Workshops Archive

AI for Science with Python on AMD Supercomputers


Workshop: High Performance Python for Science at Scale

Authors: Giacomo Capodaglio, Nicholas Malaya, Jose Noudohouenou, Sidafa Conde, Alessandro Fanfarillo, and Jonathan Belof (Advanced Micro Devices (AMD))

Abstract: Running Python workloads at scale is becoming increasingly important in the scientific landscape of today, especially given the rise of AI as a partner in scientific development and discovery. We will show various ways of deploying AI workloads (including AI agents) at scale running on AMD GPUs: our focus will be on AI for science on heterogeneous types of systems with variety of deployments (i.e. wheels and containers) and job schedulers (Slurm and Kubernetes). We will provide a walkthrough of AMD Inference Microservices (AIM) and how this inference engine can be used by scientists to deploy their Python AI workloads.


Back to High Performance Python for Science at Scale Archive Listing

Back to Full Workshop Archive Listing