Dagger.jl VS JET.jl

Firstly, I'll say that we already have work started to implement out-of-core directly in Dagger: https://github.com/JuliaParallel/Dagger.jl/pull/289.

With that PR in place, it should be possible to define a "storage device" which is backed by a database. I haven't had a chance to actually try this, since the PR still needs quite some work and testing, but it's definitely something on my radar!

I'm not sure that's necessarily the domain of a low-level package like CUDA.jl though (which I assume you're referring to). That kind of interface is more the domain of higher-level packages like https://github.com/JuliaParallel/Dagger.jl/ and to a lesser extent https://juliagpu.github.io/KernelAbstractions.jl/stable/. Moreover, the jury is still out on whether the built-in Distributed module is an ideal abstraction for every use-case (clusters, heterogeneous compute, etc.)

WRT Nx, my biggest question is how they'll crack the problem of still needing big balls of C++ and the shims everywhere to get acceleration. Creating a compiler that generates efficient GPU or other accelerator code is a massive research project with no clear winners, never mind the challenge of reconciling the very mutation-heavy needs of GPU compute with a mostly immutable language model.

An informative discussion on julia forum. Have you tried using https://github.com/aviatesk/JET.jl to minimize type instabilities?

For instance, https://github.com/aviatesk/JET.jl is still in its relative infancy, but it's played a big role in detecting quite a few potential bugs that had never been reported to use by users or caught in our testing infrastructure. There's also been a lot developments like interfaces to RR the time travelling debugger https://rr-project.org/ which helps us better understand and catch some very hard to debug non-deterministic bugs.

Have you seen Shuhei Tadowaki's work on JET.jl (?)

If you're curious: https://github.com/aviatesk/JET.jl

This may seem more about performance (than IDE development) but Shuhei is one of the driving contributors behind developing the capabilities to use compiler capabilities for IDE integration -- and indeed JET.jl contains the kernel of a number of these capabilities.

This is sort of being done right now, as dynamic languages have begun to adopt gradual typing... at least Python and Julia, that I know of.

If something like [JET.jl](https://github.com/aviatesk/JET.jl) become ubiquitous in Julia, one could add a function that pointed out all the places in the code where types are not fully inferred by the compiler.

It'll never be quite the same level of safety as a static language, however.

- Pattern matching (sometimes you don't want the overhead of a method lookup)

[1]: https://github.com/aviatesk/JET.jl

You can use the `@code_warntype` macro to check for type stability, which is very helpful for detecting such performance pitfalls on single function level. In the future, https://github.com/aviatesk/JET.jl may give a more powerful way to do it.