A novel Cloud High Performance Computing infrastructure for user
summary
Cloud HPC runs OpenStack on a partition of the Galileo100 Tier‑1 system, using Intel Platinum Xeon 8260 nodes with ~700 GB RAM per node, a 100 Gb/s low‑latency network and 1 PB CEPH NVMe/SSD storage. Deployment uses Ansible/Crowbar-style automation with three controller and ~68 compute hosts, enforces no resource oversubscription, and exposes Barbican and Magnum for encryption and container orchestration. Production workloads include the COSMO weather model feeding metab.mraportal and GDPR‑compliant genomic and federated‑learning pipelines that scale workers on demand and run end‑to‑end in the cloud.
AI-written summary · about
chapters (5)
- 0:00Introduction and project overview
- 3:55Cloud infrastructure Galileo 100
- 8:40Use cases in weather and genomics
- 13:52COVID-19 data analysis platform
- 15:17Summary and Q&A
related talks
15mThe CERN Cloud Infrastructure: 5 years in 15 minutes
36mIntroduction to High Performance Computing on Google Cloud Platform
47mCloud Native and Sustainable, Reproducible Scientific Computing
43mSupporting NASA Science with OpenStack
21mKeynote: CERN Experiences - Ricardo Rocha & Clenimar Filemon
42mBuilding a green supercomputer in the cloud
15mAn open cloud for european researchers, staff and students
10mLightning Talk: Does Cloud Elasticity Pay off for HPC Workloads?