Updates to the one-loop provider NLOX
Figueroa, Diogenes ; Quackenbush, Seth ; Reina, Laura and Reuschle, Christian LU
(2022)
In Computer Physics Communications
270.
- Abstract
In this release note we describe the 1.2 update to NLOX, a computer program for calculations in high-energy particle physics. New features since the 1.0 release and other changes are described, along with usage documentation. New version program summary: Program title: NLOX Developer's repository link: http://www.hep.fsu.edu/~nlox Licensing provisions: CC BY NC 3.0 Programming language: C++. Fortran interface available, and Fortran compiler required for dependencies. Journal reference of previous version: Comput. Phys. Commun. 257 (2020) 107284 Does the new version supersede the previous version?: Yes Reasons for the new version: We have added several new features to NLOX, and made significant stability and efficiency improvements. The... (More)
In this release note we describe the 1.2 update to NLOX, a computer program for calculations in high-energy particle physics. New features since the 1.0 release and other changes are described, along with usage documentation. New version program summary: Program title: NLOX Developer's repository link: http://www.hep.fsu.edu/~nlox Licensing provisions: CC BY NC 3.0 Programming language: C++. Fortran interface available, and Fortran compiler required for dependencies. Journal reference of previous version: Comput. Phys. Commun. 257 (2020) 107284 Does the new version supersede the previous version?: Yes Reasons for the new version: We have added several new features to NLOX, and made significant stability and efficiency improvements. The new features provide new types of output, and more flexibility for user input. These features are largely motivated to simplify interfacing NLOX with other automated tools, Monte Carlo and real-radiation providers in particular. Occasionally in a Monte Carlo run using NLOX instabilities are discovered in the loop results of NLOX in certain corners of phase space. While the original version had methods to detect these instabilities, and partially correct for them, the latest version has additional methods of detecting instabilities, and existing methods are improved. This version has better reliability and specificity of detection, and a higher success rate of correction. Summary of revisions: We have added new output modes. The first new mode allows the user to isolate contributing pieces in the amplitude-squared in powers of the constants αe and αs. NLOX is also now capable of producing color-correlated output for Born-level matrix elements, required for interfacing to real-radiation providers. We have added support for the Les Houches standard interface function, OLP_SetParameter, implemented as NLOX_OLP_SetParameter. This allows users to set certain parameters, for our purposes coupling constants, masses, widths, the number of light and heavy quark flavors, and the renormalization scale. Previously these were runtime constants in NLOX, requiring editing an input file, or specialized functions to set after program initialization. We have improved the numerical stability of NLOX substantially in this release. First, the stability tests performed internally by the TRed library have been improved. In particular, we have devised a better metric for determining configurations of phase space where the Gram determinant is small, and therefore the default reduction numerically unstable, by comparing the determinant, a measure of the “volume” of the matrix, to a maximal volume spanned by the vectors of the matrix. Second, we have interfaced with higher-precision external libraries where available and necessary. The thresholds for recomputation in higher precision and warning flags for internal failure have been adjusted accordingly, allowing for fewer recomputations, a higher sensitivity to phase space points giving numerically incorrect results, and more specificity to flagging points as failed to the user. Finally, we have implemented new routines that compute the IR pole structure of the real emission corresponding to the virtual routines of NLOX, as a final check of the amplitude squared. These routines are used automatically to determine an estimated accuracy of the result returned through the standard Les Houches flag acc of the user function NLOX_OLP_EvalSubProcess. We have retooled the generation of processes provided in downloadable packages to reduce expression sizes and eliminate some common structures, resulting in significantly smaller code sizes, sometimes a factor of two smaller than in the 1.0 version. For a more complete discussion of the changes in this release, please see the accompanying document in the package, nlox-1.2.0.pdf. Nature of problem: The computation of higher-order terms in the coupling expansion of Standard Model scattering amplitudes is required for precision studies in collider experiments. Techniques for computing the first corrections are well-known, and are now suited to automation. We wish to provide code that calculates virtual (one-loop) quantum chromodynamics and electroweak corrections for desired amplitudes using a package that automates the production of this code. Solution method: We use Python scripts and a computer algebra system, FORM, to reduce virtual amplitudes to C++ code and data based on Feynman rules of the Standard Model. The scripts perform a tensor decomposition of the one loop integrals to reduce each amplitude to a dependence on tensor integral coefficients. These coefficients are called at runtime by the provided library TRed, which performs tensor reduction into base (scalar) coefficients at runtime. The scripts identify repeated structures to be calculated once in the produced code for efficiency. The tensor reduction code is designed such that needed tensor coefficients need to be computed only once per evaluation of the desired amplitude, and are built recursively from other needed coefficients. Additional comments including restrictions and unusual features: The code-producing scripts are not provided in this release, only fixed libraries such as TRed and required interface code for pre-generated processes. Some processes are provided with this release, with others available upon request. • Required External Dependencies: QCDLOOP 1.95, ONELOOP 3.6 (available for download in utility tarball). • Required Compilers: gcc 4.6 or higher. Interface to certain optional libraries requires C++ 11 support found in gcc 4.7 or higher. • Operating System: Linux, MacOS.
(Less)
- author
- Figueroa, Diogenes
; Quackenbush, Seth
; Reina, Laura
and Reuschle, Christian
LU
- organization
- publishing date
- 2022-01
- type
- Contribution to journal
- publication status
- published
- subject
- keywords
- Higher-order calculations, NLO QCD and EW automation, One-loop provider
- in
- Computer Physics Communications
- volume
- 270
- article number
- 108150
- publisher
- Elsevier
- external identifiers
-
- scopus:85114384127
- ISSN
- 0010-4655
- DOI
- 10.1016/j.cpc.2021.108150
- language
- English
- LU publication?
- yes
- id
- 75a4f417-0f29-460a-a441-f27edf7da027
- date added to LUP
- 2021-10-04 11:55:36
- date last changed
- 2024-04-20 12:11:54
@article{75a4f417-0f29-460a-a441-f27edf7da027,
abstract = {{<p>In this release note we describe the 1.2 update to NLOX, a computer program for calculations in high-energy particle physics. New features since the 1.0 release and other changes are described, along with usage documentation. New version program summary: Program title: NLOX Developer's repository link: http://www.hep.fsu.edu/~nlox Licensing provisions: CC BY NC 3.0 Programming language: C++. Fortran interface available, and Fortran compiler required for dependencies. Journal reference of previous version: Comput. Phys. Commun. 257 (2020) 107284 Does the new version supersede the previous version?: Yes Reasons for the new version: We have added several new features to NLOX, and made significant stability and efficiency improvements. The new features provide new types of output, and more flexibility for user input. These features are largely motivated to simplify interfacing NLOX with other automated tools, Monte Carlo and real-radiation providers in particular. Occasionally in a Monte Carlo run using NLOX instabilities are discovered in the loop results of NLOX in certain corners of phase space. While the original version had methods to detect these instabilities, and partially correct for them, the latest version has additional methods of detecting instabilities, and existing methods are improved. This version has better reliability and specificity of detection, and a higher success rate of correction. Summary of revisions: We have added new output modes. The first new mode allows the user to isolate contributing pieces in the amplitude-squared in powers of the constants α<sub>e</sub> and α<sub>s</sub>. NLOX is also now capable of producing color-correlated output for Born-level matrix elements, required for interfacing to real-radiation providers. We have added support for the Les Houches standard interface function, OLP_SetParameter, implemented as NLOX_OLP_SetParameter. This allows users to set certain parameters, for our purposes coupling constants, masses, widths, the number of light and heavy quark flavors, and the renormalization scale. Previously these were runtime constants in NLOX, requiring editing an input file, or specialized functions to set after program initialization. We have improved the numerical stability of NLOX substantially in this release. First, the stability tests performed internally by the TRed library have been improved. In particular, we have devised a better metric for determining configurations of phase space where the Gram determinant is small, and therefore the default reduction numerically unstable, by comparing the determinant, a measure of the “volume” of the matrix, to a maximal volume spanned by the vectors of the matrix. Second, we have interfaced with higher-precision external libraries where available and necessary. The thresholds for recomputation in higher precision and warning flags for internal failure have been adjusted accordingly, allowing for fewer recomputations, a higher sensitivity to phase space points giving numerically incorrect results, and more specificity to flagging points as failed to the user. Finally, we have implemented new routines that compute the IR pole structure of the real emission corresponding to the virtual routines of NLOX, as a final check of the amplitude squared. These routines are used automatically to determine an estimated accuracy of the result returned through the standard Les Houches flag acc of the user function NLOX_OLP_EvalSubProcess. We have retooled the generation of processes provided in downloadable packages to reduce expression sizes and eliminate some common structures, resulting in significantly smaller code sizes, sometimes a factor of two smaller than in the 1.0 version. For a more complete discussion of the changes in this release, please see the accompanying document in the package, nlox-1.2.0.pdf. Nature of problem: The computation of higher-order terms in the coupling expansion of Standard Model scattering amplitudes is required for precision studies in collider experiments. Techniques for computing the first corrections are well-known, and are now suited to automation. We wish to provide code that calculates virtual (one-loop) quantum chromodynamics and electroweak corrections for desired amplitudes using a package that automates the production of this code. Solution method: We use Python scripts and a computer algebra system, FORM, to reduce virtual amplitudes to C++ code and data based on Feynman rules of the Standard Model. The scripts perform a tensor decomposition of the one loop integrals to reduce each amplitude to a dependence on tensor integral coefficients. These coefficients are called at runtime by the provided library TRed, which performs tensor reduction into base (scalar) coefficients at runtime. The scripts identify repeated structures to be calculated once in the produced code for efficiency. The tensor reduction code is designed such that needed tensor coefficients need to be computed only once per evaluation of the desired amplitude, and are built recursively from other needed coefficients. Additional comments including restrictions and unusual features: The code-producing scripts are not provided in this release, only fixed libraries such as TRed and required interface code for pre-generated processes. Some processes are provided with this release, with others available upon request. • Required External Dependencies: QCDLOOP 1.95, ONELOOP 3.6 (available for download in utility tarball). • Required Compilers: gcc 4.6 or higher. Interface to certain optional libraries requires C++ 11 support found in gcc 4.7 or higher. • Operating System: Linux, MacOS.</p>}},
author = {{Figueroa, Diogenes and Quackenbush, Seth and Reina, Laura and Reuschle, Christian}},
issn = {{0010-4655}},
keywords = {{Higher-order calculations; NLO QCD and EW automation; One-loop provider}},
language = {{eng}},
publisher = {{Elsevier}},
series = {{Computer Physics Communications}},
title = {{Updates to the one-loop provider NLOX}},
url = {{http://dx.doi.org/10.1016/j.cpc.2021.108150}},
doi = {{10.1016/j.cpc.2021.108150}},
volume = {{270}},
year = {{2022}},
}