lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [day] [month] [year] [list]
Message-ID: <20161107024810.GD21529@yexl-desktop>
Date:   Mon, 7 Nov 2016 10:48:10 +0800
From:   kernel test robot <xiaolong.ye@...el.com>
To:     Huang Ying <ying.huang@...el.com>
Cc:     Linus Torvalds <torvalds@...ux-foundation.org>,
        Rik van Riel <riel@...hat.com>,
        Hugh Dickins <hughd@...gle.com>, Shaohua Li <shli@...nel.org>,
        Minchan Kim <minchan@...nel.org>,
        Mel Gorman <mgorman@...hsingularity.net>,
        Tejun Heo <tj@...nel.org>,
        Wu Fengguang <fengguang.wu@...el.com>,
        Dave Hansen <dave.hansen@...el.com>,
        Andrew Morton <akpm@...ux-foundation.org>,
        LKML <linux-kernel@...r.kernel.org>, lkp@...org
Subject: [lkp] [mm]  371a096edf:  vm-scalability.throughput 6.1% improvement


Greeting,

FYI, we noticed a 6.1% improvement of vm-scalability.throughput due to commit:


commit 371a096edf43a8c71844cf71c20765c8b21d07d9 ("mm: don't use radix tree writeback tags for pages in swap cache")
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git master

in testcase: vm-scalability
on test machine: 72 threads Intel(R) Xeon(R) CPU E5-2699 v3 @ 2.30GHz with 128G memory
with following parameters:

	runtime: 300
	thp_enabled: never
	thp_defrag: never
	nr_task: 16
	nr_ssd: 1
	test: swap-w-seq
	cpufreq_governor: performance

The motivation behind this suite is to exercise functions and regions of the mm/ of the Linux kernel which are of interest to us.

In addition to that, the commit also has significant impact on the following tests:

+------------------+-----------------------------------------------------------------------+
| testcase: change | vm-scalability: vm-scalability.throughput 14.1% improvement           |
| test machine     | 72 threads Intel(R) Xeon(R) CPU E5-2699 v3 @ 2.30GHz with 128G memory |
| test parameters  | cpufreq_governor=performance                                          |
|                  | nr_ssd=1                                                              |
|                  | nr_task=64                                                            |
|                  | runtime=300                                                           |
|                  | test=swap-w-seq                                                       |
|                  | thp_defrag=never                                                      |
|                  | thp_enabled=always                                                    |
+------------------+-----------------------------------------------------------------------+


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.

Details are as below:
-------------------------------------------------------------------------------------------------->


To reproduce:

        git clone git://git.kernel.org/pub/scm/linux/kernel/git/wfg/lkp-tests.git
        cd lkp-tests
        bin/lkp install job.yaml  # job file is attached in this email
        bin/lkp run     job.yaml

=========================================================================================
compiler/cpufreq_governor/kconfig/nr_ssd/nr_task/rootfs/runtime/tbox_group/test/testcase/thp_defrag/thp_enabled:
  gcc-6/performance/x86_64-rhel-7.2/1/16/debian-x86_64-2016-08-31.cgz/300/lkp-hsw-ep4/swap-w-seq/vm-scalability/never/never

commit: 
  1d8bf926f8 ("mm/bootmem.c: replace kzalloc() by kzalloc_node()")
  371a096edf ("mm: don't use radix tree writeback tags for pages in swap cache")

1d8bf926f8739bd3 371a096edf43a8c71844cf71c2 
---------------- -------------------------- 
       fail:runs  %reproduction    fail:runs
           |             |             |    
         %stddev     %change         %stddev
             \          |                \  
   2213672 ±  0%      +6.1%    2348422 ±  1%  vm-scalability.throughput
    534544 ±  1%      +6.8%     570794 ±  4%  vm-scalability.time.involuntary_context_switches
 1.458e+08 ±  0%      +7.0%   1.56e+08 ±  1%  vm-scalability.time.minor_page_faults
      4991 ± 50%   +3615.9%     185462 ± 31%  vm-scalability.time.voluntary_context_switches
      0.08 ± 27%   +3440.8%       2.69 ± 50%  turbostat.CPU%c3
    229747 ±  5%     +47.7%     339402 ± 11%  softirqs.RCU
    270638 ±  1%     +13.0%     305899 ±  2%  softirqs.SCHED
   1277270 ±  0%      +7.8%    1377440 ±  1%  vmstat.io.bo
   1277264 ±  0%      +7.8%    1377434 ±  1%  vmstat.swap.so
      5148 ±  0%     +50.8%       7765 ±  9%  vmstat.system.cs
    110151 ±  0%     +30.6%     143842 ± 11%  meminfo.SUnreclaim
      9703 ± 20%     -32.8%       6516 ± 28%  meminfo.Shmem
    783039 ±  5%     -26.3%     577083 ± 17%  meminfo.SwapCached
      1167 ± 27%  +27840.3%     326175 ± 61%  meminfo.Writeback
    637164 ±  5%     -25.4%     475615 ±  9%  numa-meminfo.node0.FilePages
    243.40 ± 36%  +19536.4%      47794 ± 58%  numa-meminfo.node0.Writeback
     50438 ±  4%     +31.7%      66437 ± 12%  numa-meminfo.node1.SUnreclaim
     83150 ±  4%     +17.5%      97667 ±  8%  numa-meminfo.node1.Slab
      1045 ± 24%  +25899.2%     271899 ± 60%  numa-meminfo.node1.Writeback
    159271 ±  5%     -25.4%     118746 ±  9%  numa-vmstat.node0.nr_file_pages
     59.60 ± 37%  +19680.4%      11789 ± 61%  numa-vmstat.node0.nr_writeback
     60.00 ± 37%  +19544.0%      11786 ± 61%  numa-vmstat.node0.nr_zone_write_pending
     12609 ±  4%     +31.5%      16579 ± 12%  numa-vmstat.node1.nr_slab_unreclaimable
    265.30 ± 19%  +25322.4%      67445 ± 61%  numa-vmstat.node1.nr_writeback
    266.80 ± 19%  +25179.8%      67446 ± 61%  numa-vmstat.node1.nr_zone_write_pending
   7424364 ± 50%    +715.3%   60527951 ± 47%  cpuidle.C1-HSW.time
     85779 ±  8%    +394.3%     423973 ± 36%  cpuidle.C1-HSW.usage
  11362590 ± 16%   +1931.5%  2.308e+08 ± 58%  cpuidle.C1E-HSW.time
     58721 ± 12%   +1054.1%     677675 ± 52%  cpuidle.C1E-HSW.usage
  60495231 ± 13%   +1096.4%  7.237e+08 ± 46%  cpuidle.C3-HSW.time
    181445 ± 13%    +599.7%    1269496 ± 43%  cpuidle.C3-HSW.usage
      2126 ±  9%    +798.9%      19117 ± 46%  cpuidle.POLL.usage
  40653369 ± 70%     -88.3%    4756647 ± 58%  proc-vmstat.compact_migrate_scanned
    370154 ±  3%     -13.7%     319306 ±  8%  proc-vmstat.nr_file_pages
     27536 ±  0%     +30.4%      35899 ± 11%  proc-vmstat.nr_slab_unreclaimable
  71520720 ±  5%     +12.4%   80357490 ±  3%  proc-vmstat.nr_vmscan_write
    292.40 ± 31%  +27493.5%      80683 ± 62%  proc-vmstat.nr_writeback
    294.00 ± 31%  +27342.9%      80682 ± 62%  proc-vmstat.nr_zone_write_pending
 1.262e+08 ±  3%     +15.9%  1.463e+08 ±  1%  proc-vmstat.numa_pte_updates
  87331985 ±  1%     +33.7%  1.168e+08 ±  3%  proc-vmstat.pgrotated
  28607602 ±  2%     +22.0%   34898577 ±  2%  proc-vmstat.pgscan_kswapd
  14766402 ±  4%     +20.4%   17782412 ±  4%  proc-vmstat.pgsteal_kswapd
   1828938 ±  1%     +52.5%    2790009 ±  9%  perf-stat.context-switches
     17676 ±  6%    +309.2%      72326 ± 29%  perf-stat.cpu-migrations
      0.06 ± 11%    +112.1%       0.13 ± 23%  perf-stat.dTLB-load-miss-rate%
 9.969e+08 ± 13%     +93.7%  1.931e+09 ± 18%  perf-stat.dTLB-load-misses
      0.22 ±  2%     +10.0%       0.24 ±  5%  perf-stat.dTLB-store-miss-rate%
  8.69e+08 ±  4%     +18.2%  1.027e+09 ±  9%  perf-stat.dTLB-store-misses
     36.04 ±  2%      +8.4%      39.05 ±  2%  perf-stat.iTLB-load-miss-rate%
 1.331e+08 ±  3%     +19.8%  1.594e+08 ±  6%  perf-stat.iTLB-load-misses
 1.463e+08 ±  0%      +7.0%  1.565e+08 ±  1%  perf-stat.minor-faults
     60.76 ±  1%      +2.8%      62.46 ±  0%  perf-stat.node-store-miss-rate%
 1.463e+08 ±  0%      +7.0%  1.565e+08 ±  1%  perf-stat.page-faults
    178.50 ± 14%    +201.1%     537.40 ± 14%  slabinfo.bdev_cache.active_objs
    178.50 ± 14%    +201.1%     537.40 ± 14%  slabinfo.bdev_cache.num_objs
    444.80 ± 12%     +66.6%     741.10 ± 13%  slabinfo.file_lock_cache.active_objs
    444.80 ± 12%     +66.6%     741.10 ± 13%  slabinfo.file_lock_cache.num_objs
      4019 ±  1%    +100.4%       8054 ± 13%  slabinfo.kmalloc-1024.active_objs
    128.10 ±  0%    +100.5%     256.90 ± 14%  slabinfo.kmalloc-1024.active_slabs
      4100 ±  0%    +100.8%       8232 ± 14%  slabinfo.kmalloc-1024.num_objs
    128.10 ±  0%    +100.5%     256.90 ± 14%  slabinfo.kmalloc-1024.num_slabs
      7578 ±  0%     +21.8%       9232 ±  7%  slabinfo.kmalloc-192.active_objs
      7609 ±  0%     +21.4%       9238 ±  7%  slabinfo.kmalloc-192.num_objs
      4829 ±  1%     +47.7%       7134 ± 11%  slabinfo.kmalloc-2048.active_objs
    312.10 ±  2%     +45.2%     453.10 ± 11%  slabinfo.kmalloc-2048.active_slabs
      4898 ±  1%     +47.9%       7246 ± 11%  slabinfo.kmalloc-2048.num_objs
    312.10 ±  2%     +45.2%     453.10 ± 11%  slabinfo.kmalloc-2048.num_slabs
     16028 ±  6%    +479.7%      92924 ± 51%  slabinfo.kmalloc-256.active_objs
    351.20 ±  4%   +1071.3%       4113 ± 56%  slabinfo.kmalloc-256.active_slabs
     16315 ±  5%    +487.8%      95897 ± 51%  slabinfo.kmalloc-256.num_objs
    351.20 ±  4%   +1071.3%       4113 ± 56%  slabinfo.kmalloc-256.num_slabs
    751.10 ±  6%     +74.9%       1313 ± 14%  slabinfo.nsproxy.active_objs
    751.10 ±  6%     +74.9%       1313 ± 14%  slabinfo.nsproxy.num_objs
     37639 ±  4%     -16.0%      31603 ±  6%  slabinfo.radix_tree_node.active_objs
    705.70 ±  4%     -17.4%     583.10 ±  7%  slabinfo.radix_tree_node.active_slabs
     39498 ±  3%     -17.4%      32627 ±  7%  slabinfo.radix_tree_node.num_objs
    705.70 ±  4%     -17.4%     583.10 ±  7%  slabinfo.radix_tree_node.num_slabs
    146695 ±  1%     -28.0%     105561 ± 17%  sched_debug.cfs_rq:/.exec_clock.max
     57.20 ± 73%    +574.1%     385.58 ± 26%  sched_debug.cfs_rq:/.exec_clock.min
     50781 ±  2%     -26.8%      37173 ± 18%  sched_debug.cfs_rq:/.exec_clock.stddev
    150.31 ±  5%     -25.0%     112.67 ± 17%  sched_debug.cfs_rq:/.load_avg.stddev
   2792970 ±  4%     -40.8%    1653864 ± 29%  sched_debug.cfs_rq:/.min_vruntime.max
    918436 ±  5%     -38.0%     569297 ± 31%  sched_debug.cfs_rq:/.min_vruntime.stddev
    842.53 ±  5%     -31.2%     579.83 ± 18%  sched_debug.cfs_rq:/.runnable_load_avg.max
    134.83 ±  5%     -34.0%      89.01 ± 24%  sched_debug.cfs_rq:/.runnable_load_avg.stddev
    918795 ±  5%     -37.9%     570175 ± 31%  sched_debug.cfs_rq:/.spread0.stddev
    380.58 ±  2%     -14.8%     324.26 ± 10%  sched_debug.cfs_rq:/.util_avg.stddev
     42.86 ± 17%   +1021.2%     480.60 ± 88%  sched_debug.cpu.clock.stddev
     42.86 ± 17%   +1021.2%     480.60 ± 88%  sched_debug.cpu.clock_task.stddev
    843.67 ±  5%     -31.3%     579.83 ± 18%  sched_debug.cpu.cpu_load[0].max
    134.98 ±  5%     -34.0%      89.02 ± 24%  sched_debug.cpu.cpu_load[0].stddev
    136.36 ±  5%     -30.6%      94.67 ± 26%  sched_debug.cpu.cpu_load[3].stddev
    136.89 ±  5%     -31.3%      94.00 ± 26%  sched_debug.cpu.cpu_load[4].stddev
    441.52 ±  5%     -16.4%     369.26 ± 13%  sched_debug.cpu.curr->pid.avg
      0.00 ± 10%    +846.3%       0.00 ± 86%  sched_debug.cpu.next_balance.stddev
    155148 ±  0%     -16.9%     128991 ± 11%  sched_debug.cpu.nr_load_updates.max
     38237 ±  2%     -27.9%      27553 ± 19%  sched_debug.cpu.nr_load_updates.stddev
     12989 ±  1%     +66.7%      21655 ± 11%  sched_debug.cpu.nr_switches.avg
     38611 ±  9%     +50.9%      58253 ± 13%  sched_debug.cpu.nr_switches.max
      1120 ± 32%    +350.3%       5042 ± 29%  sched_debug.cpu.nr_switches.min
     10487 ±  3%     +27.6%      13383 ±  9%  sched_debug.cpu.nr_switches.stddev
      0.01 ± 14%   +1005.7%       0.09 ± 38%  sched_debug.cpu.nr_uninterruptible.avg
     22.03 ± 19%    +124.5%      49.47 ± 28%  sched_debug.cpu.nr_uninterruptible.max
    -25.55 ±-32%    +173.8%     -69.95 ±-27%  sched_debug.cpu.nr_uninterruptible.min
      7.95 ± 16%    +144.6%      19.46 ± 18%  sched_debug.cpu.nr_uninterruptible.stddev
     13007 ±  2%     +66.2%      21615 ± 11%  sched_debug.cpu.sched_count.avg
    664.22 ± 53%    +590.3%       4584 ± 32%  sched_debug.cpu.sched_count.min
      1797 ±  3%    +189.0%       5196 ± 21%  sched_debug.cpu.sched_goidle.avg
    112.57 ± 48%   +1648.7%       1968 ± 35%  sched_debug.cpu.sched_goidle.min
      6018 ±  1%     +76.9%      10646 ± 12%  sched_debug.cpu.ttwu_count.avg
     17865 ±  3%    +107.9%      37145 ± 21%  sched_debug.cpu.ttwu_count.max
     93.22 ± 36%    +600.3%     652.75 ± 27%  sched_debug.cpu.ttwu_count.min
      5368 ±  2%     +69.2%       9082 ± 14%  sched_debug.cpu.ttwu_count.stddev
      4852 ±  1%     +36.6%       6626 ±  8%  sched_debug.cpu.ttwu_local.avg
     55.75 ± 24%    +310.6%     228.88 ± 16%  sched_debug.cpu.ttwu_local.min



***************************************************************************************************
lkp-hsw-ep4: 72 threads Intel(R) Xeon(R) CPU E5-2699 v3 @ 2.30GHz with 128G memory
=========================================================================================
compiler/cpufreq_governor/kconfig/nr_ssd/nr_task/rootfs/runtime/tbox_group/test/testcase/thp_defrag/thp_enabled:
  gcc-6/performance/x86_64-rhel-7.2/1/64/debian-x86_64-2016-08-31.cgz/300/lkp-hsw-ep4/swap-w-seq/vm-scalability/never/always

commit: 
  1d8bf926f8 ("mm/bootmem.c: replace kzalloc() by kzalloc_node()")
  371a096edf ("mm: don't use radix tree writeback tags for pages in swap cache")

1d8bf926f8739bd3 371a096edf43a8c71844cf71c2 
---------------- -------------------------- 
       fail:runs  %reproduction    fail:runs
           |             |             |    
         %stddev     %change         %stddev
             \          |                \  
   2125843 ±  0%     +14.1%    2426196 ±  1%  vm-scalability.throughput
    741119 ±  3%      +6.0%     785919 ±  2%  vm-scalability.time.involuntary_context_switches
 1.074e+08 ±  1%     +16.5%  1.251e+08 ±  1%  vm-scalability.time.minor_page_faults
      6243 ±  0%     -40.2%       3733 ± 10%  vm-scalability.time.percent_of_cpu_this_job_got
     21748 ±  0%     -41.0%      12831 ± 10%  vm-scalability.time.system_time
    315.07 ±  1%     -10.9%     280.79 ±  3%  vm-scalability.time.user_time
    333774 ±  2%     +12.3%     374973 ±  3%  softirqs.SCHED
  11885858 ±  0%     -39.4%    7205748 ±  9%  softirqs.TIMER
  40607746 ±  1%     +25.4%   50919220 ± 12%  numa-numastat.node0.numa_foreign
  40597356 ±  1%     +25.4%   50910950 ± 12%  numa-numastat.node1.numa_miss
      0.25 ±173%   +1200.0%       3.25 ± 70%  numa-numastat.node1.other_node
   1211829 ±  1%     +16.0%    1405484 ±  1%  vmstat.io.bo
   4939033 ±  3%     +16.1%    5733178 ±  6%  vmstat.memory.free
   2.1e+08 ±  1%     +14.9%  2.412e+08 ±  2%  vmstat.memory.swpd
      0.00 ±  0%      +Inf%      31.50 ± 17%  vmstat.procs.b
     67.00 ±  0%     -42.5%      38.50 ± 10%  vmstat.procs.r
   1211824 ±  1%     +16.0%    1405478 ±  1%  vmstat.swap.so
     92.51 ±  0%     -39.5%      56.01 ±  9%  turbostat.%Busy
      2583 ±  0%     -39.4%       1566 ±  9%  turbostat.Avg_MHz
      6.96 ±  2%    +406.3%      35.26 ± 12%  turbostat.CPU%c1
      0.01 ±  0%  +24025.0%       2.41 ± 19%  turbostat.CPU%c3
      0.52 ± 21%   +1125.7%       6.31 ± 14%  turbostat.CPU%c6
    258.28 ±  0%     -12.0%     227.24 ±  1%  turbostat.PkgWatt
     55.36 ±  0%      +1.1%      55.98 ±  0%  turbostat.RAMWatt
  28505239 ± 10%     -31.5%   19524763 ± 15%  meminfo.AnonHugePages
   5537897 ±  9%     -14.4%    4740114 ±  9%  meminfo.DirectMap2M
   5822789 ±  2%     +13.2%    6594234 ±  5%  meminfo.MemAvailable
   6222814 ±  2%     +12.5%    7001717 ±  4%  meminfo.MemFree
    709649 ±  0%     +12.6%     798807 ±  5%  meminfo.PageTables
    111455 ±  0%     +55.1%     172917 ±  6%  meminfo.SUnreclaim
    169503 ±  0%     +37.7%     233321 ±  4%  meminfo.Slab
   1504686 ±  3%     -24.8%    1131932 ± 15%  meminfo.SwapCached
      3028 ± 10%  +22884.4%     695969 ± 20%  meminfo.Writeback
  43194655 ± 11%    +619.4%  3.107e+08 ± 16%  cpuidle.C1-HSW.time
    622561 ± 17%    +104.2%    1271319 ± 17%  cpuidle.C1-HSW.usage
  14115248 ± 15%   +9297.9%  1.327e+09 ± 16%  cpuidle.C1E-HSW.time
     84418 ±  8%   +2439.5%    2143769 ± 16%  cpuidle.C1E-HSW.usage
  53083465 ± 15%   +2556.0%   1.41e+09 ± 21%  cpuidle.C3-HSW.time
    154076 ± 13%   +1115.8%    1873320 ± 20%  cpuidle.C3-HSW.usage
 1.803e+09 ±  3%    +350.2%  8.117e+09 ± 10%  cpuidle.C6-HSW.time
   1936956 ±  3%    +347.5%    8667200 ± 11%  cpuidle.C6-HSW.usage
  14995046 ± 10%   +2237.0%  3.504e+08 ± 14%  cpuidle.POLL.time
     13351 ± 17%    +202.3%      40357 ± 15%  cpuidle.POLL.usage
  22648511 ± 14%     -36.8%   14307991 ± 19%  numa-meminfo.node0.AnonHugePages
   1045808 ±  5%     -47.9%     544659 ± 15%  numa-meminfo.node0.FilePages
     58493 ±  4%     +57.0%      91837 ± 11%  numa-meminfo.node0.SUnreclaim
     85451 ±  4%     +40.6%     120107 ±  9%  numa-meminfo.node0.Slab
      1041 ± 12%  +15041.9%     157627 ± 13%  numa-meminfo.node0.Writeback
   5767065 ±  4%     -32.4%    3896120 ± 36%  numa-meminfo.node1.AnonHugePages
    293938 ±  5%     +50.9%     443468 ± 27%  numa-meminfo.node1.PageTables
     52953 ±  4%     +53.2%      81112 ± 14%  numa-meminfo.node1.SUnreclaim
     17882 ±  4%     -55.2%       8018 ± 59%  numa-meminfo.node1.Shmem
     84052 ±  5%     +34.8%     113281 ± 11%  numa-meminfo.node1.Slab
      2236 ±  8%  +23862.8%     535987 ± 24%  numa-meminfo.node1.Writeback
     11242 ± 13%     -38.0%       6974 ± 21%  numa-vmstat.node0.nr_anon_transparent_hugepages
    262464 ±  4%     -48.5%     135174 ± 17%  numa-vmstat.node0.nr_file_pages
   1215804 ±  3%     +13.8%    1383278 ±  4%  numa-vmstat.node0.nr_free_pages
      1287 ± 17%    +102.0%       2599 ± 24%  numa-vmstat.node0.nr_isolated_anon
     14625 ±  4%     +54.0%      22521 ± 12%  numa-vmstat.node0.nr_slab_unreclaimable
  10793807 ±  9%     +36.2%   14705195 ±  5%  numa-vmstat.node0.nr_vmscan_write
    233.00 ±  7%  +15561.6%      36491 ± 15%  numa-vmstat.node0.nr_writeback
  10793782 ±  9%     +35.9%   14668927 ±  5%  numa-vmstat.node0.nr_written
    233.00 ±  8%  +15561.5%      36491 ± 15%  numa-vmstat.node0.nr_zone_write_pending
  22566322 ±  3%     +31.3%   29625222 ± 12%  numa-vmstat.node0.numa_foreign
      2872 ±  4%     -27.1%       2093 ± 23%  numa-vmstat.node1.nr_anon_transparent_hugepages
    404362 ±  9%     +34.0%     541983 ± 19%  numa-vmstat.node1.nr_free_pages
    156.00 ± 11%     +95.4%     304.75 ± 15%  numa-vmstat.node1.nr_pages_scanned
     13235 ±  4%     +50.2%      19874 ± 14%  numa-vmstat.node1.nr_slab_unreclaimable
  49196401 ±  1%     +24.1%   61044573 ±  7%  numa-vmstat.node1.nr_vmscan_write
    572.50 ± 14%  +21710.0%     124862 ± 28%  numa-vmstat.node1.nr_writeback
  49196179 ±  1%     +23.8%   60919929 ±  7%  numa-vmstat.node1.nr_written
    574.25 ± 14%  +21643.7%     124863 ± 28%  numa-vmstat.node1.nr_zone_write_pending
  22550830 ±  3%     +31.3%   29618804 ± 12%  numa-vmstat.node1.numa_miss
 4.086e+12 ±  0%     -28.4%  2.928e+12 ±  5%  perf-stat.branch-instructions
      0.10 ±  0%     +52.0%       0.15 ±  6%  perf-stat.branch-miss-rate%
 4.119e+09 ±  0%      +8.5%  4.471e+09 ±  1%  perf-stat.branch-misses
     30.75 ±  0%     -13.1%      26.72 ±  3%  perf-stat.cache-miss-rate%
 1.046e+10 ±  1%      +9.9%  1.149e+10 ±  1%  perf-stat.cache-misses
   3.4e+10 ±  0%     +26.5%  4.303e+10 ±  2%  perf-stat.cache-references
 6.545e+13 ±  0%     -37.9%  4.065e+13 ±  7%  perf-stat.cpu-cycles
    195546 ±  4%     +32.7%     259393 ±  6%  perf-stat.cpu-migrations
      0.04 ± 12%     +84.7%       0.08 ± 10%  perf-stat.dTLB-load-miss-rate%
 4.077e+12 ±  0%     -33.5%  2.713e+12 ±  8%  perf-stat.dTLB-loads
      0.18 ±  4%     +16.4%       0.21 ±  2%  perf-stat.dTLB-store-miss-rate%
  6.92e+08 ±  3%     +17.8%  8.149e+08 ±  9%  perf-stat.dTLB-store-misses
     37.76 ±  2%     -41.7%      22.01 ±  5%  perf-stat.iTLB-load-miss-rate%
  42983189 ±  3%     +36.3%   58591746 ±  8%  perf-stat.iTLB-load-misses
  70857240 ±  2%    +194.4%  2.086e+08 ± 11%  perf-stat.iTLB-loads
 1.656e+13 ±  0%     -27.9%  1.194e+13 ±  5%  perf-stat.instructions
    385661 ±  3%     -46.6%     205896 ± 12%  perf-stat.instructions-per-iTLB-miss
      0.25 ±  0%     +16.3%       0.29 ±  2%  perf-stat.ipc
 1.079e+08 ±  1%     +16.4%  1.256e+08 ±  1%  perf-stat.minor-faults
     85.52 ±  0%      -4.7%      81.49 ±  0%  perf-stat.node-load-miss-rate%
 9.377e+08 ±  2%     +28.9%  1.208e+09 ±  4%  perf-stat.node-loads
     55.12 ±  0%      +2.7%      56.60 ±  1%  perf-stat.node-store-miss-rate%
 2.191e+09 ±  0%      +4.7%  2.294e+09 ±  1%  perf-stat.node-store-misses
 1.079e+08 ±  1%     +16.4%  1.256e+08 ±  1%  perf-stat.page-faults
    174.75 ± 11%    +206.4%     535.50 ± 38%  slabinfo.bdev_cache.active_objs
    174.75 ± 11%    +206.4%     535.50 ± 38%  slabinfo.bdev_cache.num_objs
      6222 ±  1%     +18.0%       7341 ±  5%  slabinfo.cred_jar.active_objs
      6222 ±  1%     +18.0%       7341 ±  5%  slabinfo.cred_jar.num_objs
    453.00 ± 11%     +65.0%     747.25 ± 16%  slabinfo.file_lock_cache.active_objs
    453.00 ± 11%     +65.0%     747.25 ± 16%  slabinfo.file_lock_cache.num_objs
      4046 ±  1%    +101.9%       8169 ±  3%  slabinfo.kmalloc-1024.active_objs
    126.50 ±  0%    +106.5%     261.25 ±  3%  slabinfo.kmalloc-1024.active_slabs
      4071 ±  0%    +105.6%       8372 ±  3%  slabinfo.kmalloc-1024.num_objs
    126.50 ±  0%    +106.5%     261.25 ±  3%  slabinfo.kmalloc-1024.num_slabs
      7564 ±  0%     +17.1%       8853 ±  1%  slabinfo.kmalloc-192.active_objs
      7615 ±  0%     +16.3%       8860 ±  1%  slabinfo.kmalloc-192.num_objs
      4766 ±  1%     +69.8%       8094 ±  8%  slabinfo.kmalloc-2048.active_objs
    307.00 ±  2%     +67.9%     515.50 ±  8%  slabinfo.kmalloc-2048.active_slabs
      4826 ±  1%     +70.5%       8227 ±  8%  slabinfo.kmalloc-2048.num_objs
    307.00 ±  2%     +67.9%     515.50 ±  8%  slabinfo.kmalloc-2048.num_slabs
     14592 ±  1%    +934.5%     150965 ± 34%  slabinfo.kmalloc-256.active_objs
    342.25 ±  1%   +1924.0%       6927 ± 34%  slabinfo.kmalloc-256.active_slabs
     14906 ±  1%    +934.8%     154254 ± 34%  slabinfo.kmalloc-256.num_objs
    342.25 ±  1%   +1924.0%       6927 ± 34%  slabinfo.kmalloc-256.num_slabs
     12576 ±  3%     +51.5%      19055 ±  7%  slabinfo.kmalloc-512.active_objs
    218.50 ± 15%     +56.3%     341.50 ± 10%  slabinfo.kmalloc-512.active_slabs
     12727 ±  3%     +54.0%      19595 ±  9%  slabinfo.kmalloc-512.num_objs
    218.50 ± 15%     +56.3%     341.50 ± 10%  slabinfo.kmalloc-512.num_slabs
    765.50 ±  4%     +69.1%       1294 ± 21%  slabinfo.nsproxy.active_objs
    765.50 ±  4%     +69.1%       1294 ± 21%  slabinfo.nsproxy.num_objs
      1612 ± 13%     -69.5%     492.25 ± 96%  proc-vmstat.compact_fail
      1619 ± 12%     -69.1%     501.25 ± 95%  proc-vmstat.compact_stall
     13831 ±  9%     -31.2%       9522 ± 16%  proc-vmstat.nr_anon_transparent_hugepages
    140989 ±  6%     +25.8%     177310 ±  4%  proc-vmstat.nr_dirty_background_threshold
    282323 ±  6%     +25.8%     355056 ±  4%  proc-vmstat.nr_dirty_threshold
    471239 ±  3%     -20.9%     372789 ± 11%  proc-vmstat.nr_file_pages
   1496965 ±  6%     +24.3%    1860851 ±  3%  proc-vmstat.nr_free_pages
      5139 ±  4%     +21.5%       6246 ±  2%  proc-vmstat.nr_isolated_anon
    178332 ±  1%     +13.1%     201722 ±  5%  proc-vmstat.nr_page_table_pages
    147.75 ±  8%    +100.8%     296.75 ± 18%  proc-vmstat.nr_pages_scanned
     27862 ±  0%     +52.7%      42558 ±  6%  proc-vmstat.nr_slab_unreclaimable
  60269851 ±  1%     +25.1%   75424523 ±  5%  proc-vmstat.nr_vmscan_write
    774.25 ±  2%  +21072.7%     163929 ± 22%  proc-vmstat.nr_writeback
 1.073e+08 ±  1%     +15.5%  1.239e+08 ±  2%  proc-vmstat.nr_written
    775.50 ±  2%  +21038.5%     163928 ± 22%  proc-vmstat.nr_zone_write_pending
  46256859 ±  1%     +26.7%   58617918 ±  8%  proc-vmstat.numa_foreign
     13246 ±  6%     -46.1%       7141 ± 17%  proc-vmstat.numa_hint_faults
      8187 ±  2%     -42.4%       4713 ± 20%  proc-vmstat.numa_hint_faults_local
  46256859 ±  1%     +26.7%   58617918 ±  8%  proc-vmstat.numa_miss
 1.394e+08 ±  0%     +13.0%  1.575e+08 ±  1%  proc-vmstat.pgalloc_normal
 1.084e+08 ±  1%     +15.4%  1.251e+08 ±  2%  proc-vmstat.pgdeactivate
 1.079e+08 ±  1%     +16.4%  1.257e+08 ±  1%  proc-vmstat.pgfault
 1.377e+08 ±  1%     +13.3%  1.559e+08 ±  1%  proc-vmstat.pgfree
      3711 ± 22%     -67.7%       1198 ± 32%  proc-vmstat.pgmigrate_fail
 4.293e+08 ±  1%     +15.5%  4.958e+08 ±  2%  proc-vmstat.pgpgout
  76556208 ±  1%     +21.6%   93113321 ±  2%  proc-vmstat.pgrefill
  61504865 ±  2%     +85.7%  1.142e+08 ±  5%  proc-vmstat.pgrotated
 1.754e+08 ±  1%     +26.7%  2.223e+08 ±  4%  proc-vmstat.pgscan_direct
   8178772 ±  3%     +61.6%   13217350 ± 18%  proc-vmstat.pgscan_kswapd
 1.007e+08 ±  1%     +12.8%  1.136e+08 ±  2%  proc-vmstat.pgsteal_direct
   6418690 ±  7%     +57.7%   10125267 ± 23%  proc-vmstat.pgsteal_kswapd
 1.073e+08 ±  1%     +15.5%  1.239e+08 ±  2%  proc-vmstat.pswpout
    209541 ±  1%     +16.5%     244206 ±  1%  proc-vmstat.thp_fault_fallback
    142799 ±  0%     -43.9%      80147 ±  5%  sched_debug.cfs_rq:/.exec_clock.avg
    131668 ±  1%     -69.0%      40852 ± 14%  sched_debug.cfs_rq:/.exec_clock.min
      5099 ± 13%    +520.8%      31657 ± 20%  sched_debug.cfs_rq:/.exec_clock.stddev
    227.88 ±  4%     +35.0%     307.70 ± 15%  sched_debug.cfs_rq:/.load_avg.min
    246.74 ± 26%     -53.7%     114.30 ± 23%  sched_debug.cfs_rq:/.load_avg.stddev
   9367426 ±  0%     -58.8%    3856380 ± 10%  sched_debug.cfs_rq:/.min_vruntime.avg
  10027888 ±  0%     -46.2%    5391954 ±  6%  sched_debug.cfs_rq:/.min_vruntime.max
   8536208 ±  2%     -70.5%    2518516 ± 18%  sched_debug.cfs_rq:/.min_vruntime.min
    308749 ±  8%    +188.8%     891811 ± 21%  sched_debug.cfs_rq:/.min_vruntime.stddev
      0.82 ±  2%     -31.3%       0.56 ± 15%  sched_debug.cfs_rq:/.nr_running.avg
      0.23 ± 13%     +43.2%       0.33 ±  8%  sched_debug.cfs_rq:/.nr_running.stddev
     36.70 ± 13%     -67.1%      12.06 ± 20%  sched_debug.cfs_rq:/.nr_spread_over.avg
    123.42 ± 19%     -68.2%      39.20 ± 35%  sched_debug.cfs_rq:/.nr_spread_over.max
      8.92 ± 14%     -64.9%       3.13 ± 52%  sched_debug.cfs_rq:/.nr_spread_over.min
     23.11 ± 20%     -74.8%       5.83 ± 27%  sched_debug.cfs_rq:/.nr_spread_over.stddev
     32.16 ±  6%     -37.6%      20.08 ± 33%  sched_debug.cfs_rq:/.runnable_load_avg.avg
    653879 ± 47%    +135.8%    1541932 ± 15%  sched_debug.cfs_rq:/.spread0.max
    315599 ±  9%    +184.2%     896957 ± 21%  sched_debug.cfs_rq:/.spread0.stddev
    898.96 ±  1%     -27.4%     652.99 ± 15%  sched_debug.cfs_rq:/.util_avg.avg
    673946 ±  3%     +24.3%     837742 ±  4%  sched_debug.cpu.avg_idle.avg
     49038 ±  8%    +259.6%     176360 ± 60%  sched_debug.cpu.avg_idle.min
    301898 ±  6%     -25.5%     225046 ± 21%  sched_debug.cpu.avg_idle.stddev
    190551 ±  0%      +9.0%     207609 ±  2%  sched_debug.cpu.clock.max
    233.98 ±  3%   +2164.8%       5299 ± 53%  sched_debug.cpu.clock.stddev
    190551 ±  0%      +9.0%     207609 ±  2%  sched_debug.cpu.clock_task.max
    233.98 ±  3%   +2164.8%       5299 ± 53%  sched_debug.cpu.clock_task.stddev
     32.11 ±  6%     -37.8%      19.98 ± 33%  sched_debug.cpu.cpu_load[0].avg
     35.42 ±  9%     -37.8%      22.02 ± 26%  sched_debug.cpu.cpu_load[1].avg
    133.36 ± 15%     -39.1%      81.19 ± 37%  sched_debug.cpu.cpu_load[1].stddev
     34.65 ±  7%     -38.1%      21.45 ± 27%  sched_debug.cpu.cpu_load[2].avg
    126.41 ±  9%     -38.0%      78.32 ± 38%  sched_debug.cpu.cpu_load[2].stddev
     34.35 ±  6%     -38.4%      21.15 ± 28%  sched_debug.cpu.cpu_load[3].avg
    123.58 ±  6%     -38.4%      76.07 ± 38%  sched_debug.cpu.cpu_load[3].stddev
     34.45 ±  5%     -38.7%      21.11 ± 28%  sched_debug.cpu.cpu_load[4].avg
    123.29 ±  4%     -38.7%      75.54 ± 38%  sched_debug.cpu.cpu_load[4].stddev
      1367 ±  2%     -31.7%     934.09 ± 15%  sched_debug.cpu.curr->pid.avg
      0.00 ±  2%   +2051.8%       0.01 ± 53%  sched_debug.cpu.next_balance.stddev
    148590 ±  1%     -25.6%     110559 ±  3%  sched_debug.cpu.nr_load_updates.avg
    139700 ±  1%     -43.5%      78865 ±  5%  sched_debug.cpu.nr_load_updates.min
      3733 ± 13%    +519.4%      23123 ± 21%  sched_debug.cpu.nr_load_updates.stddev
      0.84 ±  2%     -31.9%       0.57 ± 17%  sched_debug.cpu.nr_running.avg
      0.28 ± 10%     +28.5%       0.36 ± 10%  sched_debug.cpu.nr_running.stddev
     14395 ±  9%     +34.0%      19289 ± 23%  sched_debug.cpu.nr_switches.min
      0.11 ± 43%    +470.4%       0.62 ± 52%  sched_debug.cpu.nr_uninterruptible.avg
      7400 ± 21%     +31.3%       9718 ± 15%  sched_debug.cpu.sched_goidle.avg
      1050 ±  7%    +245.9%       3632 ± 38%  sched_debug.cpu.sched_goidle.min
     40897 ± 23%     +43.3%      58610 ±  7%  sched_debug.cpu.ttwu_count.max
     10300 ± 11%     -51.3%       5018 ± 29%  sched_debug.cpu.ttwu_count.min
      5838 ± 26%     +99.6%      11652 ± 12%  sched_debug.cpu.ttwu_count.stddev
      7617 ±  7%     +18.5%       9024 ±  5%  sched_debug.cpu.ttwu_local.avg
      4191 ± 25%     +36.8%       5733 ±  9%  sched_debug.cpu.ttwu_local.stddev



Thanks,
Xiaolong

View attachment "config-4.8.0-10066-g371a096" of type "text/plain" (153641 bytes)

View attachment "job-script" of type "text/plain" (7082 bytes)

View attachment "job.yaml" of type "text/plain" (4740 bytes)

View attachment "reproduce" of type "text/plain" (575 bytes)

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ