CudaRuntimeAPIError: [100] Call to cudaRuntimeGetVersion results in CUDA_ERROR_NO_DEVICE

Hello! I have a problem Anaconda and GPU accelerating on WSL2 Ubuntu 22.04 with RapidsAI libs.
Conda create GPU envirovment:

conda create -n rapids-22.10 -c rapidsai -c conda-forge -c nvidia  \
    rapids=22.10 python=3.9 cudatoolkit=11.2 \

I have installed Cuda by Nvidia FAQ - CUDA on WSL :: CUDA Toolkit Documentation then I have tested it.
My workstation with installed Windows GPU drivers includes Nvidia GeForce RXT 3060.

nvidia-smi is Ok.

| NVIDIA-SMI 525.46       Driver Version: 526.86       CUDA Version: 12.0     |
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|                               |                      |               MIG M. |
|   0  NVIDIA GeForce ...  On   | 00000000:03:00.0 Off |                  N/A |
|  0%   38C    P8     9W / 170W |    878MiB / 12288MiB |      1%      Default |
|                               |                      |                  N/A |

| Processes:                                                                  |
|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |
|        ID   ID                                                   Usage      |
|    0   N/A  N/A        26      G   /Xwayland                       N/A      |

But /dev not includes GPU, I am trying “ls /dev | grep nvidia*” and it is haven’t results
Then I started Cuda samples ./deviceQuery and it successful.

CUDA Device Query (Runtime API) version (CUDART static linking)

Detected 1 CUDA Capable device(s)

Device 0: "NVIDIA GeForce RTX 3060"
  CUDA Driver Version / Runtime Version          12.0 / 11.8
  CUDA Capability Major/Minor version number:    8.6
  Total amount of global memory:                 12287 MBytes (12884246528 bytes)
  (028) Multiprocessors, (128) CUDA Cores/MP:    3584 CUDA Cores
  GPU Max Clock rate:                            1852 MHz (1.85 GHz)
  Memory Clock rate:                             7501 Mhz
  Memory Bus Width:                              192-bit
  L2 Cache Size:                                 2359296 bytes
  Maximum Texture Dimension Size (x,y,z)         1D=(131072), 2D=(131072, 65536), 3D=(16384, 16384, 16384)
  Maximum Layered 1D Texture Size, (num) layers  1D=(32768), 2048 layers
  Maximum Layered 2D Texture Size, (num) layers  2D=(32768, 32768), 2048 layers
  Total amount of constant memory:               65536 bytes
  Total amount of shared memory per block:       49152 bytes
  Total shared memory per multiprocessor:        102400 bytes
  Total number of registers available per block: 65536
  Warp size:                                     32
  Maximum number of threads per multiprocessor:  1536
  Maximum number of threads per block:           1024
  Max dimension size of a thread block (x,y,z): (1024, 1024, 64)
  Max dimension size of a grid size    (x,y,z): (2147483647, 65535, 65535)
  Maximum memory pitch:                          2147483647 bytes
  Texture alignment:                             512 bytes
  Concurrent copy and kernel execution:          Yes with 5 copy engine(s)
  Run time limit on kernels:                     Yes
  Integrated GPU sharing Host Memory:            No
  Support host page-locked memory mapping:       Yes
  Alignment requirement for Surfaces:            Yes
  Device has ECC support:                        Disabled
  Device supports Unified Addressing (UVA):      Yes
  Device supports Managed Memory:                Yes
  Device supports Compute Preemption:            Yes
  Supports Cooperative Kernel Launch:            Yes
  Supports MultiDevice Co-op Kernel Launch:      No
  Device PCI Domain ID / Bus ID / location ID:   0 / 3 / 0
  Compute Mode:
     < Default (multiple host threads can use ::cudaSetDevice() with device simultaneously) >

deviceQuery, CUDA Driver = CUDART, CUDA Driver Version = 12.0, CUDA Runtime Version = 11.8, NumDevs = 1
Result = PASS

My code python:

import os
import cudf
os.environ['CUDA_VISIBLE_DEVICES'] = "0"
df = cudf.Series([1,2,3])

But my code executes with error
numba.cuda.cudadrv.runtime.CudaRuntimeAPIError: [100] Call to cudaRuntimeGetVersion results in CUDA_ERROR_NO_DEVICE

Hey @sterrchov , There are two ways to possibly rememdy this. If you installed the drivers uto your WSL2 instance, which you sould not have, please ninstall the drivers/redo you WSL2 istance (WSLg/Cuda suddenly broken due to nvidia-smi unable to find GPU · Issue #9099 · microsoft/WSL · GitHub)

If you did not install drivers directly into your WSL2 instance, but only your host side (Windows) please uninstall your driver and CUDA version, then install the previous version. If it fixes it, this driver may have an issue where it looks like it has CUDA 12.0, which is incompatible with RAPIDS. You need to have CUDA 11.x to use CUDA’s compatibility layer for cudatoolkit 11.x. For now, please download and use the 522.30 version of the driver here Official Advanced Driver Search | NVIDIA, or you can try 526.98 and see if it works. That’s the last driver I personally used in my tests.

Please let us know which cause and solution works for you!

Thanks! Downgrade driver to 522.30 sloved it.

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.