Greeting, FYI, we noticed a 130.6% improvement of fxmark.ssd_f2fs_DWTL_72_directio.works/sec due to commit: commit: cfd66bb715fd11fde3338d0660cffa1396adc27d ("f2fs: fix deadloop in foreground GC") 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: DWTL fstype: f2fs directio: directio cpufreq_governor: performance test-description: FxMark is a filesystem benchmark that test multicore scalability. test-url: https://github.com/sslab-gatech/fxmark Details are as below: ========================================================================================= compiler/cpufreq_governor/directio/disk/fstype/kconfig/media/rootfs/tbox_group/test/testcase: gcc-11/performance/directio/1SSD/f2fs/x86_64-rhel-8.3/ssd/debian-11.1-x86_64-20220510.cgz/lkp-icl-2sp5/DWTL/fxmark commit: 25f8236213 ("f2fs: fix to do sanity check on block address in f2fs_do_zero_range()") cfd66bb715 ("f2fs: fix deadloop in foreground GC") 25f8236213a91efd cfd66bb715fd11fde3338d0660c ---------------- --------------------------- %stddev %change %stddev \ | \ 0.63 ± 2% -27.9% 0.45 ± 3% fxmark.ssd_f2fs_DWTL_18_directio.idle_sec 0.01 ± 31% -100.0% 0.00 fxmark.ssd_f2fs_DWTL_18_directio.iowait_sec 1.32 ± 31% -100.0% 0.00 fxmark.ssd_f2fs_DWTL_18_directio.iowait_util 0.05 -21.9% 0.04 ± 2% fxmark.ssd_f2fs_DWTL_18_directio.real_sec 0.02 -61.1% 0.01 fxmark.ssd_f2fs_DWTL_18_directio.secs 14.50 ± 10% +24.7% 18.08 ± 6% fxmark.ssd_f2fs_DWTL_18_directio.sys_util 12.06 ± 5% +32.0% 15.91 ± 2% fxmark.ssd_f2fs_DWTL_18_directio.user_util 595640 +157.3% 1532350 fxmark.ssd_f2fs_DWTL_18_directio.works/sec 0.05 ± 10% -100.0% 0.00 fxmark.ssd_f2fs_DWTL_1_directio.iowait_sec 38.86 ± 8% -100.0% 0.00 fxmark.ssd_f2fs_DWTL_1_directio.iowait_util 0.12 -44.7% 0.07 ± 5% fxmark.ssd_f2fs_DWTL_1_directio.real_sec 0.09 -61.1% 0.04 ± 7% fxmark.ssd_f2fs_DWTL_1_directio.secs 38.99 ± 11% +56.3% 60.95 ± 7% fxmark.ssd_f2fs_DWTL_1_directio.sys_util 22.15 ± 15% +57.2% 34.82 ± 12% fxmark.ssd_f2fs_DWTL_1_directio.user_util 110377 +158.9% 285786 ± 7% fxmark.ssd_f2fs_DWTL_1_directio.works/sec 0.01 ± 31% -100.0% 0.00 fxmark.ssd_f2fs_DWTL_2_directio.iowait_sec 9.63 ± 26% -100.0% 0.00 fxmark.ssd_f2fs_DWTL_2_directio.iowait_util 0.06 ± 3% -19.5% 0.05 ± 2% fxmark.ssd_f2fs_DWTL_2_directio.real_sec 0.03 ± 7% -41.1% 0.02 fxmark.ssd_f2fs_DWTL_2_directio.secs 23.54 ± 10% +30.2% 30.66 ± 6% fxmark.ssd_f2fs_DWTL_2_directio.user_util 357310 ± 6% +69.1% 604308 fxmark.ssd_f2fs_DWTL_2_directio.works/sec 1.26 ± 9% -26.1% 0.94 ± 2% fxmark.ssd_f2fs_DWTL_36_directio.idle_sec 0.02 ± 40% -100.0% 0.00 fxmark.ssd_f2fs_DWTL_36_directio.iowait_sec 1.13 ± 43% -100.0% 0.00 fxmark.ssd_f2fs_DWTL_36_directio.iowait_util 0.05 ± 7% -18.6% 0.04 fxmark.ssd_f2fs_DWTL_36_directio.real_sec 0.02 ± 13% -53.9% 0.01 ± 7% fxmark.ssd_f2fs_DWTL_36_directio.secs 15.56 ± 11% +39.5% 21.71 ± 6% fxmark.ssd_f2fs_DWTL_36_directio.sys_util 11.19 ± 7% +17.1% 13.11 ± 7% fxmark.ssd_f2fs_DWTL_36_directio.user_util 566930 ± 17% +112.6% 1205463 ± 6% fxmark.ssd_f2fs_DWTL_36_directio.works/sec 0.10 ± 5% -21.7% 0.08 ± 4% fxmark.ssd_f2fs_DWTL_4_directio.idle_sec 0.01 -100.0% 0.00 fxmark.ssd_f2fs_DWTL_4_directio.iowait_sec 5.27 ± 3% -100.0% 0.00 fxmark.ssd_f2fs_DWTL_4_directio.iowait_util 0.05 -22.3% 0.04 ± 3% fxmark.ssd_f2fs_DWTL_4_directio.real_sec 0.02 ± 2% -56.9% 0.01 ± 3% fxmark.ssd_f2fs_DWTL_4_directio.secs 582453 ± 2% +132.3% 1353108 ± 3% fxmark.ssd_f2fs_DWTL_4_directio.works/sec 1.88 ± 6% -25.2% 1.40 fxmark.ssd_f2fs_DWTL_54_directio.idle_sec 69.13 ± 4% -9.5% 62.59 fxmark.ssd_f2fs_DWTL_54_directio.idle_util 0.05 ± 2% -17.4% 0.04 fxmark.ssd_f2fs_DWTL_54_directio.real_sec 0.02 ± 11% -51.6% 0.01 ± 6% fxmark.ssd_f2fs_DWTL_54_directio.secs 19.12 ± 18% +27.6% 24.39 ± 3% fxmark.ssd_f2fs_DWTL_54_directio.sys_util 9.78 ± 9% +19.4% 11.68 ± 9% fxmark.ssd_f2fs_DWTL_54_directio.user_util 505108 ± 13% +104.2% 1031595 ± 6% fxmark.ssd_f2fs_DWTL_54_directio.works/sec 2.86 ± 7% -35.1% 1.86 ± 3% fxmark.ssd_f2fs_DWTL_72_directio.idle_sec 70.68 ± 4% -14.2% 60.66 ± 3% fxmark.ssd_f2fs_DWTL_72_directio.idle_util 0.03 ± 55% -100.0% 0.00 fxmark.ssd_f2fs_DWTL_72_directio.iowait_sec 0.69 ± 53% -100.0% 0.00 fxmark.ssd_f2fs_DWTL_72_directio.iowait_util 0.06 ± 5% -24.3% 0.04 ± 4% fxmark.ssd_f2fs_DWTL_72_directio.real_sec 0.02 ± 10% -56.4% 0.01 ± 12% fxmark.ssd_f2fs_DWTL_72_directio.secs 16.69 ± 19% +50.2% 25.06 ± 7% fxmark.ssd_f2fs_DWTL_72_directio.sys_util 10.34 ± 12% +26.6% 13.08 ± 3% fxmark.ssd_f2fs_DWTL_72_directio.user_util 423632 ± 9% +130.6% 976850 ± 12% fxmark.ssd_f2fs_DWTL_72_directio.works/sec 50.20 -10.9% 44.72 fxmark.time.elapsed_time 50.20 -10.9% 44.72 fxmark.time.elapsed_time.max 34124 ± 8% -51.9% 16405 fxmark.time.file_system_inputs 682953 -3.1% 661972 fxmark.time.file_system_outputs 10.33 ± 4% +16.1% 12.00 fxmark.time.percent_of_cpu_this_job_got 5.138e+09 ± 2% -13.1% 4.462e+09 cpuidle..time 2143 ± 4% -23.4% 1642 ± 5% meminfo.Active 2.23 ± 3% +0.2 2.47 ± 7% mpstat.cpu.all.sys% 500512 ± 10% -20.3% 399144 ± 20% numa-numastat.node0.numa_hit 935.00 ± 3% -51.1% 457.00 ± 3% vmstat.io.bi 18107 ± 3% +9.9% 19896 ± 3% vmstat.io.bo 7339 ± 3% +14.7% 8417 ± 3% vmstat.system.cs 1262 ± 10% -20.5% 1004 ± 8% numa-meminfo.node0.Active 2619 ± 12% +26.1% 3302 ± 8% numa-meminfo.node0.PageTables 28944 ± 32% -50.5% 14326 ± 18% numa-meminfo.node1.AnonHugePages 2408 ± 15% -28.2% 1730 ± 17% numa-meminfo.node1.PageTables 126.83 ± 23% -76.7% 29.50 ± 6% numa-vmstat.node0.nr_active_file 654.00 ± 12% +26.2% 825.67 ± 8% numa-vmstat.node0.nr_page_table_pages 126.83 ± 23% -76.7% 29.50 ± 6% numa-vmstat.node0.nr_zone_active_file 601.67 ± 15% -28.2% 432.00 ± 17% numa-vmstat.node1.nr_page_table_pages 33.51 ±128% -0.6 32.93 ±141% perf-profile.calltrace.cycles-pp.exit_to_user_mode_loop.exit_to_user_mode_prepare.syscall_exit_to_user_mode.do_syscall_64.entry_SYSCALL_64_after_hwframe 11.25 ±130% -8.8 2.47 ±144% perf-profile.children.cycles-pp.mutex_lock 33.51 ±128% -3.6 29.94 ±141% perf-profile.children.cycles-pp.__fput 33.51 ±128% -2.6 30.87 ±141% perf-profile.children.cycles-pp.task_work_run 33.51 ±128% -0.6 32.93 ±141% perf-profile.children.cycles-pp.syscall_exit_to_user_mode 33.51 ±128% -0.6 32.93 ±141% perf-profile.children.cycles-pp.exit_to_user_mode_prepare 33.51 ±128% -0.6 32.93 ±141% perf-profile.children.cycles-pp.exit_to_user_mode_loop 8.20 ±121% -5.9 2.26 ±149% perf-profile.self.cycles-pp.mutex_lock 6533 ± 3% +7.7% 7037 ± 2% perf-stat.i.context-switches 89575 -3.5% 86448 perf-stat.i.cpu-clock 5357 +8.9% 5835 perf-stat.i.minor-faults 51.62 -2.8 48.79 ± 4% perf-stat.i.node-load-miss-rate% 5365 +8.9% 5842 perf-stat.i.page-faults 89575 -3.5% 86448 perf-stat.i.task-clock 6392 ± 3% +7.6% 6876 ± 2% perf-stat.ps.context-switches 87853 -3.8% 84536 perf-stat.ps.cpu-clock 5240 +8.8% 5702 perf-stat.ps.minor-faults 5247 +8.8% 5709 perf-stat.ps.page-faults 87853 -3.8% 84536 perf-stat.ps.task-clock 149.83 ± 19% -74.4% 38.33 ± 35% proc-vmstat.nr_active_file 163134 -11.3% 144709 proc-vmstat.nr_dirtied 7977 +1.8% 8119 proc-vmstat.nr_mapped 163103 -11.3% 144678 proc-vmstat.nr_written 149.83 ± 19% -74.4% 38.33 ± 35% proc-vmstat.nr_zone_active_file 241.83 ± 76% +140.3% 581.17 ± 23% proc-vmstat.nr_zone_write_pending 756870 -3.0% 734345 proc-vmstat.numa_hit 640117 -3.5% 617496 proc-vmstat.numa_local 10158 -70.5% 2998 proc-vmstat.pgactivate 758086 -3.0% 735540 proc-vmstat.pgalloc_normal 3845 ± 9% -74.2% 993.33 ± 3% proc-vmstat.pgdeactivate 507575 -2.6% 494609 proc-vmstat.pgfault 548605 -4.0% 526561 proc-vmstat.pgfree 48618 -59.1% 19882 proc-vmstat.pgpgin 938208 -7.9% 864503 proc-vmstat.pgpgout 31716 -15.7% 26741 proc-vmstat.pgrotated 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. -- 0-DAY CI Kernel Test Service https://01.org/lkp