The SubMIT team recently held its monthly users meeting in May with updates from both the project team and user community.

SubMIT Overview & News
Marianne Moore, our project team member, summarized recent upgrades of HTCondor, Slurm, Ceph storage, and the AlmaLinux operating system. She shared news about our monitoring of job efficiencies on HTCondor and Slurm. She also talked about our new versions of A2rchi, the AI agent to suggest solutions for user requests.

Open discussion
In a following discussion session, a question about using SubMIT in education for lectures and exercises was raised. SubMIT is very well suited for this and has successfully been used in Physics 802 and Astrophysics classes.

Users talk
Alexander Avdoshkin, a member of the SubMIT project team and a theoretical physicist working on quantum many-body systems, condensed matter, and machine learning methods for physics, presented recent work done using SubMIT, introducing a machine-learning-based framework to study interacting fermionic systems with spin. The method combines variational Monte Carlo with transformer-based neural networks to efficiently model complex quantum states with both spatial and spin degrees of freedom, enabling applications to systems such as electron gases and quantum magnets.

You can find slides and video recordings of the talks on indico.