[Paper Review] The SM and NLO multileg working group: Summary report

T. Binoth, G. Dissertori+93 more|University of North Texas Digital Library (University of North Texas)|Mar 5, 2010

Particle physics theoretical and experimental studiesPhysics and Astronomy4 references114 citations

TL;DR

This paper presents a standardized interface between Monte Carlo event generators and one-loop matrix element programs, demonstrating its successful implementation between BlackHat, Rocket, and MadFKS for NLO QCD calculations in e+e− collisions. The LHA interface enables seamless integration across different programming languages and tools, significantly improving automation and reproducibility in high-energy physics precision calculations at hadron and lepton colliders.

ABSTRACT

This report summarizes the activities of the SM and NLO Multileg Working Group of the Workshop "Physics at TeV Colliders", Les Houches, France 8-26 June, 2009.

Motivation & Objective

• To establish a standardized, interoperable interface between Monte Carlo event generators and one-loop matrix element programs for precision QCD calculations.
• To address the challenge of integrating diverse software tools written in different languages (C++ and Fortran) in multi-leg NLO computations.
• To improve automation, reproducibility, and collaboration in high-energy physics by defining a common output format and calling convention.
• To validate the interface through practical implementations using BlackHat, Rocket, and MadFKS in e+e− → jets processes.
• To support future extensions to higher orders and more complex processes in the context of LHC and future collider physics.

Proposed method

• The LHA (Les Houches Accords) interface is used to define a common calling convention between Monte Carlo programs and one-loop matrix element programs.
• C++ and Fortran codes are linked by handling differences in argument passing (by reference), array memory layout (row-major vs. column-major), and name mangling (underscores and case sensitivity).
• A null-terminated string is passed from Fortran to C++ to ensure correct file path handling for contract files.
• The interface uses 'extern "C"' in C++ to prevent symbol name mangling, ensuring correct linkage with Fortran subroutines.
• The virtual matrix element squared is computed and returned via a standardized subroutine call with momenta, scale, and coupling constants as inputs.
• The implementation is validated using 2-, 3-, and 4-jet production in e+e− collisions at √s = MZ with the kT algorithm and E-scheme recombination.

Experimental results

Research questions

• RQ1 How can one achieve robust and portable integration between C++ and Fortran-based one-loop matrix element programs and Monte Carlo generators?
• RQ2 What are the key technical challenges in linking tools written in different languages for NLO calculations?
• RQ3 Can a standardized interface like LHA ensure consistent and reproducible results across different software frameworks?
• RQ4 How does the interface perform in practice for multi-jet processes in e+e− collisions at NLO?
• RQ5 What are the implications of this interface for future automation and standardization in precision QCD calculations?

Key findings

• The LHA interface successfully enables interoperability between BlackHat (C++) and MadFKS (Fortran), demonstrating correct linking and execution across different languages.
• The interface handles language-specific issues such as argument passing by reference, array layout differences, and symbol name mangling with consistent results.
• NLO predictions for inclusive cosθ and thrust distributions in e+e− → 2 and 3 jets show significant corrections compared to LO, validating the numerical accuracy of the implementation.
• The C and D parameters in e+e− → 3 and 4 jets at NLO are computed and compared with LO, showing improved agreement with experimental data.
• The interface is robust and extensible, with no additional issues expected when linking other OLP or MC codes using the same standard.
• The results confirm that standardized, automated, and portable NLO calculations are now feasible across major software frameworks in high-energy physics.