Need help in connecting VGPU instance to my Cuda-Q python VScode

[Amogh1506]

Dear Team,

I have procured VGPU(A100) through NVIDIA Brev. I want to run Cuda-Q application on this GPU. In my linux code, brev list'' command on ubundtu shows the GPU instance as well. But when I run my python file with cudaq.set_target("nvidia") or this cudaq.set_target('nvidia', option = 'fp64'), It doesnt work. IT certainly works with CPU : cudaq.set_target("qpp-cpu") .

How do I configure the VGPU to my code. Please help.

Regards,
Amogh

[bmhowe23]

Hi @Amogh1506 - could you please run the nvidia-smi command from the same session where you are attempting to run your Python scripts? If nvidia-smi cannot see any GPUs, then CUDA-Q will also not be able to see any GPU(s), so I'd first like to validate that nvidia-smi can indeed see your GPU.

[Amogh1506]

Thanks @bmhowe23 for quick response. Below is the msg

NVIDIA-SMI has failed because it couldn't communicate with the NVIDIA driver. Make sure that the latest NVIDIA driver is installed and running.

[bmhowe23]

I suspect you need to reach out to NVIDIA Brev support because it sounds like you don't have access to the GPU.