Hi, Michal Just FYI, this is another bisect result captured by 0day. This is actually the same as pervious "[mm, vmscan] 5e56dfbd83: fsmark.files_per_sec -11.1% regression", just commit sha change due to linux-next rebase. commit: 74e6b057d2406d596f60fbdd279d1a4e6b830aac ("mm, vmscan: consider eligible zones in get_scan_count") https://git.kernel.org/cgit/linux/kernel/git/next/linux-next.git master in testcase: fsmark on test machine: 48 threads Intel(R) Xeon(R) CPU E5-2697 v2 @ 2.70GHz with 64G memory with following parameters: iterations: 1 nr_threads: 64 disk: 3HDD md: RAID5 fs: btrfs filesize: 4M test_size: 130G sync_method: NoSync cpufreq_governor: performance test-description: The fsmark is a file system benchmark to test synchronous write workloads, for example, mail servers workload. test-url: https://sourceforge.net/projects/fsmark/ 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 testcase/path_params/tbox_group/run: fsmark/1-64-3HDD-RAID5-btrfs-4M-130G-NoSync-performance/ivb44 588efaf242c620c2 74e6b057d2406d596f60fbdd27 ---------------- -------------------------- %stddev change %stddev \ | \ 57.60 -11% 51.20 fsmark.files_per_sec 608 10% 668 fsmark.time.elapsed_time 608 10% 668 fsmark.time.elapsed_time.max 139 6% 148 fsmark.time.system_time 617366 631564 fsmark.time.voluntary_context_switches 14194 ± 14% -22% 11017 ± 5% fsmark.time.involuntary_context_switches 325773 ± 4% 10% 357414 ± 7% interrupts.CAL:Function_call_interrupts 207718 -9% 188656 vmstat.io.bo 93 -5% 89 turbostat.Avg_MHz 5.17 -6% 4.84 turbostat.RAMWatt 3.22 ± 4% -7% 3.00 turbostat.%Busy 1154955 10% 1267383 perf-stat.page-faults 1154939 10% 1267363 perf-stat.minor-faults 4098974 8% 4428264 perf-stat.context-switches 50.11 -16% 41.96 perf-stat.cache-miss-rate% 0.52 ± 9% -16% 0.44 ± 11% perf-stat.dTLB-load-miss-rate% 1.521e+09 ± 11% -16% 1.274e+09 ± 16% perf-stat.dTLB-load-misses 9.988e+09 -17% 8.291e+09 ± 6% perf-stat.cache-misses 4.616e+09 -18% 3.776e+09 ± 10% perf-stat.node-stores 207602 -9% 188553 iostat.md0.wkB/s 12.05 ± 5% 47% 17.75 ± 10% iostat.sda.rrqm/s 103778 -9% 94296 iostat.sda.wkB/s 25724 -9% 23312 iostat.sda.wrqm/s 931 -15% 789 iostat.sda.avgrq-sz 118.65 -37% 74.29 ± 7% iostat.sda.r_await 26.37 -53% 12.39 ± 6% iostat.sda.avgqu-sz 118.68 -54% 54.60 ± 6% iostat.sda.await 118.43 -54% 54.33 ± 6% iostat.sda.w_await 15.16 ± 6% 36% 20.58 ± 6% iostat.sdb.rrqm/s 103754 -9% 94275 iostat.sdb.wkB/s 25719 -9% 23308 iostat.sdb.wrqm/s 931 -15% 788 iostat.sdb.avgrq-sz 122.91 -38% 76.07 ± 4% iostat.sdb.r_await 27.43 -52% 13.08 ± 6% iostat.sdb.avgqu-sz 123.40 -53% 57.69 ± 7% iostat.sdb.await 123.18 -53% 57.43 ± 7% iostat.sdb.w_await 11.91 ± 4% 45% 17.24 ± 11% iostat.sdc.rrqm/s 228 18% 270 iostat.sdc.w/s 103780 -9% 94299 iostat.sdc.wkB/s 25723 -9% 23303 iostat.sdc.wrqm/s 905 -21% 716 iostat.sdc.avgrq-sz 115.29 -40% 69.74 ± 4% iostat.sdc.r_await 26.95 -51% 13.22 ± 6% iostat.sdc.avgqu-sz 118.48 -54% 54.83 ± 6% iostat.sdc.await 118.28 -54% 54.62 ± 6% iostat.sdc.w_await perf-stat.page-faults 1.28e+06 ++-O-------------------------------------------------------------+ O O O | 1.26e+06 ++ | | O | 1.24e+06 ++ | | | 1.22e+06 ++ | | | 1.2e+06 ++ | | | 1.18e+06 ++ | | .*.. .*.. .*. .*.*.. *.. | 1.16e+06 ++.*. .* *. .*..* .*.. .*. *..*.*. + .*.. .* *. *. *..* * * * * * | 1.14e+06 ++---------------------------------------------------------------+ perf-stat.minor-faults 1.28e+06 ++-O-------------------------------------------------------------+ O O O | 1.26e+06 ++ | | O | 1.24e+06 ++ | | | 1.22e+06 ++ | | | 1.2e+06 ++ | | | 1.18e+06 ++ | | .*.. .*.. .*. .*.*.. *.. | 1.16e+06 ++.*. .* *. .*..* .*.. .*. *..*.*. + .*.. .* *. *. *..* * * * * * | 1.14e+06 ++---------------------------------------------------------------+ perf-stat.cache-miss-rate_ 52 ++---------------------------------------------------------------------+ | .*.. * *.*..| 50 ++ .*. + + *.*.. *.. *.. *.. + | *..* *.. + + .. .. + .*..*.. + *.. + * | *.* * * *. + * .* * | 48 ++ * *. | | | 46 ++ | | | 44 ++ | | O | O | 42 ++ O O | | O | 40 ++---------------------------------------------------------------------+ fsmark.time.elapsed_time 680 ++--------------------------------------------------------------------+ O | 670 ++ O O O | 660 ++ O | | | 650 ++ | | | 640 ++ | | | 630 ++ | 620 ++ | | .*.. | 610 *+. .*.*.. .*..*.*..*.. .*.*..*..*.*. *.. .*.. .*.*..| | *.*..*. *. *.*..*. * *. * 600 ++--------------------------------------------------------------------+ fsmark.time.elapsed_time.max 680 ++--------------------------------------------------------------------+ O | 670 ++ O O O | 660 ++ O | | | 650 ++ | | | 640 ++ | | | 630 ++ | 620 ++ | | .*.. | 610 *+. .*.*.. .*..*.*..*.. .*.*..*..*.*. *.. .*.. .*.*..| | *.*..*. *. *.*..*. * *. * 600 ++--------------------------------------------------------------------+ fsmark.files_per_sec 58 ++---------------------------------------------------------------------+ *..*.*..*..*..*.*..*..*..*. .*..*..*.*..*..*.*..* *.*..*..*..*.*..* 57 ++ *. + + | | + + | 56 ++ * | | | 55 ++ | | | 54 ++ | | | 53 ++ | | | 52 ++ | | | 51 O+-O-O--O--O-----------------------------------------------------------+ [*] 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, Xiaolong