Analyzing type stability for performance

[yeesian]

Got inspired reading http://staff.mmcs.sfedu.ru/~ulysses/Papers/2021-julia-type-stability.pdf (Updated link: Type Stability in Julia: Avoiding Performance Pathologies in JIT Compilation (Extended Version) (ArXiv link).

Imight be nice to profile this package for type stability and groundedness and see if there are any opportunities for improvement (e.g. #246).

[yeesian]

Putting this within scope for v1.0 in case it might require breaking changes to the API.

[rafaqz]

I was thinking about this yesterday for ArchGDAL. Better type stability should also help with compilation time, which could also be profiled at the same time.

[evetion]

@yeesian Do you have another link to that paper or a full citation? Can't access that server.

[mathieu17g]

@evetion With a search with the following keywords: Type Stability in Julia acm, you should find it easily

• https://www.researchgate.net/publication/354400703_Type_Stability_in_Julia_Avoiding_Performance_Pathologies_in_JIT_Compilation_Extended_Version/fulltext/6136f4cf2b40ec7d8bed5aa3/Type-Stability-in-Julia-Avoiding-Performance-Pathologies-in-JIT-Compilation-Extended-Version.pdf
• https://www.researchgate.net/publication/354400703_Type_Stability_in_Julia_Avoiding_Performance_Pathologies_in_JIT_Compilation_Extended_Version