[WIP] Cuda12 rebuild again

[RaulPPelaez]

Checklist

• Used a personal fork of the feedstock to propose changes
• Bumped the build number (if the version is unchanged)
• Reset the build number to 0 (if the version changed)
• Re-rendered with the latest conda-smithy (Use the phrase @conda-forge-admin, please rerender in a comment in this PR for automated rerendering)
• Ensured the license file is being packaged.

There were some conflicts in my previous PR #181 , I think I was able to solve them.

[conda-forge-webservices]

Hi! This is the friendly automated conda-forge-linting service.

I wanted to let you know that I linted all conda-recipes in your PR (recipe) and found some lint.

Here's what I've got...

For recipe:

• Selectors are suggested to take a <two spaces>#<one space>[<expression>] form. See lines [94, 116, 147]

[conda-forge-webservices]

Hi! This is the friendly automated conda-forge-linting service.

I wanted to let you know that I linted all conda-recipes in your PR (recipe) and found some lint.

Here's what I've got...

For recipe:

• Selectors are suggested to take a <two spaces>#<one space>[<expression>] form. See lines [95, 117, 148]

[RaulPPelaez]

OSX builds run into this error pytorch/pytorch#99035
Now trying to silence it with -Wno-unused-command-line-argument

[RaulPPelaez]

Silencing causes this error in CMake:

-- Performing Test COMPILER_SUPPORTS_AVX512 - Failed
  CMake Error at third_party/fbgemm/CMakeLists.txt:76 (message):
    A compiler with AVX512 support is required.


  -- Configuring incomplete, errors occurred!
  error: subprocess-exited-with-error
  

[conda-forge-webservices]

Hi! This is the friendly automated conda-forge-linting service.

I just wanted to let you know that I linted all conda-recipes in your PR (recipe) and found it was in an excellent condition.

[jakirkham]

Interesting wonder why AVX512 support is a hard requirement? Not all CPUs have those instructions

[Tobias-Fischer]

I think we should just set USE_FBGEMM=OFF for osx.

The check for avx512 is different in pytorch where they decide whether to build fbgemm (https://github.com/pytorch/pytorch/blob/71632d4d24616ddad6685814aae4ae54c981c0d2/cmake/MiscCheck.cmake#L43-L83 - this succeeds) and within fbgemm itself (https://github.com/pytorch/FBGEMM/blob/d43717ad9f62c30a89a31d537303e7d5970c1018/CMakeLists.txt#L111-L122 - fails).

[conda-forge-webservices]

Hi! This is the friendly automated conda-forge-linting service.

I was trying to look for recipes to lint for you, but it appears we have a merge conflict.
Please try to merge or rebase with the base branch to resolve this conflict.

Please ping the 'conda-forge/core' team (using the @ notation in a comment) if you believe this is a bug.

[conda-forge-webservices]

Hi! This is the friendly automated conda-forge-linting service.

I just wanted to let you know that I linted all conda-recipes in your PR (recipe) and found it was in an excellent condition.

[RaulPPelaez]

@conda-forge-admin, please rerender

[RaulPPelaez]

The lingering issues are:

• In some OSX build succeeds, test fail with this error:

Traceback (most recent call last):
  File "backends/xeon/test_launch.py", line 65, in <module>
    run_tests()
  File "/Users/runner/miniforge3/conda-bld/pytorch-recipe_1693911031252/_test_env_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold/lib/python3.8/site-packages/torch/testing/_internal/common_utils.py", line 765, in run_tests
    from xmlrunner.result import _XMLTestResult  # type: ignore[import]
ModuleNotFoundError: No module named 'xmlrunner.result'
FINISHED PRINTING LOG FILE of backends/xeon/test_launch (/Users/runner/miniforge3/conda-bld/pytorch-recipe_1693911031252/test_tmp/test/test-reports/backends-xeon-test_launch_kar0386k.log)

backends/xeon/test_launch failed!
Traceback (most recent call last):
  File "./test/run_test.py", line 1394, in <module>
    main()
  File "./test/run_test.py", line 1352, in main
    raise RuntimeError(
RuntimeError: backends/xeon/test_launch failed!

But the xmlrunner package is being installed, and this error is not present in many other osx builds.

• In some cpu only linux builds build succeeds, but test fails to install the env with this error:

with channels:

The reported errors are:
- Encountered problems while solving:
-   - package pytorch-2.0.0-cpu_generic_py38h1a61d8b_1 requires mkl >=2022.2.1,<2023.0a0, but none of the providers can be installed
- 

weird because the mkl package is installed during build just fine.

[jakirkham]

Now that PR ( #195 ) is in, should we close this?