CUDA Questions Please Help.

Accelerated Computing CUDA CUDA Programming and Performance
1

Hi guys,
Im new to the scene. Just have some general support questions.

Setup:
1/ Motherboatd model - X8DTG-DF
2/ Tesla Cards - Nvidia C1060
3/ CPU - 2x Intel Xeon E5520 Quad-Core’s
4/ RAM , 6x 2GB DDR3-1333 Registered ECC Hynix
5/ Onboard VGA - Matrox G200e
6/ OS - Win XP 64bit

6/ Nvidia Card used - Nvidia NVS 285

  1. I was trying to get my Tesla c1060 cards working with an non-nvidia onboard VGA which is used as primary video output. I couldn’t get the Tesla cards to show up as it said no CUDA device detected. The thing is that the 2 tesla cards have shown up in device manager. So, my conclusion is that windows XP doesnt support two graphic card drivers. The primary video card has to be NVIDIA chipset? Is this correct.

  2. I disabled the onboard vga port and have put in a Nvidia NVS 285 card to be the primary video output. And i installed the CUDA driver instead of the normal nvidia driver. After i installed the CUDA driver and toolkit and the sdk sample all version 2.3, then i ran the same sample program as in my first question and it ran!. But the thing is that the Nvidia NVS 285 card is not CUDA enabled… Why is this so? I am able to change the NVS 285 settings on the control manager found in control panel.

3.my last question is if windows cannot support two or multiple device drivers is any linux distro able to support this?

Please i am not a programmer, I have just started in this scene.
Thank you in advance.

2

Can anyone please suggest anything?? Please.

3

Hi AsIzzy,

A google search turned up this web page:

http://www.nvidia.com/object/tesla_build_your_own.html

“Windows requires NVIDIA GPU in one of the PCI-e slots” when using CUDA, but on-board graphics “works with Linux.”

Another google search turned up this forum post:

http://forums.nvidia.com/index.php?showtopic=75599

which in turn links to

http://www.nvidia.com/object/cuda_learn_products.html

Which indicates that the NVS 285 does not support CUDA. I don’t know why, but the reason probably has to do with the age of the GPU.

I can’t tell what “settings” you are referring to in the control panel, but I wasn’t aware there were any CUDA settings that could be changed in the control panel. Perhaps you are confusing CUDA with graphics?

If the SDK sample ran correctly, then it was probably using your Tesla card.

In the future, you should try to do some google searches before asking questions on the forum. People are more likely to take their own time to help you if you’ve demonstrated that you’ve already tried to help yourself. Also, you are also more likely to get helpful responses to your posts if the topic is more specific and descriptive than “CUDA Questions Please Help.” Though written kind of harshly, maybe look at this, and the sections below it? http://catb.org/~esr/faqs/smart-questions.html#before

Best of luck,

Gabriel

4

Thank you for your words of wisdom. External Image
I have already been to those sites you have mentioned, And i double read one of the sites there and it answered some of my other questions i had which was not written on this thread. So thank you for putting the sites there which made me read through it again External Image
I thought about giving the title a more specific one but my questions are more general. :ph34r: