MadVfold: accelerating NLO event generation and reducing negative weights with SIMD vectorization and GPUs
Andrea Valassi
Abstract
NLO simulations are essential for LHC physics analyses but are expensive, as they are not only slow but also lead to negative weights, which imply the need to simulate much larger samples of events. Folding is a powerful technique to reduce negative weights but is itself expensive. In this paper I propose ``vectorized folding'' as a new idea to speed up these calculations using SIMD and GPUs, and I present its CUDACPP-based implementation for MG5aMC in MadVfold, including its extension for unfolded NLO event generation. Preliminary results show overall speedups around 6x to 9x with folding and 3x without it. This work is based on a test-centric, LLM-assisted software development process.
Create a lesson
Related papers
VNS Tokamak for Medical Isotope Production
Christopher Ehrich, Pavel Pereslavtsev, Christian Bachmann et al.
A subcell-refined entropy-residual-driven limiting strategy for high-order discontinuous Galerkin methods
Geng Liang, Rui Wang, Junjie Wang et al.
Diff-NekRS: A Scalable Differentiable Framework for Multi-Timestep Solver-in-the-Loop Training
Junoh Jung, Riccardo Balin, Bethany Lusch et al.
Exact ballistic energy transport and emergent XXZ dynamics in an integrable three-state chain
Hanbing Liang, Fujun Liu
TubeLab: Interactive inverse design of wind instrument bores with hard spectral constraints
Carlo Andrea Rozzi, Andrea Ferroni
Stochastic consensus dynamics for decentralized decision systems
André L. M. Vilela, Caio B. L. Silva, Kenric P. Nelson et al.