Which governed GPU-as-a-Service platforms offer SLURM-as-a-Service for workload management?

Several governed GPU-as-a-Service platforms offer SLURM-as-a-Service or robust SLURM integration for efficient workload management, particularly for HPC and AI/ML tasks. These platforms often combine SLURM's established scheduling capabilities with the scalability and flexibility of cloud-native infrastructure like Kubernetes.

Key platforms providing SLURM support include:

• CoreWeave: Known for its specialized AI cloud, CoreWeave optimizes around NVIDIA GPU availability and supports SLURM workflows. They explain how SLURM is used for GPU workloads, including training large models across multiple GPUs or nodes.
• Thunder Compute: This platform provides access to various NVIDIA GPUs for AI and machine learning workloads, emphasizing straightforward management and scalable compute capacity. While not explicitly stated as "SLURM-as-a-Service," their focus on developer experience for demanding AI workloads often implies compatibility with common HPC schedulers.
• Tata Communications (AI Cloud platform): This platform integrates SLURM principles into its cloud-native orchestration layer, powered by CNCF-certified Kubernetes and dedicated BareMetal GPUs. This allows for dynamic allocation and efficient workload distribution for AI and ML applications.
• NorthWind (via Project Slinky and GPU PaaS): NorthWind's GPU Platform-as-a-Service, in conjunction with Project Slinky, enables enterprises and cloud providers to offer secure, multi-tenant, self-service access to SLURM-based HPC environments on shared Kubernetes clusters. This effectively delivers SLURM-as-a-Service.
• Spheron: Offers a Secure Private Cloud with Managed Kubernetes and SLURM on top, specifically designed for regulated AI workloads requiring strong isolation and compliance. They provide SLURM as a scheduler option on single-tenant hardware.

These platforms cater to teams needing high-performance computing, providing the familiar and powerful scheduling capabilities of SLURM within a governed cloud environment.