Re: [PATCH][RFC] make cpu_sibling_map a cpumask_t

From: Ingo Molnar (mingo_at_redhat.com)
Date: 12/09/03

  • Next message: Greg KH: "Re: State of devfs in 2.6?" 
    Date:	Mon, 8 Dec 2003 18:36:28 -0500 (EST)
To: Nick Piggin <piggin@cyberone.com.au>

    the thing that makes balancing-only driven SMT possible with current 2.6.0
    is the overbalancing we do (to have cross-CPU fairness). Previous
    iterations of the O(1) scheduler (all the 2.4 backports) didnt do this so
    all the add-on SMT schedulers tended to have a problem achieving good SMT
    distribution. Now that we more agressively balance, this isnt such a big
    issue anymore.

    so i tend to lean towards your SMT patch, it's much easier to maintain
    than my runqueue-sharing approach. The performance is equivalent as far as
    i can see (around 20%, and a stabilization of the distribution of
    runtimes) - but please also boot my patch and repeat the exact same
    measurements you did.

            Ingo
    -
    To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
    the body of a message to majordomo@vger.kernel.org
    More majordomo info at http://vger.kernel.org/majordomo-info.html
    Please read the FAQ at http://www.tux.org/lkml/

  • Next message: Greg KH: "Re: State of devfs in 2.6?"

    Relevant Pages

    • Re: 2.6.4-mm1
      ... > For SMT it is a less complex than shared runqueues, ... > It is also more flexible than shared runqueues in that you can still ... > For Opteron type NUMA, it actually balances much more aggressively ... > than the default NUMA scheduler, especially when a CPU is idle. ...
      (Linux-Kernel)
    • Re: 2.6.4-mm1
      ... >>For SMT it is a less complex than shared runqueues, ... and done in setup __init code. ... >>example would probably be more difficult to do with shared runqueues. ... >>than the default NUMA scheduler, especially when a CPU is idle. ...
      (Linux-Kernel)
    • Re: 2.6.9-rc1-mm2
      ... Some seriously bad scheduler ... > performance with SMT and HT was fixed up, ... Any particular reason you felt like relaying akpm's 2.6.9-rc1-mm2 ... send the line "unsubscribe linux-kernel" in ...
      (Linux-Kernel)
    • Re: help needed pls. scheduler(kernel 2.6) + hyperthreaded related questions?
      ... You shouldn't be able to select SMT if SMP is not enabled. ... If SMT and SMP is selected, then the scheduler will recognise ... send the line "unsubscribe linux-kernel" in ...
      (Linux-Kernel)
    • Re: 2.6.4-mm1
      ... >>and yet keeps the common scheduler code simple. ... For SMT it is a less complex than shared runqueues, ... It is also more flexible than shared runqueues in that you can still ... For Opteron type NUMA, it actually balances much more aggressively ...
      (Linux-Kernel)