AI-generated JIT compilers from interpreters: Are we there yet?

3 pointsprotocontrol1y ago2 comments

Have you seen someone already using generative AI to automatically generate optimized JIT compiler frontends from interpreter definitions?

2 comments

I see no reason for any sort of AI to be applied here.

There's an interesting guide on converting an interpreter into a static binary translator: https://gtoal.com/sbt/

It would be great to see this expanded into a JIT compiler (using a generic translated code management/caching and optimization framework).

In a world where AI has crept into virtually every activity, it is hard to believe that AI cannot help in compiler generation.

Why would the field of JIT meta-compilers be exempt from AI?

j / k navigate · click thread line to collapse

3 pointsprotocontrol1y ago2 comments

Have you seen someone already using generative AI to automatically generate optimized JIT compiler frontends from interpreter definitions?

I see no reason for any sort of AI to be applied here.

There's an interesting guide on converting an interpreter into a static binary translator: https://gtoal.com/sbt/

It would be great to see this expanded into a JIT compiler (using a generic translated code management/caching and optimization framework).

In a world where AI has crept into virtually every activity, it is hard to believe that AI cannot help in compiler generation.

Why would the field of JIT meta-compilers be exempt from AI?

j / k navigate · click thread line to collapse