The work of MCnet is divided into several projects.

Pythia, Herwig and Sherpa are all projects to develop the physics models and implementations of general-purpose event generators and support their use by the LHC experiments and elsewhere.

MadGraph and Ariadne are projects to develop implementations and alternative approaches to two of the most important steps of event generation: automated hard-process generation and parton-showering evolution, particularly in the high-energy limit.

CEDAR is a more general project aimed to provide general-purpose facilities to do high energy physics analysis and phenomenology. Among its many sub-projects, Rivet and Professor are particularly important for MCnet, providing generator-independent frameworks for comparison of event generator predictions with corrected data and for tuning event generator parameters. Currently active projects that build on Rivet’s event-analysis database include Contur, a toolkit for re-interpretation of published collider data via MC modelling of new-physics signatures.

Pythia and Vincia

Pythia has its roots in the Jetset program, begun more than 30 years ago. It has been used by generations of physicists, notably at LEP and the Tevatron, and for LHC preparations, but also in areas such as astroparticle physics. Traditionally Pythia has put strong emphasis on soft physics, such as hadronization, minimum-bias physics and the soft underlying event. At the same time it contains a wide selection of Standard Model (SM) and Beyond-Standard Model (BSM) processes, and modern dipole-style transverse- momentum-ordered parton showers.

The Pythia project includes Vincia as a subproject. Vincia is plugin to Pythia 8 for dipole-antenna showers. Tree-level matrix elements (e.g., from MadGraph) can be incorporated in the shower evolution, using the GKS matching scheme. Vincia also has extensive possibilities for uncertainty estimates, via systematic variations of shower parameters. The development focus in the near future is on extensions of the formalism to hadron collisions, the inclusion of NLO matrix elements in the shower evolution, and further automation of the uncertainty variations.


The Herwig event generator has been used in the planning and data analysis of high energy physics experiments for over twenty years. Following a major rewrite, the current version simulates lepton and hadron collisions including: careful modeling of parton showers, including coherence effects and a description of uncertainties; a detailed model of the underlying event; generation of processes accurate at next-to-leading order and the matching of leading-order processes to higher multiplicity matrix elements in an improvement to the original approach of Catani, Krauss, Kuhn and Webber (CKKW); simulation of a range of BSM models with spin correlations; a detailed model of hadron and tau decays.


Sherpa is a fully-fledged event generator comprising solutions to all aspects of event simulation. Its hallmark feature is the automated merging of leading-order matrix elements for multijet production with parton showers, provided by two independent matrix element generators for large final state multiplicities in various models, and interfaced to an automated Feynman rule generator. It also provides an automated method to treat the infrared divergent structures emerging in NLO QCD corrections, which is being used for state-of-the-art NLO calculations. First steps towards a truly automated version of the matching of parton showers with NLO matrix elements have met success - Sherpa by now includes all methods necessary to perform such a matching within the Powheg formalism, tested in a large range of processes. At the same time various theoretical problems with both the Powheg and the MC@NLO method have been found, which led the authors to concentrate on the further development of MC@NLO methods alone.


MadGraph provides a world-wide on-line service for the generation of events relevant for studies at present and future high-energy colliders. The capabilities of the current version 5 include: event generation via the web for any Standard Model process as well as for new physics models via FeynRules; fully-staged simulation chain, from model building to detector simulation; multi-jet sample production and merging; interfaces with the software frameworks of the main experiments at the LHC and the possibility of event generation over the computing GRID. Version 5 now distributes publicly the automatic computation of loop amplitudes (MadLoop), the combination with the real corrections (MadFKS) and the automatic interface via the MC@NLO method (aMC@NLO) to Herwig and Pythia for parton showering and hadronization.


The general-purpose event generators, Pythia, Herwig and Sherpa, work well for most processes at the LHC, but there is still room for improvement. The development of models is often done outside the three main programs, although the resulting code typically is interfaced to plug in to one or more of them. A special class of such programs involves parton shower models where the standard collinear evolution is amended by so-called small-x resummation. Two such programs are explicitly included in this project:

  • Ariadne/DIPSY is based on dipole formulation of parton showers where the initial-state shower is based on evolutionin impact-parameter space. This program is also able to model collisions involving heavy-ion collision. Such collisions are well integrated parts on the LHC program, but none of the three main programs have the capability of fully simulating them.
  • HEJ is based on a high-energy version of collinear factorization, where small-x resummation is added in the limit of multi-Regge kinematics. The program needs to be interfaced to final state showers to get the full modeling of the resulting jets. So far there exists a simplified interface to Ariadne, and another more sophisticated interface is being developed. Work is also ongoing on merging NLO partonic samples within the HEJ formalism.

Besides developing these two programs, the Plugin project will act as an interface to theorists in the wider community working on special models for parts of the event generating chain. The project is led by Leif Lonnblad in Lund, with Jeppe Andersen and Jennifer Smillie of the Durham team.


Several activities are collected together as the CEDAR project, with the general theme of providing tools to interlink the other projects, for validation and tuning of the generators, and linking with the users of MC generators. While led from UCL and Glasgow, the project also has key people in CERN, Durham, and Fermilab.

The CEDAR teams are responsible for the Rivet (Robust Independent Validation of Experiment & Theory) package, which provides generator-independent analysis tools that enable MC-model comparisons to the measurements in the HepData database. HepData functions as the central archive of measurements from previous and current colliders, including the LHC, and is used by all generator projects for tuning and validation of models. A major goal is to facilitate implementation of key LHC data analyses in Rivet as soon as they become available, and to provide all the functionality needed for analysis logic preservation.

The output from Rivet can be coupled with a mechanism for generator tuning. The Professor package provides an automated tool to do this, and is central to the MC tunes constructed and used by the LHC experiments. Rivet is also the central component of the Contur project, which uses measurement analyses from collider experiments to place statistical limits on the viability of new-physics (or Beyond Standard Model) models, complementing and often pre-empting dedicated direct-search analyses.

CEDAR also developed and supports community tools, e.g. the software-development platform HepForge; the LHAPDF library of parton densities; the HepMC event record standard and code library; the Les Houches accords for data exchange between generators; and the LHC Computing Grid “Generator Services” project. CEDAR includes members of the LHC and other experiments, to provide support, liaison and training.