Home » Mailing list archives » linux-kernel » 2008 » February » 11

Re: Scheduler(?) regression from 2.6.22 to 2.6.24 for short-lived threads

Previous message: [thread] [date] [author]
Next message: [thread] [date] [author]
[view in full thread]
From: Olof Johansson
Subject: Re: Scheduler(?) regression from 2.6.22 to 2.6.24 for short-lived threads
Date: Monday, February 11, 2008 - 10:26 am

On Mon, Feb 11, 2008 at 09:15:55AM +0100, Mike Galbraith wrote:

The real workload doesn't see much of an improvement. The changes I did
when tinkering yesterday seem like they're better at modelling just
what's going on with that one.

I've added the new testcase I'm using for the numbers I posted last
night at http://lixom.net/~olof/threadtest/new/, including numbers for
the various kernels. I also included the binaries I built. (with "gcc
-DLOOPS=<size> testcase.c -lpthread").

I tried graphing the numbers as well, it looks like for larger workloads
that 2.6.22 has a fixed benefit over newer kernels. I.e. it seems quicker
at rebalancing, but once things balance out they're obviously doing ok
independent of kernel version.

Graph at http://lixom.net/olof/threadtest/new/schedgraph.pdf. I couldn't
figure out how to make the X axis linear, so it obviously looks odd with
the current powers-of-two at the end instead of linear, but the
differences can still be seen clearly.


-Olof

--
Previous message: [thread] [date] [author]
Next message: [thread] [date] [author]
Messages in current thread:
Scheduler(?) regression from 2.6.22 to 2.6.24 for short-li ..., Olof Johansson, (Fri Feb 8, 5:04 pm)
Re: Scheduler(?) regression from 2.6.22 to 2.6.24 for shor ..., Ingo Molnar, (Fri Feb 8, 5:08 pm)
Re: Scheduler(?) regression from 2.6.22 to 2.6.24 for shor ..., Olof Johansson, (Fri Feb 8, 5:32 pm)
Re: Scheduler(?) regression from 2.6.22 to 2.6.24 for shor ..., Mike Galbraith, (Sat Feb 9, 12:58 am)
Re: Scheduler(?) regression from 2.6.22 to 2.6.24 for shor ..., Willy Tarreau, (Sat Feb 9, 1:03 am)
Re: Scheduler(?) regression from 2.6.22 to 2.6.24 for shor ..., Mike Galbraith, (Sat Feb 9, 3:58 am)
Re: Scheduler(?) regression from 2.6.22 to 2.6.24 for shor ..., Willy Tarreau, (Sat Feb 9, 4:40 am)
Re: Scheduler(?) regression from 2.6.22 to 2.6.24 for shor ..., Mike Galbraith, (Sat Feb 9, 6:37 am)
Re: Scheduler(?) regression from 2.6.22 to 2.6.24 for shor ..., Willy Tarreau, (Sat Feb 9, 9:19 am)
Re: Scheduler(?) regression from 2.6.22 to 2.6.24 for shor ..., Mike Galbraith, (Sat Feb 9, 10:33 am)
Re: Scheduler(?) regression from 2.6.22 to 2.6.24 for shor ..., Olof Johansson, (Sat Feb 9, 10:29 pm)
Re: Scheduler(?) regression from 2.6.22 to 2.6.24 for shor ..., Willy Tarreau, (Sat Feb 9, 11:15 pm)
Re: Scheduler(?) regression from 2.6.22 to 2.6.24 for shor ..., Olof Johansson, (Sun Feb 10, 12:00 am)
Re: Scheduler(?) regression from 2.6.22 to 2.6.24 for shor ..., Willy Tarreau, (Sun Feb 10, 12:58 am)
Re: Scheduler(?) regression from 2.6.22 to 2.6.24 for shor ..., Mike Galbraith, (Mon Feb 11, 1:15 am)
Re: Scheduler(?) regression from 2.6.22 to 2.6.24 for shor ..., Olof Johansson, (Mon Feb 11, 10:26 am)
Re: Scheduler(?) regression from 2.6.22 to 2.6.24 for shor ..., Mike Galbraith, (Mon Feb 11, 12:58 pm)
Re: Scheduler(?) regression from 2.6.22 to 2.6.24 for shor ..., Olof Johansson, (Mon Feb 11, 1:31 pm)
Re: Scheduler(?) regression from 2.6.22 to 2.6.24 for shor ..., Bill Davidsen, (Mon Feb 11, 2:45 pm)
Re: Scheduler(?) regression from 2.6.22 to 2.6.24 for shor ..., Mike Galbraith, (Mon Feb 11, 9:30 pm)
Re: Scheduler(?) regression from 2.6.22 to 2.6.24 for shor ..., Mike Galbraith, (Tue Feb 12, 2:23 am)
Re: Scheduler(?) regression from 2.6.22 to 2.6.24 for shor ..., Mike Galbraith, (Tue Feb 12, 10:49 pm)