Update oneTBB, oneMKL, oneDAL, MPI and OpenMP versions

[Alexsandruss]

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.

[conda-forge-admin]

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/meta.yaml) and found some lint.

Here's what I've got...

For recipe/meta.yaml:

• When defining a source/url please add a sha256, sha1 or md5 checksum (sha256 preferably).
• When defining a source/url please add a sha256, sha1 or md5 checksum (sha256 preferably).
• When defining a source/url please add a sha256, sha1 or md5 checksum (sha256 preferably).
• When defining a source/url please add a sha256, sha1 or md5 checksum (sha256 preferably).
• When defining a source/url please add a sha256, sha1 or md5 checksum (sha256 preferably).
• When defining a source/url please add a sha256, sha1 or md5 checksum (sha256 preferably).
• When defining a source/url please add a sha256, sha1 or md5 checksum (sha256 preferably).
• When defining a source/url please add a sha256, sha1 or md5 checksum (sha256 preferably).
• When defining a source/url please add a sha256, sha1 or md5 checksum (sha256 preferably).
• When defining a source/url please add a sha256, sha1 or md5 checksum (sha256 preferably).
• When defining a source/url please add a sha256, sha1 or md5 checksum (sha256 preferably).
• When defining a source/url please add a sha256, sha1 or md5 checksum (sha256 preferably).
• When defining a source/url please add a sha256, sha1 or md5 checksum (sha256 preferably).
• When defining a source/url please add a sha256, sha1 or md5 checksum (sha256 preferably).
• When defining a source/url please add a sha256, sha1 or md5 checksum (sha256 preferably).
• When defining a source/url please add a sha256, sha1 or md5 checksum (sha256 preferably).
• When defining a source/url please add a sha256, sha1 or md5 checksum (sha256 preferably).
• When defining a source/url please add a sha256, sha1 or md5 checksum (sha256 preferably).
• When defining a source/url please add a sha256, sha1 or md5 checksum (sha256 preferably).
• Jinja2 variable definitions are suggested to take a {%<one space>set<one space><variable name><one space>=<one space><expression><one space>%} form. See lines [46, 51]

[Alexsandruss]

@conda-forge-admin, please rerender

[Alexsandruss]

Waiting for 2022 version of oneTBB in conda-forge (https://github.com/conda-forge/tbb-feedstock)

[ZzEeKkAa]

Thank you @Alexsandruss !

There is also intel-pti package added with this release, but I have scripts to build it from source:
conda-forge/pti-gpu-unitrace-feedstock#1

I guess I'll update it just for now so it produces intel-pti without split.

But we need to wait for conda-forge/intel-compiler-repack-feedstock#52

[ZzEeKkAa]

@conda-forge-admin, please rerender

[ZzEeKkAa]

@conda-forge-admin, please rerender

[ZzEeKkAa]

I'm working on this PR right now. I've fixed build locally and currently working on proper dependencies.

[ZzEeKkAa]

Do not merge yet, still working on it

[ZzEeKkAa]

Hmm,

024-11-07T00:51:59.0898228Z Traceback (most recent call last):
2024-11-07T00:51:59.0900327Z   File "D:\Miniforge\Scripts\conda-build-script.py", line 10, in <module>
2024-11-07T00:51:59.0908273Z     sys.exit(execute())
2024-11-07T00:51:59.0909693Z              ^^^^^^^^^
2024-11-07T00:51:59.0910660Z   File "D:\Miniforge\Lib\site-packages\conda_build\cli\main_build.py", line 589, in execute
2024-11-07T00:51:59.0927579Z     api.build(
2024-11-07T00:51:59.0928734Z   File "D:\Miniforge\Lib\site-packages\conda_build\api.py", line 209, in build
2024-11-07T00:51:59.0935634Z     return build_tree(
2024-11-07T00:51:59.0936360Z            ^^^^^^^^^^^
2024-11-07T00:51:59.0937291Z   File "D:\Miniforge\Lib\site-packages\conda_build\build.py", line 3670, in build_tree
2024-11-07T00:51:59.1012625Z     test(pkg, config=metadata.config.copy(), stats=stats)
2024-11-07T00:51:59.1013833Z   File "D:\Miniforge\Lib\site-packages\conda_build\build.py", line 3422, in test
2024-11-07T00:51:59.1024978Z     environ.create_env(
2024-11-07T00:51:59.1025938Z   File "D:\Miniforge\Lib\site-packages\conda_build\environ.py", line 1050, in create_env
2024-11-07T00:51:59.1051828Z     _execute_actions(prefix, precs)
2024-11-07T00:51:59.1052876Z   File "D:\Miniforge\Lib\site-packages\conda_build\environ.py", line 1338, in _execute_actions
2024-11-07T00:51:59.1062982Z     progressive_fetch_extract.execute()
2024-11-07T00:51:59.1064083Z   File "D:\Miniforge\Lib\site-packages\conda\core\package_cache_data.py", line 900, in execute
2024-11-07T00:51:59.1086495Z     raise CondaMultiError(not_cancelled)
2024-11-07T00:51:59.1087261Z conda.CondaMultiError: [Errno 28] No space left on device
2024-11-07T00:51:59.1087963Z [Errno 28] No space left on device
2024-11-07T00:51:59.1088463Z 

On windows we are out of space...

[ZzEeKkAa]

One thing we've just discussed with @oleksandr-pavlyk is that neither openmp, neither tbb are required in mkl. It's up to end user to chose what dependencies they want to use. So we would want to move it to separate discussion to remove it from mkl dependency. In a meantime. I've accidentally added openmp mutex to all onemkl-sycl-* packages. I need to fix it before merge.

[ZzEeKkAa]

LGMT! But I would wait for @oleksandr-pavlyk approval.

[h-vetinari]

Couple of questions here, I believe some things may have gone wrong. PTAL @ZzEeKkAa @Alexsandruss

[h-vetinari]

Don't use >= at build time, but pin to a specific version that you need. This is because the version of sysroot_linux at build time will create a corresponding runtime requirement. For example, the current MKL 2025.0 builds are not installable on a linux system with __glibc <2.28 anymore. Is that intentional?

[ZzEeKkAa]

Yes, this is intentional for onemkl-sycl-* packages because sycl runtime/compiler requires glibc >= 2.28. Here it looks like I've accidentally set sycl like requirement on the glibc version

[h-vetinari]

In particular, I don't understand why the sycl components would affect MKL? And if it's just for building, then it should be enough to bump the image version, but leave c_stdlib_version untouched.