Multi-threaded adaptive extrapolation procedure for Feynman loop integrals in the physical region

24th IUPAP Conference on Computational Physics

物理学;计算机科学

De Doncker, E.^1 ; Yuasa, F.^2 ; Assaf, R.^1

Department of Computer Science, Western Michigan University, Kalamazoo MI 49008, United States^1

High Energy Accelerator Research Organization (KEK), Oho 1-1, Ibaraki, 305-0801, Tsukuba, Japan^2

关键词: Adaptive extrapolations;    Adaptive integration;    Adaptive procedure;    Feynman loop integrals;    Iterated integrals;    Quantum field theory;    Sequential methods;    Shared memory parallelization;   

Others  :  https://iopscience.iop.org/article/10.1088/1742-6596/454/1/012082/pdf
DOI  :  10.1088/1742-6596/454/1/012082

学科分类：计算机科学（综合）

来源: IOP

【 摘 要 】

Feynman loop integrals appear in higher order corrections of interaction cross section calculations in perturbative quantum field theory. The integrals are computationally intensive especially in view of singularities which may occur within the integration domain. For the treatment of threshold and infrared singularities we developed techniques using iterated (repeated) adaptive integration and extrapolation. In this paper we describe a shared memory parallelization and its application to one- and two-loop problems, by multi-threading in the outer integrations of the iterated integral. The implementation is layered over OpenMP and retains the adaptive procedure of the sequential method exactly. We give performance results for loop integrals associated with various types of diagrams including one-loop box, pentagon, two-loop self-energy and two-loop vertex diagrams.