compute, gpu, t1-le-45: cannot get nvidia driver to work

hi there,

new user here, currently checking out ovh in the trial period. i hope you guys can help me out.

for a project i need to have a gpu'd vm. i created an instance of a t1-le-45 machine, this one is equipped with a nvidia tesla v100 gpu. (sometimes those are referred to by the name volta-architecture.) so far i don't understand the variety of available drivers in all detail. but having installed the proprietary cuda-drivers version 5.25 (`apt-get install cuda-drivers-525`) gives me promising results.

the instance is running ubuntu 22.04lts.

after rebooting with cuda drivers 5.25 (525.147.05-1) i can see a v100 detected in dmesg and syslog. but nvidia-smi, this is sort of the central device manager tool, shows:
`No devices were found`

so looking into it, there is one error message in the kernel log (dmesg), that might be the culprit:
`NVRM: GPU 0000:00:06.0: GPU does not have the necessary power cables connected.`

the pci-id found in dmesg points to the very same card that is listed by lspci:
`00:06.0 3D controller: NVIDIA Corporation GV100GL [Tesla V100 PCIe 16GB] (rev a1)`

since this is a template based pcie-passthrough vm, i guess the driver is not right when saying cables are not connected correctly. ;)

so my question: is anyone using this kind of vm along with ubuntu and has the driver working? what did you do? what versions are you running? how did you proceed?

it'd be great, if anyone could point me in the right direction. or in case i am doing it all wrong, tell me what to do. any insight or idea is much appreciated.

thank you!
best,
sebastian