New Ways to Compile Julia | Bezanson, Baraldi | JuliaCon 2024

[@carpenecopinum1665]

I feel this will be huge. I've been running into several road blocks where the JIT/JAoT compilation model of julia won't work (e.g. a service for a raspberry pi) and being able to build a proper executable from a julia program would make it feasible again

[@csctyf]

Love to heard that is possible soon

[@SomeTechGuy666]

So friggin awesome. Next step: get the code running on an embedded processor.

[@DAV30371]

I always wanted to build binary from Julia!

[@HitsInSandbox]

I guess I am moving back over to Julia, Time to relearn again.... Will be nice to not have to look at any other language in the future.

[@biona002]

Very cool. Which julia version will support this?

[@shalokshalom]

@@biona002 It's open, so basically when its done. ;)

[@chrisrackauckasofficial]

It looks like it's going to make the v1.12 feature freeze

[@meryplays8952]

I think all the R books should be re-written to Julia.

[@BananthahallyVijay]

Good work.👏 But going by the talk, it seems tricky getting a reasonably sized executable for just a "Hello World". That means for an early stage Julia user like me, a concise, stand alone executable is not going to happen soon as my Julia codes are built with lots of dependencies and multiple dispatch at many places.

[@NeverTrust298]

Wow that's pretty cool really smart people!

[@pabloe1802]

🎉🎉🎉

[@xael821]

is there a link to how to use this tooling today?

[@TheJuliaLanguage]

The content of this talk is likely to make it into Julia 1.12, which is currently the development branch of Julia. This pull request: github.com/JuliaLang/julia/pull/55047 contains some *but not all* of the ongoing development

[@tychoides]

I have been checking Julia since release. In my humble opinion, Julia has always been the best for writing simulations. You can get more performance in C but needs a lot of knowledge of the underlying machine. However, Julia never quite make it as a data analysis scripting language. Python is usually enough. And Python has lots of binding to numerous APIs. Python is everywhere. And Julia JIT approach doesn't provide any advantage and is less performant for short scripts. However, static compilation is a game changer. You don't need to compete with Python anymore, you can finally make a library and make a Python API. You can use Julia in Julia or any other language. If this comes to fruition Julia have a chance against Rust or Mojo. Julia is not perfect but GPU stuff is much better than those so far.

[@HitsInSandbox]

Interesting video even watching Gabriel watching the micro expressions and body language of Jeff showing the signs of being nervous with public speaking. Jeff remember in the back of your mind your audience are all sitting their in the nude that should help calm your speech to flow without the nervous traits we as humans give off. Also another tip when doing speeches stay away from striped shirts go with a solid color even if it has a logo that is ok. As the stripes make it easier for the crowd to pick up on the nervous movements as the brains of onlookers is able to use the stripes as a baseline stability index which makes the visual cortex quicker to pickup on more less obvious nervous movements that it would not otherwise unless one trains themselves to watch and track them.

[@StevenSiew2]

ANY considered bloatful.