As the topic say, what is the cheapest Nvidia Grid gpu I can get to test out vgpu and "slicing" of a gpu to different vm's and still use nvidia driver with cuda in the vm?
(not assigning 1 gpu to 1 vm).
Trying to setup a lab here, do do a cheap poc that I can present at work.
The newest grid cards are not an option, as this is something I have to fund myself
I was looking at the Grid K1 and K2, as it seems they dont need a license to run a vgpu.
(Quote: NVIDIA GRID K1 and K2 GPUs do not require a license to run vGPU.)
However, I dont know if they are the cheapest option
I dont need much performance for the lab, just so I can present a poc for cuda accelerated workloads.
No video/graphics usage, only computational use.
Hopefully I wont need Horizon, or any other VDI solutions, as I want to keep the cost down for the solution I hopefully can present.
(Got vmug advantage, so licenses is not an option for the lab testing)
-Jannis
(not assigning 1 gpu to 1 vm).
Trying to setup a lab here, do do a cheap poc that I can present at work.
The newest grid cards are not an option, as this is something I have to fund myself
I was looking at the Grid K1 and K2, as it seems they dont need a license to run a vgpu.
(Quote: NVIDIA GRID K1 and K2 GPUs do not require a license to run vGPU.)
However, I dont know if they are the cheapest option
I dont need much performance for the lab, just so I can present a poc for cuda accelerated workloads.
No video/graphics usage, only computational use.
Hopefully I wont need Horizon, or any other VDI solutions, as I want to keep the cost down for the solution I hopefully can present.
(Got vmug advantage, so licenses is not an option for the lab testing)
-Jannis