Nvidia Quadro FX 580 cudaErrorInsufficientDriver(35) problem


I have Nvidia Quadro FX 580 and i need to make some computation on it.
Drivers (342.00-quadro-grid-desktop-notebook-win8-win7-64bit-international-whql.exe) are installed… during the installation of drivers there were no problems.
The graphic card is on Devices List.

I tried to install CUDA Toolkit 9.2 (cuda_9.2.148_windows_network.exe) but during installation warning occurs

“This graphics driver could not find compatible graphics hardrware. You may continue installation, but you may not be able to run CUDA application with this driver. This may occur with graphics hardware that is newer than this toolkit. In that case, it is suggested that you keep your existing driver and install the remaining portions of CUDA Toolkit”

I follow the suggestion and uncheck driver during installation (also make another clean installation with drivers from the toolkit, the result was the same).

When i want to make some Sample CUDA application i get cudaStatus cudaErrorInsufficientDriver(35) and output “cudaSetDevice failed! DO you have a CUDA-capable GPU installed?”.

What should I do??

The last CUDA version that supported that GPU was CUDA 6.5

You can find it on the CUDA toolkit archive page. (hint: google “cuda toolkit archive”)

Your 342.xx driver will work with that CUDA toolkit.

I was looking for a solution for over a week!! Thank you very much!!

Installation goes smoothly but I can’t open Samples in VS2017 Community :(
Is this a problem for old Toolkit to run on newest VS

Yes. The compatibility matrix for a toolkit is contained in the install guide for that toolkit version/OS. So the windows install guide for CUDA 6.5 will tell you which Visual Studio versions it is compatible with.

If 6.5 is not compatible with VS2017 is there any solution to run it on free version of VS ??
Only newest Community version is avaiable for free… now 2017.

google turned this up:


Ooo it must be something new because i wanted to install older version of VS a couple months ago and only the newest Was avaiable…

Last question… Where can i find information which version of CUDA supports specific graphics hardware ??

Because i have also NVIDIA GeForce 210