TensorRT v9.2 installation error due to CUDA version. ""-- The CUDA compiler identification is unknown""

[jamil-z]

Hello everyone, I'm trying to install TensorRT version 9.2. I followed the requirements and switched to CUDA 12.2 on a 16GB GPU machine running Ubuntu 20.

When I run the command: cmake .. -DTRT_LIB_DIR=$TRT_LIBPATH -DTRT_OUT_DIR=pwd/out

I get the following error:
(venv_tensorrt) azureuser@myUser1:~/cloudfiles/code/Users/myUser/tensorrt/TensorRT/build$ cmake .. -DTRT_LIB_DIR=$TRT_LIBPATH -DTRT_OUT_DIR=pwd`/out
Building for TensorRT version: 8.6.1, library version: 8
-- 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/local/share/cmake-3.15/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/azureuser/cloudfiles/code/Users/myUser/tensorrt/TensorRT/build/CMakeFiles/CMakeTmp

Run Build Command(s):/usr/bin/make cmTC_a28cd/fast && /usr/bin/make -f CMakeFiles/cmTC_a28cd.dir/build.make CMakeFiles/cmTC_a28cd.dir/build
make[1]: Entering directory '/mnt/batch/tasks/shared/LS_root/mounts/clusters/myUser1/code/Users/myUser/tensorrt/TensorRT/build/CMakeFiles/CMakeTmp'
Building CUDA object CMakeFiles/cmTC_a28cd.dir/main.cu.o
/usr/bin/nvcc     -x cu -c /home/azureuser/cloudfiles/code/Users/myUser/tensorrt/TensorRT/build/CMakeFiles/CMakeTmp/main.cu -o CMakeFiles/cmTC_a28cd.dir/main.cu.o
sh: 1: cicc: not found
make[1]: *** [CMakeFiles/cmTC_a28cd.dir/build.make:66: CMakeFiles/cmTC_a28cd.dir/main.cu.o] Error 127
make[1]: Leaving directory '/mnt/batch/tasks/shared/LS_root/mounts/clusters/myUser1/code/Users/myUser/tensorrt/TensorRT/build/CMakeFiles/CMakeTmp'
make: *** [Makefile:121: cmTC_a28cd/fast] Error 2

CMake will not be able to correctly generate this project.
Call Stack (most recent call first):
CMakeLists.txt:75 (project)

-- Configuring incomplete, errors occurred!
See also "/home/azureuser/cloudfiles/code/Users/myUser/tensorrt/TensorRT/build/CMakeFiles/CMakeOutput.log".
See also "/home/azureuser/cloudfiles/code/Users/myUser/tensorrt/TensorRT/build/CMakeFiles/CMakeError.log".`

The output of my command: nvcc --version

is as follows:
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2023 NVIDIA Corporation
Built on Tue_Jun_13_19:16:58_PDT_2023
Cuda compilation tools, release 12.2, V12.2.91
Build cuda_12.2.r12.2/compiler.32965470_0

From what I understand, it's an error related to CUDA not recognizing its installation correctly, but I've been trying to solve it and can't pinpoint the specific cause. Please, if anyone could help me with this, I would appreciate it.

[zerollzeng]

@rajeevsrao ^ ^

[ttyio]

@jamil-z , have you tried compile using the docker environment follow the README? thanks!

[ttyio]

closing since no activity for more than 3 weeks, pls reopen if you still have question, thanks all!