Re: [RFC] Scheduler queue implementation ...

Pavel Machek (pavel@suse.cz)
Fri, 21 Dec 2001 20:21:13 +0100


Hi!

> > I'd guess that if cpu-bound software wants to use clone(CLONE_VM) to
> > gain some performance, it should better work "mostly" in different
> > memory areas on different cpus... But I could be wrong.
>
> Lots of people use shared mm objects and threads for things like UI
> rather

But ... UI performance should not matter much. And that is *abuse* of
threads.

> than just for cpu hogging. If they have multiple cpu hogs doing that then
> they want punishing (or better yet sending a copy of a document on
> caches)

If I have a raytracer, and want to explore 8 cpus, how do I do that?
Separate scene into 8 pieces, make sure no r/w data are shared, and
clone(CLONE_VM). Are there other ways? [I do not know if people are
really doing it like that. *I* would do it that way. Is it bad?]

Pavel

-- 
"I do not steal MS software. It is not worth it."
                                -- Pavel Kankovsky
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/