Re: More ULE bugs fixed.

From: Jeff Roberson (jroberson_at_chesapeake.net)
Date: 10/27/03

  • Next message: Jon Mini: "Re: More ULE bugs fixed."
    Date: Mon, 27 Oct 2003 02:22:39 -0500 (EST)
    To: Bruce Evans <bde@zeta.org.au>
    
    

    On Fri, 17 Oct 2003, Bruce Evans wrote:

    > On Fri, 17 Oct 2003, Jeff Roberson wrote:
    >
    > > On Fri, 17 Oct 2003, Bruce Evans wrote:
    > >
    > > > How would one test if it was an improvement on the 4BSD scheduler? It
    > > > is not even competitive in my simple tests.
    > > > ...
    > >
    > > At one point ULE was at least as fast as 4BSD and in most cases faster.
    > > This is a regression. I'll sort it out soon.
    >
    > How much faster?
    >

    make kernel on UP seems to be within 1% of 4BSD now. I actually had some
    runs which showed lower system time. I think I can still improve the
    situation some. Anyway, I found some bugs relating to idle prio tasks,
    and also ULE had been doing almost twice as many context switches as 4BSD.
    Now it's doing about 8% more. I'm still tracking this down.

    Anyhow, it should be much closer now. I still have some plans for SMP
    that should improve things quite a bit there but UP is looking good.

    Cheers,
    Jeff

    _______________________________________________
    freebsd-current@freebsd.org mailing list
    http://lists.freebsd.org/mailman/listinfo/freebsd-current
    To unsubscribe, send any mail to "freebsd-current-unsubscribe@freebsd.org"


  • Next message: Jon Mini: "Re: More ULE bugs fixed."

    Relevant Pages

    • [PATCH 1/8] lib/sort: Heapsort implementation of sort()
      ... This patch adds a generic array sorting library routine. ... which has two problem areas for kernel use. ... +void sort(void *base, size_t num, size_t size, ... * A fast, small, non-recursive Osort for the Linux kernel ...
      (Linux-Kernel)
    • Re: [PATCH] (Re: regression tracking (Re: Linux 2.6.21))
      ... >> The problem is that we aren't able to handle the many regression reports ... >> spent much time and energy in writing a good bug report. ... >> source of quality problems in the Linux kernel. ...
      (Linux-Kernel)
    • Re: [Bugme-new] [Bug 7065] New: Devices no longer automount
      ... Summary: Devices no longer automount ... A post-2.6.17 regression, nicely bisected down to a particular changeset ... plug in a USB storage device e.g USB memory stick. ... Last bit of output from udevmonitor for working kernel ...
      (Linux-Kernel)
    • Re: [PATCH] x86/paravirt: revert exports to restore old behaviour
      ... Christoph Hellwig objects to this patch on the grounds that modules ... CONFIG_PARAVIRT and non-CONFIG_PARAVIRT configurations. ... Current practice in the kernel is that when something that should not be ... It cannot be a regression since the kernel does not have a stable API ...
      (Linux-Kernel)
    • Re: [regression] Re: brk randomization breaks columns
      ... I believe it shows clear regression in latest 2.6.25 kernel. ... disables or enables it. ... Those "invalid assumptions" crept into documentation. ...
      (Linux-Kernel)