"CMakeCUDACompilerId.cu" failed.

[paulgnz]

I get this error when trying to cmake

------ System -----

Intel i9, RTX 3070
Windows 10
nvcc --version = V10.2.89
python --version = Python 3.10.2

Developer Command Prompt for VS 2019

Tried uninstalling everything, installing from scratch etc

------ Error -----

C:\ngp>cd instant-ngp

C:\ngp\instant-ngp>cmake . -B build
-- Selecting Windows SDK version 10.0.19041.0 to target Windows 10.0.19044.
CMake Error at C:/Program Files/CMake/share/cmake-3.24/Modules/CMakeDetermineCompilerId.cmake:739 (message):
Compiling the CUDA compiler identification source file
"CMakeCUDACompilerId.cu" failed.

Compiler:

Build flags:

Id flags: --keep;--keep-dir;tmp -v

The output was:

1

Microsoft (R) Build Engine version 16.11.2+f32259642 for .NET Framework

Copyright (C) Microsoft Corporation. All rights reserved.

Build started 9/2/2022 10:00:34 PM.

Project
"C:\ngp\instant-ngp\build\CMakeFiles\3.24.1\CompilerIdCUDA\CompilerIdCUDA.vcxproj"
on node 1 (default targets).

PrepareForBuild:

Creating directory "Debug\".
Creating directory "Debug\CompilerIdCUDA.tlog\".

InitializeBuildStatus:

Creating "Debug\CompilerIdCUDA.tlog\unsuccessfulbuild" because "AlwaysCreate" was specified.

AddCudaCompileDeps:

C:\Program Files (x86)\Microsoft Visual Studio\2019\Community\VC\Tools\MSVC\14.29.30133\bin\HostX64\x64\cl.exe /E /nologo /showIncludes /TP /D__CUDACC__ /D_MBCS /I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.2\include" /I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.2\bin" /I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.2\include" /I. /FIcuda_runtime.h /c C:\ngp\instant-ngp\build\CMakeFiles\3.24.1\CompilerIdCUDA\CMakeCUDACompilerId.cu

Project
"C:\ngp\instant-ngp\build\CMakeFiles\3.24.1\CompilerIdCUDA\CompilerIdCUDA.vcxproj"
(1) is building
"C:\ngp\instant-ngp\build\CMakeFiles\3.24.1\CompilerIdCUDA\CompilerIdCUDA.vcxproj"
(1:2) on node 1 (CudaBuildCore target(s)).

CudaBuildCore:

Compiling CUDA source file CMakeCUDACompilerId.cu...
cmd.exe /C "C:\Users\Big Red\AppData\Local\Temp\tmp57729dd2ac15416dbec043081fdf7ebf.cmd"
"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.2\bin\nvcc.exe" -gencode=arch=compute_30,code=\"sm_30,compute_30\" --use-local-env -ccbin "C:\Program Files (x86)\Microsoft Visual Studio\2019\Community\VC\Tools\MSVC\14.29.30133\bin\HostX64\x64" -x cu  -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.2\include" -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.2\include"  -G   --keep-dir x64\Debug -maxrregcount=0  --machine 64 --compile -cudart static -v -g   -D_MBCS -Xcompiler "/EHsc /W0 /nologo /Od /FdDebug\vc142.pdb /FS /Zi /RTC1 /MDd " -o Debug\CMakeCUDACompilerId.cu.obj "C:\ngp\instant-ngp\build\CMakeFiles\3.24.1\CompilerIdCUDA\CMakeCUDACompilerId.cu"

C:\ngp\instant-ngp\build\CMakeFiles\3.24.1\CompilerIdCUDA>"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.2\bin\nvcc.exe" -gencode=arch=compute_30,code=\"sm_30,compute_30\" --use-local-env -ccbin "C:\Program Files (x86)\Microsoft Visual Studio\2019\Community\VC\Tools\MSVC\14.29.30133\bin\HostX64\x64" -x cu  -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.2\include" -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.2\include"  -G   --keep-dir x64\Debug -maxrregcount=0  --machine 64 --compile -cudart static -v -g   -D_MBCS -Xcompiler "/EHsc /W0 /nologo /Od /FdDebug\vc142.pdb /FS /Zi /RTC1 /MDd " -o Debug\CMakeCUDACompilerId.cu.obj "C:\ngp\instant-ngp\build\CMakeFiles\3.24.1\CompilerIdCUDA\CMakeCUDACompilerId.cu"

C:\Program Files (x86)\Microsoft Visual
Studio\2019\Community\MSBuild\Microsoft\VC\v160\BuildCustomizations\CUDA
10.2.targets(764,9): error MSB3721: The command ""C:\Program Files\NVIDIA
GPU Computing Toolkit\CUDA\v10.2\bin\nvcc.exe"
-gencode=arch=compute_30,code="sm_30,compute_30" --use-local-env -ccbin
"C:\Program Files (x86)\Microsoft Visual
Studio\2019\Community\VC\Tools\MSVC\14.29.30133\bin\HostX64\x64" -x cu
-I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.2\include"
-I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.2\include" -G
--keep-dir x64\Debug -maxrregcount=0 --machine 64 --compile -cudart static
-v -g -D_MBCS -Xcompiler "/EHsc /W0 /nologo /Od /FdDebug\vc142.pdb /FS /Zi
/RTC1 /MDd " -o Debug\CMakeCUDACompilerId.cu.obj
"C:\ngp\instant-ngp\build\CMakeFiles\3.24.1\CompilerIdCUDA\CMakeCUDACompilerId.cu""
exited with code 1.
[C:\ngp\instant-ngp\build\CMakeFiles\3.24.1\CompilerIdCUDA\CompilerIdCUDA.vcxproj]

Done Building Project
"C:\ngp\instant-ngp\build\CMakeFiles\3.24.1\CompilerIdCUDA\CompilerIdCUDA.vcxproj"
(CudaBuildCore target(s)) -- FAILED.

Done Building Project
"C:\ngp\instant-ngp\build\CMakeFiles\3.24.1\CompilerIdCUDA\CompilerIdCUDA.vcxproj"
(default targets) -- FAILED.

Build FAILED.

"C:\ngp\instant-ngp\build\CMakeFiles\3.24.1\CompilerIdCUDA\CompilerIdCUDA.vcxproj"
(default target) (1) ->

"C:\ngp\instant-ngp\build\CMakeFiles\3.24.1\CompilerIdCUDA\CompilerIdCUDA.vcxproj"
(CudaBuildCore target) (1:2) ->

(CudaBuildCore target) ->

C:\Program Files (x86)\Microsoft Visual Studio\2019\Community\MSBuild\Microsoft\VC\v160\BuildCustomizations\CUDA 10.2.targets(764,9): error MSB3721: The command ""C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.2\bin\nvcc.exe" -gencode=arch=compute_30,code=\"sm_30,compute_30\" --use-local-env -ccbin "C:\Program Files (x86)\Microsoft Visual Studio\2019\Community\VC\Tools\MSVC\14.29.30133\bin\HostX64\x64" -x cu  -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.2\include" -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.2\include"  -G   --keep-dir x64\Debug -maxrregcount=0  --machine 64 --compile -cudart static -v -g   -D_MBCS -Xcompiler "/EHsc /W0 /nologo /Od /FdDebug\vc142.pdb /FS /Zi /RTC1 /MDd " -o Debug\CMakeCUDACompilerId.cu.obj "C:\ngp\instant-ngp\build\CMakeFiles\3.24.1\CompilerIdCUDA\CMakeCUDACompilerId.cu"" exited with code 1. [C:\ngp\instant-ngp\build\CMakeFiles\3.24.1\CompilerIdCUDA\CompilerIdCUDA.vcxproj]



  0 Warning(s)
  1 Error(s)

Time Elapsed 00:00:01.52

Call Stack (most recent call first):
C:/Program Files/CMake/share/cmake-3.24/Modules/CMakeDetermineCompilerId.cmake:6 (CMAKE_DETERMINE_COMPILER_ID_BUILD)
C:/Program Files/CMake/share/cmake-3.24/Modules/CMakeDetermineCompilerId.cmake:48 (__determine_compiler_id_test)
C:/Program Files/CMake/share/cmake-3.24/Modules/CMakeDetermineCUDACompiler.cmake:307 (CMAKE_DETERMINE_COMPILER_ID)
CMakeLists.txt:11 (project)

-- Configuring incomplete, errors occurred!
See also "C:/ngp/instant-ngp/build/CMakeFiles/CMakeOutput.log".
See also "C:/ngp/instant-ngp/build/CMakeFiles/CMakeError.log".

C:\ngp\instant-ngp>

[Pangolin112]

I also met this problem, still wondering what the problem is.

[nepfaff]

Adding -D CMAKE_CUDA_COMPILER=$(which nvcc) to cmake fixed this for me:
cmake . -D TCNN_CUDA_ARCHITECTURES=86 -D CMAKE_CUDA_COMPILER=$(which nvcc) -B build

[Pangolin112]

This way for my device(win11, cuda 11.7, 3070Ti[laptop]) the compiler can be found, but the error remains the same.

[MichalSzczekocki]

Adding -D CMAKE_CUDA_COMPILER=$(which nvcc) to cmake fixed this for me: cmake . -D TCNN_CUDA_ARCHITECTURES=86 -D CMAKE_CUDA_COMPILER=$(which nvcc) -B build

Hi. I'm also trying to fix it as they're dependencies that have to be resolved. I had the same problem as @Pangolin112 but after using command you recommended I got this error.

-- The CUDA compiler identification is NVIDIA 10.2.89
-- Detecting CUDA compiler ABI info
-- Detecting CUDA compiler ABI info - done
-- Check for working CUDA compiler: /usr/local/cuda/bin/nvcc - skipped
-- Detecting CUDA compile features
-- Detecting CUDA compile features - done
-- No release type specified. Setting to 'Release'.
-- Automatic GPU detection failed. Building for common architectures.
-- Obtained target architecture from CMake variable TCNN_CUDA_ARCHITECTURES=86
CMake Warning at dependencies/tiny-cuda-nn/CMakeLists.txt:118 (message):
CUDA version 10.2 is too low for architecture 86. Targeting architecture
75 instead.

-- Targeting GPU architectures: 75
-- Module support is disabled.
-- Version: 9.1.1
-- Build type: Release
CMake Error at dependencies/tiny-cuda-nn/dependencies/fmt/CMakeLists.txt:234 (target_compile_features):
target_compile_features no known features for CXX compiler

"GNU"

version 8.4.0.

-- Found Vulkan: /usr/lib/x86_64-linux-gnu/libvulkan.so (found version "1.3.204") missing components: glslc glslangValidator
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD - Success
-- Found Threads: TRUE
-- Using X11 for window creation
-- Found X11: /usr/include
-- Looking for XOpenDisplay in /usr/lib/x86_64-linux-gnu/libX11.so;/usr/lib/x86_64-linux-gnu/libXext.so
-- Looking for XOpenDisplay in /usr/lib/x86_64-linux-gnu/libX11.so;/usr/lib/x86_64-linux-gnu/libXext.so - found
-- Looking for gethostbyname
-- Looking for gethostbyname - found
-- Looking for connect
-- Looking for connect - found
-- Looking for remove
-- Looking for remove - found
-- Looking for shmat
-- Looking for shmat - found
-- Found GLEW: /usr/include (found version "2.2.0")
-- Found OpenMP_C: -fopenmp (found version "4.5")
-- Found OpenMP_CXX: -fopenmp (found version "4.5")
-- Found OpenMP: TRUE (found version "4.5")
-- OptiX headers (optix.h and friends) not found.
CMake Warning at CMakeLists.txt:194 (message):
OptiX was not found. Neural graphics primitives will still compile and run
correctly. However, SDF training in 'raystab' and 'pathescape' modes will
be significantly slower.

-- Found Python: /usr/bin/python3.10 (found suitable version "3.10.6", minimum required is "3.7") found components: Interpreter Development Development.Module Development.Embed
-- pybind11 v2.10.1
-- Performing Test HAS_FLTO
-- Performing Test HAS_FLTO - Success
-- Configuring incomplete, errors occurred!
See also "/home/michaels/CLionProjects/instant-ngp/build/CMakeFiles/CMakeOutput.log".
See also "/home/michaels/CLionProjects/instant-ngp/build/CMakeFiles/CMakeError.log".

But after deleting the build directory and typing the command once again it seems it has built successfully:

michaels@michaels-320S:~/CLionProjects/instant-ngp$ cmake . -D TCNN_CUDA_ARCHITECTURES=86 -D CMAKE_CUDA_COMPILER=$(which nvcc) -B build
-- The C compiler identification is GNU 8.4.0
-- The CXX compiler identification is GNU 8.4.0
-- The CUDA compiler identification is NVIDIA 10.2.89
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Check for working C compiler: /usr/bin/cc - skipped
-- Detecting C compile features
-- Detecting C compile features - done
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Check for working CXX compiler: /usr/bin/c++ - skipped
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Detecting CUDA compiler ABI info
-- Detecting CUDA compiler ABI info - done
-- Check for working CUDA compiler: /usr/local/cuda/bin/nvcc - skipped
-- Detecting CUDA compile features
-- Detecting CUDA compile features - done
-- No release type specified. Setting to 'Release'.
-- Automatic GPU detection failed. Building for common architectures.
-- Obtained target architecture from CMake variable TCNN_CUDA_ARCHITECTURES=86
CMake Warning at dependencies/tiny-cuda-nn/CMakeLists.txt:118 (message):
CUDA version 10.2 is too low for architecture 86. Targeting architecture
75 instead.

-- Targeting GPU architectures: 75
-- Module support is disabled.
-- Version: 9.1.1
-- Build type: Release
-- Found Vulkan: /usr/lib/x86_64-linux-gnu/libvulkan.so (found version "1.3.204") missing components: glslc glslangValidator
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD - Success
-- Found Threads: TRUE
-- Using X11 for window creation
-- Found X11: /usr/include
-- Looking for XOpenDisplay in /usr/lib/x86_64-linux-gnu/libX11.so;/usr/lib/x86_64-linux-gnu/libXext.so
-- Looking for XOpenDisplay in /usr/lib/x86_64-linux-gnu/libX11.so;/usr/lib/x86_64-linux-gnu/libXext.so - found
-- Looking for gethostbyname
-- Looking for gethostbyname - found
-- Looking for connect
-- Looking for connect - found
-- Looking for remove
-- Looking for remove - found
-- Looking for shmat
-- Looking for shmat - found
-- Found GLEW: /usr/include (found version "2.2.0")
-- Found OpenMP_C: -fopenmp (found version "4.5")
-- Found OpenMP_CXX: -fopenmp (found version "4.5")
-- Found OpenMP: TRUE (found version "4.5")
-- OptiX headers (optix.h and friends) not found.
CMake Warning at CMakeLists.txt:194 (message):
OptiX was not found. Neural graphics primitives will still compile and run
correctly. However, SDF training in 'raystab' and 'pathescape' modes will
be significantly slower.

-- Found Python: /usr/bin/python3.10 (found suitable version "3.10.6", minimum required is "3.7") found components: Interpreter Development Development.Module Development.Embed
-- pybind11 v2.10.1
-- Performing Test HAS_FLTO
-- Performing Test HAS_FLTO - Success
-- Configuring done
-- Generating done
-- Build files have been written to: /home/michaels/CLionProjects/instant-ngp/build

[rulke]

Adding to cmake fixed this for me:-D CMAKE_CUDA_COMPILER=$(which nvcc) cmake . -D TCNN_CUDA_ARCHITECTURES=86 -D CMAKE_CUDA_COMPILER=$(which nvcc) -B build

amazing

[MikePelton]

Some confusion in the comments here as to whether this is Windows or Linux the OP is interested in; for those who fell into this hole trying to build under Windows I found setting an environment variable

set CUDATOOLKITDIR = C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.1

..fixed it for me - this was with CMake 3.25.3 and CUDA 12.1 - change your setting of the variable according to your version of the CUDA Toolkit.

[Uniformed9]

This way for my device(win11, cuda 11.7, 3070Ti[laptop]) the compiler can be found, but the error remains the same.

I ran into the same problem.Have you solved it?

[YuehChuan]

@Uniformed9 @MichalSzczekocki
@Pangolin112 @paulgnz

Check here!
https://github.com/YuehChuan/cudaVStemplate

The main issue is IDE.
Visual Studio 2022 supports CUDA 11.6 above.

[arithy]

This issue can be caused by too long a value in the environment variable PATH.
In my case, it was resolved by removing unnecessary entries from PATH.