[haiku-development] Re: Scheduler algorithm improvements

  • From: "Axel Dörfler" <axeld@xxxxxxxxxxxxxxxx>
  • To: haiku-development@xxxxxxxxxxxxx
  • Date: Sat, 07 Nov 2009 13:25:41 +0100

Mikhail Panasyuk <otoko@xxxxxxxxx> wrote:
> 06.11.09, 17:33, "Axel Dörfler" <axeld@xxxxxxxxxxxxxxxx>:
> > > in it's current state violates scheduling rule declared in Be
> > > Book /
> > > Be Newsletter
> > > (see volume III issue 45).
> > I couldn't find any scheduling related content in that issue.
> The rule I refer to:
> 1) The likeliness for a thread to be scheduled increases with a factor
> of two for such unit of priority.

I just searched for it: it's in volume 4, issue 45, not volume 3.

[...]
> > > http://otoko.narod.ru/files/haiku/haiku_scheduler_part1.html
> > > http://otoko.narod.ru/files/haiku/haiku_scheduler_part2.html
> > There is obviously something wrong with the encoding set, so that
> > Google
> > cannot translate any of it. Would be nice if you could fix this.
> Hosting problem. Document in UTF-8 with correct content-type meta-tag
> but
> server reports Windows-1251. Maybe because there is floating ad window
> added by hoster... Anyway, I've converted htmls to Windows-1251,
> should be
> ok now.

Indeed, thanks!
Now I at least understand what you mean by "better": fairness. From you
numbers it looks like the Haiku scheduler gives lower priority threads
less chance to run if a higher priority thread wants to run.
In the case of Haiku, it skips to a lower priority thread in 20% of the
case, while the BeOS/Dano schedulers seems to be more fair. It might
certainly be a good idea to check if the fairer approach feels good as
well, or even better.

> In real world even current Haiku implementation or old BeOS one should
> be
> good enough in majority of cases and they are. But I believe there is
> room for
> improvements.

There is definitely room for improvements, I'm not saying that our
scheduler is perfect; it just seem to work well enough.

Bye,
   Axel.


Other related posts: