Re: Linux 2.6.23

Previous message: [thread] [date] [author]
Next message: [thread] [date] [author]
From: Nick Piggin
Date: Thursday, October 11, 2007 - 2:16 am

On Wednesday 10 October 2007 20:14, Ingo Molnar wrote:

;) I think you snipped the important bit:

"the peak is terrible but it has virtually no dropoff and performs
better under load than the default 2.6.21 scheduler." (verbatim)

The dropoff under load was due to trivially avoided mmap_sem
contention in the kernel and glibc (and not-very-scalable mysql
heap locking), rather than specifically anything the scheduler
was doing wrong, I think (when the scheduler chose to start
preempting threads holding locks, then performance would tank.
Exactly when that point was reached, and what happens afterwards
was probably just luck.)
-
Previous message: [thread] [date] [author]
Next message: [thread] [date] [author]

Messages in current thread:
Linux 2.6.23, Linus Torvalds, (Tue Oct 9, 1:54 pm)
Re: Linux 2.6.23, Nicholas Miell, (Tue Oct 9, 11:12 pm)
Re: Linux 2.6.23, René Rebe, (Wed Oct 10, 12:44 am)
Re: Linux 2.6.23, Alexey Dobriyan, (Wed Oct 10, 1:37 am)
Re: Linux 2.6.23, Michael Tokarev, (Wed Oct 10, 2:12 am)
Re: Linux 2.6.23, Ingo Molnar, (Wed Oct 10, 3:14 am)
Re: Linux 2.6.23, Alexey Dobriyan, (Wed Oct 10, 3:36 am)
Re: Linux 2.6.23, Jan Engelhardt, (Wed Oct 10, 3:53 am)
Re: Linux 2.6.23, Michael Tokarev, (Wed Oct 10, 4:13 am)
Re: Linux 2.6.23, Ingo Molnar, (Wed Oct 10, 12:14 pm)
Re: Linux 2.6.23, Michael Tokarev, (Wed Oct 10, 12:26 pm)
Re: Linux 2.6.23, Andi Kleen, (Wed Oct 10, 1:04 pm)
Re: Linux 2.6.23, Krzysztof Halasa, (Wed Oct 10, 4:27 pm)
Re: Linux 2.6.23, Nicholas Miell, (Wed Oct 10, 6:20 pm)
Re: Linux 2.6.23, Zhang, Yanmin, (Wed Oct 10, 7:34 pm)
Re: Linux 2.6.23, Nick Piggin, (Thu Oct 11, 2:16 am)
Re: Linux 2.6.23, Ingo Molnar, (Thu Oct 11, 6:32 am)
Re: Linux 2.6.23, Nick Piggin, (Thu Oct 11, 7:15 am)
Re: Linux 2.6.23, Ingo Molnar, (Thu Oct 11, 10:46 pm)
Re: Linux 2.6.23, Bill Davidsen, (Fri Oct 12, 5:21 am)