Re: [LKP] [btrfs] ca6dee6b79: fxmark.ssd_btrfs_MWRM_72_bufferedio.works/sec -8.4% regression

[Yujie Liu <yujie.liu@intel.com>]

Hi Filipe,

We noticed that this case was reported when the patch was in linux-next.
Thanks for your comment that it is an expected result due to heavy rename.

https://lore.kernel.org/all/Ysb4T7Z8hKgdvPRk@xsang-OptiPlex-9020/

This report is due to the patch being merged into mainline, if it is still
the same case, please ignore this duplicate report. Sorry for the inconvenience.

--
Thanks,
Yujie

On 8/30/2022 11:17, kernel test robot wrote:
> Greeting,
> 
> FYI, we noticed a -8.4% regression of fxmark.ssd_btrfs_MWRM_72_bufferedio.works/sec due to commit:
> 
> 
> commit: ca6dee6b7946794fa340a7290ca399a50b61705f ("btrfs: balance btree dirty pages and delayed items after a rename")
> https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git master
> 
> in testcase: fxmark
> on test machine: 128 threads 2 sockets Intel(R) Xeon(R) Platinum 8358 CPU @ 2.60GHz (Ice Lake) with 128G memory
> with following parameters:
> 
>      disk: 1SSD
>      media: ssd
>      test: MWRM
>      fstype: btrfs
>      directio: bufferedio
>      cpufreq_governor: performance
>      ucode: 0xd000363
> 
> test-description: FxMark is a filesystem benchmark that test multicore scalability.
> test-url: https://github.com/sslab-gatech/fxmark
> 
> 
> =========================================================================================
> compiler/cpufreq_governor/directio/disk/fstype/kconfig/media/rootfs/tbox_group/test/testcase/ucode:
>    gcc-11/performance/bufferedio/1SSD/btrfs/x86_64-rhel-8.3/ssd/debian-11.1-x86_64-20220510.cgz/lkp-icl-2sp5/MWRM/fxmark/0xd000363
> 
> commit:
>    b8bea09a45 ("btrfs: add trace event for submitted RAID56 bio")
>    ca6dee6b79 ("btrfs: balance btree dirty pages and delayed items after a rename")
> 
> b8bea09a456fc31a ca6dee6b7946794fa340a7290ca
> ---------------- ---------------------------
>           %stddev     %change         %stddev
>               \          |                \
>     1821853           -13.9%    1568247 ±  3%  fxmark.ssd_btrfs_MWRM_36_bufferedio.works
>       36436           -13.9%      31362 ±  3%  fxmark.ssd_btrfs_MWRM_36_bufferedio.works/sec
>     1675102           -14.0%    1439994 ±  7%  fxmark.ssd_btrfs_MWRM_54_bufferedio.works
>       33497           -14.0%      28796 ±  7%  fxmark.ssd_btrfs_MWRM_54_bufferedio.works/sec
>     1572332            -8.4%    1440600 ±  6%  fxmark.ssd_btrfs_MWRM_72_bufferedio.works
>       31445            -8.4%      28809 ±  6%  fxmark.ssd_btrfs_MWRM_72_bufferedio.works/sec
>      356010           +80.0%     640832        fxmark.time.involuntary_context_switches
>       68.50           -24.1%      52.00        fxmark.time.percent_of_cpu_this_job_got
>      630.47           -24.0%     479.23        fxmark.time.system_time
>   1.335e+10           +49.8%      2e+10        cpuidle..time
>        1045 ±  4%     +11.8%       1168        uptime.idle
>       31.54           +50.2%      47.37        iostat.cpu.idle
>       64.16           -24.7%      48.29        iostat.cpu.system
>       31.17           +50.3%      46.83        vmstat.cpu.id
>       12.83 ±  5%     -55.8%       5.67 ±  8%  vmstat.procs.r
>       32.13           +15.8       47.95        mpstat.cpu.all.idle%
>        0.47 ±  7%      +0.1        0.53 ±  3%  mpstat.cpu.all.iowait%
>       63.37           -16.1       47.31        mpstat.cpu.all.sys%
>       10.04 ±  3%     +13.5%      11.39 ±  3%  perf-stat.i.metric.K/sec
>      869.81 ± 10%     -16.2%     728.74 ± 15%  perf-stat.i.node-loads
>      871.23 ± 10%     -16.2%     730.49 ± 15%  perf-stat.ps.node-loads
>        3004 ±  8%     -52.1%       1440 ±  6%  numa-meminfo.node0.Active(anon)
>     1218568           -10.8%    1086453        numa-meminfo.node0.Inactive
>      351812 ±  3%     -29.0%     249640 ± 12%  numa-meminfo.node0.Inactive(anon)
>      120150           -79.3%      24861 ±  3%  numa-meminfo.node0.Shmem
>        3489 ±  8%     -45.0%       1919 ±  2%  meminfo.Active(anon)
>      492107           -19.0%     398809        meminfo.Committed_AS
>      382253           -24.6%     288151        meminfo.Inactive(anon)
>      124727           -76.8%      28886 ±  2%  meminfo.Shmem
>        2050 ±  4%     -10.5%       1834 ±  5%  meminfo.Writeback
>      750.83 ±  8%     -52.1%     360.00 ±  6%  numa-vmstat.node0.nr_active_anon
>       87951 ±  3%     -29.0%      62408 ± 12%  numa-vmstat.node0.nr_inactive_anon
>       30038           -79.3%       6216 ±  3%  numa-vmstat.node0.nr_shmem
>      750.83 ±  8%     -52.1%     360.00 ±  6%  numa-vmstat.node0.nr_zone_active_anon
>       87951 ±  3%     -29.0%      62408 ± 12%  numa-vmstat.node0.nr_zone_inactive_anon
>     7554028 ±  3%     -71.2%    2174126 ±  5%  sched_debug.cfs_rq:/.min_vruntime.avg
>     7640393 ±  3%     -70.5%    2254050 ±  5%  sched_debug.cfs_rq:/.min_vruntime.max
>     7291209 ±  3%     -73.6%    1926973 ±  5%  sched_debug.cfs_rq:/.min_vruntime.min
>      873.62 ±  7%     -19.2%     705.68 ± 10%  sched_debug.cfs_rq:/.runnable_avg.avg
>      790.32 ±  7%     -21.4%     621.34 ± 12%  sched_debug.cfs_rq:/.runnable_avg.min
>      747.11 ±  3%     -22.7%     577.37 ±  3%  sched_debug.cfs_rq:/.util_avg.avg
>      670.92 ±  5%     -25.2%     501.70 ±  2%  sched_debug.cfs_rq:/.util_avg.min
>      409.44 ±  9%     -35.1%     265.80 ± 21%  sched_debug.cfs_rq:/.util_est_enqueued.avg
>      789.44 ±  3%     -20.1%     630.53 ±  5%  sched_debug.cfs_rq:/.util_est_enqueued.max
>        0.00 ± 13%     -67.3%       0.00 ± 22%  sched_debug.cpu.next_balance.stddev
>      872.67 ±  8%     -45.0%     479.83 ±  2%  proc-vmstat.nr_active_anon
>     1801345            -1.7%    1771330        proc-vmstat.nr_file_pages
>       95550           -24.6%      72037        proc-vmstat.nr_inactive_anon
>        8752            -3.7%       8426        proc-vmstat.nr_mapped
>       31169           -76.8%       7221 ±  2%  proc-vmstat.nr_shmem
>      872.67 ±  8%     -45.0%     479.83 ±  2%  proc-vmstat.nr_zone_active_anon
>       95550           -24.6%      72037        proc-vmstat.nr_zone_inactive_anon
>        9553 ± 10%     -16.8%       7950 ±  3%  proc-vmstat.numa_hint_faults
>    18886391            -3.6%   18207624        proc-vmstat.numa_hit
>    18770999            -3.6%   18091363        proc-vmstat.numa_local
>     7398756            -4.0%    7105675        proc-vmstat.pgactivate
>    18885154            -3.6%   18206666        proc-vmstat.pgalloc_normal
>     7248262            -4.3%    6933915 ±  2%  proc-vmstat.pgdeactivate
>    18894473            -3.4%   18243898        proc-vmstat.pgfree
>     7829962            -3.0%    7596447 ±  2%  proc-vmstat.pgrotated
> 
> 
> If you fix the issue, kindly add following tag
> Reported-by: kernel test robot <yujie.liu@intel.com>
> 
> 
> To reproduce:
> 
>          git clone https://github.com/intel/lkp-tests.git
>          cd lkp-tests
>          sudo bin/lkp install job.yaml           # job file is attached in this email
>          bin/lkp split-job --compatible job.yaml # generate the yaml file for lkp run
>          sudo bin/lkp run generated-yaml-file
> 
>          # if come across any failure that blocks the test,
>          # please remove ~/.lkp and /lkp dir to run from a clean state.
> 
> 
> 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.
> 
> 
> #regzbot introduced: ca6dee6b79
> 
> 
> 
> _______________________________________________
> LKP mailing list -- lkp@lists.01.org
> To unsubscribe send an email to lkp-leave@lists.01.org