Can't pass gpu benchmark in Ubuntu18.04 by WSL2

Accelerated Computing CUDA CUDA on Windows Subsystem for Linux
1

I have passed deviceQuery and BlackScholes
windows build:21337.1010
docker version:v20.10.5
wsl --update shows no update available and kernel version 5.4.91

+-----------------------------------------------------------------------------+
| NVIDIA-SMI 470.05       Driver Version: 470.05       CUDA Version: 11.3     |
|-------------------------------+----------------------+----------------------+
| GPU  Name            TCC/WDDM | Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|                               |                      |               MIG M. |
|===============================+======================+======================|
|   0  NVIDIA GeForce ... WDDM  | 00000000:01:00.0  On |                  N/A |
| 30%   41C    P8    19W / 175W |    824MiB /  8192MiB |     24%      Default |
|                               |                      |                  N/A |
+-------------------------------+----------------------+----------------------+

i cant find /dev/dxg
i have use docker desktop and enable Use the WSL 2 based engine

error:
docker run --gpus all nvcr.io/nvidia/k8s/cuda-sample:nbody nbody -gpu -benchmark
Unable to find image ‘nvcr.io/nvidia/k8s/cuda-sample:nbody’ locally
nbody: Pulling from nvidia/k8s/cuda-sample
d519e2592276: Pull complete
d22d2dfcfa9c: Pull complete
b3afe92c540b: Pull complete
b25f8d7adb24: Pull complete
ddb025f124b9: Pull complete
fe72fda9c19e: Pull complete
c6a265e4ffa3: Pull complete
c931a9542ebf: Pull complete
f7eb321dd245: Pull complete
d67fd954fbd5: Pull complete
Digest: sha256:a2117f5b8eb3012076448968fd1790c6b63975c6b094a8bd51411dee0c08440d
Status: Downloaded newer image for nvcr.io/nvidia/k8s/cuda-sample:nbody
docker: Error response from daemon: OCI runtime create failed: container_linux.go:367: starting container process caused: process_linux.go:495: container init caused: Running hook #0:: error running hook: exit status 1, stdout: , stderr: nvidia-container-cli: requirement error: unsatisfied condition: cuda>=11.2, please update your driver to a newer version, or use an earlier cuda container: unknown.
ERRO[0040] error waiting for container: context canceled

then i uninstall windows docker desktop and use
curl https://get.docker.com | sh to install docker

still not work

install code
curl https://get.docker.com | sh
% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
100 13857 100 13857 0 0 12070 0 0:00:01 0:00:01 --:–:-- 12070
# Executing docker install script, commit: 3d8fe77c2c46c5b7571f94b42793905e5b3e42e4

WSL DETECTED: We recommend using Docker Desktop for Windows.
Please get Docker Desktop from https://www.docker.com/products/docker-desktop


You may press Ctrl+C now to abort this script.
+ sleep 20



+ sudo -E sh -c apt-get update -qq >/dev/null
+ sudo -E sh -c DEBIAN_FRONTEND=noninteractive apt-get install -y -qq apt-transport-https ca-certificates curl >/dev/null
+ sudo -E sh -c curl -fsSL "https://download.docker.com/linux/ubuntu/gpg" | apt-key add -qq - >/dev/null
Warning: apt-key output should not be parsed (stdout is not a terminal)
+ sudo -E sh -c echo "deb [arch=amd64] https://download.docker.com/linux/ubuntu bionic stable" > /etc/apt/sources.list.d/docker.list
+ sudo -E sh -c apt-get update -qq >/dev/null
+ [ -n  ]
+ sudo -E sh -c apt-get install -y -qq --no-install-recommends docker-ce >/dev/null

If you would like to use Docker as a non-root user, you should now consider
adding your user to the "docker" group with something like:

  sudo usermod -aG docker ricar

Remember that you will have to log out and back in for this to take effect!

WARNING: Adding a user to the "docker" group will grant the ability to run
         containers which can be used to obtain root privileges on the
         docker host.
         Refer to https://docs.docker.com/engine/security/security/#docker-daemon-attack-surface
         for more information.
2

I am having exactly the same problem.

Other images in the samples repo works though. For example:

docker run --rm --gpus all nvcr.io/nvidia/k8s/cuda-sample:vectoradd-cuda10.2

Looks like driver is behind, Does the 470.14 driver support CUDA 11?

3

yes，i have passed cuda10.2 either，shou i install cuda tookit==10.2?

4

my windwos CUDA version is 10.1.243,should i update to >=11.2?

5

This command works
docker run --gpus all --env NVIDIA_DISABLE_REQUIRE=1 nvcr.io/nvidia/k8s/cuda-sample:nbody nbody -gpu -benchmark

2 Likes
6

it works! but which version dirver and library should i install?

7

my wsl2 details, that works without issues

  1. uname -a: 5.10.16.3-microsoft-standard-WSL2
  2. Driver Version: 470.14
  3. CUDA Version: 11.3
  4. OS build: 21370.1
8

Did you have to alter anything to get this to work, I have all the same versions except for my OS Build is at 21382. I am also running a 3070 gpu. devicequery and Blackholes work but nbody fails. I also am not able to launch a docker container.

9

Same