Re: [ck] Re: Scheduler benchmarks - a follow-up

!MAILaRCHIVE_VOTE_RePLACE
Previous message: [thread] [date] [author]
Next message: [thread] [date] [author]
To: Jos Poortvliet <jospoortvliet@...>
Cc: Rob Hussey <robjhussey@...>, <ck@...>, <linux-kernel@...>
Date: Monday, September 17, 2007 - 10:12 am

* Jos Poortvliet <jospoortvliet@gmail.com> wrote:


lower is better, but you have to take the thing below into account:


on my box the TSC overhead has hit CFS quite hard, i'm not sure that's 
true on Rob's box. So i'd expect them to be in roughly the same area.


yeah, it's much less pronounced in real life - a context-switch rate 
above 10,000/sec is already excessive - while for example the lat_ctx 
test generates close to a million context switches a second.


i dont think so - we want precise/accurate scheduling before 
performance. (otherwise tasks working off the timer tick could steal 
away cycles without being accounted for them fairly, and could starve 
out all other tasks.) Unless the difference was really huge in real life 
- but it isnt.

	Ingo
-
Previous message: [thread] [date] [author]
Next message: [thread] [date] [author]

Messages in current thread:
Scheduler benchmarks - a follow-up, Rob Hussey, (Mon Sep 17, 5:21 am)
Re: Scheduler benchmarks - a follow-up, Ingo Molnar, (Mon Sep 17, 9:05 am)
Re: Scheduler benchmarks - a follow-up, Rob Hussey, (Mon Sep 17, 9:44 pm)
Re: Scheduler benchmarks - a follow-up, Ingo Molnar, (Mon Sep 17, 4:36 pm)
Re: Scheduler benchmarks - a follow-up, Rob Hussey, (Tue Sep 18, 12:30 am)
Re: Scheduler benchmarks - a follow-up, Ingo Molnar, (Tue Sep 18, 4:48 am)
Re: Scheduler benchmarks - a follow-up, Rob Hussey, (Tue Sep 18, 5:45 am)
Re: Scheduler benchmarks - a follow-up, Ingo Molnar, (Tue Sep 18, 5:48 am)
Re: Scheduler benchmarks - a follow-up, Ingo Molnar, (Tue Sep 18, 2:40 am)
Re: Scheduler benchmarks - a follow-up, Rob Hussey, (Tue Sep 18, 4:23 am)
Re: Scheduler benchmarks - a follow-up, Willy Tarreau, (Tue Sep 18, 12:53 am)
Re: Scheduler benchmarks - a follow-up, Rob Hussey, (Tue Sep 18, 12:58 am)
Re: [ck] Re: Scheduler benchmarks - a follow-up, Jos Poortvliet, (Mon Sep 17, 10:01 am)
Re: [ck] Re: Scheduler benchmarks - a follow-up, Ingo Molnar, (Mon Sep 17, 10:12 am)
Re: Scheduler benchmarks - a follow-up, Ingo Molnar, (Mon Sep 17, 7:27 am)
Re: Scheduler benchmarks - a follow-up, Ed Tomlinson, (Mon Sep 17, 7:12 am)
Re: Scheduler benchmarks - a follow-up, Ingo Molnar, (Mon Sep 17, 4:22 pm)
Re: Scheduler benchmarks - a follow-up, Ingo Molnar, (Mon Sep 17, 7:47 am)