Has anyone used Nvidia a10 /30 /40 with opennebula successfully? Specifically the resource slicing so multiple vms can share the GPUs resources. Looking to spec out an hpc vm box.
Hello @slapplebags, are you looking for a use case, or a user that have had experience with the GPU Resources?
If not, you can go to our Documentation: NVIDIA vGPU support — OpenNebula 6.6.2 documentation
According to this: Add support for virtualized Nvidia GPUs · Issue #5351 · OpenNebula/one · GitHub
Nvidia is supported.
Let me know if you are looking for other type of documentation, or as I said above, if you are looking for users that have successfully deployed this.
Regards.
The VGPU support feature was validated on an A10 GPU. So that series should be especially successful.
I’m looking to see if anyone has actually set it up and gotten it working correctly with OpenNebula 6.4 CE. I have setup GPU passthrough in the past but when i tried with 6.2 CE i was unable to get an nvidia A30 to passthrough at all, either with vGPU or without.
Great thats exactly what i was looking. based on that I imagine the a30 will also work since i believe the a10 and a30 are part of the same series.