Re: HTT and SMP question

From: Tom (tom_at_sdf.com)
Date: 04/20/04

  • Next message: FreeBSD Tinderbox: "[releng_4 tinderbox] failure on i386/i386"
    Date: Tue, 20 Apr 2004 10:30:05 -0700 (PDT)
    To: Kris Kennaway <kris@obsecurity.org>
    
    

    On Tue, 20 Apr 2004, Kris Kennaway wrote:

    ...
    > My parallel package builds are slower on build machines with HTT.
    > It's not a magic bullet.

      It is more complicated than that. There isn't just one kind of HTT.
    Intel has been tweaking the HTT feature since it was first released.
    First generation HTT was known to have some performance irregularities.

      HTT is designed to utilize more instruction units on the CPU at the same
    time by running two threads of execution. So at worst, it shouldn't cause
    any slow down versus non-HTT. However, first generation HTT definitely
    did slow down on some work loads. Intel says they have fixed this in the
    newer generation processors.

      There are also FreeBSD factors: FreeBSD 4 compiled with SMP, avoids a
    lot of locking code. FreeBSD 4 with SMP has a lot of extra locking code,
    and that code has cost associated with it. The cost of SMP locking in
    FreeBSD 4, may out way the benefit of HTT. With FreeBSD 5, the SMP
    locking is done quite differently, and it is done all the time anyhow
    (since FreeBSD 5 is SMP by default). So HTT on FreeBSD 5 is more likely
    to be beneficial than HTT on FreeBSD 4 (with a non SMP kernel).

    > Kris

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


  • Next message: FreeBSD Tinderbox: "[releng_4 tinderbox] failure on i386/i386"

    Relevant Pages

    • Re: Kudos to everyone who worked on the gcc-4.2 merge
      ... load time decrease) in my server (P4 HTT with SMP). ... This will be one of the great changes that will put FreeBSD back on ... and let me know if you need testing with gcc ...
      (freebsd-current)
    • Re: still spurious interrupts with ICH5 SATA
      ... My system works just fine with HTT and SMP now. ... > If I boot with apic disabled - doesn't make difference ... > HTT is disabled in bios, and SMP not compiled in kernel. ...
      (freebsd-current)
    • Re: Possible Threading problem with -CURRENT / MySQL?
      ... On Wed, 16 Jun 2004, Robert Watson wrote: ... A chat with Scott Long this evening suggests that HTT ... > the SMP numbers a fair amount due to additional IPIs without benefit. ... > configuration is marginally faster than it was last month, ...
      (freebsd-current)
    • Re: Intel Hyperthreading and 5.3-RELEASE
      ... > The hardware release notes and SMP man page are a little vague about ... It is strongly dependent on your desired workload. ... WRT HTT, you have to try it yourself and see how it works for you - ...
      (comp.unix.bsd.freebsd.misc)
    • Re: FBSD and and the sk() network driver
      ... it isn't really a true SMP. ... You might compare HTT to a Winmodem. ... The only things I have added to my kernel that differs from 'GENERIC' are: ... # CPUS if needed. ...
      (comp.unix.bsd.freebsd.misc)