Upgrade embree to 4.3.1

[lumurillo]

Upgrade the embree version to 4.3.1.

Type

• Bug fix (non-breaking change which fixes an issue): Fixes #
• 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

Use of the version 4 of the embree API.

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

• Changed the embree tarball version to 4.3.1
• Adapt the use of embree API acording to this notes: https://github.com/embree/embree/tree/master?tab=readme-ov-file#upgrading-from-embree-3-to-embree-4

[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.

[ssheorey]

@lumurillo for style check:

pip install clang-format~=10.0
python util/check_style.py --apply

[ssheorey]

• Please update to the latest available embree (4.3.1) - supports macOS ARM.
• We will need to test GPU functionality.

[ssheorey]

Plan for embree upgrade:

• Upgrade embree to latest and update API.
• Pass CI on CPU.

Separate PR [copy this comment over to new PR]:

• Add iGPU (SYCL) support and pass CI on iGPU.
• Add dGPU (SYCL) support for Intel GPU and update / pass CI on dGPU.
• [Future PR / Optional] Add dGPU (SYCL) support for Nvidia GPU and pass CI on dGPU. (device: ext_oneapi_cuda:gpu:0) with Codeplay oneAPI CUDA plugin.

[ssheorey]

Hi @lumurillo did you notice any performance difference with the new embree, especially after the change in API (rtcInterset1M -> rtcIntersect1). It would be good to do a quick benchmark on a large mesh (~1M+ vertices).