AMFlow 2.0: significant algorithmic and software improvements for Feynman integral evaluation

Abstract

We present significant improvements to the AMFlow package for the numerical computation of dimensionally regularized Feynman integrals. Several new features are introduced to reduce computational cost, including an alternative recursion mode, a high-performance differential equation solver, support for state-of-the-art integration-by-parts reducers and other useful improvements. We benchmark the new version on a three-loop five-point topology and find that both the symbolic and numerical performance are significantly improved.