criterium VS JWM

(criterium does not work here yet b.t.w., but it probably will be working soon)

criterium, specifically the quick-bench function, will actually run multiple samples an provide a mean runtime (as well as other useful stats) so you can get an idea of what a jit'd warmed up performance looks like. time is great in a pinch, but you end up needing to run it multiple times to ensure optimizations are kicking and and other artifacts (like gc) aren't throwing the results.

I'm going to leave tooling out of this and run everything through a repl on the command line right from the jar. One other thing I want to do is include the incredible criterium library so we can profile. I'm deliberately including criterium separately like this because you shouldn't have a dev-time tool like criterium in an uberjar. And knowing how to easily combine other jars with your real production jar can be very helpful. I grabbed the jar from my .m2 cache.

I am just going to leave this here - https://github.com/hugoduncan/criterium

where crit is criterium. As you can see, you're spending most of your time in the seq transformation part.

It's better to benchmark with something like criterium. time is a bit inaccurate. Though, if it's really 15 seconds, I guess will not be that big of a difference

>>> One of Clojure's biggest weaknesses in practice is that breaking in to those functional structures to figure out where the time is being spent or to debug them is harder than in other languages. This is a natural trade-off of developing a terse and powerful language.

Not that hard if you use something like YourKit. There's also a quite good Clojure library https://github.com/hugoduncan/criterium .

Criterium (the benchmarking library used here) uses multiple runs to obtain tighter bounds on amortized performance, as well as techniques to amortize the effects of garbage collection and JIT compilation. See https://github.com/hugoduncan/criterium for a brief overview, as well as links to the pitfalls and statistical techniques involved in JVM benchmarking.

sigh

Yeah. I am very bullish on Kotlin. Think it's probably the most exciting language evolving right now.

I went on a few-tweet minirant here about why:

https://twitter.com/GavinRayDev/status/1443279425311805440

But the tl;dr is that:

- There is Jetpack Compose currently, for Desktop, Web, and Android

- And Kotlin Native putting a large portion of resources into Skia bindings (JetBrains calls the lib "Skiko" for Kotlin Native https://github.com/JetBrains/skiko and "Skija")

It's very clear (and there are some employees which have confirmed this IIRC) that they are working on "Jetpack Compose Everywhere" that runs on iOS as well, from a single codebase.

There's the big Kotlin event going on right now, where they just announced the new WASM backend and changes in their compiler + IR commonizing/restructuring ("K2").

- https://blog.jetbrains.com/kotlin/2021/10/the-road-to-the-k2...

- https://www.youtube.com/watch?v=-pqz9sKXatw

The net result is that you wind up with a single language that you can use to write your backend API, your UI code (Jetpack Compose app deployed across Web/Android/iOS/Mac/Win/Linux, or transpile to JS/TS if you just want a web app, etc) and with Kotlin Native even your native, low-level code to integrate with existing C/C++ etc ecosystem.

KN already does automatic bindgen for C and Swift headers, they have direct C++ interop (like Swift does) on their future roadmap as a potential "todo".

All of this is mostly possible already -- I can do the same thing using IE Java, GraalVM, and a transpiler like Google's j2cl or bck2brwser (which is what Gluon uses for JavaFX on the web). Including the "native" part.

IE, here's a contribution I made to get GraalVM producing native binaries using Skia from the JVM + JNI Jetbrains Skia library:

https://github.com/HumbleUI/JWM/issues/158

But Kotlin is pushing the hardest to make this whole platform/stack from native <-> desktop <-> mobile <-> browser a seamless, unified experience. And you can feel it, when you try to do the "whole stack, every platform, one language" thing.

Sorry for the rant and wall of text!