linaro-pm-wg team mailing list archive

[Daniel Lezcano <1179414@xxxxxxxxxxxxxxxxxx>]

** Changed in: linaro-power-idlestat
       Status: New => Fix Released

-- 
You received this bug notification because you are a member of Linaro
Power Management Working Group, which is subscribed to Linaro Power
Idlestat.
Matching subscriptions: PMWG-idlestat
https://bugs.launchpad.net/bugs/1179414

Title:
  Wrong statistics results for some cpus

Status in Idlestat - optimise CPU idling:
  Fix Released

Bug description:
  I have just tested the last version of idlestat
  (f0b53ce34662525e5fc2022fbb79c97d34446c71)and i'm a bit surprised by
  some results.

  I have tested it on 3 traces and the results for mp3-2.dat file looks wrong. A quick look at the trace with kernelshark
  will show you why i have reached such conclusion.

  I have put below the results of idlestat for mp3-2.dat

  $ ./idlestat trace-cpuidle-2.dat
  Log is 10.015449 secs long with 51456 events
  cpu0/state0, 12 hits, total 7942.00us, avg 661.83us, min 45.00us, max 1583.00us
  cpu0/state1, 2265 hits, total 8194621.00us, avg 3617.93us, min 354.00us, max 4390.00us
  cpu1/state0, 0 hits, total 0.00us, avg 0.00us, min 0.00us, max 0.00us
  cpu1/state1, 161 hits, total 9893397.00us, avg 61449.67us, min 6686.00us, max 201103.00us
  cpu2/state0, 0 hits, total 0.00us, avg 0.00us, min 0.00us, max 0.00us
  cpu2/state1, 2 hits, total 10013913.00us, avg 5006956.50us, min 999999.00us, max 8736870.00us
  cpu3/state0, 0 hits, total 0.00us, avg 0.00us, min 0.00us, max 0.00us
  cpu3/state1, 8 hits, total 500972.00us, avg 62621.50us, min 6556.00us, max 434479.00us
  cpu4/state0, 0 hits, total 0.00us, avg 0.00us, min 0.00us, max 0.00us
  cpu4/state1, 16 hits, total 7486588.00us, avg 467911.75us, min 2182.00us, max 2237485.00us
  cluster/state0, 0 hits, total 0.00us, avg 0.00us, min 0.00us, max 0.00us
  cluster/state1, 127 hits, total 403596.00us, avg 3177.92us, min 74.00us, max 4337.00us

  The cpu3, cpu4 and cluster statistics look particularly suspicious

  I have put my results below which look more aligned with the
  kernelshark output

  system not busy 8196776us, 7942 switch,
  system in C1 state 8188834us, 2359 switch, average 3471µs

  I have attached the trace file mp3-2.dat

To manage notifications about this bug go to:
https://bugs.launchpad.net/linaro-power-idlestat/+bug/1179414/+subscriptions