dstack vs Moab HPC Suite
dstack
dstack.aiMoab HPC Suite
adaptivecomputing.comSide by side
| Figure | dstack | Moab HPC Suite |
|---|---|---|
| Price from | Free | — |
| Free plan | Yes | — |
| Free trial | — | — |
| Open source | — | — |
| Web | Yes | — |
| Windows | Yes | — |
| macOS | Yes | — |
| Linux | Yes | Yes |
| API | Yes | Yes |
| Self-hosted | Yes | Yes |
| Plans published | 2 | 1 |
| Spec score · GPU Cluster Management Software | 8.9 (#1) | 7.2 (#14) |
| API | dstack offers an HTTP API for functionality not available in the CLI and for integrations that need to call the server directly. | Moab Web Services provides a RESTful interface for programmatic interaction with Moab Workload Manager and its jobs, nodes, virtual machines, and reservations. |
| Company | The terms identify dstack Inc. as a Delaware corporation with offices in Dover, Delaware, United States. | Adaptive Computing says it was founded in 2001 and lists its address as 1100 5th Ave S., Suite 201, Naples, Florida 34102, United States. |
| Deployment | The server can run on a laptop or another environment with access to the cloud and on-prem clusters being used. | The suite documentation describes manual and RPM installation approaches and recommends testing upgrades in an environment that mirrors production. |
| Purpose | dstack is an open-source orchestration layer for AI workloads on heterogeneous accelerators, including GPU clouds, Kubernetes, VMs, and bare-metal clusters. | Moab HPC Suite optimizes resource utilization and job scheduling across high-performance computing infrastructure. |
| Support | The dstack Sky page directs users with questions or needing help to Discord or to contact the company directly. | Customers with a support services contract can submit unresolved support cases by email for help from a technical support specialist. |
Plans
In their own words
dstack
dstack is an open-source orchestration layer for AI workloads across GPU clouds, Kubernetes, virtual machines, and bare-metal clusters. YAML configurations define fleets, development environments, tasks, services, presets, and volumes, while dstack handles provisioning and job scheduling, including auto-scaling, port forwarding, and ingress. Supported accelerators include NVIDIA, AMD, TPU, and Tenstorrent. Documented backends include AWS, Azure, GCP, Kubernetes, GPU cloud providers, remote SSH hosts, and experimental Slurm. Tasks can use frameworks such as accelerate, torchrun, Ray, and Spark. Users manage resources through the CLI or HTTP API, and the server can run on a laptop or another environment that can access the clusters in use. Services can publish inference endpoints through gateways with HTTPS, custom domains, auto-scaling, and rate limits. The self-hosted open-source stack is free. Server data and project secrets are stored in plaintext by default unless administrators configure AES-256-GCM encryption. TPU support is limited to single-host instances with up to eight cores. GPU Marketplace usage is prepaid, with prices that vary by GPU and provider and appear in the console before provisioning.
Moab HPC Suite
Moab HPC Suite is self-hosted software for scheduling workloads and managing resources across high-performance computing infrastructure. It runs on Linux and exposes an API. Scheduling algorithms place jobs and assign resources to improve cluster utilization and reduce waiting. Automatic workload balancing and priority management direct capacity toward critical jobs. Administrators can define scheduling, allocation, and user-access policies, including quota controls. The suite provides cluster-performance and resource-use reporting, GPU utilization metrics, and management of multiple HPC clusters through one interface with unified monitoring and control. Its high-availability architecture includes failover capabilities for continuous operation of critical workloads. Enterprise Edition components include Moab Workload Manager, Moab Web Services, Moab Accounting Manager, and TORQUE Resource Manager. Moab Web Services provides REST access to jobs, nodes, virtual machines, and reservations. Documentation lists TORQUE 5.0.1 and SLURM support, manual and RPM installation, and guides for Red Hat, SUSE, and Ubuntu-based systems. Pricing is available on request.