Скачать или смотреть HPC Containers on ALPS

Content
Linux containers provide a convenient way to package applications along with their dependencies into portable, self-contained units that can be reliably deployed across different computing systems. On the Alps Research Infrastructure, containerized HPC workflows are supported through the Container Engine toolset, which integrates containers seamlessly into the platform while maintaining scalability, performance, and compatibility with standard container formats.

This webinar introduces the Container Engine on Alps and walks participants through key capabilities, including building images directly on compute nodes, running containerized workloads through the Slurm Workload Manager, and harnessing advanced hardware features like GPUs and the Slingshot interconnect. We will also share recommended practices and guidelines for use cases that rely on MPI, NCCL, and other communication libraries.

Trainer
Alberto Madonna (Senior Software Engineer, CSCS)

Target audience
Members of the CSCS User Community