Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[MKL] Add IntelOpenMP as dependency #260

Merged
merged 3 commits into from
Jan 16, 2020
Merged

Conversation

giordano
Copy link
Member

@giordano giordano commented Dec 4, 2019

This reorganises the builder like #259 in order to properly work with Azure Pipelines and adds IntelOpenMP as dependency. Requires JuliaRegistries/General#6255 to be merged.

@giordano
Copy link
Member Author

giordano commented Dec 4, 2019

Here audit takes quite a lot of time, should we disable it, like it was before?

@giordano giordano changed the title [MKL] Use should_build_platform to be able to build on Azure [MKL] Add IntelOpenMP as dependency Dec 4, 2019
@giordano giordano closed this Dec 4, 2019
@giordano giordano reopened this Dec 4, 2019
@giordano
Copy link
Member Author

giordano commented Dec 4, 2019

Interesting: after adding IntelOpenMP as dependency audit is failing 😕

@staticfloat
Copy link
Member

I don't know why, but it's putting the libraries into lib/lib/libmkl_core.so instead of lib/libmkl_core.so, so it can't find them at the end.

@giordano
Copy link
Member Author

Maybe I should check the mv commands

@giordano
Copy link
Member Author

giordano commented Dec 17, 2019

Now ${libdir} already exists and

mv lib ${libdir}

puts lib into ${libdir}.

After we finish the LLVM builds I'll update the script to:

cd $WORKSPACE/srcdir
if [[ ${target} == *mingw* ]]; then
    cp -r Library/bin/* ${libdir}
else
    cp -r lib/* ${libdir}
fi
install_license info/*.txt

@giordano
Copy link
Member Author

giordano commented Jan 16, 2020

Good to go?

I've marked the artifacts to be lazy because, even though the license of MKL allows redistribution, the library is not free and open source software and is also quite large. This sparked some legitimate complaints in JuliaMath/FFTW.jl#133. By making the artifacts lazy we require the users to opt-in the download of the tarball, like setting an environment variable for FFTW.jl.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants