NVIDIA
Hypervisor selection
Dear Community, would like to hear your vision on the matter of hypervisor selection if multiple VMs are needed (up to 8) with 2Mb vGPU per each. I was doing some exploration in regards to available hypervisor features paired with grid and found out that WMware vSphere is limited to only 2 options like vSGA with max 512Mb vGPU and vDGA for full GPU capacity pass through. At the same time Citrix Xen appears to be more flexible in terms of vGPU allocation to virtual machines. The question is whether my direction is right or I can consider some other options? Thnx!
Dear Community,

would like to hear your vision on the matter of hypervisor selection if multiple VMs are needed (up to 8) with 2Mb vGPU per each.
I was doing some exploration in regards to available hypervisor features paired with grid and found out that WMware vSphere is limited to only 2 options like vSGA with max 512Mb vGPU and vDGA for full GPU capacity pass through.
At the same time Citrix Xen appears to be more flexible in terms of vGPU allocation to virtual machines.
The question is whether my direction is right or I can consider some other options?
Thnx!

#1
Posted 01/17/2017 12:53 PM   
Hi ondskan_, This sounds out of date / wrong: vSGA: only VMware - API intercept so not terribly performant and limite in OpenGL/DX support but no limit on numbers of users can attempt to talking about vGPU profiles not relevant vDGA/passthrough/DDA: Same tech available on ESXi/XenServer/Hyper-V respectively - one physical GPU per VM vGPU: with all the profiles available on both ESXi and XenServer. http://images.nvidia.com/content/grid/pdf/161207-GRID-Packaging-and-Licensing-Guide.pdf http://images.nvidia.com/content/grid/pdf/GRID-Licensing-Guide.pdf
Hi ondskan_,

This sounds out of date / wrong:
vSGA: only VMware - API intercept so not terribly performant and limite in OpenGL/DX support but no limit on numbers of users can attempt to talking about vGPU profiles not relevant
vDGA/passthrough/DDA: Same tech available on ESXi/XenServer/Hyper-V respectively - one physical GPU per VM
vGPU: with all the profiles available on both ESXi and XenServer.


http://images.nvidia.com/content/grid/pdf/161207-GRID-Packaging-and-Licensing-Guide.pdf

http://images.nvidia.com/content/grid/pdf/GRID-Licensing-Guide.pdf

#2
Posted 01/17/2017 03:44 PM   
Scroll To Top

Add Reply