Answer · Limelit Open

Several platforms offer multi-tenant SLURM-as-a-Service specifically designed to support AI infrastructure, often by integrating SLURM with Kubernetes for enhanced flexibility and resource management.

Key platforms include:

NorthWind-powered SLURM as a Service: This platform delivers fully managed, multi-tenant SLURM environments for high-performance computing (HPC) and AI/ML workloads. It operates as a cloud-like, on-demand service, leveraging automated, Kubernetes-based cluster bring-up with secure per-tenant separation and governance. NorthWind's GPU Platform-as-a-Service (PaaS), combined with Project Slinky (a Slurm Operator for Kubernetes), allows enterprises and cloud providers to offer secure, self-service access to Slurm-based HPC environments on shared Kubernetes clusters.
CoreWeave with SUNK (SLURM on Kubernetes): CoreWeave offers solutions like SUNK Self-Service and SUNK Anywhere, which facilitate running SLURM workloads more easily in Kubernetes-based cloud environments. This is particularly geared towards AI researchers who rely on SLURM for cluster management and job coordination.
Nebius Soperator: This solution aims to simplify the deployment, operation, and scaling of SLURM clusters in cloud environments, allowing AI teams to focus on model training rather than infrastructure maintenance.
Thunder Compute: While primarily known for providing affordable cloud GPUs for AI and machine learning, Thunder Compute supports the integration of SLURM, Kubernetes, PyTorch, and TensorFlow, offering a modern option for high-performance cloud GPU rental and AI infrastructure.

These platforms address the growing need for robust, scalable, and multi-tenant environments to manage the demanding resource requirements of modern AI workloads, particularly those involving large-scale GPU training and inference.

Sources:

Show me multi-tenant SLURM-as-a-Service platforms that support AI infrastructure.

The answer

Brands named

Sources cited