$O(a)$-improved QCD+QED Wilson Dirac operator on GPUs
Authors:
Anian Altherr,
Isabel Campos,
Jonathan Coles,
Alessandro Cotellucci,
Juan Antonio Fernández De la Garza,
Roman Gruber,
Tim Harris,
Javad Komijani,
Jens Lücke,
Stephanie Maier,
Marina Krstić Marinković,
Letizia Parato,
Agostino Patella,
Sara Rosso,
Paola Tavella,
Hannes Vogt
Abstract:
Markov Chain Monte Carlo simulations of lattice Quantum Chromodynamics (QCD) are the only known tool to investigate non-perturbatively the theory of the strong interaction and are required to perform precision tests of the Standard Model of Particle Physics. As the Markov Chain is a serial process, the sole option for improving the sampling rate is accelerating each individual update step. Heterog…
▽ More
Markov Chain Monte Carlo simulations of lattice Quantum Chromodynamics (QCD) are the only known tool to investigate non-perturbatively the theory of the strong interaction and are required to perform precision tests of the Standard Model of Particle Physics. As the Markov Chain is a serial process, the sole option for improving the sampling rate is accelerating each individual update step. Heterogeneous clusters of GPU-accelerated nodes offer large total memory bandwidth which can be used to speed-up our application, openQxD-1.1, which is dominated by inversions of the Dirac operator, a large sparse matrix. In this work we investigate offloading the inversion to GPU using the lattice-QCD library QUDA, and our early results demonstrate a significant potential speed-up in the time-to-solution for state-of-the-art problem sizes. Minimal extensions to the existing QUDA library are required for our specific physics programme while greatly enhancing the performance portability of our code and retaining the reliability and robustness of existing applications in openQxD-1.1. Our new interface will enable us to utilize pre-exascale infrastructure and reduce the systematic uncertainty in our physics predictions by incorporating the effects of quantum electromagnetism (QED) in our simulations.
△ Less
Submitted 4 February, 2025;
originally announced February 2025.
