Re: 2.6.1 and irq balancing

From: Ethan Weinstein (lists_at_stinkfoot.org)
Date: 01/13/04

  • Next message: Martin J. Bligh: "[Bug 1855] New: de4x5 does not work with SMP" 
    Date:	Tue, 13 Jan 2004 01:50:17 -0500
To: "Nakajima, Jun" <jun.nakajima@intel.com>

    Nakajima, Jun wrote:

    >> Admittedly, the machine's load was not high when I took this sample.
    >> However, creating a great deal of load does not change these statistics
    >> at all. Being that there are patches available for 2.4.x kernels to
    >> fix this, I don't think this at all by design, but what do I know? =)
    >>

    > 2.6 kernels don't need a patch to it as far as I understand. Are you
    > saying that with significant amount of load, you did not see any
    > distribution of interrupts? Today's threshold in the kernel is high
    > because we found moving around interrupts frequently rather hurt the
    > cache and thus lower the performance compared to "do nothing". Can you
    > try to create significant load with your network (eth0 and eh1) and see
    > what happens?
    >
    > Jun

    Here's the situation two days later, I created some brief periods of
    high load on eth1 and I see we have some change:

                CPU0 CPU1 CPU2 CPU3
       0: 184932542 0 2592511 0 IO-APIC-edge timer
       1: 1875 0 0 0 IO-APIC-edge i8042
       2: 0 0 0 0 XT-PIC cascade
       3: 3046103 0 0 0 IO-APIC-edge serial
       8: 2 0 0 0 IO-APIC-edge rtc
       9: 0 0 0 0 IO-APIC-level acpi
      14: 76 0 0 0 IO-APIC-edge ide0
      16: 2978264 0 0 0 IO-APIC-level sym53c8xx
      22: 7838940 0 0 0 IO-APIC-level eth0
      48: 916078 0 125150 0 IO-APIC-level aic79xx
      49: 1099375 0 0 0 IO-APIC-level aic79xx
      54: 51484241 316 50560879 279 IO-APIC-level eth1
    NMI: 0 0 0 0
    LOC: 187530735 187530988 187530981 187530986
    ERR: 0
    MIS: 0

    My argument is (see below). This is an old 2x pentium2 @400, also
    running 2.6, an old Compaq Proliant to be exact. This machine obviously
    has no HT, so why the balanced load?

                CPU0 CPU1
       0: 1066522197 1117196193 IO-APIC-edge timer
       1: 42 19 IO-APIC-edge i8042
       2: 0 0 XT-PIC cascade
       5: 23523428 23510845 IO-APIC-level TLAN
       8: 0 4 IO-APIC-edge rtc
       9: 15 15 IO-APIC-level sym53c8xx
      10: 6874323 6809042 IO-APIC-level sym53c8xx
      11: 7545802 7509034 IO-APIC-level ida0
      14: 8 2 IO-APIC-edge ide0
    NMI: 0 0
    LOC: 2183867261 2183867237
    ERR: 0
    MIS: 0

    Ethan
    -
    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: Martin J. Bligh: "[Bug 1855] New: de4x5 does not work with SMP"

    Relevant Pages

    • Re: 2.6.1 and irq balancing
      ... >> 2.6 kernels don't need a patch to it as far as I understand. ... >> saying that with significant amount of load, ... >> because we found moving around interrupts frequently rather hurt the ... send the line "unsubscribe linux-kernel" in ...
      (Linux-Kernel)
    • Understanding I/O behaviour
      ... load goes up to 12 and responsiveness goes down. ... echo "Finish 3 dd processes: ... The "felt" responsiveness was best with the last two kernels, ...
      (Linux-Kernel)
    • Re: [RFC] Kernel version numbering scheme change
      ... Upstream kernels can be a bit iffy especially on 32bit boxes with lots ... At minimum with 2.6.x under high load you also need Arjan van de Ven's ... amounts of I/O make sure your I/O devices are 64bit capable - particular ...
      (Linux-Kernel)
    • 2.6.8.1-mm3 and above hang with iptables and sound
      ... When initscripts reach the load of stored iptables rules, the box hangs. ... Previuos kernels work fine. ...
      (Linux-Kernel)
    • unknown symbols loading modules under 2.6.x
      ... They load just fine under the 2.4.x kernels. ... The loadable module these symbols are ... send the line "unsubscribe linux-kernel" in ...
      (Linux-Kernel)