build error : ptxas fatal : Value 'sm_30' is not defined for option 'gpu-name'

[pmankyu]

A problem occurred during build

The build msg is:

`pmk@pmk-ubuntu:~/yolo-tensorrt/build$ cmake .. -- The C compiler identification is GNU 7.5.0 -- The CXX compiler identification is GNU 7.5.0 -- Check for working C compiler: /usr/bin/cc -- Check for working C compiler: /usr/bin/cc -- works -- Detecting C compiler ABI info -- Detecting C compiler ABI info - done -- Detecting C compile features -- Detecting C compile features - done -- Check for working CXX compiler: /usr/bin/c++ -- Check for working CXX compiler: /usr/bin/c++ -- works -- Detecting CXX compiler ABI info -- Detecting CXX compiler ABI info - done -- Detecting CXX compile features -- Detecting CXX compile features - done -- The CUDA compiler identification is unknown -- Check for working CUDA compiler: /usr/bin/nvcc -- Check for working CUDA compiler: /usr/bin/nvcc -- broken CMake Error at /usr/share/cmake-3.10/Modules/CMakeTestCUDACompiler.cmake:46 (message): The CUDA compiler

"/usr/bin/nvcc"

is not able to compile a simple test program.

It fails with the following output:

Change Dir: /home/pmk/yolo-tensorrt/build/CMakeFiles/CMakeTmp

Run Build Command:"/usr/bin/make" "cmTC_14f93/fast"
/usr/bin/make -f CMakeFiles/cmTC_14f93.dir/build.make CMakeFiles/cmTC_14f93.dir/build
make[1]: 디렉터리 '/home/pmk/yolo-tensorrt/build/CMakeFiles/CMakeTmp' 들어감
Building CUDA object CMakeFiles/cmTC_14f93.dir/main.cu.o
/usr/bin/nvcc     -x cu -c /home/pmk/yolo-tensorrt/build/CMakeFiles/CMakeTmp/main.cu -o CMakeFiles/cmTC_14f93.dir/main.cu.o
ptxas fatal   : Value 'sm_30' is not defined for option 'gpu-name'
CMakeFiles/cmTC_14f93.dir/build.make:65: recipe for target 'CMakeFiles/cmTC_14f93.dir/main.cu.o' failed
make[1]: *** [CMakeFiles/cmTC_14f93.dir/main.cu.o] Error 255
make[1]: 디렉터리 '/home/pmk/yolo-tensorrt/build/CMakeFiles/CMakeTmp' 나감
Makefile:126: recipe for target 'cmTC_14f93/fast' failed
make: *** [cmTC_14f93/fast] Error 2

CMake will not be able to correctly generate this project. Call Stack (most recent call first): CMakeLists.txt:3 (enable_language)

-- Configuring incomplete, errors occurred! `

[enazoe]

@pmankyu set the nvcc path to your environment variables

[pmankyu]

@enazoe I have already setting in bashrc

export PATH=/usr/local/cuda-11.4/bin:$PATH export LD_LIBRARY_PATH=/usr/local/cuda-11.4/lib64:$LD_LIBRARY_PATH export CUDA_HOME=/usr/local/cuda-11.4

[Wilson-shuo]

I also meet this bug. Here is the way how i sloved it. Add extra additional instructions behind your cmake order like below： sudo cmake .. -DCMAKE_CUDA_COMPILER:PATH=/usr/local/cuda-10.2/bin/nvcc (Remember to change your own CUDA path) That could appoint compiler of cudacxx. After that, you can compile successfully

[pmankyu]

@Wilson-shuo Thank you! cmake problem is solved, but the following error occurs when "make"

[Wilson-shuo]

It seemed that you are compiling this repo on a x86/x64 plantform, not on Jetson series. I could not understand the Korean words behind the "Fatal Error". I guess that means the cmake could not find nvinfer.h? This file belong to TensorRT sdk. So Please make sure you have installed "Tensor RT SDK" correctly on your computer.

[pmankyu]

@Wilson-shuo thank you and my system is ubuntu 18.04, cuda 11.3, opencv 4.2.0, tensorrt 8.0.1, cudnn 8.2.1 As you said, I installed Tensorrt and when I "make" it fails with the error below. What's the problem?

[enazoe]

@pmankyu checkout the branch to TRT8

[pmankyu]

@enazoe Thank you! When I checked out and "make", the following error occurred I also installed tensorRT normally, but I don't know what the problem is.

[pmankyu]

@enazoe I update latest cmake 3.21.1 and when I do 'cmake' at TRT8 branch

A warning message appears

[pmankyu]

@enazoe I solve the problem fatal error NvInfer.h

change include path

include "/usr/include/x86_64-linux-gnu/NvInfer.h" at /modules/plugin_factory.h

thank you all