DynamiaDynamia
COMMUNITY

Build together in the open.
Make compute work better.

Discover HAMi, learn how Dynamia supports the community and connect with people solving similar problems.

01 / HAMi

What is HAMi?

HAMi is open-source middleware for heterogeneous GPU virtualization and sharing. Its lightweight, non-intrusive deployment works across public, private and hybrid clouds, enabling flexible, reliable, on-demand GPU sharing.

HAMi improves device utilization through GPU memory and compute sharing, resource isolation and scheduling. It supports heterogeneous chips including NVIDIA, Ascend, MetaX, Cambricon, Hygon, Moore Threads and Iluvatar CoreX, integrating with existing Kubernetes infrastructure.

HAMi open ecosystem diagram showing AI workloads, cloud-native schedulers, heterogeneous accelerators and runtime observability.
02 / DYNAMIA × HAMi

How are Dynamia and HAMi connected?

Dynamia was founded by the creators and core maintainers of HAMi, a CNCF project, to build the open-source core of AI infrastructure.

We help the AI team to make the Heterogeneous GPU and the AI chip resources more accessible, efficient and manageable, turning complex compute resourcees into shared, commotion, demand-based infrastructure capacity.

About Dynamia
03 / STAY CONNECTED

Where can you find us?

Watch technical talks, share practical experience and bring your questions and ideas to the community.

Image viewer