Hello,
I am evaluating the Jetson AGX Orin for use in a distributed compute application involving multiple customers.
Specifically, I would like to determine whether it is possible to partition the underlying hardware resources (i.e., CPU, GPU, and storage), so that multiple customers can have their own isolated workloads running on the same physical Jetson instance. So far, my research suggests two techniques for achieving multi-tenancy that may apply here: (1) VMs, and (2) containers. However, it’s still very unclear to me what kind of software and hardware support are required to enable these technologies.
To this end, I am wondering:
- Does NVIDIA support any hypervisors on the Jetson AGX Orin?
- Does the Jetson AGX Orin have hardware virtualization support?
- Does the Jetson AGX Orin support partitioning its GPU (i.e., multi-instance GPUs/MIGs)?
- Is there any other NVIDIA hardware that you think would be more suitable for this application?
I appreciate your help.