[sched] 517b8b57fb5: -31.6% cpuidle.C3-NHM.usage

[sched] 517b8b57fb5: -31.6% cpuidle.C3-NHM.usage

[Aaron Lu]

[-- Attachment #1: Type: text/plain, Size: 3052 bytes --]

FYI, we noticed the below changes on

git://bee.sh.intel.com/git/ydu19/linux cc-v4.7-on-3.16-rc3
commit 517b8b57fb5e6ceb175439474d4e539dfeb74bfe ("sched: Remove update_rq_runnable_avg")

test case: xps2/pigz/100%-128K

      v3.16-rc3  517b8b57fb5e6ceb175439474  
---------------  -------------------------  
       776 ~22%     -31.6%        531 ~16%  TOTAL cpuidle.C3-NHM.usage
      0.26 ~ 5%     -25.5%       0.19 ~ 5%  TOTAL turbostat.%c6
   6904073 ~ 2%     -23.4%    5290138 ~ 2%  TOTAL cpuidle.C6-NHM.time
      7762 ~ 0%     +22.1%       9481 ~ 1%  TOTAL proc-vmstat.nr_active_anon
     31071 ~ 0%     +22.1%      37939 ~ 1%  TOTAL meminfo.Active(anon)
      7847 ~ 0%     +21.9%       9563 ~ 1%  TOTAL proc-vmstat.nr_anon_pages
     31404 ~ 0%     +21.9%      38271 ~ 1%  TOTAL meminfo.AnonPages
      2369 ~ 1%     -15.5%       2002 ~ 1%  TOTAL proc-vmstat.pgactivate
     10420 ~ 2%     +16.1%      12102 ~ 0%  TOTAL time.maximum_resident_set_size
       970 ~ 3%     +12.6%       1092 ~ 0%  TOTAL slabinfo.kmalloc-96.num_objs
       970 ~ 3%     +12.6%       1092 ~ 0%  TOTAL slabinfo.kmalloc-96.active_objs
      4.56 ~ 4%     +13.6%       5.18 ~ 8%  TOTAL perf-profile.cpu-cycles.copy_user_generic_string.copy_page_from_iter.pipe_write.new_sync_write.vfs_write
    108172 ~ 0%     +10.4%     119392 ~ 0%  TOTAL meminfo.Active
     12905 ~ 0%     +11.6%      14408 ~ 0%  TOTAL time.minor_page_faults

Legend:
	~XX%    - stddev percent
	[+-]XX% - change percent


                               time.minor_page_faults

  15000 ++------------------------------------------------------------------+
        O    O               O                                              |
        |         O     O                                                   |
  14500 ++                        O                    O          O         O
        |                              O     O    O         O          O    |
        |                                                                   |
  14000 ++                                                                  |
        |                                                                   |
  13500 ++                                                                  |
        |                                                                   |
        |                                                                   |
  13000 ++                           ..*.....*....                          |
        *....*....*.....*....*....*..             *....*                    |
        |                                                                   |
  12500 ++------------------------------------------------------------------+


	[*] bisect-good sample
	[O] bisect-bad  sample


Disclaimer:
Results have been estimated based on internal Intel analysis and are provided
for informational purposes only. Any difference in system hardware or software
design or configuration may affect actual performance.

Thanks,
Aaron

[-- Attachment #2: reproduce.ksh --]
[-- Type: text/plain, Size: 585 bytes --]

echo performance > /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor
echo performance > /sys/devices/system/cpu/cpu1/cpufreq/scaling_governor
echo performance > /sys/devices/system/cpu/cpu2/cpufreq/scaling_governor
echo performance > /sys/devices/system/cpu/cpu3/cpufreq/scaling_governor
echo performance > /sys/devices/system/cpu/cpu4/cpufreq/scaling_governor
echo performance > /sys/devices/system/cpu/cpu5/cpufreq/scaling_governor
echo performance > /sys/devices/system/cpu/cpu6/cpufreq/scaling_governor
echo performance > /sys/devices/system/cpu/cpu7/cpufreq/scaling_governor