-1

I have installed latest cudatoolkit and driver today however when I try to build and run matrixMul program using visual studio 2019, I get the following error:

[Matrix Multiply Using CUDA] - Starting… CUDA error at C:\ProgramData\NVIDIA Corporation\CUDA Samples\v10.2\common\inc\helper_cuda.h:775 code=35(cudaErrorInsufficientDriver) “cudaGetDeviceCou nt(&device_count)” C:\ProgramData\NVIDIA Corporation\CUDA Samples\v10.2\0_Simple\matrixMul…/…/bin/win64/Debug/matrixMul.exe (process 7140) exited with code 1.

More information about the setup: 1: per Nvidia control panel driver version is 391.35

2: GPU GeForce GT 420M which is cuda 2.1 as per https://developer.nvidia.com/cuda-gpus#compute

3: Visual Studio 2019

4: The program i am trying to build/run is C:\ProgramData\NVIDIA Corporation\CUDA Samples\v10.2\0_Simple\matrixMul\matrixMul_vs2019.sln

5: with a bit of debugging it appears that program is failing at line checkCudaErrors(cudaGetDeviceCount(&device_count)); inside cuda_runtime_api.h @ line 1288. The function is supposed to returns the number of devices with compute capability

greater or equal to 2.0 Apparently it sounds like GeForce GT420M is Cuda 2.1 capable but current runtime is not recognizing it and failing. Could someone please help me to resolve this error?

talonmies
  • 70,661
  • 34
  • 192
  • 269
User852
  • 35
  • 7
  • 2
    "I have installed latest cudatoolkit and driver" -- that GPU is too old to be supported by the latest toolkit and driver. Please read the release notes for what hardware is supported – talonmies Apr 15 '20 at 12:06

2 Answers2

1

Your device (compute capability 2.1) is not supported by CUDA 10.2. You need to install a lower version of CUDA toolkit that supports it. The last CUDA version that supports compute capability 2.x is CUDA 8.

heapoverflow
  • 264
  • 2
  • 12
  • 2
    Is that last sentence actually true? I haven't ever tried it on unsupported hardware – talonmies Apr 15 '20 at 13:06
  • You are right, it is not correct. https://stackoverflow.com/questions/53422407/different-cuda-versions-shown-by-nvcc-and-nvidia-smi , max CUDA version output in nvidia-smi is added after the driver version 410.x . I will edit my answer. Thanks for pointing out. – heapoverflow Apr 15 '20 at 13:32
0

Many thanks for your input. Apologies for late reply as I was trying to make system work.

Yes reverting back to cuda 8 and corresponding tools and drivers I was able to build and run the samples code. The other additional thing i had to do is to increase the value of TdrDelay in registry to avoid frequent interrupts (which then allows sufficient time to complete program).

Thanks again.

User852
  • 35
  • 7