用 nvcc 编译 cuda_rasterizer/backward.cu.出现问题

Question

(gaussian) E:\3D Gaussian\gaussian-splatting>pip install submodules/diff-gaussian-rasterization
Looking in indexes: https://pypi.tuna.tsinghua.edu.cn/simple
Processing e:\3d gaussian\gaussian-splatting\submodules\diff-gaussian-rasterization
  Preparing metadata (setup.py) ... done
Building wheels for collected packages: diff_gaussian_rasterization
  Building wheel for diff_gaussian_rasterization (setup.py) ... error
  error: subprocess-exited-with-error

  × python setup.py bdist_wheel did not run successfully.
  │ exit code: 1
  ╰─> [10 lines of output]
      running bdist_wheel
      running build
      running build_py
      creating build\lib.win-amd64-cpython-38\diff_gaussian_rasterization
      copying diff_gaussian_rasterization\__init__.py -> build\lib.win-amd64-cpython-38\diff_gaussian_rasterization
      running build_ext
      building 'diff_gaussian_rasterization._C' extension
      creating build\temp.win-amd64-cpython-38\Release\cuda_rasterizer
      "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.8\bin\nvcc" -c cuda_rasterizer/backward.cu -o build\temp.win-amd64-cpython-38\Release\cuda_rasterizer/backward.obj -ID:\anaconda\envs\gaussian\lib\site-packages\torch\include -ID:\anaconda\envs\gaussian\lib\site-packages\torch\include\torch\csrc\api\include -ID:\anaconda\envs\gaussian\lib\site-packages\torch\include\TH -ID:\anaconda\envs\gaussian\lib\site-packages\torch\include\THC "-IC:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.8\include" -ID:\anaconda\envs\gaussian\include -ID:\anaconda\envs\gaussian\Include "-IC:\Program Files (x86)\Microsoft Visual Studio\2019\Community\VC\Tools\MSVC\14.29.30133\ATLMFC\include" "-IC:\Program Files (x86)\Microsoft Visual Studio\2019\Community\VC\Tools\MSVC\14.29.30133\include" "-IE:\Windows Kits\10\include\10.0.19041.0\ucrt" "-IE:\Windows Kits\10\include\10.0.19041.0\shared" "-IE:\Windows Kits\10\include\10.0.19041.0\um" "-IE:\Windows Kits\10\include\10.0.19041.0\winrt" "-IE:\Windows Kits\10\include\10.0.19041.0\cppwinrt" -Xcudafe --diag_suppress=dll_interface_conflict_dllexport_assumed -Xcudafe --diag_suppress=dll_interface_conflict_none_assumed -Xcudafe --diag_suppress=field_without_dll_interface -Xcudafe --diag_suppress=base_class_has_different_dll_interface -Xcompiler /EHsc -Xcompiler /wd4190 -Xcompiler /wd4018 -Xcompiler /wd4275 -Xcompiler /wd4267 -Xcompiler /wd4244 -Xcompiler /wd4251 -Xcompiler /wd4819 -Xcompiler /MD -D__CUDA_NO_HALF_OPERATORS__ -D__CUDA_NO_HALF_CONVERSIONS__ -D__CUDA_NO_BFLOAT16_CONVERSIONS__ -D__CUDA_NO_HALF2_OPERATORS__ --expt-relaxed-constexpr "-IE:\3D Gaussian\gaussian-splatting\submodules\diff-gaussian-rasterization\third_party/glm/" -DTORCH_API_INCLUDE_EXTENSION_H -DTORCH_EXTENSION_NAME=_C -D_GLIBCXX_USE_CXX11_ABI=0 -gencode=arch=compute_89,code=compute_89 -gencode=arch=compute_89,code=sm_89 --use-local-env
      error: command 'C:\\Program Files\\NVIDIA GPU Computing Toolkit\\CUDA\\v11.8\\bin\\nvcc' failed with exit code 1
      [end of output]

  note: This error originates from a subprocess, and is likely not a problem with pip.
  ERROR: Failed building wheel for diff_gaussian_rasterization
  Running setup.py clean for diff_gaussian_rasterization
Failed to build diff_gaussian_rasterization
ERROR: ERROR: Failed to build installable wheels for some pyproject.toml based projects (diff_gaussian_rasterization)

(gaussian) E:\3D Gaussian\gaussian-splatting>nvcc --version
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2022 NVIDIA Corporation
Built on Wed_Sep_21_10:41:10_Pacific_Daylight_Time_2022
Cuda compilation tools, release 11.8, V11.8.89
Build cuda_11.8.r11.8/compiler.31833905_0**

求指导

Answer 1

以下回复参考：皆我百晓生、券券喵儿等免费微信小程序作答:

从给出的错误信息来看，问题似乎与CUDA编译有关。这可能是由几个原因导致的，比如CUDA版本不兼容、CUDA环境配置问题或代码本身的问题。以下是一些可能的解决方案：

1. 检查CUDA版本兼容性：确保你的CUDA版本与项目要求的版本相匹配。如果项目需要特定版本的CUDA，请确保已正确安装该版本。
2. 检查CUDA安装和配置：确保CUDA已正确安装并配置在你的系统中。你可以通过运行nvcc --version来检查NVCC的版本，确认CUDA工具集已正确安装。
3. 检查Python和PyTorch版本：确保你使用的Python和PyTorch版本与项目的要求相匹配。项目可能需要特定版本的Python和PyTorch来正常工作。
4. 检查代码问题：如果以上步骤都没问题，可能是代码本身的问题。检查cuda_rasterizer/backward.cu文件是否有编译错误。
5. 尝试清理并重新构建：运行pip install命令之前，尝试清理之前的构建并重新构建项目。可以使用python setup.py clean命令清理之前的构建。
6. 查看详细错误日志：错误信息可能只是问题的冰山一角。查看完整的错误日志可能会提供更多关于问题的线索。你可以通过查看构建过程中的详细输出来获取更多信息。

由于错误信息较多且涉及CUDA编译的复杂过程，解决方案可能需要一些调试和尝试。你可以尝试上述步骤，并根据具体情况进行调整。如果问题仍然存在，建议查阅相关项目的文档或寻求更专业的帮助。