1
votes

Aucun CMAKE_CUDA_COMPILER n'a pu être trouvé

J'essaie de générer un système de compilation pour CMake, mais cela me lance une erreur, que je n'ai pas réussi à corriger depuis des jours. J'ai installé la dernière version de VC 2017 et CUDA.

Dans mon répertoire, j'ai créé le répertoire de construction et j'ai exécuté:

    Compiling the CUDA compiler identification source file "CMakeCUDACompilerId.cu" failed.
Compiler:  
Build flags: 
Id flags:  -v;--keep;--keep-dir;tmp

The output was:
1
Microsoft (R) Build Engine version 15.9.21+g9802d43bc3 for .NET Framework
Copyright (C) Microsoft Corporation. All rights reserved.

Build started 07-Mar-19 14:14:03.
Project "C:\Users\lenovo\Documents\Fax\4\Lasersko skeniranje\Projekat\Open3D-PointNet2-Semantic3D\tf_ops\build\CMakeFiles\3.13.3\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\2017\Community\VC\Tools\MSVC\14.16.27023\bin\HostX86\x86\cl.exe /E /nologo /showIncludes /TP /D__CUDACC__ /D_MBCS /I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\include" /I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\bin" /I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\include" /I. /FIcuda_runtime.h /c "C:\Users\lenovo\Documents\Fax\4\Lasersko skeniranje\Projekat\Open3D-PointNet2-Semantic3D\tf_ops\build\CMakeFiles\3.13.3\CompilerIdCUDA\CMakeCUDACompilerId.cu" 
Project "C:\Users\lenovo\Documents\Fax\4\Lasersko skeniranje\Projekat\Open3D-PointNet2-Semantic3D\tf_ops\build\CMakeFiles\3.13.3\CompilerIdCUDA\CompilerIdCUDA.vcxproj" (1) is building "C:\Users\lenovo\Documents\Fax\4\Lasersko skeniranje\Projekat\Open3D-PointNet2-Semantic3D\tf_ops\build\CMakeFiles\3.13.3\CompilerIdCUDA\CompilerIdCUDA.vcxproj" (1:2) on node 1 (CudaBuildCore target(s)).
CudaBuildCore:
  Compiling CUDA source file CMakeCUDACompilerId.cu...
  cmd.exe /C "C:\Users\lenovo\AppData\Local\Temp\tmp5b923de4e22e4f04be6ba6f85d34ca98.cmd"
  "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\bin\nvcc.exe" -gencode=arch=compute_35,code=\"sm_35,compute_35\" --use-local-env -ccbin "C:\Program Files (x86)\Microsoft Visual Studio\2017\Community\VC\Tools\MSVC\14.16.27023\bin\HostX86\x86" -x cu  -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\include" -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\include"  -G   --keep-dir Debug -maxrregcount=0  --machine 32 --compile -cudart static  -g   -D_MBCS -Xcompiler "/EHsc /W0 /nologo /Od /FdDebug\vc141.pdb /FS /Zi /RTC1 /MDd " -o Debug\CMakeCUDACompilerId.cu.obj "C:\Users\lenovo\Documents\Fax\4\Lasersko skeniranje\Projekat\Open3D-PointNet2-Semantic3D\tf_ops\build\CMakeFiles\3.13.3\CompilerIdCUDA\CMakeCUDACompilerId.cu"

  C:\Users\lenovo\Documents\Fax\4\Lasersko skeniranje\Projekat\Open3D-PointNet2-Semantic3D\tf_ops\build\CMakeFiles\3.13.3\CompilerIdCUDA>"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\bin\nvcc.exe" -gencode=arch=compute_35,code=\"sm_35,compute_35\" --use-local-env -ccbin "C:\Program Files (x86)\Microsoft Visual Studio\2017\Community\VC\Tools\MSVC\14.16.27023\bin\HostX86\x86" -x cu  -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\include" -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\include"  -G   --keep-dir Debug -maxrregcount=0  --machine 32 --compile -cudart static  -g   -D_MBCS -Xcompiler "/EHsc /W0 /nologo /Od /FdDebug\vc141.pdb /FS /Zi /RTC1 /MDd " -o Debug\CMakeCUDACompilerId.cu.obj "C:\Users\lenovo\Documents\Fax\4\Lasersko skeniranje\Projekat\Open3D-PointNet2-Semantic3D\tf_ops\build\CMakeFiles\3.13.3\CompilerIdCUDA\CMakeCUDACompilerId.cu" 
  nvcc fatal   : 32 bit compilation is only supported for Microsoft Visual Studio 2013 and earlier
C:\Program Files (x86)\Microsoft Visual Studio\2017\Community\Common7\IDE\VC\VCTargets\BuildCustomizations\CUDA 10.1.targets(757,9): error MSB3721: The command ""C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\bin\nvcc.exe" -gencode=arch=compute_35,code=\"sm_35,compute_35\" --use-local-env -ccbin "C:\Program Files (x86)\Microsoft Visual Studio\2017\Community\VC\Tools\MSVC\14.16.27023\bin\HostX86\x86" -x cu  -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\include" -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\include"  -G   --keep-dir Debug -maxrregcount=0  --machine 32 --compile -cudart static  -g   -D_MBCS -Xcompiler "/EHsc /W0 /nologo /Od /FdDebug\vc141.pdb /FS /Zi /RTC1 /MDd " -o Debug\CMakeCUDACompilerId.cu.obj "C:\Users\lenovo\Documents\Fax\4\Lasersko skeniranje\Projekat\Open3D-PointNet2-Semantic3D\tf_ops\build\CMakeFiles\3.13.3\CompilerIdCUDA\CMakeCUDACompilerId.cu"" exited with code 1. [C:\Users\lenovo\Documents\Fax\4\Lasersko skeniranje\Projekat\Open3D-PointNet2-Semantic3D\tf_ops\build\CMakeFiles\3.13.3\CompilerIdCUDA\CompilerIdCUDA.vcxproj]
Done Building Project "C:\Users\lenovo\Documents\Fax\4\Lasersko skeniranje\Projekat\Open3D-PointNet2-Semantic3D\tf_ops\build\CMakeFiles\3.13.3\CompilerIdCUDA\CompilerIdCUDA.vcxproj" (CudaBuildCore target(s)) -- FAILED.
Done Building Project "C:\Users\lenovo\Documents\Fax\4\Lasersko skeniranje\Projekat\Open3D-PointNet2-Semantic3D\tf_ops\build\CMakeFiles\3.13.3\CompilerIdCUDA\CompilerIdCUDA.vcxproj" (default targets) -- FAILED.

Build FAILED.

"C:\Users\lenovo\Documents\Fax\4\Lasersko skeniranje\Projekat\Open3D-PointNet2-Semantic3D\tf_ops\build\CMakeFiles\3.13.3\CompilerIdCUDA\CompilerIdCUDA.vcxproj" (default target) (1) ->
"C:\Users\lenovo\Documents\Fax\4\Lasersko skeniranje\Projekat\Open3D-PointNet2-Semantic3D\tf_ops\build\CMakeFiles\3.13.3\CompilerIdCUDA\CompilerIdCUDA.vcxproj" (CudaBuildCore target) (1:2) ->
(CudaBuildCore target) -> 
  C:\Program Files (x86)\Microsoft Visual Studio\2017\Community\Common7\IDE\VC\VCTargets\BuildCustomizations\CUDA 10.1.targets(757,9): error MSB3721: The command ""C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\bin\nvcc.exe" -gencode=arch=compute_35,code=\"sm_35,compute_35\" --use-local-env -ccbin "C:\Program Files (x86)\Microsoft Visual Studio\2017\Community\VC\Tools\MSVC\14.16.27023\bin\HostX86\x86" -x cu  -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\include" -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\include"  -G   --keep-dir Debug -maxrregcount=0  --machine 32 --compile -cudart static  -g   -D_MBCS -Xcompiler "/EHsc /W0 /nologo /Od /FdDebug\vc141.pdb /FS /Zi /RTC1 /MDd " -o Debug\CMakeCUDACompilerId.cu.obj "C:\Users\lenovo\Documents\Fax\4\Lasersko skeniranje\Projekat\Open3D-PointNet2-Semantic3D\tf_ops\build\CMakeFiles\3.13.3\CompilerIdCUDA\CMakeCUDACompilerId.cu"" exited with code 1. [C:\Users\lenovo\Documents\Fax\4\Lasersko skeniranje\Projekat\Open3D-PointNet2-Semantic3D\tf_ops\build\CMakeFiles\3.13.3\CompilerIdCUDA\CompilerIdCUDA.vcxproj]

    0 Warning(s)
    1 Error(s)

Time Elapsed 00:00:02.74


Compiling the CUDA compiler identification source file "CMakeCUDACompilerId.cu" failed.
Compiler:  
Build flags: 
Id flags:  -v;--keep;--keep-dir;tmp

The output was:
1
Microsoft (R) Build Engine version 15.9.21+g9802d43bc3 for .NET Framework
Copyright (C) Microsoft Corporation. All rights reserved.

Build started 07-Mar-19 14:14:06.
Project "C:\Users\lenovo\Documents\Fax\4\Lasersko skeniranje\Projekat\Open3D-PointNet2-Semantic3D\tf_ops\build\CMakeFiles\3.13.3\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\2017\Community\VC\Tools\MSVC\14.16.27023\bin\HostX86\x86\cl.exe /E /nologo /showIncludes /TP /D__CUDACC__ /D_MBCS /I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\include" /I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\bin" /I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\include" /I. /FIcuda_runtime.h /c "C:\Users\lenovo\Documents\Fax\4\Lasersko skeniranje\Projekat\Open3D-PointNet2-Semantic3D\tf_ops\build\CMakeFiles\3.13.3\CompilerIdCUDA\CMakeCUDACompilerId.cu" 
Project "C:\Users\lenovo\Documents\Fax\4\Lasersko skeniranje\Projekat\Open3D-PointNet2-Semantic3D\tf_ops\build\CMakeFiles\3.13.3\CompilerIdCUDA\CompilerIdCUDA.vcxproj" (1) is building "C:\Users\lenovo\Documents\Fax\4\Lasersko skeniranje\Projekat\Open3D-PointNet2-Semantic3D\tf_ops\build\CMakeFiles\3.13.3\CompilerIdCUDA\CompilerIdCUDA.vcxproj" (1:2) on node 1 (CudaBuildCore target(s)).
CudaBuildCore:
  Compiling CUDA source file CMakeCUDACompilerId.cu...
  cmd.exe /C "C:\Users\lenovo\AppData\Local\Temp\tmp6a6113e32a54449682bdc11103da68f9.cmd"
  "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\bin\nvcc.exe" -gencode=arch=compute_35,code=\"sm_35,compute_35\" --use-local-env -ccbin "C:\Program Files (x86)\Microsoft Visual Studio\2017\Community\VC\Tools\MSVC\14.16.27023\bin\HostX86\x86" -x cu  -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\include" -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\include"  -G   --keep-dir Debug -maxrregcount=0  --machine 32 --compile -cudart static  -g   -D_MBCS -Xcompiler "/EHsc /W0 /nologo /Od /FdDebug\vc141.pdb /FS /Zi /RTC1 /MDd " -o Debug\CMakeCUDACompilerId.cu.obj "C:\Users\lenovo\Documents\Fax\4\Lasersko skeniranje\Projekat\Open3D-PointNet2-Semantic3D\tf_ops\build\CMakeFiles\3.13.3\CompilerIdCUDA\CMakeCUDACompilerId.cu"

  C:\Users\lenovo\Documents\Fax\4\Lasersko skeniranje\Projekat\Open3D-PointNet2-Semantic3D\tf_ops\build\CMakeFiles\3.13.3\CompilerIdCUDA>"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\bin\nvcc.exe" -gencode=arch=compute_35,code=\"sm_35,compute_35\" --use-local-env -ccbin "C:\Program Files (x86)\Microsoft Visual Studio\2017\Community\VC\Tools\MSVC\14.16.27023\bin\HostX86\x86" -x cu  -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\include" -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\include"  -G   --keep-dir Debug -maxrregcount=0  --machine 32 --compile -cudart static  -g   -D_MBCS -Xcompiler "/EHsc /W0 /nologo /Od /FdDebug\vc141.pdb /FS /Zi /RTC1 /MDd " -o Debug\CMakeCUDACompilerId.cu.obj "C:\Users\lenovo\Documents\Fax\4\Lasersko skeniranje\Projekat\Open3D-PointNet2-Semantic3D\tf_ops\build\CMakeFiles\3.13.3\CompilerIdCUDA\CMakeCUDACompilerId.cu" 
  nvcc fatal   : 32 bit compilation is only supported for Microsoft Visual Studio 2013 and earlier
C:\Program Files (x86)\Microsoft Visual Studio\2017\Community\Common7\IDE\VC\VCTargets\BuildCustomizations\CUDA 10.1.targets(757,9): error MSB3721: The command ""C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\bin\nvcc.exe" -gencode=arch=compute_35,code=\"sm_35,compute_35\" --use-local-env -ccbin "C:\Program Files (x86)\Microsoft Visual Studio\2017\Community\VC\Tools\MSVC\14.16.27023\bin\HostX86\x86" -x cu  -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\include" -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\include"  -G   --keep-dir Debug -maxrregcount=0  --machine 32 --compile -cudart static  -g   -D_MBCS -Xcompiler "/EHsc /W0 /nologo /Od /FdDebug\vc141.pdb /FS /Zi /RTC1 /MDd " -o Debug\CMakeCUDACompilerId.cu.obj "C:\Users\lenovo\Documents\Fax\4\Lasersko skeniranje\Projekat\Open3D-PointNet2-Semantic3D\tf_ops\build\CMakeFiles\3.13.3\CompilerIdCUDA\CMakeCUDACompilerId.cu"" exited with code 1. [C:\Users\lenovo\Documents\Fax\4\Lasersko skeniranje\Projekat\Open3D-PointNet2-Semantic3D\tf_ops\build\CMakeFiles\3.13.3\CompilerIdCUDA\CompilerIdCUDA.vcxproj]
Done Building Project "C:\Users\lenovo\Documents\Fax\4\Lasersko skeniranje\Projekat\Open3D-PointNet2-Semantic3D\tf_ops\build\CMakeFiles\3.13.3\CompilerIdCUDA\CompilerIdCUDA.vcxproj" (CudaBuildCore target(s)) -- FAILED.
Done Building Project "C:\Users\lenovo\Documents\Fax\4\Lasersko skeniranje\Projekat\Open3D-PointNet2-Semantic3D\tf_ops\build\CMakeFiles\3.13.3\CompilerIdCUDA\CompilerIdCUDA.vcxproj" (default targets) -- FAILED.

Build FAILED.

"C:\Users\lenovo\Documents\Fax\4\Lasersko skeniranje\Projekat\Open3D-PointNet2-Semantic3D\tf_ops\build\CMakeFiles\3.13.3\CompilerIdCUDA\CompilerIdCUDA.vcxproj" (default target) (1) ->
"C:\Users\lenovo\Documents\Fax\4\Lasersko skeniranje\Projekat\Open3D-PointNet2-Semantic3D\tf_ops\build\CMakeFiles\3.13.3\CompilerIdCUDA\CompilerIdCUDA.vcxproj" (CudaBuildCore target) (1:2) ->
(CudaBuildCore target) -> 
  C:\Program Files (x86)\Microsoft Visual Studio\2017\Community\Common7\IDE\VC\VCTargets\BuildCustomizations\CUDA 10.1.targets(757,9): error MSB3721: The command ""C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\bin\nvcc.exe" -gencode=arch=compute_35,code=\"sm_35,compute_35\" --use-local-env -ccbin "C:\Program Files (x86)\Microsoft Visual Studio\2017\Community\VC\Tools\MSVC\14.16.27023\bin\HostX86\x86" -x cu  -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\include" -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\include"  -G   --keep-dir Debug -maxrregcount=0  --machine 32 --compile -cudart static  -g   -D_MBCS -Xcompiler "/EHsc /W0 /nologo /Od /FdDebug\vc141.pdb /FS /Zi /RTC1 /MDd " -o Debug\CMakeCUDACompilerId.cu.obj "C:\Users\lenovo\Documents\Fax\4\Lasersko skeniranje\Projekat\Open3D-PointNet2-Semantic3D\tf_ops\build\CMakeFiles\3.13.3\CompilerIdCUDA\CMakeCUDACompilerId.cu"" exited with code 1. [C:\Users\lenovo\Documents\Fax\4\Lasersko skeniranje\Projekat\Open3D-PointNet2-Semantic3D\tf_ops\build\CMakeFiles\3.13.3\CompilerIdCUDA\CompilerIdCUDA.vcxproj]

    0 Warning(s)
    1 Error(s)

Time Elapsed 00:00:02.65

Après cela, il a commencé à le créer mais il n'a pas trouvé le compilateur CUDA

C:\Users\lenovo\Documents\Fax\4\Lasersko skeniranje\Projekat\Open3D-PointNet2-Semantic3D\tf_ops\build>cmake ..

-- Building for: Visual Studio 15 2017
-- The CXX compiler identification is MSVC 19.16.27027.1
-- The CUDA compiler identification is unknown
-- Check for working CXX compiler: C:/Program Files (x86)/Microsoft Visual Studio/2017/Community/VC/Tools/MSVC/14.16.27023/bin/Hostx86/x86/cl.exe
-- Check for working CXX compiler: C:/Program Files (x86)/Microsoft Visual Studio/2017/Community/VC/Tools/MSVC/14.16.27023/bin/Hostx86/x86/cl.exe -- works
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Detecting CXX compile features
-- Detecting CXX compile features - done
CMake Error at CMakeLists.txt:5 (project):

>  No CMAKE_CUDA_COMPILER could be found.

>-- Configuring incomplete, errors occurred!
See also "C:/Users/lenovo/Documents/Fax/4/Lasersko skeniranje/Projekat/Open3D-PointNet2-Semantic3D/tf_ops/build/CMakeFiles/CMakeOutput.log".

Le fichier journal dit:

cmake ..

Une idée? Je n'arrive pas à résoudre le problème ... J'utilise Win10

cmake cuda

4 commentaires

Votre problème semble être nvcc fatal: la compilation 32 bits n'est prise en charge que pour Microsoft Visual Studio 2013 et versions antérieures . Vous devriez vérifier vos fichiers CMake s'il essaie de construire en 32 bits au lieu de 64 bits

Oh désolé, j'ai lu trop vite, le problème vient de CMakeCUDACompilerId.cu . Peut-être essayer de mettre à jour CMake? Je n'ai pas vu la version CMake dans les journaux

La version installée était 3.13 (installée avec pip). Mais maintenant, j'ai installé le Cmake GUI 3.14 et ai exécuté avec lui, il a reconnu le compilateur. Merci de votre aide

Aucun problème. J'en ferai une réponse pour qu'il ait une réponse acceptée

3 Réponses :

2
votes

Selon les journaux, le problème est nvcc fatal: la compilation 32 bits n'est prise en charge que pour Microsoft Visual Studio 2013 et versions antérieures lors de la compilation de CMakeCUDACompilerId.cu , qui est utilisé en interne par CMake pour s'assurer que le compilateur fonctionne.

Pour certaines raisons, CMake a décidé de compiler le fichier en 32 bits, ce qui n'est plus supporté. Comme la version de CMake n'a pas été spécifiée, je suggérerais de mettre à jour CMake.

0 commentaires

2
votes

Essayez cmake -G "Microsoft Visual Studio 15 2017 Win64" ..

1 commentaires

Ou simplement cmake -A x64 .. , comme suggéré à cmake.org/cmake/help/latest/generator/... .

1
votes

Il semble que vous deviez spécifier le chemin complet vers nvcc à la main . Je n'aime pas l'idée, cependant.

Vous faites cela de cette façon:

cmake -DCMAKE_CUDA_COMPILER:PATH=/usr/local/cuda/bin/nvcc ...other cmake options...

Cela compilera les choses comme prévu, sauf que cela n'ajoute pas automatiquement l'inclusion dossier. Je cherche toujours à gérer cette deuxième partie sans inclure le fichier FindCUDA.cmake.

0 commentaires