-art in air shower simulations, with a focus on the status and progress of CORSIKA 8.

      • 10:50
        DynStack / Memory managment - implementation and features 20m

        Overview of different features and possibilites of an extensive Stack/Memory managment implementation.

        Speaker: Dominik Baack (TU Dortmund)
      • 11:10
        Run control and steering 45m

        As part of this, I would also discuss "Workspace environment"

        Speaker: Dr. Lukas Nellen (I de CIencias Nucleares, Universidad Nacional Autonoma de Mexico)
      • 11:55
        Speedup of extensive air shower simulations with neural networks 20m

        The time complexity of extensive air shower simulations rises approximately linearly with the incident particle energy for the CORSIKA 7 framework. The range of cosmic ray energies observed on earth covers several orders of magnitude. In order to simulate the highest energies in the cosmic ray spectrum, one has to introduce some sort of heuristic (e.g. thinning) which reduces runtime and preserves the shower properties to leading order. The physical content on higher order effects, like shower-to-shower fluctuations, is usually reduced. In this talk I am going to present my ideas on how to supplement current heuristics by training neural networks on CORSIKA simulations.

        Speaker: Marcel Köpke (Karlsruhe Institute of Technology)
    • 12:25 12:45
      Conclusion and end of workshop 20m
      Speaker: Ralf Ulrich (KIT)
    • 12:45 14:10
      Lunch 1h 25m
    • 14:10 16:05
      Working session on CORSIKA 8 (optional, post-workshop) 1h 55m

      This is a post-workshop event. Interested people should contact so that I can organize a room large enough for us. This session extends to Friday, and will also continue the following complete week (Max is visiting from Argentina).

      Post-workshop CORSIKA discussions, 20th June 2019

      Stephan, Alexander, Lukas, Felix, Max, Hans, Ralf


      • First public demonstration of Stack via dynstack
        ~1 week by Doiminik
      • Suggest to add "Particle" class, evtl. only as facade
      • rename IsEmpty() -> empty(), see also below
      • Particles delete:
        Only set flag, physically purge particles when needed,


      • keep gcc7 as minimum now, gcc8 crystalizes as standard in a while
      • look at "ranges", but don't overdo interface work
      • Use std compatible naming: NOT CamelCase !  →decided
      • Variable names: small  →decided
      • Member variables with postfix “_”   → decided
      • Getters: GetX() -> x()   → decided
      • Setters: SetX(const&) -> set_x(const&) → decided
      • Method/function parameter names: no special signature ?? not clear yet
      • Typenames in templates: keep the "T" prefix, not clear yet.


      Vectorisierung, is important, but requires later discussion

      Profiling needed

      -> time scale 6...9month to start



      Nothing about configuration has been finalized or decided yet. This is still all under discussion. Example would be extremely useful.

      • Need global baseline definition of Particle cuts and thresholds (property of particle-properties??)
      • Configuration: singleton class, map of map (several layers), python access,
      • Init functions should disappear -> constructors are better
      • Constructor -> single-config search in config maps -> fill once locally, during construction
      • Config object with main functionality in c++. Not in python. Python is just the interface.
      • Config object must be able to stream data in/out. Validation when? With Schema probably not; directly in c++?
      • Problem with unit testing?
      • Config object may count how often entries are read. If zero: might indicate typo or mistake

      Preliminary Conclusions:

      • Need test implementation.

      • Python needs to handle units.

      • Need for validation procedures, need for debugging

      • Check impact on unit testing

      • Global state (singleton) is bad by default… investigate different options

      -> Marcel will send configuration code example



      Coordinate system pointers change in reference systems -> template types combined with dynamic configuration

      -> Lukas will look at this


      • Open issues 2h

      • Further planning and tasks 1h
Your browser is out of date!

Update your browser to view this website correctly. Update my browser now