Re: 2.6.5-mjb1

From: Marcus Hartig (m.f.h_at_web.de)
Date: 04/10/04

  • Next message: Chris Friesen: "want to clarify powerpc assembly conventions in head.S and entry.S" 
    Date:	Sat, 10 Apr 2004 09:19:31 +0200
To: linux-kernel@vger.kernel.org

    Hi,

    tested with latencytest-0.5.2. Sched stats turned off in the kernel.
    Look at "2.6.5 latencytest +aa5 +mm3 as/cfq scheduler" to compare.

    ---------- 2.6.5-mjb1 AS scheduler ----------
    x11.log:
    cpu_hz = 2.08792e+09
    cpu_load=0.800000 loops per run = 537513
    total latency = 1.953125 ms
    cpu latency = 1.562500 ms
    max diff = 3.03868 ms (overrun 0)
    within 1ms = 33088, factor = 99.9788%
    within 2ms = 33095, factor = 100%
    deviation = 0.0196589

    proc.log:
    cpu_hz = 2.08792e+09
    cpu_load=0.800000 loops per run = 537547
    total latency = 1.953125 ms
    cpu latency = 1.562500 ms
    max diff = 3.05196 ms (overrun 0)
    within 1ms = 15385, factor = 99.9805%
    within 2ms = 15388, factor = 100%
    deviation = 0.0205876

    diskwrite.log:
    cpu_hz = 2.08792e+09
    cpu_load=0.800000 loops per run = 537500
    total latency = 1.953125 ms
    cpu latency = 1.562500 ms
    max diff = 4.40418 ms (overrun 1)
    within 1ms = 10989, factor = 99.9273%
    within 2ms = 10996, factor = 99.9909%
    deviation = 0.0545299

    diskcopy.log:
    cpu_hz = 2.08792e+09
    cpu_load=0.800000 loops per run = 537552
    total latency = 1.953125 ms
    cpu latency = 1.562500 ms
    max diff = 5.08319 ms (overrun 3)
    within 1ms = 27949, factor = 99.9321%
    within 2ms = 27965, factor = 99.9893%
    deviation = 0.0385951

    diskread.log:
    cpu_hz = 2.08792e+09
    cpu_load=0.800000 loops per run = 537564
    total latency = 1.953125 ms
    cpu latency = 1.562500 ms
    max diff = 3.15495 ms (overrun 0)
    within 1ms = 28029, factor = 99.9715%
    within 2ms = 28037, factor = 100%
    deviation = 0.0249104

    Best regards,

    Marcus

    -
    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: Chris Friesen: "want to clarify powerpc assembly conventions in head.S and entry.S"

    Relevant Pages

    • 2.6.5 latencytest +aa5 +mm3 as/cfq scheduler
      ... see which scheduler/kernel has a good or low latency. ... total latency = 1.953125 ms ... cpu latency = 1.562500 ms ... deviation = 0.021617 ...
      (Linux-Kernel)
    • [PATCH] x86: enable preemption in delay
      ... The RT team has been searching for a nasty latency. ... I also noticed that this always seemed to happen during a __delay call. ... static void delay_tsc(unsigned long loops) ... unsigned bclock, now; ...
      (Linux-Kernel)
    • [PATCH -v4] x86: enable preemption in delay
      ... The RT team has been searching for a nasty latency. ... I also noticed that this always seemed to happen during a __delay call. ... static void delay_tsc(unsigned long loops) ... unsigned bclock, now; ...
      (Linux-Kernel)
    • Re: XP Pro file sharing (SMB) performance problem
      ... I've narrowed down the problem to amazingly bad latency in responding ... to network requests: I've got ping times from the server with mean of ... Other machines on my network (including ... have a mean delay of 3ms with less than 0.3ms deviation. ...
      (microsoft.public.windowsxp.perform_maintain)
    • Re: 2.6.5-mjb1
      ... Now, in the second try it has the loops running as the others, I ... total latency = 1.953125 ms ... cpu latency = 1.562500 ms ... deviation = 0.0216797 ...
      (Linux-Kernel)