runtimelab VS sdk

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.

[DllImport("libc")] private static extern uint geteuid(); public bool IsCurrentProcessElevated() { if (RuntimeInformation.IsOSPlatform(OSPlatform.Windows)) { // https://github.com/dotnet/sdk/blob/v6.0.100/src/Cli/dotnet/Installer/Windows/WindowsUtils.cs#L38 using var identity = WindowsIdentity.GetCurrent(); var principal = new WindowsPrincipal(identity); return principal.IsInRole(WindowsBuiltInRole.Administrator); } // https://github.com/dotnet/maintenance-packages/blob/62823150914410d43a3fd9de246d882f2a21d5ef/src/Common/tests/TestUtilities/System/PlatformDetection.Unix.cs#L58 // 0 is the ID of the root user return geteuid() == 0; }

> There's definitely things they tried to improve on that... weren't really improvements. The way "assemblies" are matched in .NET is much more sophisticated- the goal there was to try to kill DLL hell. It evolved into the Global Assembly Cache, which is sort of the Windows Registry of DLLs. Not a huge fan of those bits.

The Global Assembly Cache did not make the jump to the modern .NET (Core). There was the thing called `dotnet store`, but it’s broken since .NET 6: https://github.com/dotnet/sdk/issues/24752

The assembly redirection hell has also been greatly reduced there.

I do the same.

I have a small write-up here: https://chrlschn.dev/blog/2023/10/end-to-end-type-safety-wit...

You get end-to-end type safety (even better once you connect it to EF Core since you get it all ways to your DB).

With this setup with hot-reload (currently broken in .NET 8 [0]), productivity is really, really good. Like tRPC but with one of the most powerful ORMs out there right now.

[0] https://github.com/dotnet/sdk/issues/36918

Noticed an open issue just now.

You're thinking of https://github.com/dotnet/sdk/issues/22247

That's a twisted and wrong narrative

Unity like refers to a Editor driven approach

Unity became popular with its moonscript language (javascript like), they then ditched it to focus on C#, but what propelled unity to what it is today is the Editor driven approach, not c#, not DOTS

They are forced to transpile C# to C++ via IL2CPP as a result to target consoles/mobiles

C# is a disease when it comes to console/mobile support

It's a substantial dependency, quite heavy

And you are not free of unity like fuck ups, it's a microsoft language after all:

https://github.com/dotnet/sdk/issues/22247

And let's not forget when they changed the license of their debugger overnight to prevent people from using it in their products (jetbrains for example)

And them deprecating open source tooling to a proprietary/closed one for vscode (c# devkit)

Let's be careful when we recommend evil as an alternative to evil ;)

The script can be found at the end of the thread here https://github.com/dotnet/sdk/issues/8742

Another difference you might be surprised by is that the .NET tooling by default collects various data from your system and sends it to Microsoft [1]. If you want to avoid this (and still want to use .NET) you'll have to make sure that the environment variable DOTNET_CLI_TELEMETRY_OPTOUT is set in all contexts before touching anything.

[1] https://github.com/dotnet/sdk/issues/6145

SDK Pull Request

Pretty sure the final file referenced here https://github.com/dotnet/sdk/issues/8742 is the one I am thinking of.