This intermediate-level training will discuss two topics in depth. First, we’ll discuss how the Slurm scheduler works and how to use it effectively to get youx`r jobs to run as quickly as possible, including jobs that use parallelized code and jobs that use GPUs. Second, We’ll discuss the use of Conda (and Mamba) to install software (including, but not limited to, Python packages) in a reproducible and isolated fashion. For both topics we’ll cover a variety of tips and tricks that will help you better use Savio and similar HPC systems.
This hybrid event will be held in person in 117 Dwinelle Hall (Academic Innovation Studio) and also on Zoom. Snacks/drinks will be available afterwards for those attending in-person, and attendees will be able to talk with Savio user consultants in person afterwards.