Messages in this thread |  | | Date | Fri, 25 Aug 2000 11:49:14 +0200 | From | Jamie Lokier <> | Subject | Re: SCO: "thread creation is about a thousand times faster than on native Linux" |
| |
Theodore Y. Ts'o wrote: > because they think it's amusing to create (in some caes) hundreds of > thousands of threads, and a 1:1: model simply falls over in the face > of such a (unreasonable?) number of threds.
A large number of threads is unreasonable if you're, say, a web server.
But if you can get rid of the stacks, and you _can_ get rid of the stacks sometimes, then why not have one thread per widget in a GUI? Or one thread per animated objected on a web page? Some notions of "thread" are light enough to compare with straightforward list processing and function calls, because that's how they're implemented.
The cherry on that pie is when you can have lightweight threads that have the _option_ to go off and do some heavy system calls or processing that ties up a stack, without blocking the rest of the lightweight threads. Obviously if they all do some heavy processing, lots of resources are tied up, but they don't all do that at the same time.
_That_ requires a little kernel support. As you point out, Solaris tried to do it with a 1:n model, and it didn't work too well.
-- Jamie - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org Please read the FAQ at http://www.tux.org/lkml/
|  |