Re: [Jackit-devel] Re: Statistical methods for latency profiling

From: Ingo Molnar (mingo_at_redhat.com)
Date: 08/01/04

  • Next message: Ingo Molnar: "Re: [patch] voluntary-preempt-2.6.8-rc2-M5" 
    Date:	Sun, 1 Aug 2004 07:21:17 -0400 (EDT)
To: Lee Revell <rlrevell@joe-job.com>

    On Sun, 1 Aug 2004, Lee Revell wrote:

    > So stressing the filesystem moves the center to the right a bit, from
    > 6-7 to 9-10, and *drastically* lengthens the 'tail'.

    basically each codepath has a typical latency distribution, and when a
    workload uses multiple codepaths then the latencies get intermixed almost
    linearly.

    > These numbers suggest to me that a lot of the latencies from 47 usecs
    > and up are caused by one code path, because they are so uniformly
    > distributed over the upper part of the histogram. The prime suspect of
    > course being the ide io completions. I tested this theory by lowering
    > max_sectors_kb from 64 to 32:

    > These numbers all point to the ide sg completion code as the only thing
    > on the system generating latencies over ~42 usecs.

    yep, that's a fair assumption. Once the IO-APIC irq-redirection problems
    are solved i'll try to further thread the IDE completion IRQ to remove
    that ~100 usecs latency.

            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: Ingo Molnar: "Re: [patch] voluntary-preempt-2.6.8-rc2-M5"

    Relevant Pages

    • RE: Consistent lock up 2.6.10-rc1-bk7 (mutex/SCHED_RR bug?)
      ... Alan: ... The application in question is not "hard RT" and I am using SCHED_RR to improve latency, ... You are right to suspect futexes don't deal with hard real time but the ... send the line "unsubscribe linux-kernel" in ...
      (Linux-Kernel)
    • latency.c [was: Re: 2.6.9-rc1-mm1]
      ... [- snip -] ... > With these two things in hand, you should be able to measure the latency ... which is handy if you need to monitor processes that exit ... send the line "unsubscribe linux-kernel" in ...
      (Linux-Kernel)
    • Re: [PATCH] speed up SATA
      ... observation of device throughput and latency. ... latency" if he has gobs and gobs of memory and knows his configuration well. ... send the line "unsubscribe linux-kernel" in ...
      (Linux-Kernel)
    • Re: [patch] max-sectors-2.6.9-rc1-bk14-A0
      ... > solve the latency. ... > (I'll push the hardirq patches later on too but their acceptance should ... I'm quite fine with the patch. ... send the line "unsubscribe linux-kernel" in ...
      (Linux-Kernel)
    • Flames go here (was Re: Attempted summary of "RT patch acceptance" thread)
      ... >> Some of the comments from various folks are just intolerably paranoid ... > linux-RTOS has a guaranteed worst case latency of 50usec. ... send the line "unsubscribe linux-kernel" in ...
      (Linux-Kernel)