The Hadronic Matter Group has been awarded 11.8 million CPU hours for the JETSCAPE project on the CSD3 system (High Performance Computer at the University of Cambridge) via the STFC IRIS e-infrastructure programme (https://www.iris.ac.uk/). This award, led by Jaime Norman as PI, is for 2025-2026 and up significantly from only 500k hours secured in the previous year to perform benchmark calculations. A follow up request for 2026-2027 has been submitted recently.
The Liverpool group is an associate member of the JETSCAPE Collaboration in the USA (https://jetscape.org/), an interdisciplinary team of physicists, computer scientists, and statisticians from several institutions, which develops and runs a comprehensive software framework – JETSCAPE – a modular, multi-stage Monte Carlo event generator used to simulate ultra-relativistic heavy-ion collisions.
These simulations are compared with data from the Relativistic Heavy-Ion Collider at Brookhaven and the Large Hadron Collider at CERN via Bayesian inference to extract the transport and thermodynamic properties of the hot state of QCD matter created in these collisions – the Quark-Gluon Plasma (QGP). To determine properties of the QGP, JETSCAPE model parameters are constrained via Bayesian inference, where simulations of heavy-ion collisions are run varying model parameters to sample the parameter space (where one parameter set corresponds to a ‘design point’). Direct evaluation of the model through parameter space requires sampling multiple design points, which requires significant computing resources.
Two broad categories of simulations are performed, bulk-evolution (hydrodynamics) and jet-propagation simulations. The former are used to constrain global QGP properties (e.g., shear viscosity, initial temperature) while the latter are used to constrain energy-loss parameters of high energy quarks or gluons, collectively known as ‘jets’, which propagate through the QGP. To fully capture correlations and allow feedback between the two stages of these calculations, future analyses will run both parts of the analysis concurrently. The use of these allocated computing resources will be the first time that a full Bayesian analysis will be performed in this setup.
For further details on JETSCAPE see the recent paper, including Jaime Norman and Roy Lemmon as authors:
Bayesian inference analysis of jet quenching using inclusive jet and hadron suppression measurements, published PHYSICAL REVIEW C 111, 054913 (2025).
