Dynamia Enterprise for HAMi
For teams with existing Kubernetes / PaaS: enhanced GPU virtualization, production hardening and enterprise support built on open-source HAMi.
Compare Dynamia Enterprise for HAMi and Dynamia AI Platform for HAMi.
Choose your path to production-grade heterogeneous compute.

HAMi is a CNCF Incubating project for heterogeneous AI compute virtualization, initiated and maintained by Dynamia’s founding team. Enterprise products build on open-source HAMi with production hardening, operations and enterprise support.
Explore our open-source communityEnhance an existing Kubernetes platform or build unified heterogeneous compute management.
For teams with existing Kubernetes / PaaS: enhanced GPU virtualization, production hardening and enterprise support built on open-source HAMi.
Built on Dynamia Enterprise for HAMi, add a multi-cluster GUI, observability, tenant quotas, metering, billing and OpenAPI to manage clusters, resources and teams together.
Compare open-source HAMi, Dynamia Enterprise for HAMi and Dynamia AI Platform for HAMi against your current needs.
| Product capability | Open-source HAMi | Dynamia Enterprise for HAMi | Dynamia AI Platform for HAMi |
|---|---|---|---|
| GPU Virtualization Core | |||
| GPU memory sharing | ✓ | ✓ | ✓ |
| GPU compute sharing | ✓ | ✓ | ✓ |
| Memory hard isolation | — | ✓ | ✓ |
| Compute hard isolation | — | ✓ | ✓ |
| Memory & compute oversubscription | — | ✓ | ✓ |
| HAMi Turbo mode | — | ✓ | ✓ |
| Auto-scaling without restart | — | ✓ | ✓ |
| Task priority & preemption | ✓ | ✓ | ✓ |
| Idle-time analytics & memory profiling | — | ✓ | ✓ |
| Optimized scheduler engine (parallel) | — | ✓ | ✓ |
| Binpack / Spread strategies | ✓ | ✓ | ✓ |
| NVIDIA topology-aware affinity | ✓ | ✓ | ✓ |
| Volcano deep integration (preemption / vGPU joint) | — | ✓ | ✓ |
| Heterogeneous device resource pools | — | — | ✓ |
| Heterogeneous GPU Vendor Support | |||
| NVIDIA (full lineup) | ✓ | ✓ | ✓ |
| AMD | ✓ | ✓ | ✓ |
| Ascend (full lineup) | ✓ | ✓ | ✓ |
| MetaX | ✓ | ✓ | ✓ |
| Kunlunxin | ✓ | ✓ | — |
| Hygon | ✓ | ✓ | ✓ |
| Iluvatar (full lineup) | ✓ | ✓ | ✓ |
| Cambricon | ✓ | ✓ | ✓ |
| Enflame | ✓ | ✓ | ✓ |
| Moore Threads | ✓ | ✓ | ✓ |
| Alibaba T-Head (full lineup) | — | ✓ | ✓ |
| Multi-cluster Control Plane | |||
| Multi-cluster observability (cluster / app / GPU) | — | — | ✓ |
| GPU hardware monitoring | — | — | ✓ |
| Comprehensive alert policies | — | ✓ | ✓ |
| Turn-key GUI (multi-cluster dashboard, quota, workload) | — | — | ✓ |
| Visual Binpack / Spread strategy editor | — | — | ✓ |
| Multi-cluster lifecycle management (LCM) | — | — | ✓ |
| One-click trial model deployment | — | — | ✓ |
| Enterprise Operations | |||
| Enterprise multi-tenant management | — | — | ✓ |
| Metering & billing | — | — | ✓ |
| OpenAPI & automation | — | — | ✓ |
| Visual license management | — | — | ✓ |
| OEM white-label branding | — | — | ✓ |
| SLA-backed support (24/7) | — | ✓ | ✓ |
| Air-gap / offline deployment | — | ✓ | ✓ |
The capability matrix follows Dynamia’s product overview. Specific chip models, enhancements and delivery scope depend on the selected version and confirmed compatibility.
Read public user case studies to explore how different businesses use the technology.
Tell us about your devices, clusters and business goals to identify the right product, deployment approach and validation path.