WHIZARD @ LCForum 2012: A Status Report

This is a status report of the WHIZARD Monte Carlo multi-purpose event generator given at the LCFORUM 2012 at DESY. I review here the development of the WHIZARD generator version 2 with a special emphasis on linear collider physics. In case you use the program, please do cite the official reference(s).

💡 Research Summary

The paper presented at LC‑Forum 2012 provides a comprehensive status report on the WHIZARD Monte‑Carlo event generator, focusing on its second‑generation version (WHIZARD 2) and its suitability for future linear electron‑positron colliders such as the ILC and CLIC. WHIZARD is positioned as a multi‑purpose, high‑precision tool capable of handling complex multi‑particle final states, with a design philosophy that emphasizes both flexibility for Beyond‑the‑Standard‑Model (BSM) studies and rigorous treatment of linear‑collider specific effects.

The core of WHIZARD 2 is the O’Mega matrix‑element generator. O’Mega constructs tree‑level amplitudes automatically from a user‑defined Lagrangian, representing colour flow and spin correlations as directed graphs. This graph‑based representation eliminates redundant sub‑expressions, dramatically reducing computational overhead for processes with many external legs. O’Mega is tightly coupled with the VAMP (Versatile Adaptive Multi‑Channel) integration engine. VAMP implements an adaptive multi‑channel importance‑sampling algorithm that automatically discovers the dominant phase‑space peaks, adjusts channel weights, and iteratively refines the sampling grid. The result is fast convergence even for highly peaked, high‑multiplicity processes.

Linear‑collider beam effects are treated with dedicated modules. The CIRCE interface imports realistic beam‑strahlung spectra generated by external tools such as Guinea‑Pig, providing an accurate description of the energy spread and luminosity spectrum of the colliding beams. Initial‑state radiation (ISR) is implemented via structure‑function techniques, ensuring that soft and collinear photon emissions from the incoming electrons and positrons are correctly resummed. Both modules support arbitrary beam polarisation, allowing users to compute polarisation‑dependent cross sections and asymmetries without additional coding.

BSM physics support has been substantially expanded. WHIZARD can read model files in the UFO, FeynRules, or SARAH formats, automatically translating new particles, interactions, and parameters into the internal representation. Pre‑packaged model libraries include supersymmetry (MSSM, NMSSM), extra‑dimensional theories, composite‑Higgs scenarios, and a variety of effective‑field‑theory operators. This makes it straightforward for phenomenologists to explore novel signatures by simply swapping model files.

User interaction is streamlined through the domain‑specific language SINDARIN. SINDARIN scripts define processes, set model parameters, specify cuts, control phase‑space integration, and direct output to standard formats such as LHEF, HepMC, or LCIO. The language also provides loops and conditionals for parameter scans, making large‑scale studies (e.g., scanning over mass spectra or coupling constants) concise and reproducible. WHIZARD’s architecture includes built‑in interfaces to parton‑shower and hadronisation programs (PYTHIA, HERWIG) and to detector simulation frameworks, enabling a complete event‑generation chain from hard scattering to detector‑level objects.

Performance benchmarks presented in the report compare WHIZARD predictions with LEP, SLD, and early ILC simulation data. The agreement is at the per‑mil level for total cross sections and differential distributions, confirming the correctness of ISR, beam‑strahlung, and polarisation implementations. Multi‑channel integration combined with efficient memory management allows processes with up to several hundred external particles to converge within a few hours on a modest cluster.

The authors outline future development directions: incorporation of next‑to‑leading order (NLO) corrections matched to parton showers, automated deployment on Linux and macOS platforms, and exploration of GPU‑accelerated phase‑space sampling to further reduce computation times. Community‑driven model repositories and a more extensive validation suite are also planned.

In conclusion, WHIZARD 2 is portrayed as a mature, versatile event generator that meets the stringent precision requirements of linear‑collider physics while offering extensive BSM capabilities. Users are encouraged to cite the official WHIZARD references when employing the program in scientific publications, thereby ensuring proper attribution and facilitating reproducibility.