Accelerating Research with GPU on High Performance Computing Cluster (HPC)

Hands-on tutorials to help build skills required in Precision Child Health, data science and bioinformatics. The training is prepared for staff and trainees at SickKids and beyond by the Centre for Computational Medicine (CCM).

This is a hybrid event offered both in-person and online.

Topics include:

• High Performance Computing at SickKids
• Introduction to GPUs and their applications
• How to access GPUs available in the HPC via Slurm
• Managing and monitoring GPU resources
• Using GPUs to accelerate training of deep learning models in PyTorch

This 1-hour tutorial will cover the essentials of using GPUs within the SickKids HPC environment. Participants will learn how to set up and run GPU-accelerated applications on the HPC and manage GPU resources effectively. The session will also provide tips on troubleshooting common issues and best practices for maximizing GPU utilization. Participants who have access to the SickKids HPC can follow along with the live demo.

Sponsors

This course is supported by:

• Centre for Computational Medicine, SickKids
• Digital Research Alliance of Canada and Compute Ontario