CUDA Support on Dell Studio XPS 13

Accelerated Computing CUDA CUDA Programming and Performance
1

I own a Dell Studio XPS 13 with Nvidia 9400M and 9200MGS in SLI Mode. My driver installed is 179.82(Vista 64) which has been downloaded from Dell website.
However, I am unable to use these drivers for CUDA. the nvcuda.dll has not been installed. However, for CUDA 2.0 the supported driver mentioned is found to be 178.08 or later which means my driver should be supported. However I find that nvcuda.dll is not installed in my system.

I cannot upgrade the drivers since Dell XPS 13 has not yet finished testing 185.85 drivers for CUDA 2.2 as well… When I try to install this driver after downloading from nvidia, the message appears " no supported hardware found "… So what could be the reason ?

2

I too have the same problem, and contacted Dell regarding this. However they are not ready to say when they would have updates :( and are pointing me contact Nvidia instead! If any one from Nvidia could clarify the matter it would be good.

3

exact same problem when i try to use badaboom lol nways it shall be done when dell finalizez drivers for windows 7 i use cyber link media show expresso and it uses my video card just fine ( i too have the hybrid SLI ) 9500 M forget any response from NVIDIA :magic: External Image

4

I just finished the installation of cuda 2.2 on my dell xps 13.

Here is the way to do it (Remember to take your os into acount. I’m using vista 64 bit):

  • Download the 186 driver for the GeForce 9M series for notebooks.
  • Remove your old nvidia driver using control panel > remove program. You only want to remove the nvidia display driver.
  • Install the 186 driver. (It includes the 2.2 version of NVCUDA.dll)
  • Verify that you now have cuda 2.2 installed: Nvidia control panel > system info > components
  • Download the cuda toolkit 2.2 and cuda sdk 2.2.1 for your os
  • Install it as described in the Getting started guide.
  • Verify the whole installation according to the same guide.
  • Run some cool demos from the sdk :-)

The SLI setup seems to cause no problem since both GPU’s support cuda.

Good luck!