Hi
I have a question:

• What are the best practices for optimizing vGPU?
• and also Is there any way to apply the over-subscription or over-provision concept to vGPU?

Please note that I'm using 2 NVIDIA P100 (16GB).