GeneralJing

yes. here is the version info. (base) root@156f417d667f:/# gcc --version gcc (Ubuntu 7.5.0-3ubuntu1~18.04) 7.5.0 Copyright (C) 2017 Free Software Foundation, Inc. This is free software; see the source for copying...

i use this docker env to train my deep learning models, so I skipped the cuda env checking steps. in the container, i can use nvidia-smi, but not nvcc -V....

thank you, i will try that.

i add the env variables in the .bashrc file, and now nvcc -V can print the normal info.Then i re-execute the .sh file, got the same question. nvcc: NVIDIA (R)...

even i install the 11.2 cuda version, it's still the same question. below that, there are a lot of these errors: /usr/local/cuda/targets/x86_64-linux/include/cuda_runtime.h:1:1: error: source file is not valid UTF-8 [clang-diagnostic-error]

the last two steps, i didn't execute the related commands. i don't know if this is the reason. i will try that tomorrow on my computer. thanks for the advice.

can you provide a dockerfile that i can reproduce the env? or a docker image that i can docker pull from the docker hub?

i'm stuck in this question for a few days. thank you for your kind advice. i will try that today. If I have any new information, I will reply here.

Scanning dependencies of target dogm_spec Scanning dependencies of target demo [ 91%] Building CXX object test/CMakeFiles/dogm_spec.dir/dogm_spec.cpp.o [ 91%] Building CXX object demo/CMakeFiles/demo.dir/main.cpp.o [ 93%] Linking CUDA device code CMakeFiles/dogm_spec.dir/cmake_device_link.o [...

the difference i think is the installation of cuda. i use the installation commands on the nvidia website. I'm not sure what's the difference between the two installation methods. because...