Thank you to Alvin and the team for this wonderful news. I hope that the GPU extension for Windows Admin Center is fully functional and tested, to make it easy to adopt this feature without hassle.
Unfortunately my preferred OEM doesn't offer GPU Support on NVMe Intel based solution, unlike their AMD solution, due technical restrictions of Intel (too few PCIe lanes), but this should be adressed with the next Generation of Intel Server CPUs.
Allow me some in detail questions on this:
1. since Azure Stack HCI supports GPU Pooling, can VMs with assigned GPU-P be live migrated, or isn't this possible due SR-IOV
2. How easy is it to update the Nvidia drivers? Do we need to update the Hyper-V Hosts and VM drivers, similar to VMware, or is it sufficient to only Update VMs with Nvidia drivers (feature and security fixes) ?
3. How is the licensing for Nvidia on Azure Stack HCI? On VMware additional licensing from Nvidia is needed and this also needs a licensing server (Windows VM) to handle the licensing.
4. What's the minimum Citrix version to support GPU-P on Azure Stack HCI with Azure Virtual Desktop?
5. Will the Azure Stack HCI size be updated to reflect GPU / GPU-P support, so one get proper with different workload profiles + User amount and scaling.
Examples:
- Office / Teams / Edge Browser / Video playback.
- CAD/ CAM, video editing.
- ML
and GPU sizing / PSU sizing).
Thank you for your answers in advance!