1. 12 3月, 2010 1 次提交
    • M
      sched: Rate-limit nohz · 39c0cbe2
      Mike Galbraith 提交于
      Entering nohz code on every micro-idle is costing ~10% throughput for netperf
      TCP_RR when scheduling cross-cpu.  Rate limiting entry fixes this, but raises
      ticks a bit.  On my Q6600, an idle box goes from ~85 interrupts/sec to 128.
      
      The higher the context switch rate, the more nohz entry costs.  With this patch
      and some cycle recovery patches in my tree, max cross cpu context switch rate is
      improved by ~16%, a large portion of which of which is this ratelimiting.
      Signed-off-by: NMike Galbraith <efault@gmx.de>
      Signed-off-by: NPeter Zijlstra <a.p.zijlstra@chello.nl>
      LKML-Reference: <1268301003.6785.28.camel@marge.simson.net>
      Signed-off-by: NIngo Molnar <mingo@elte.hu>
      39c0cbe2
  2. 11 3月, 2010 3 次提交
  3. 02 3月, 2010 36 次提交