tealr VS Enzyme

working more on tealr more specifically tealr_doc_gen

Fixing some short comings in tealr that people discovered. Also, making hv_lua work with tealr and ideally start adding that fishfight.

Fantastic! Very happy to see this move forward at such a fast pace. We’re implementing scripting in Fish Fight via the tealr crate, which wraps mlua (and rlua).

Hopefully the last hand on https://github.com/lenscas/tealr for a new release. Mostly going over the documentation. Laying less emphasis on teal and more on "Hey, using this you can express a more typesafe api to lua and can actually easily document it.".

tealr, my wrapper around rlua and mlua to focus on (among other things) being better able to describe your api in the form of types: Getting it to a point to release a new version. Which is all about allowing people to document they API that they created for lua/teal. This documentation gets written to the definition files when generating them. In addition, you can also ask tealr to generate a instance.help(page:Option) method on your types. When doing this, the documentation gets made available through this function. By default it prints out "type level" documentation for this type, as well as what other pages are availalbe. When given a string, it shows the documentation for that page instead. Pages that belong to methods also automatically include the type signature.

This requires some new features in my Rust <-> Teal FFI library tealr https://github.com/lenscas/tealr which is now also close to getting a new release, with the only thing I want to do is fix how its TypedFunction work (right now, it always contains a lua function, even if it was made inside of Rust, which puts some needless limits on it).

After that it is time to finish mlua support for tealr . As well as making it better at dealing with api's like tealsql has made.

In case anyone is interested: Issue about it here https://github.com/lenscas/tealr/issues/17

Can you try removing the println! at https://github.com/lenscas/tealr/blob/cee2363a14c53b995a69efb2b4cbd84010905276/tealr_derive/src/lib.rs#L155?

Automatic differentiation is done using https://enzyme.mit.edu/

lattner left google and was the primary reason they chose swift, so they lost interest.

if you're asking from an ML perspective, i believe the original motivation was to incorporate automatic differentiation in the swift compiler. i believe enzyme is the spiritual successor.

https://github.com/EnzymeAD/Enzyme

https://ispc.github.io/ispc.html

For the auto-differentiation when I need performance or memory, I currently use tapenade ( http://tapenade.inria.fr:8080/tapenade/index.jsp ) and/or manually written gradient when I need to fuse some kernel, but Enzyme ( https://enzyme.mit.edu/ ) is also very promising.

MPI for parallelization across machines.

For autodiff, we are currently working again on publishing a new Enzyme (https://enzyme.mit.edu) Frontend for Rust which can also handle pure Rust types, first version should be done in ~ a week.

Have you taken a look at enzymeAD? There is a group porting it to rust.

Enzyme dev here, so take everything I say as being a bit biased:

While, by design Enzyme is able to run very fast by operating within the compiler (see https://proceedings.neurips.cc/paper/2020/file/9332c513ef44b... for details) -- it aggressively prioritizes correctness. Of course that doesn't mean that there aren't bugs (we're only human and its a large codebase [https://github.com/EnzymeAD/Enzyme], especially if you're trying out newly-added features).

Notably, this is where the current rough edges for Julia users are -- Enzyme will throw an error saying it couldn't prove correctness, rather than running (there is a flag for "making a best guess, but that's off by default"). The exception to this is garbage collection, for which you can either run a static analysis, or stick to the "officially supported" subset of Julia that Enzyme specifies.

Incidentally, this is also where being a cross-language tool is really nice -- namely we can see edge cases/bug reports from any LLVM-based language (C/C++, Fortran, Swift, Rust, Python, Julia, etc). So far the biggest code we've handled (and verified correctness for) was O(1million) lines of LLVM from some C++ template hell.

I will also add that while I absolutely love (and will do everything I can to support) Enzyme being used throughout arbitrary Julia code: in addition to exposing a nice user-facing interface for custom rules in the Enzyme Julia bindings like Chris mentioned, some Julia-specific features (such as full garbage collection support) also need handling in Enzyme.jl, before Enzyme can be considered an "all Julia AD" framework. We are of course working on all of these things (and the more the merrier), but there's only a finite amount of time in the day. [^]

[^] Incidentally, this is in contrast to say C++/Fortran/Swift/etc, where Enzyme has much closer to whole-language coverage than Julia -- this isn't anything against GC/Julia/etc, but we just have things on our todo list.

Idk, Enzyme is pretty next gen, all the way down to LLVM code.

https://github.com/EnzymeAD/Enzyme

I'm working on merging my build-tool for (oxide)-enzyme into Enzyme itself. Also looking into improving the documentation.

that seems one of the points of enzyme[1], which was mentioned in the article.

[1] - https://enzyme.mit.edu/

being able in effect do interprocedural cross language analysis seems awesome.