since Nvidia decided 2021 to change their policy to enable geforce customers to passthrough their GPU for VMs, I was wondering if there are plans to enable geforce gpus to use vGPU for enthusiasts to use. Or is their hardware not capable to use sr-iov?
Just to clarify, I’m not asking to support geforce gpus but it would be amazing if the feature was unlocked.
Any statement on that matter by nvidia would be appreciated.
Sorry if this was asked before but I couldn’t find anything recent, especially for Ampere or Ada GPUs.
There is a huge difference between PT and vGPU. vGPU requires a software stack on host and guest and is therefore a product that requires licensing to get the required software and support. In addition, there are features like motion support which needs effort to develop and maintain.
I don’t expect we will ever see vGPU for consumer grade GPUs.
thanks for your response.
I am aware that Nvidia has no interest to bring vGPU to consumer grade GPUs as part of the officially supported hardware since there would be a huge effort which would only serve a niche market and therefore won’t be affordable.
Since I don’t know about details of the difference in hardware architecture and limitations of consumer grad GPUs, I’m asking more about possibilities to enable the community to use the existing software solution to achieve a basic functionality.
As far as I understand for example the RTX 3090 uses the same base architecture as the RTX A6000 so I was wondering if the RTX 3090 has other hardware limitations or differences which would make it impossible or is this a locked feature in the firmware of the device because there is no official support from the software?
I don’t expect a detailed answer, I rather want to suggest, if there are possibilities to enable the community to achieve basic working results by themselves without Nvidia spending a lot of resources this would be highly appreciated.
Thanks for your time.