Cuda Error Insuficient Driver (Error 35)

I’m sorry if there is another post with this same content, but I’m trying to solve this problem and I didn’t find a solution yet.

When I try to run the sample projects, I receive this Error 35: Cuda Error Insuficient Driver. I’m using Ubuntu 18.04.2 LTS and a GeForce GT 620.

I will post some information bellow that I found in another posts


nvidia-smi

±----------------------------------------------------------------------------+
| NVIDIA-SMI 390.87 Driver Version: 390.87 |
|-------------------------------±---------------------±---------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
|===============================+======================+======================|
| 0 GeForce GT 620 Off | 00000000:01:00.0 N/A | N/A |
| 52% 49C P12 N/A / N/A | 188MiB / 1984MiB | N/A Default |
±------------------------------±---------------------±---------------------+

±----------------------------------------------------------------------------+
| Processes: GPU Memory |
| GPU PID Type Process name Usage |
|=============================================================================|
| 0 Not Supported |
±----------------------------------------------------------------------------+


nvcc -V

nvcc: NVIDIA ® Cuda compiler driver
Copyright © 2005-2017 NVIDIA Corporation
Built on Fri_Nov__3_21:07:56_CDT_2017
Cuda compilation tools, release 9.1, V9.1.85


ldconfig -p | grep libcuda

libcudart.so.9.1 (libc6,x86-64) => /usr/lib/x86_64-linux-gnu/libcudart.so.9.1
libcudart.so (libc6,x86-64) => /usr/lib/x86_64-linux-gnu/libcudart.so
libcuda.so.1 (libc6,x86-64) => /usr/lib/x86_64-linux-gnu/libcuda.so.1
libcuda.so.1 (libc6) => /usr/lib/i386-linux-gnu/libcuda.so.1
libcuda.so (libc6,x86-64) => /usr/lib/x86_64-linux-gnu/libcuda.so
libcuda.so (libc6) => /usr/lib/i386-linux-gnu/libcuda.so


The GeForce GT 620 appears to be a GPU with compute capability 2.1. Recent and current CUDA versions require GPUs with compute capability >= 3.0.

You would want to use a newer GPU, and should be able to find low-cost options among Pascal-family devices such as the GeForce GT 1030 (~ $90), with compute capability 6.x.

The last CUDA version that supported Fermi (cc2.x) devices is CUDA 8.0

Thanks very much for the help

Hi All,

I have the same error but not quite the same environment so I am not sure what the issue is in my case.

When running deviceQuerry from CUDA Samples v10.1 (as well as v10.0) I get:

cudaGetDeviceCount returned 35
-> CUDA driver version is insufficient for CUDA runtime version
Result = FAIL

I am running Windows 10 Pro, 64bit, version 1809
GPU NVS 315, Drivers 392.37 (got it fresh from NVIDIA Driver Downloads)
I got today CUDA Toolkit 10.1 from https://developer.nvidia.com/cuda-downloads
and the drivers from https://www.nvidia.co.uk/Download/index.aspx?lang=en-uk

nvidia-smi gives
±----------------------------------------------------------------------------+
| NVIDIA-SMI 392.37 Driver Version: 392.37 |
|-------------------------------±---------------------±---------------------+
| GPU Name TCC/WDDM | Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
|===============================+======================+======================|
| 0 NVS 315 WDDM | 00000000:01:00.0 N/A | N/A |
| 30% 36C P8 N/A / N/A | 272MiB / 1024MiB | N/A Default |
±------------------------------±---------------------±---------------------+

±----------------------------------------------------------------------------+
| Processes: GPU Memory |
| GPU PID Type Process name Usage |
|=============================================================================|
| 0 Not Supported |
±----------------------------------------------------------------------------+

Would anyone know what the issue is here?

Thank you very much for your help

Posts #2 and #3 apply to the NVS 315, which is a Fermi-architecture GPU from around 2012.

OK, thx njuffa, good to know! Regards