High-Performance Computing for Artificial Intelligence at the University of Turin
We are open for collaborations in research projects (EU, National, Regional). UNITO is currently running 125+ EU H2020 projects (27 coordinated, 7 ERC) The Computer Science Department, with its 80+ professors, covers the major research areas: HPC, AI (Deep/Federated Learning, Reasoning, Perception, …), BigData, UI/UX, Blockchain, Algorithms, Edge/IoT.
HPC4AI implements a cloudifyed HPC system targeting AI workloads.
Designed for research, allowing end-to-end configuration options: from HW to OS to cloud to service. The portfolio of services includes: VMs+Storage, virtual clusters with access to OpenStack console (IaaS & PaaS).
Advanced Cloud Services
Private Kubernetes instances, hybrid cloud-HPC workflow management, cloud bursting, private segregated virtual cluster with full admin privileges.
The insfracture is eligible to receive Piedmont region voucher program supporting research costs up to 70%.
Tier-III datacentre with 2000+ CPU cores and 80+ V100/T4 GPUs
A 250KVA Tier-III adiabatic green datacenter with 16 racks, Intel, 40-cores compute nodes with 4 GPUs per node (T4 or V100-SMX2), three storage classes (300TB SSD hyper-converged, 500TB hybrid multi-tenant HW-encrypted scale-out NAS, backup), 25-100GB/s SDN networking.
Multi-tenant hybrid edge-cloud-HPC system
A novel hybrid cloud system operated with UNITO’s “secure HPC tenant” technologies, which is based on research products such as Deployment-as-a-service and Streamflow cross-cloud workflow management system. No lock-in thanks to segregated privates kubernetes instances.
15+ Academic and Industrial Research Projects
Edge-Cloud engineering, Federated Learning, Genetics, Medical imagining, Radiomics, Oncology, Complex networks, Astrophysics, Social sciences, Gaming, …
Designed for research on HPC & AI
Flexibility-by-design at all level of the software stack. Possibility to experiment at Metal/Infrastructure/ Platform/Service/Deployment-as-Service levels. Configured by HPC people for AI workloads.