Re: CFS and O(1) scheduler
- From: Michael Schnell <mschnell_at_bschnell_dot_de@xxxxxxx>
- Date: Mon, 03 Dec 2007 21:43:10 +0100
http://people.redhat.com/mingo/cfs-scheduler/sched-design-CFS.txt
Thanks for the interesting explanation on the CFS.
The article is wrong when stating that there is no processor that does completely fair scheduling in hardware. This is _exactly_ how explicit multithreaded processors are designed.
See the ip5K processor (www.ubicom.com). Same has ten hardware "threads" and with each clock cycle a scheduling table decide which of the active ones is executed. So you can make several processes run at exactly the same speed or define a different scheduling (e.g. 50%, 25%, 12,5% 12,5% and such). Of course there also is a simple priority scheme additionally.
-Michael
.
- Follow-Ups:
- Re: CFS and O(1) scheduler
- From: Juergen Beisert
- Re: CFS and O(1) scheduler
- References:
- Re: CFS and O(1) scheduler
- From: Juergen Beisert
- Re: CFS and O(1) scheduler
- Prev by Date: Re: CFS and O(1) scheduler
- Next by Date: Re: CFS and O(1) scheduler
- Previous by thread: Re: CFS and O(1) scheduler
- Next by thread: Re: CFS and O(1) scheduler
- Index(es):
Relevant Pages
|
|