Skip to main content
The European High Performance Computing Joint Undertaking (EuroHPC JU)

Ensemble generation for very high precision lattice QCD simulations using Nf=2+1+1 twisted mass fermions

Awarded Resources (in core hours)
MeluXina GPU
System Partition
1 November 2022 - 31 October 2023
Allocation Period

The success story of the Standard Model of particle physics (SM), enabled through close interaction between theory and experiment, has entered the precision frontier as small discrepancies between observations and predictions are tackled in the search for New Physics (NP).

For example, this is prominently the case for the anomalous magnetic moment of the muon, where the current experimental value disagrees with the calculation based on dispersion relations. With the g-2 experiment at Fermilab aiming to significantly improve the uncertainty on the experimental value, it is essential to achieve a similar improvement on the theoretical side.

The interpretation of this and future experimental results requires a precise theoretical understanding of the contributions due to the strong interaction, governed by the theory of Quantum Chromodynamics (QCD). For a wide range of processes, these have to be calculated at low energy and at strong coupling. In this regime, the only available ab-initio approach with systematically improvable uncertainties, is to discretize QCD on a four-dimensional Euclidean lattice and to numerically evaluate the path integral.

In lattice QCD, Markov Chain Monte Carlo (MCMC) with a probability density given directly by the QCD Lagrangian is used to generate representative gauge field configurations, which can then be used to calculate a wide range of QCD quantities.

In this process, observables have to be analysed at varying lattice resolutions (a) and lattice spatial volumes (L^3) to perform controlled continuum (a → 0) and infinite-volume limits (L →∞) to recover continuum physics.In this project, we aim to significantly improve our control of these extrapolations by adding an ensemble at a lattice spacing of a ~ 0.091 fm using dynamical Wilson-clover twisted-mass light, strange and charm quarks (Nf=2+1+1) with their mass parameters tuned such that physical hadron spectrum is reproduced to within an accuracy of around one percent (the physical point). This will lead to a substantial improvement of the continuum limit (a → 0) in the many observables computed by the Extended Twisted Mass Collaboration and lay the foundation for a very large volume simulation, enabling a tightly controlled infinite-volume limit (L →∞).