Building issues [no CUDA toolset found]

[ph57fe]

I have building issues with the following system settings:

Windows 10 Visual Studio 2019 (Community Edition) Version 16.5.4 cmake 3.17.1 python 3.7 CUDA 10.1

Even though CUDA is installed, but cmake is not able to locate it. The VCTargets is not created in the VC folder: C:/Program Files (x86)/Microsoft Visual Studio/2017/Community/Common7/IDE/VC/

Ankit@DESKTOP-KVF0V7L MINGW64 ~/localgitrepository/mitsuba2 (master)
$ cmake -G "Visual Studio 16 2019" -A x64
CMake Warning:
  No source or binary directory provided.  Both will be assumed to be the
  same as the current working directory, but note that this warning will
  become a fatal error in future CMake releases.

-- Selecting Windows SDK version 10.0.18362.0 to target Windows 10.0.18363.
-- The C compiler identification is MSVC 19.25.28614.0
-- The CXX compiler identification is MSVC 19.25.28614.0
-- Check for working C compiler: C:/Program Files (x86)/Microsoft Visual Studio/2019/Community/VC/Tools/MSVC/14.25.28610/bin/Hostx64/x64/cl.exe
-- Check for working C compiler: C:/Program Files (x86)/Microsoft Visual Studio/2019/Community/VC/Tools/MSVC/14.25.28610/bin/Hostx64/x64/cl.exe - 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: C:/Program Files (x86)/Microsoft Visual Studio/2019/Community/VC/Tools/MSVC/14.25.28610/bin/Hostx64/x64/cl.exe
-- Check for working CXX compiler: C:/Program Files (x86)/Microsoft Visual Studio/2019/Community/VC/Tools/MSVC/14.25.28610/bin/Hostx64/x64/cl.exe - works
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Found PythonInterp: C:/Program Files (x86)/Microsoft Visual Studio/Shared/Python37_64/python.exe (found version "3.7.5")
-- Found PythonLibs: C:/Program Files (x86)/Microsoft Visual Studio/Shared/Python37_64/libs/Python37.lib
-- Building the following variants of Mitsuba:
--  * scalar_rgb
--  * scalar_spectral
--  * gpu_rgb
--  * gpu_autodiff_rgb
-- Mitsuba: building the Python plugin.
CMake Error at C:/Program Files/CMake/share/cmake-3.17/Modules/CMakeDetermineCompilerId.cmake:367 (message):
  No CUDA toolset found.
Call Stack (most recent call first):
  C:/Program Files/CMake/share/cmake-3.17/Modules/CMakeDetermineCompilerId.cmake:32 (CMAKE_DETERMINE_COMPILER_ID_BUILD)
  C:/Program Files/CMake/share/cmake-3.17/Modules/CMakeDetermineCUDACompiler.cmake:72 (CMAKE_DETERMINE_COMPILER_ID)
  ext/enoki/CMakeLists.txt:20 (enable_language)

-- Configuring incomplete, errors occurred!
See also "C:/Users/Ankit/localgitrepository/mitsuba2/CMakeFiles/CMakeOutput.log".

Do you have an idea? Thanks!

[Speierers]

Hi,

is nvcc in your PATH?

E.g. if you try to run:

nvcc --version

[ph57fe]

This is what I get with nvcc --version :

Ankit@DESKTOP-KVF0V7L MINGW64 ~/localgitrepository3/mitsuba2 (master)
$ nvcc --version
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2019 NVIDIA Corporation
Built on Fri_Feb__8_19:08:26_Pacific_Standard_Time_2019
Cuda compilation tools, release 10.1, V10.1.105

Got the issue solved by manually copying the files from C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\extras\visual_studio_integration\MSBuildExtensions to C:\Program Files (x86)\Microsoft Visual Studio\2019\Community\MSBuild\Microsoft\VC\v160\BuildCustomizations.

[Speierers]

Looking at the environment variables in my system. Do you have something similar?

[ph57fe]

Yes, I have these.

[Speierers]

Could you try to pull latest master and update the submodule. We fixed something in enoki recently which might be related to this.

git pull --rebase origin master
git submodule update --recursive

[hjoonpark]

Is my problem similar to this? I get this CMAKE configure error:

CMake Warning: No source or binary directory provided. Both will be assumed to be the same as the current working directory, but note that this warning will become a fatal error in future CMake releases.

-- Building the following variants of Mitsuba: -- scalar_rgb -- gpu_autodiff_rgb -- Mitsuba: building the Python plugin. CMake Error at C:/Users/hjoon/Documents/cmake-3.16.0-win64-x64/share/cmake-3.16/Modules/CMakeDetermineCompilerId.cmake:351 (message): No CUDA toolset found. Call Stack (most recent call first): C:/Users/hjoon/Documents/cmake-3.16.0-win64-x64/share/cmake-3.16/Modules/CMakeDetermineCompilerId.cmake:32 (CMAKE_DETERMINE_COMPILER_ID_BUILD) C:/Users/hjoon/Documents/cmake-3.16.0-win64-x64/share/cmake-3.16/Modules/CMakeDetermineCUDACompiler.cmake:72 (CMAKE_DETERMINE_COMPILER_ID) ext/enoki/CMakeLists.txt:28 (enable_language)

-- Configuring incomplete, errors occurred!

Here are what I've done so far. I pulled the latest master using:

git pull --rebase origin master git submodule update --recursive

I checked nvcc is installed:

nvcc: NVIDIA (R) Cuda compiler driver Copyright (c) 2005-2019 NVIDIA Corporation Built on Sun_Jul_28_19:12:52_Pacific_Daylight_Time_2019 Cuda compilation tools, release 10.1, V10.1.243

and I set the environment variables:

CMAKE_CUDA_COMPILER = C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\bin\nvcc CUDA_PATH = C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1 CUDA_PATH_V10_1 = C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1

I also followed this:

Got the issue solved by manually copying the files from C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\extras\visual_studio_integration\MSBuildExtensions to C:\Program Files (x86)\Microsoft Visual Studio\2019\Community\MSBuild\Microsoft\VC\v160\BuildCustomizations.

but still not resolved..

[Speierers]

Would you be able to update CUDA on your system? (e.g. v10.2)

Not sure this would solve your problem, but it is worth a try!

[ph57fe]

Hi @Speierers

Thanks, I got all the previous issues resolved by updating to CUDA 10.2. I'm facing now facing an error when integrating with PyTorch.

I have the following: Python 3.7.6 Pytorch 1.5.0 conda 4.8.0 cudatoolkit 10.2.89 torchvision 0.6.0

I am facing the following error.

Traceback (most recent call last):
  File "C:/Users/Ankit/localgitrepository/mitsuba2/dist/scene_render.py", line 13, in <module>
    import torch
  File "C:\Users\Ankit\Anaconda3\lib\site-packages\torch\__init__.py", line 81, in <module>
    ctypes.CDLL(dll)
  File "C:\Users\Ankit\Anaconda3\lib\ctypes\__init__.py", line 364, in __init__
    self._handle = _dlopen(self._name, mode)
OSError: [WinError 127] The specified procedure could not be found

Do you have an idea? Thanks!

[Speierers]

This seems to be related to Pytorch and not Mitsuba: See this: https://github.com/pytorch/pytorch/issues/37209

[KnightYao]

mark

[Theo-Wu]

Is my problem similar to this? I get this CMAKE configure error:

CMake Warning: No source or binary directory provided. Both will be assumed to be the same as the current working directory, but note that this warning will become a fatal error in future CMake releases. -- Building the following variants of Mitsuba: -- scalar_rgb -- gpu_autodiff_rgb -- Mitsuba: building the Python plugin. CMake Error at C:/Users/hjoon/Documents/cmake-3.16.0-win64-x64/share/cmake-3.16/Modules/CMakeDetermineCompilerId.cmake:351 (message): No CUDA toolset found. Call Stack (most recent call first): C:/Users/hjoon/Documents/cmake-3.16.0-win64-x64/share/cmake-3.16/Modules/CMakeDetermineCompilerId.cmake:32 (CMAKE_DETERMINE_COMPILER_ID_BUILD) C:/Users/hjoon/Documents/cmake-3.16.0-win64-x64/share/cmake-3.16/Modules/CMakeDetermineCUDACompiler.cmake:72 (CMAKE_DETERMINE_COMPILER_ID) ext/enoki/CMakeLists.txt:28 (enable_language) -- Configuring incomplete, errors occurred!

Here are what I've done so far. I pulled the latest master using:

git pull --rebase origin master git submodule update --recursive

I checked nvcc is installed:

nvcc: NVIDIA (R) Cuda compiler driver Copyright (c) 2005-2019 NVIDIA Corporation Built on Sun_Jul_28_19:12:52_Pacific_Daylight_Time_2019 Cuda compilation tools, release 10.1, V10.1.243

and I set the environment variables:

CMAKE_CUDA_COMPILER = C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\bin\nvcc CUDA_PATH = C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1 CUDA_PATH_V10_1 = C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1

I also followed this:

Got the issue solved by manually copying the files from C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\extras\visual_studio_integration\MSBuildExtensions to C:\Program Files (x86)\Microsoft Visual Studio\2019\Community\MSBuild\Microsoft\VC\v160\BuildCustomizations.

but still not resolved..

I am using CUDA 10.2 but get the same problem too. Can you help me with this?

Solved by reinstalling CUDA 10.2. Then deleted CmakeFiles folder and CmakeCache.txt, and built successfully.

[Chenguoz]

60\BuildCustomizations。

Thanks, I tossed and turned all day until I saw your reply and I solved the problem!

[giodude12]

Hey, i'm still getting the issue. My path seems right and i have the files manually copied into visua studio. Does anybody have a fix?

[SHPEYOLO]

Hey, i'm getting the issue.

nvcc fatal : Could not set up the environment for Microsoft Visual Studio using 'D:/VS2019/Community/VC/Tools/MSVC/14.29.30133/bin/HostX64/x64/../../../ ../../../../VC/Auxiliary/Build/vcvars64.bat' D:\VS2019\Community\MSBuild\Microsoft\VC\v160\BuildCustomizations\CUDA 11.6.targets(879,9): error MSB3721: 命令“"G:\CUDA_manger\CUDA11.6\CUDA2\bin\nvcc.exe" - dlink -o ngp.dir\RelWithDebInfo\ngp.device-link.obj -Xcompiler "/EHsc /W1 /nologo /O2 /FdF:\Tutorial\ngp\instant-ngp\build\RelWithDebInfo\ngp.pdb /Zi /MD / GR" -LG:\CUDA_manger\CUDA11.6\CUDA2\bin/crt -LG:\CUDA_manger\CUDA11.6\CUDA2\lib\x64 opengl32.lib dependencies\glfw\src\glfw_objects.dir\RelWithDebInfo\conte xt.obj dependencies\glfw\src\glfw_objects.dir\RelWithDebInfo\init.obj dependencies\glfw\src\glfw_objects.dir\RelWithDebInfo\input.obj dependencies\glfw\src\ glfw_objects.dir\RelWithDebInfo\monitor.obj dependencies\glfw\src\glfw_objects.dir\RelWithDebInfo\vulkan.obj dependencies\glfw\src\glfw_objects.dir\RelWithD ebInfo\window.obj dependencies\glfw\src\glfw_objects.dir\RelWithDebInfo\win32_init.obj dependencies\glfw\src\glfw_objects.dir\RelWithDebInfo\win32_joystick. obj dependencies\glfw\src\glfw_objects.dir\RelWithDebInfo\win32_monitor.obj dependencies\glfw\src\glfw_objects.dir\RelWithDebInfo\win32_time.obj dependencie s\glfw\src\glfw_objects.dir\RelWithDebInfo\win32_thread.obj dependencies\glfw\src\glfw_objects.dir\RelWithDebInfo\win32_window.obj dependencies\glfw\src\glf w_objects.dir\RelWithDebInfo\wgl_context.obj dependencies\glfw\src\glfw_objects.dir\RelWithDebInfo\egl_context.obj dependencies\glfw\src\glfw_objects.dir\Re lWithDebInfo\osmesa_context.obj dependencies\glfw\src\glfw_objects.dir\RelWithDebInfo\glfw.res "dependencies\tiny-cuda-nn\src\RelWithDebInfo\tiny-cuda-nn.li b" cuda.lib cublas.lib cudadevrt.lib cudart_static.lib kernel32.lib user32.lib gdi32.lib winspool.lib comdlg32.lib advapi32.lib shell32.lib ole32.lib oleaut 32.lib uuid.lib odbc32.lib odbccp32.lib cudart.lib -forward-unknown-to-host-compiler -Wno-deprecated-gpu-targets -gencode=arch=compute_86,code=compute_86 - gencode=arch=compute_86,code=sm_86 --machine 64 ngp.dir\RelWithDebInfo\camera_path.obj ngp.dir\RelWithDebInfo\common_device.obj ngp.dir\RelWithDebInfo\marc hing_cubes.obj ngp.dir\RelWithDebInfo\nerf_loader.obj ngp.dir\RelWithDebInfo\render_buffer.obj ngp.dir\RelWithDebInfo\testbed.obj ngp.dir\RelWithDebInfo\tes tbed_image.obj ngp.dir\RelWithDebInfo\testbed_nerf.obj ngp.dir\RelWithDebInfo\testbed_sdf.obj ngp.dir\RelWithDebInfo\testbed_volume.obj ngp.dir\RelWithDebIn fo\tinyexr_wrapper.obj ngp.dir\RelWithDebInfo\triangle_bvh.obj”已退出，返回代码为 1。 [F:\Tutorial\ngp\instant-ngp\build\ngp.vcxproj]

[merlinND]

Hi @SHPEYOLO,

Your error log seems to refer to instant-ngp rather than Mitsuba, did you post in the right repo?

[Shy-fawn]

您好!!!!這是我以下碰到的問題: 我需要在命令提示字元輸入: cmake . -B build cmake --build build --config RelWithDebInfo -j 16

但是一直顯示這樣的錯誤: D:\Tutorial\ngp\instant-ngp>cmake . -B build -- Selecting Windows SDK version 10.0.19041.0 to target Windows 10.0.22000. -- Automatic GPU detection failed. Building for common architectures. -- Targeting GPU architectures: 53;60;61;70;75;80;86 CMake Warning at dependencies/tiny-cuda-nn/CMakeLists.txt:140 (message): Fully fused MLPs do not support GPU architectures of 70 or less. Falling back to CUTLASS MLPs. Remove GPU architectures 70 and lower to allow maximum performance

CMake Error at dependencies/tiny-cuda-nn/CMakeLists.txt:188 (add_subdirectory): The source directory

D:/Tutorial/ngp/instant-ngp/dependencies/tiny-cuda-nn/dependencies/fmt

does not contain a CMakeLists.txt file.

-- Could NOT find Vulkan (missing: Vulkan_LIBRARY Vulkan_INCLUDE_DIR) (found version "") CMake Warning at CMakeLists.txt:118 (message): Vulkan was not found. Neural graphics primitives will still compile and run correctly, but DLSS will not be supported.

-- Using Win32 for window creation -- OptiX_INSTALL_DIR value: C:\ProgramData\NVIDIA Corporation\OptiX SDK 7.6.0 -- pybind11 v2.7.1 -- Configuring incomplete, errors occurred! See also "D:/Tutorial/ngp/instant-ngp/build/CMakeFiles/CMakeOutput.log". See also "D:/Tutorial/ngp/instant-ngp/build/CMakeFiles/CMakeError.log".

D:\Tutorial\ngp\instant-ngp>cmake --build build --config RelWithDebInfo -j 16 Microsoft (R) Build Engine for .NET Framework 16.11.2+f32259642 版 Copyright (C) Microsoft Corporation. 著作權所有，並保留一切權利。

MSBUILD : error MSB1009: 專案檔不存在。 參數: ALL_BUILD.vcxproj

請問我是有什麼沒安裝好，還是此步驟其實是可以忽略的?

[merlinND]

Hello @Shy-fawn,

Same as above: your error log seems to refer to instant-ngp rather than Mitsuba, did you post in the right repo?

[Shy-fawn]

@merlinND 不好意思再打擾您 請問您是指這部分嗎? 請問可以查看到我付上的圖片嗎?

[wjakob]

@Shy-fawn 您在錯誤的存儲庫上發布。 請用英文發帖，否則我們很遺憾需要封鎖您的帳戶。

[zeynepiskenderoglu]

'cmake' is not recognized as an internal or external command, operable program or batch file. i am having the problem. _Got the issue solved by manually copying the files from C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\extras\visual_studiointegration\MSBuildExtensions to C:\Program Files (x86)\Microsoft Visual Studio\2019\Community\ MSBuild\Microsoft\VC\v160\BuildCustomizations I am trying to implement the process, but there is no such path on my computer; C:\Program Files (x86)\Microsoft Visual Studio\2019\Community\MSBuild\Microsoft\VC\v160\BuildCustomizations instead there are these what should I do?

[wjakob]

@zeynepiskenderoglu -- please use Mitsuba 3, this project is no longer supported.

[aakash-chaddha]

If still facing the issue on windows:

Make sure cmake is path is in windows global env-variables. Steps:

1. Search "Developer Command Prompt" in all-apps, I get this "Developer Command Prompt for VS 2019"
2. Now type "where cmake" copy the path i.e "C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\Common7\IDE\CommonExtensions\Microsoft\CMake\CMake\bin"
3. Append it in Path in System Variables
4. Make sure to remove "build" folder from your working dir where cmake build command failed.
5. run cmake build command