Hacker News Reader

Comment by maxbond | Hacker News Reader

To be fair though M:N threads aren't the provenance of research language, Go and Rust both have them, and I'm sure some other languages.

(But it's awesome Java has them now too, other languages getting the feature earlier doesn't really devalue it.)

View on HN

My point is that they're on diametrically opposite ends of the language-philosophical spectrum.

.net has had this feature for a decade++

Yeah, so has a lot of languages. See the previous comment.

you’re trying to support “judicious and conservative” but realistically java is on a lifeline like fortran, delphi and objectiveC

What are you basing this sentiment on?

That “lifeline” is apparently the same the internet is on.

What is the virtual thread feature in .NET called? I can't find any information about it.

Tasks, simple intro here https://code-maze.com/csharp-tasks-vs-threads/

That looks similar to Java FutureTasks + Executors which is a very different concept from virtual threads.

Virtual threads mean that a blocking thread can yield to any other non blocking thread seamlessly and with very little overhead. .NET Tasks cannot do this as far as I can tell.

I don't think this is the same thing. As far as I can tell, the task abstraction is a threapool where you can submit operations and return futures. If a task blocks indefinitely, the underlying threadpool OS worker thread will be blocked, and the threadpool either has to run with less resources or spawn a new worker. Virtual threads are an M:N abstraction: blocking on a virtual thread will not block the underlying OS thread.

.NET might indeed have a virtual thread abstraction and if it does you could of course implement the Task abstraction on top of either virtual threads or OS threads, but what you linked to is not a proof that it does.

Are you sure? All the discussion I can find online makes it seem to me like TPL and friends are just executing tasks on thread pools until completion. (see e.g., https://github.com/dotnet/runtime/issues/50796 for some discussion)

Fair enough.

Rust doesn’t have them, though.

In what way? async-std and tokio both support M:N.

But they can’t turn blocking calls to non-blocking which is the whole point, that requires runtime support.

Oh interesting, that's very cool, I didn't realize Java was doing that. That's a different axis than M:N though (cooperative versus preemptive) and you could definitely write a preemptive async runtime for Rust (rtic comes to mind). But the async-std and tokio runtimes are certainly cooperative.

(As a note, cooperative scheduling also requires a runtime - Rust might not "have a runtime" by default but you need to opt into one to use async.)

Reply on news.ycombinator.com