HPC Compute Service
Opis
This service provides high-performance CPU-based compute nodes optimized for parallel processing, batch execution, and large-scale scientific and AI workloads. It is intended for users who need reliable access to managed compute infrastructure for simulation, data processing, model development, and other CPU-intensive tasks without having to operate the underlying systems themselves. The service is based on a managed HPC environment with Linux compute nodes, a batch scheduler and queuing system, user access management, workload monitoring, and performance-tuning support. The platform is designed for parallel execution and integration with the wider SLAIF compute and storage environment.
Pričakovani rezultati
Efficient execution of large-scale CPU-intensive simulations, data processing pipelines, and AI workloads with predictable performance, managed access, and operational support.
Kartica storitve
| Kategorija | Infrastructure and data-related services |
| Domena | HPC and AI Compute Infrastructure |
| Vertikala | Cross-domain |
| Ciljni uporabniki | AI Researchers, Data Scientists, HPC Users, SMEs, and Startups |
| Tip storitve | Infrastructure Service / Managed |
| Rezultat | Managed access to HPC CPU compute queues and execution environments for batch and parallel jobs. |
| Zahteve | Project registration, approved access credentials, and workload preparation for CPU-based HPC execution. |
| Omejitve | Best suited for CPU-centric workloads; GPU-dependent training or inference workloads should use the HPC Accelerated Compute Service. |
| Na voljo od | Available (new system Q3/2027) |
| Licenca | |
| Slika | |
| Oznaka slike |