As CERN’s Large Hadron Collider (LHC) prepares to restart this March, a team of Fermilab physicists are using powerful Department of Energy supercomputing resources to reduce the risks and costs associated with producing these high intensity particle beams.
Led by Fermilab physicist James Amundson, the team is working with the Argonne Leadership Computing Facility (ALCF) to port and optimize Synergia, the accelerator simulation software package developed at Fermilab, to run on ALCF’s Mira supercomputer. The hybrid Python code harnessed 100,000 cores on Mira, enabling researchers to simulate complex internal accelerator interactions. The team is especially interested in studying the effects that accelerator components exert on particles inside high-intensity, low-energy machines.
“The more particles you have, the more intensity you have, the more pronounced the effects become,” says Amundson. “So, since we want to move toward higher intensities, it’s important for us to understand these intensity-dependent effects which, of course, are the things that are computationally difficult to do.”
Thanks to the 10-petaflops Mira supercomputer, one of the world’s fastest, researchers identified an instability issue with the potential to jeopardize the operation of the beam.
This insight has important implications for Fermilab and for the next phase of CERN’s LHC program, which will be upping the intensity of experiments. It is also relevant for the upcoming high-intensity beam project called the Fermilab Proton Improvement Plan II, which is tasked with creating neutrino beams for the Long Baseline Neutrino Facility.