LHAPDF

LHAPDF
Name	LHAPDF
Developer	High Energy Physics community
Initial release	2001
Latest release	(see project distribution)
Programming language	C++, Fortran
Operating system	Unix-like, macOS, Linux, Windows (via WSL)
License	Various (see Licensing and Governance)

LHAPDF LHAPDF is a software library used in high-energy physics for accessing parton distribution functions. It provides a standardized interface employed by collaborations and experiments to retrieve quantum chromodynamics inputs, facilitating analyses across particle detectors, accelerators, and theoretical frameworks. LHAPDF interoperates with computational tools, Monte Carlo generators, and phenomenology codes to ensure reproducible comparisons among results from CERN, Fermilab, DESY, and other laboratories.

Overview

LHAPDF serves as a bridge between theoretical calculations and experimental analyses performed by collaborations such as ATLAS, CMS, LHCb, ALICE, Tevatron experiments, and theory groups at CERN, Fermilab, DESY, INFN, and SLAC National Accelerator Laboratory. It standardizes access to parton distribution functions produced by groups including MSTW, CTEQ, NNPDF, HERAPDF, ABMP, CT, and MMHT. LHAPDF supports interfaces used by generators like PYTHIA, HERWIG, SHERPA, MadGraph, and analysis frameworks such as ROOT, GEANT4, and RIVET.

History and Development

Development began in the early 2000s to replace disparate PDF access methods used by collaborations including CDF, D0, H1, and ZEUS. Major milestones involved contributions from institutions like CERN theory groups, University of Oxford phenomenologists, University of Cambridge researchers, and teams at University of Manchester. Successive releases adapted to requirements from workshops such as the Les Houches meetings and recommendations from the PDF4LHC working group. Integration efforts involved software projects including LHAPDF6 upgrades, bindings for Python, and compatibility updates to align with standards promoted at ICHEP and EPS-HEP conferences.

Architecture and Features

The architecture exposes a C++ API with Fortran compatibility layers adopted by analysis groups at Imperial College London, MIT, Princeton University, and University of California, Berkeley. Core components include data management, interpolation engines, and metadata handling used by theory collaborations such as NNPDF Collaboration and CTEQ-TEA. Features include support for Hessian and Monte Carlo error sets employed by MSTW Collaboration and NNPDF Collaboration, scale evolution hooks influenced by work from Dokshitzer, Gribov, Lipatov, Altarelli, and Parisi-inspired implementations. The design permits plugin backends used in conjunction with tools from CMake, Autotools, and package managers used at CERN IT and Software Heritage initiatives.

Supported PDF Sets and Data Formats

LHAPDF distributes sets from prominent groups including MSTW, CTEQ, NNPDF, HERAPDF, ABMP, MMHT, JAM Collaboration, and older legacy sets referenced by CTEQ6 and MRST. Data formats encompass grid files, interpolation tables, and metadata standards used by collaborations like PDF4LHC Working Group and experiments at RHIC and LHCb. The system supports LHgrid-style and native HDF5-like packaging adapted by projects such as HEPData and interfaces commonly used by APPLgrid and fastNLO for fast cross-section convolutions in analyses from groups at Brookhaven National Laboratory and CEA Saclay.

Usage and Integration

Researchers integrate LHAPDF into workflows with generators and analysis stacks including PYTHIA8, HERWIG7, SHERPA, MadGraph5_aMC@NLO, POWHEG, and detector simulation chains employing GEANT4 and AliRoot. Analysis frameworks like ROOT and validation suites such as RIVET use LHAPDF APIs to ensure consistent theory inputs across studies by teams at ATLAS Collaboration and CMS Collaboration. Continuous integration and reproducibility practices reference environments from Docker, Singularity, and package ecosystems like Conda and Spack maintained by computing centers at CERN and NERSC.

Performance and Validation

Performance optimization addresses interpolation speed and memory footprint relevant for large-scale Monte Carlo campaigns carried out by ATLAS Computing, CMS Computing, and grid infrastructures such as WLCG. Validation procedures compare cross sections and uncertainties against benchmark results from PDF4LHC recommendations and studies published by Particle Data Group. Regression testing leverages datasets and validation plots produced by collaborations and reviewed at workshops including Les Houches, ICHEP, and EPS-HEP. Parallel computing strategies draw on work at Fermilab and Brookhaven to scale evaluations in cloud and batch systems.

Licensing and Governance

Licensing depends on distributed PDF set policies from groups such as NNPDF Collaboration, CTEQ-TEA, MSTW Collaboration, and institutions like CERN and DESY; some sets impose citation or redistribution terms recognized by publishers like Physical Review Letters, Journal of High Energy Physics, and European Physical Journal C. Governance involves stewardship by working groups including PDF4LHC and coordination among maintainers affiliated with CERN Theory and university groups. Contribution workflows mirror practices from open-source projects hosted on collaborative platforms used by GitHub, with issue tracking and release coordination involving computing teams at CERN and partner laboratories.

Category:High energy physics software