I have a Docker container with ROS setup. Cuda 11 is installed, although there was a previous cuda-10 installation. I can verify cuda-11 version through the following:
- nvidia-smi
# nvidia-smi
Mon Oct 25 20:05:37 2021
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 495.29.05 Driver Version: 495.29.05 CUDA Version: 11.5 |
|-------------------------------+----------------------+----------------------+
| 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 | 00000000:01:00.0 Off | N/A |
| N/A 49C P8 3W / N/A | 485MiB / 7982MiB | 7% Default |
| | | N/A |
+-------------------------------+----------------------+----------------------+
+-----------------------------------------------------------------------------+
| Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |
|=============================================================================|
+-----------------------------------------------------------------------------+
- nvcc --version
# nvcc --version
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2021 NVIDIA Corporation
Built on Mon_Sep_13_19:13:29_PDT_2021
Cuda compilation tools, release 11.5, V11.5.50
Build cuda_11.5.r11.5/compiler.30411180_0
root@remap-msi:~/workspace/ros#
- version.json
# cat /usr/local/cuda/version.json | head -5
{
"cuda" : {
"name" : "CUDA SDK",
"version" : "11.5.0"
},
although there’s cuda-10 folder under /usr/local:
# ls -Al /usr/local
total 48
lrwxrwxrwx 1 root root 20 Oct 25 19:18 cuda -> /usr/local/cuda-11.5
drwxr-xr-x 1 root root 4096 Oct 31 2019 cuda-10.0
drwxr-xr-x 16 root root 4096 Oct 25 17:34 cuda-11.5
...
when trying to copmile zed-ros-wrapper, it fails with error:
# catkin_make -DCMAKE_BUILD_TYPE=Release
...
CMake Error at /usr/share/cmake-3.10/Modules/FindPackageHandleStandardArgs.cmake:137 (message):
Could NOT find CUDA: Found unsuitable version "10.0", but required is at
least "11" (found /usr/local/cuda)
Call Stack (most recent call first):
/usr/share/cmake-3.10/Modules/FindPackageHandleStandardArgs.cmake:376 (_FPHSA_FAILURE_MESSAGE)
/usr/share/cmake-3.10/Modules/FindCUDA.cmake:1080 (find_package_handle_standard_args)
/usr/local/zed/zed-config.cmake:72 (find_package)
zed-ros-wrapper/zed_nodelets/CMakeLists.txt:22 (find_package)
-- Configuring incomplete, errors occurred!
See also "/root/workspace/ros/build/CMakeFiles/CMakeOutput.log".
See also "/root/workspace/ros/build/CMakeFiles/CMakeError.log".
Makefile:5388: recipe for target 'cmake_check_build_system' failed
make: *** [cmake_check_build_system] Error 1
Invoking "make cmake_check_build_system" failed
I’m wondering, why cmake fails to detect cuda-11 and instead finds cuda-10? From my reading of FindCUDA.cmake, it uses nvccc to derive cuda version, which should work in my case. However it fails…