Improvements in Package Precompilation | Tim Holy & Valentin Churavy | JuliaCon 2022

Julia code can be precompiled to save time loading and/or compiling it on first execution. Precompilation is nuanced because Julia code comes in many flavours, including source text, lowered code, type-inferred code, and various stages of optimization to reach native machine code. We will summarize what has (and hasn't) previously been precompiled, some of the challenges posed by Julia's dynamism, the nature of some recent changes, and prospects for near-term extensions to precompilation. For more info on the Julia Programming Language, follow us on Twitter: https://twitter.com/JuliaLanguage and consider sponsoring us on GitHub: https://github.com/sponsors/JuliaLang Contents 00:00 Welcome! 00:30 Julia run fast and you can write code in it quickly 03:20 But compilation takes time 05:44 The two halves of Time to First Execution (TTFX) 06:13 Strategies to reduce compilation time 07:26 Why precompilation is important? 08:24 What gets saved from Julia's compilation pipeline? 12:08 Challenges ahead: invalidation 14:20 Current work on precompilations 16:26 Custom precompilation by creating Startup.jl 17:19 New package: SnoopPrecompile.jl 17:52 Comparing using Startup.jl and Packagecompiler 19:12 Looking ahead: saving native code 20:12 What's needed to support packages in storage file staticdata.c? 22:16 Challenges that we faced 24:49 GPUCompiler.jl approach and its developments 26:31 Moving from sysimage to pkgimage? 27:12 Relocating code 28:18 Diagram illustrating relocation of code 29:00 Next steps 30:18 A hopeful future 31:36 Closing remarks Want to help add timestamps to our YouTube videos to help with discoverability? Find out more here: https://github.com/JuliaCommunity/YouTubeVideoTimestamps Interested in improving the auto generated captions? Get involved here: https://github.com/JuliaCommunity/YouTubeVideoSubtitles