7

I have two laptops: a Sony vaio z-series (vocz1) & S series. The first one has Geforce with cuda model GT330M and second one has GT 640M LE. When I am trying install CUDA Geforce driver from this site http://developer.nvidia.com/cuda/cuda-downloads I am receiving below error in Windows. I am wondering would you suggest me a solution to solve this issue? all of my graphic card already installed,

Nvidia installer cannot continue This graphics card could not find compatible graphics hardware.

I am beginner in programming with CUDA, I want to know can I compile and run my CUDA program with out installing Nividia Driver? Cuda toolkit and SDK installed successfully in my machine but whenI run my program I cannot set my Cuda device. Does it mean I need to install Nividia Driver?enter image description here

cudaError_t cudaStatus1;
int deviceCount; 
cudaGetDeviceCount(&deviceCount); 
int device; 
for (device = 0; device < 10; ++device) { 
    cudaDeviceProp deviceProp; 

    cudaGetDeviceProperties(&deviceProp, device); 
    // Choose which GPU to run on, change this on a multi-GPU system.
    cudaStatus1 = cudaSetDevice(device);
    printf("Device %d has compute capability %d.%d. -  %d\n", 
            device, deviceProp.major, deviceProp.minor,cudaStatus1 ); 
}

output:

Device 0 has compute capability 3137268.3137268. -  35
Device 1 has compute capability 3137268.3137268. -  35
Device 2 has compute capability 3137268.3137268. -  35
Device 3 has compute capability 3137268.3137268. -  35
Device 4 has compute capability 3137268.3137268. -  35
Device 5 has compute capability 3137268.3137268. -  35
Device 6 has compute capability 3137268.3137268. -  35
Device 7 has compute capability 3137268.3137268. -  35
Device 8 has compute capability 3137268.3137268. -  35
Device 9 has compute capability 3137268.3137268. -  35

35 means it is not set the device , if it became 0 means device set.

after I run deviceQuery below information I received:

Microsoft Windows [Version 6.1.7601] Copyright (c) 2009 Microsoft Corporation. All rights reserved.
C:\Users\xx>"C:\ProgramData\NVIDIA Corporation\NVIDIA GPU Computing SDK 4.2\C\ bin\win64\Release\deviceQuery.exe" [deviceQuery.exe] starting... C:\ProgramData\NVIDIA Corporation\NVIDIA GPU Computing SDK 4.2\C\bin\win64\Relea se\deviceQuery.exe Starting... CUDA Device Query (Runtime API) version (CUDART static linking) Found 1 CUDA Capable device(s) Device 0: "GeForce GT 640M LE" CUDA Driver Version / Runtime Version 4.2 / 4.2 CUDA Capability Major/Minor version number: 3.0 Total amount of global memory:
1024 MBytes (1073741824 bytes) ( 2) Multiprocessors x (192) CUDA Cores/MP: 384 CUDA Cores GPU Clock rate:
405 MHz (0.41 GHz) Memory Clock rate:
900 Mhz Memory Bus Width: 128-bit L2 Cache Size: 262144 bytes Max Texture Dimension Size (x,y,z) 1D=(65536), 2D=(65536,65536), 3 D=(4096,4096,4096) Max Layered Texture Size (dim) x layers
1D=(16384) x 2048, 2D=(16384,16 384) x 2048 Total amount of constant memory: 65536 bytes Total amount of shared memory per block: 49152 bytes Total number of registers available per block: 65536 Warp size: 32
Maximum number of threads per multiprocessor: 2048 Maximum number of threads per block: 1024 Maximum sizes of each dimension of a block: 1024 x 1024 x 64 Maximum sizes of each dimension of a grid: 2147483647 x 65535 x 65535 Maximum memory pitch:
2147483647 bytes Texture alignment: 512 bytes Concurrent copy and execution: Yes with 1 copy engine(s) Run time limit on kernels: Yes
Integrated GPU sharing Host Memory: No Support host page-locked memory mapping: Yes Concurrent kernel execution:
Yes Alignment requirement for Surfaces: Yes Device has ECC support enabled: No Device is using TCC driver mode: No Device supports Unified Addressing (UVA):
No Device PCI Bus ID / PCI location ID: 1 / 0 Compute Mode: < Default (multiple host threads can use ::cudaSetDevice() with device simu ltaneously) > deviceQuery, CUDA Driver = CUDART, CUDA Driver Version = 4.2, CUDA Runtime Versi on = 4.2, NumDevs = 1, Device = GeForce GT 640M LE [deviceQuery.exe] test results... PASSED

exiting in 3 seconds: 3...2...1...done!

Amir
  • 1,919
  • 8
  • 53
  • 105
  • Your CUDA code snippet is complete nonsense and you shouldn't expect it to return anything meaningful even if you had correctly functioning CUDA installations in your laptops. – talonmies Aug 11 '12 at 09:22
  • 1
    @talonmies: I did too since your second command was rude to me. I am so disappointed with your language. Any way,I take screenshot to prove the reader all the driver installed in my computer successfully. One more thing my computer contain 2 VGA, so I tried to explain it to make it clear...Finally I am not so familiar with CUDA but you can check all the fields of class cudaError_t, you will find a field which is equal to 0 else condition means you couldnt set your device successfully. – Amir Aug 11 '12 at 19:35
  • Hey, I am using windows 7 32bit(on my mac via bootcamp), with Geforce 320M as card graphic. I have the same problem. Did you find the solution? – csuo Nov 22 '12 at 11:21

3 Answers3

1

I want to know can I compile and run my CUDA program with out installing Nividia Driver? Cuda toolkit and SDK installed successfully in my machine but whenI run my program I cannot set my Cuda device. Does it mean I need to install Nividia Driver?

You will definitely need drivers to run the program. Have you tried running deviceQuery.exe provided with the binaries. That should give you a good starting point as to what is going wrong.

Programmer
  • 6,565
  • 25
  • 78
  • 125
  • I edit my question would you review it once more? – Amir Aug 12 '12 at 04:46
  • 1
    @rima: Device query looks okay. Two questions: Why are you doing cudaSetDevice(4000) since there is no device with id 4000. I think the parameter for cudaSetDevice should be the deviceId. Try putting 0 there and see the output. Also, why does the for loop go 10 times when you have only 1 gpu device – Programmer Aug 12 '12 at 08:31
  • @rima: Please up my answer once if you like it – Programmer Aug 12 '12 at 08:32
  • Sorry that one was mistake, I already modified it. I tried with 0 but it could not set therefore I wrote those line in order to test how many device I have, they are a lot, around 64000+. Would you elaborate more what do you mean by Device query looks okay? – Amir Aug 12 '12 at 18:53
  • 1
    By device query looks okay i mean that it appears that your GPU is configured correctly and you should be able to write GPU code. – Programmer Aug 13 '12 at 12:10
1

I have a VAIO too and I had the same problem. Don't download notebook version, try Desktop version of Nvidia Driver. I also had to disable my another Graphic card (Intel). It worked for me.

Bardia
  • 393
  • 4
  • 11
  • I will try & share my experience with u by night. – Amir Aug 12 '12 at 04:50
  • It is not work at all for me. I installed cuda 5 for GT 640M and for 330 M I installed recommended driver in below page http://forum.notebookreview.com/sony/342947-my-discoveries-vaio-z-s-hybrid-graphics.html – Amir Aug 25 '12 at 21:04
1

Unfortunately, there are many NVIDIA GPUs for which the driver from the NVIDIA website will not install (especially for GPU versions that are specifically OEM'd for Sony, Lenovo, etc and the OEM wants to control the driver experience). This is most likely the case for you.

In those cases, you can edit the .inf file to add your GPU into the list of GPUs for which the driver will install. However, it is a bit tricky and typically requires editing 3 different sections of the INF file. You can search around for details on how to mod NVIDIA inf files; there are a number of sites that do that.

Of course, you have to have the appropriate CUDA driver before you can run CUDA stuff. So first things first... you've gotta get the driver installed.

arrayfire
  • 1,744
  • 12
  • 19
  • if any one have problem for z serious they can refer to below page : http://forum.notebookreview.com/sony/342947-my-discoveries-vaio-z-s-hybrid-graphics.html – Amir Aug 25 '12 at 21:02