Cuda12 rebuild

[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.

This is my take on trying to build for CUDA 12, its just a continuation of #171.
This builds locally on my machine, but is failing on testing because it does not find the mkl .so:

import: 'torch'                                                                                                                                                              
Traceback (most recent call last):                                                                                                                                           
  File "/home/conda/feedstock_root/build_artifacts/pytorch-recipe_1692196994653/test_tmp/run_test.py", line 2, in <module>                                                   
    import torch                                                                                                                                                             
  File "/home/conda/feedstock_root/build_artifacts/pytorch-recipe_1692196994653/_test_env_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_pla
cehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_p/lib/python3.11/site-packages/torch/__init__.py", line 228, in <module>              
    _load_global_deps()                                                                                                                                                      
  File "/home/conda/feedstock_root/build_artifacts/pytorch-recipe_1692196994653/_test_env_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_pla
cehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_p/lib/python3.11/site-packages/torch/__init__.py", line 187, in _load_global_deps     
    raise err
  File "/home/conda/feedstock_root/build_artifacts/pytorch-recipe_1692196994653/_test_env_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_pla
cehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_p/lib/python3.11/site-packages/torch/__init__.py", line 168, in _load_global_deps
    ctypes.CDLL(lib_path, mode=ctypes.RTLD_GLOBAL)
  File "/home/conda/feedstock_root/build_artifacts/pytorch-recipe_1692196994653/_test_env_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_pla
cehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_p/lib/python3.11/ctypes/__init__.py", line 376, in __init__
    self._handle = _dlopen(self._name, mode)
                   ^^^^^^^^^^^^^^^^^^^^^^^^^
OSError: libmkl_intel_lp64.so.1: cannot open shared object file: No such file or directory

[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.

[jakirkham]

@conda-forge-admin, please re-render

[github-actions]

Hi! This is the friendly automated conda-forge-webservice.

I tried to rerender for you, but it looks like there was nothing to do.

This message was generated by GitHub actions workflow run https://github.com/conda-forge/pytorch-cpu-feedstock/actions/runs/5882478037.

[jakirkham]

Maybe these pins could be moved to conda_build_config.yaml. Then when resolving upstream changes would prefer upstream's changes (then could readd the -devel part). IIUC -include comes with -devel

pytorch-cpu-feedstock/recipe/conda_build_config.yaml

Lines 10 to 13 in 229b447

	blas_impl:
	- mkl # [x86 or x86_64]
	- generic

[h-vetinari]

They're also really old (>2 years). Do we really need to pin mkl that tightly?

[jakirkham]

Yeah am wondering the same. It also deviates from the current global pinnings

Maybe this was needed to get the build working, in which case there probably needs to be some investigation into how to bump the versions (and fix whatever issue occurs when using more recent versions)

[RaulPPelaez]

See this comment for the rationale: #172 (comment)

[h-vetinari]

This is my take on trying to build for CUDA 12, its just a continuation of #171.
This builds locally on my machine, but is failing on testing because it does not find the mkl .so:

Thanks for picking this up @RaulPPelaez!

[h-vetinari]

The PR will also have to be rebased, or at least merged with main.

[RaulPPelaez]

I am having some success building locally. However, I am having trouble merging this with main, a lot of conflicts arise in files automatically created by the bot and I have no idea how to solve them.
I would like to somehow reopen this PR starting from the current main but I do not know how to instruct conda-smithy to trigger the CUDA 12 migration there.
Some help pls?

[RaulPPelaez]

Ok I think I got it, I am closing this and continuing in #182 .