Fix logic for adding -allow-unsupported-compiler to nvcc

[patrikhuber]

The previous logic would always add -allow-unsupported-compiler and -D_ALLOW_COMPILER_AND_STL_VERSION_MISMATCH, even when it was not needed (e.g. when a recent CUDA version is used), and also printing a (confusing) warning which is not applicable in that case.

The revised logic only adds the flags if MSVC>=1940 and CUDA<=12.4. For that, we need to use find_package(CUDAToolkit REQUIRED), like it's also done a few more lines below in the if(BUILD_COMMON_CUDA_ARCHS) code - and then we append to CMAKE_CUDA_FLAGS only when needed.

The new code also uses CMAKE_CUDA_FLAGS for the _ALLOW_COMPILER_AND_STL_VERSION_MISMATCH define, which makes CMake's try_compile (which is used when enable_language(CUDA) is called) work reliably if a CMake version >=3.29.4 is used. It doesn't work when add_compile_definitions(_ALLOW_COMPILER_AND_STL_VERSION_MISMATCH) is used.

Type

• Bug fix (non-breaking change which fixes an issue): Fixes the issue described above.
• New feature (non-breaking change which adds functionality). Resolves #
• Breaking change (fix or feature that would cause existing functionality to not work as expected) Resolves #

Motivation and Context

See above.

Checklist:

• I have run python util/check_style.py --apply to apply Open3D code style
  to my code.
• This PR changes Open3D behavior or adds new functionality.
  • Both C++ (Doxygen) and Python (Sphinx / Google style) documentation is
    updated accordingly.
  • I have added or updated C++ and / or Python unit tests OR included test
    results (e.g. screenshots or numbers) here.
• I will follow up and update the code if CI fails.
• For fork PRs, I have selected Allow edits from maintainers.

Description

See above.

[update-docs]

Thanks for submitting this pull request! The maintainers of this repository would appreciate if you could update the CHANGELOG.md based on your changes.