gc VS runtimelab

It may take some time for WasmGC to be usable by .NET. Based on the discussions the first version of WasmGC does not have a good way to handle a few .NET specific scenarios, and said scenarios are "post-post-mvp". [0]

My concern, of course, is that there is not much incentive for those features to be added if .NET is the only platform that needs them... at that point having a form of 'include' (to where a specific GC version can just be cached and loaded by another WASM assembly) would be more useful, despite the pain it would create.

[0] - https://github.com/WebAssembly/gc/issues/77

Yes, that's definitely true: a single GC will not be optimal for everything, or even possible. Atm interior pointers are not supported at all, for example, but they are on the roadmap for later:

https://github.com/WebAssembly/gc/blob/main/proposals/gc/Pos...

What launched now is enough WasmGC to support a big and useful set of languages (Java, Kotlin, Dart, OCaml, Scheme), but a lot more work will be required here!

GC proposal is from 2018: https://github.com/WebAssembly/proposals/issues/16 and there’s code: https://github.com/WebAssembly/gc/blob/master/proposals/gc/O...

Seems like an awefully long time for progress to be made, given all the possibilities it would unlock.

First is the size. Writing a server-side and client-side program is possible with Rust, and the resulting WASM package will be small enough. At the same time, Microsoft Blazor converts C# code to WASM, but the client delivery has to include the reduced .NET runtime, taking several megabytes for a script. The same is true for GoLang, even with an attempt to reduce the runtime delivery in TinyGo WASM. Developers want to work with their favorite languages, whether it is Java, Kotlin, Dart, C#, F#, Swift, Ruby, Python, C, C++, GoLang, or Rust. These languages produce groups of runtimes. For example, JVM and .NET have many common parts, Ruby and Python are dynamically interpreted at runtime, and all mentioned depend on automatic garbage collection. For smaller WASM packages, browser vendors can include extended runtime implementations, for example, by delivering a general garbage collector as part of WASM. Garbage collection support by WASM is currently in progress: WASM GC, .NET WASM Notes.

My understanding is that the main limitation is technical. WASM doens't do GC or the host system calling conventions and cannot interact directly with object from Javascript because of this. However, this is being worked[0] on and will be solved eventually. Even without this the performance overhead of bridging to JS is low enough that WASM frameworks can beat out React.

0: https://github.com/WebAssembly/gc/blob/main/proposals/gc/Ove...

I’ve just watched a video on YouTube from Google I/O 2023 on Flutter for the web. Kevin Moore explains that Flutter can compile to Wasm, but now that GC support has been added to the standard and WasmGC is supported in Chromium and Firefox, I’m quite intrigued.

I have found the main thread about using WebAssembly GC in C#: https://github.com/WebAssembly/gc/issues/77. If I understand it correctly, it is not possible to use the current prototype version of GC in C#.

The .net team has given their notes on it, the concern is more on the memory layout from what I remember. Though it may be possible still. The runtime would likely still ship some gc code, but only a subset for cases not supported by the wasm gc itself and a few more for interfacing with the gc service, which overall should still result on smaller payloads compared to current sizes.

I assume that's one of the parts of the work done at https://github.com/WebAssembly/gc - not happening any soon yet, but it'll eventually be done.

There was a "green thread" experiment for dotnet a while ago, here is the conclusion: https://github.com/dotnet/runtimelab/issues/2398

Experiment result write-up: https://github.com/dotnet/runtimelab/blob/e69dda51c7d796b812...

TLDR: The green threads experiment was a failure as it found (expected and obvious) issues that the Java applications are now getting to enjoy, joining their Go colleagues, while also requiring breaking changes. It, however, gave inspiration to subsequent re-examination of current async/await implementation and whether it can be improved by moving state machine generation and execution away from IL completely to runtime. It was a massive success as evidenced by preliminary overhead estimations in the results.

Yeah, it kind of is. There are quite a few of experiments that are conducted to see if they show promise in the prototype form and then are taken further for proper integration if they do.

Unfortunately, object stack allocation was not one of them even though DOTNET_JitObjectStackAllocation configuration knob exists today, enabling it makes zero impact as it almost never kicks in. By the end of the experiment[0], it was concluded that before investing effort in this kind of feature becomes profitable given how a lot of C# code is written, there are many other lower hanging fruits.

To contrast this, in continuation to green threads experiment, a runtime handled tasks experiment[1] which moves async state machine handling from IL emitted by Roslyn to special-cased methods and then handling purely in runtime code has been a massive success and is now being worked on to be integrated in one of the future version of .NET (hopefully 10?)

[0] https://github.com/dotnet/runtime/issues/11192

[1] https://github.com/dotnet/runtimelab/blob/feature/async2-exp...

Unlike these folks from dotnet, which tested directly on ASP for real workload

  https://github.com/dotnet/runtimelab/issues/2398?darkschemeovr=1

[1] https://github.com/dotnet/runtimelab/issues/2398

The only way to avoid it is to not build on top of Java or not adding any features on top of Java.

> To give another example with C#, there has been a lot of recent discussion about finding potential alternatives to their async-await concurrency model. They cite the level of effort it takes to maintain the async await style code and the costs that come from this.

I had a very different take-away. They did PoC with virtual threads and decided it's not worth the switch now and async-await that they have is good enough.

https://github.com/dotnet/runtimelab/issues/2398

> Some of the languages it gets compared too aren't even that old yet.

C# is old enough to drink and Scala just had its 20th birthday this week :)

It was tried and the dotnet team decided to drop it: https://github.com/dotnet/runtimelab/issues/2398

Technical details here: https://github.com/dotnet/runtimelab/blob/feature/green-thre...

Just last month .NET ended a green threading experiment, mainly because the overhead it adds to FFI was too high:

https://github.com/dotnet/runtimelab/issues/2398

Rust had green threads until late 2014, and they were removed because of their impact on performance.

Everyone has done the basic research: green threading is a convenient abstraction that comes with certain performance trade offs. It doesn't work for the kind of profile that Rust is trying to target.