Speeding up Madgraph5 aMC@NLO through CPU vectorization and GPU offloading: towards a first alpha release
Abstract
The matrix element (ME) calculation in any Monte Carlo physics event generator is an ideal fit for implementing data parallelism with lockstep processing on GPUs and vector CPUs. For complex physics processes where the ME calculation is the computational bottleneck of event generation workflows, this can lead to large overall speedups by efficiently exploiting these hardware architectures, which are now largely underutilized in HEP. In this paper, we present the status of our work on the reengineering of the Madgraph5_aMC@NLO event generator at the time of the ACAT2022 conference. The progress achieved since our previous publication in the ICHEP2022 proceedings is discussed, for our implementations of the ME calculations in vectorized C++, in CUDA and in the SYCL framework, as well as in their integration into the existing MadEvent framework. The outlook towards a first alpha release of the software supporting QCD LO processes usable by the LHC experiments is also discussed.
Keywords
Cite
@article{arxiv.2303.18244,
title = {Speeding up Madgraph5 aMC@NLO through CPU vectorization and GPU offloading: towards a first alpha release},
author = {Andrea Valassi and Taylor Childers and Laurence Field and Stephan Hageböck and Walter Hopkins and Olivier Mattelaer and Nathan Nichols and Stefan Roiser and David Smith and Jorgen Teig and Carl Vuosalo and Zenny Wettersten},
journal= {arXiv preprint arXiv:2303.18244},
year = {2023}
}
Comments
8 pages, 4 figures, 4 tables; submitted to ACAT 2022 proceedings in IOP; version 2 includes additional references and figure caption fixes as suggested by journal referee