Anyone successfully set up a GitHub action workflow with CMake's FindCUDAToolkit?

Accelerated Computing CUDA CUDA Programming and Performance
1

I’ve been having trouble setting up a GitHub Actions workflow to test-build my CUDA-related library (this one on Windows as well as Linux. Even though the workflow script manages to install CUDA - supposedly anyway; and even though CMake 3.22.1) recognizes it’s installed - the find_package(CUDAToolkit) command still fails to locate the relevant paths; the console output is attached below.

Has anyone here successfully set up a workflow on GH Actions in which CMake does manage to find all components of the CUDA installation, on a Windows 10 VM? (Naturally everything works fine on Linux.)

Run cmake -B build -G "Visual Studio 16 2019" -A x64 -DBUILD_EXAMPLES=ON
-- The CUDA compiler identification is NVIDIA 11.4.48
-- The CXX compiler identification is MSVC 19.29.30138.0
-- Detecting CUDA compiler ABI info
-- Detecting CUDA compiler ABI info - done
-- Check for working CUDA compiler: C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v11.4/bin/nvcc.exe - skipped
-- Detecting CUDA compile features
-- Detecting CUDA compile features - done
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Check for working CXX compiler: C:/Program Files (x86)/Microsoft Visual Studio/2019/Enterprise/VC/Tools/MSVC/14.29.30133/bin/Hostx64/x64/cl.exe - skipped
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Unable to find cublas_v2.h in either "C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v11.4/include" or "C:/Program Files/NVIDIA GPU Computing Toolkit/math_libs/include"
-- Found CUDAToolkit: C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v11.4/include (found version "11.4.48") 
-- Looking for C++ include pthread.h
-- Looking for C++ include pthread.h - not found
-- Found Threads: TRUE  
-- Automatic GPU detection failed. Building for common architectures.
-- Autodetected CUDA architecture(s): 5.3;6.0;6.1;7.0;7.5;8.0;8.6;8.6+PTX
-- Configuring done
CMake Error at CMakeLists.txt:39 (add_library):
  Target "nvtx" links to target "CUDA::nvToolsExt" but the target was not
  found.  Perhaps a find_package() call is missing for an IMPORTED target, or
  an ALIAS target is missing?


CMake Error at CMakeLists.txt:39 (add_library):
  Target "nvtx" links to target "CUDA::nvToolsExt" but the target was not
  found.  Perhaps a find_package() call is missing for an IMPORTED target, or
  an ALIAS target is missing?


CMake Error at CMakeLists.txt:39 (add_library):
  Target "nvtx" links to target "CUDA::nvToolsExt" but the target was not
  found.  Perhaps a find_package() call is missing for an IMPORTED target, or
  an ALIAS target is missing?


CMake Error at CMakeLists.txt:39 (add_library):
  Target "nvtx" links to target "CUDA::nvToolsExt" but the target was not
  found.  Perhaps a find_package() call is missing for an IMPORTED target, or
  an ALIAS target is missing?


-- Generating done
CMake Generate step failed.  Build files cannot be regenerated correctly.
Error: Process completed with exit code 1.
2

Bumping this for attention…

What I currently get is:

-- Found CUDAToolkit: C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v11.7/include (found suitable version "11.7.64", minimum required is "9.0") 
... snip ...
-- Automatic GPU detection failed. Building for common architectures.
-- Autodetected CUDA architecture(s): 5.3;6.0;6.1;7.0;7.5;8.0;8.6;8.6+PTX
-- Configuring done
CMake Error at CMakeLists.txt:70 (target_link_libraries):
  The link interface of target "caw_nvtx" contains:
-- Generating done

    CUDA::nvToolsExt

  but the target was not found. etc. etc.