Description of problem: The first paragraph of "Installing a vGPU on a Virtual Machine" (https://access.redhat.com/documentation/en-us/red_hat_virtualization/4.2/html-single/virtual_machine_management_guide/#Installing_a_vGPU_on_a_Virtual_Machine) reads as this: """ You can use a host with a compatible graphics processing unit (GPU) to run virtual machines with virtual GPUs (vGPUs). A virtual machine with a vGPU is better suited for graphics-intensive tasks than a virtual machine without a vGPU. A virtual machine with a vGPU can also run software that cannot run without a GPU, such as CAD. """ I think this paragraph is misleading, as it appears to point that it's possible to run software requiring 3D acceleration on RHV's VMs. While it's expected that future versions of RHV will allow streaming from a vGPU to the client using a special SPICE channel, currently (as of 4.2) it's only possible to use them in headless mode, which is mainly intended for CUDA computing.
Accepting into the 4.2.5 (z2) program, and assigning to Megan for review. Megan, it might be worth confirming with engineering and QE whether there are any additional limitations around the current implementation as well.
(In reply to Sergio Lopez from comment #0) > Description of problem: > > The first paragraph of "Installing a vGPU on a Virtual Machine" > (https://access.redhat.com/documentation/en-us/red_hat_virtualization/4.2/ > html-single/virtual_machine_management_guide/ > #Installing_a_vGPU_on_a_Virtual_Machine) reads as this: > > """ > You can use a host with a compatible graphics processing unit (GPU) to run > virtual machines with virtual GPUs (vGPUs). A virtual machine with a vGPU is > better suited for graphics-intensive tasks than a virtual machine without a > vGPU. A virtual machine with a vGPU can also run software that cannot run > without a GPU, such as CAD. > """ > > I think this paragraph is misleading, as it appears to point that it's > possible to run software requiring 3D acceleration on RHV's VMs. > > While it's expected that future versions of RHV will allow streaming from a > vGPU to the client using a special SPICE channel, currently (as of 4.2) it's > only possible to use them in headless mode, which is mainly intended for > CUDA computing. 3D acceleration does work and is used by several customers. The only "issue" is that a 3rd party remote visualization solution needs to be used. We currently test the following: - Mechdyne TGX - HP RGS - Nice DCV As you mentioned SPICE is not supported as of now, as we have issues shipping H.264 streaming due to patents.
Martin, If I understand correctly: - The original paragraph mentioned in the first comment in the bug is true and does not need to be edited. - We need to add the following to the list of vGPU requirements: - a 3rd party remote visualization solution. The following solutions have been tested: - Mechdyne TGX - HP RGS - Nice DCV NOTE: SPICE is not supported. Megan also asked about licenses, and I didn't see a response: > Finally I wanted to clarify what license/s are required from nVidia to make > this feature functional? The deep dive mentions requiring an enterprise account > to access the required drivers, is this accurate? It also mentions requiring an > nVidia GRID licensing server? Do you have any further information about this > requirement?