HPC systems with Julia support

Hey everyone! I am trying to collect information about HPC systems (from university compute clusters to supercomputers) where Julia is available as a pre-installed module and, optimally, officially supported. This is part of an effort to gather information on Julia as a HPC programming language and its prevalence in “traditional” HPC environments.

Here you will find a list of systems supporting Julia that we already have on record:

If you are a Julia HPC user, a system operator, or just someone who knows about their university’s/company’s Julia efforts, it would be great if you can contribute your system to the list. Optimally by creating a PR with the information you know (it does not have to be complete), or, alternatively, by replying to this thread. Thanks!

7 Likes

I’m very interested in the comments about the negligible performance gains of building from source vs using official binaries. I guess this is worth a thread in itself.

Yes, like this thread:

This recent pull request:
https://github.com/JuliaLang/julia/pull/45641
adds support for doing PGO and LTO to Julia build system. As I said in the thread linked above, there is some improvement in terms of compile latency (which I guess everybody would be happy about), and Julia’s own runtime, but the code generated by Julia is still the same

Thanks for sharing @giordano! I am watching this PR now and once its merged and part of at least a release candidate, I’ll give it a try in an HPC environment to see what kind of speed difference I can get.

The pull request only adds a Makefile, not sure it’s worth waiting months until it makes it in a new release, you can copy it and follow the instructions :slightly_smiling_face: