Greeting,
FYI, we noticed a 36.0% improvement of fxmark.ssd_ext4_no_jnl_dbench_client_72_directio.works/sec due to commit:
commit: 613c5a85898d1cd44e68f28d65eccf64a8ace9cf ("ext4: make directory inode spreading reflect flexbg size")
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: dbench_client
fstype: ext4_no_jnl
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/ext4_no_jnl/x86_64-rhel-8.3/ssd/debian-11.1-x86_64-20220510.cgz/lkp-icl-2sp5/dbench_client/fxmark
commit:
1940265ede ("ext4: avoid unnecessary spreading of allocations among groups")
613c5a8589 ("ext4: make directory inode spreading reflect flexbg size")
1940265ede6683f6 613c5a85898d1cd44e68f28d65e
---------------- ---------------------------
fail:runs %reproduction fail:runs
| | |
:6 100% 6:6 stderr.Couldn't_synthesize_bpf_events
%stddev %change %stddev
\ | \
323.96 ? 12% +28.7% 417.02 ? 7% fxmark.ssd_ext4_no_jnl_dbench_client_72_directio.idle_sec
7.50 ? 12% +28.9% 9.66 ? 7% fxmark.ssd_ext4_no_jnl_dbench_client_72_directio.idle_util
1583 ? 3% +30.8% 2072 ? 2% fxmark.ssd_ext4_no_jnl_dbench_client_72_directio.iowait_sec
36.66 ? 3% +31.0% 48.02 ? 2% fxmark.ssd_ext4_no_jnl_dbench_client_72_directio.iowait_util
5.60 ? 3% +27.3% 7.14 fxmark.ssd_ext4_no_jnl_dbench_client_72_directio.softirq_sec
0.13 ? 3% +27.5% 0.17 fxmark.ssd_ext4_no_jnl_dbench_client_72_directio.softirq_util
2348 ? 3% -25.4% 1752 ? 4% fxmark.ssd_ext4_no_jnl_dbench_client_72_directio.sys_sec
54.36 ? 3% -25.3% 40.61 ? 4% fxmark.ssd_ext4_no_jnl_dbench_client_72_directio.sys_util
27.58 ? 4% +21.0% 33.36 ? 3% fxmark.ssd_ext4_no_jnl_dbench_client_72_directio.user_sec
0.64 ? 4% +21.1% 0.77 ? 3% fxmark.ssd_ext4_no_jnl_dbench_client_72_directio.user_util
1685 ? 4% +36.0% 2292 ? 3% fxmark.ssd_ext4_no_jnl_dbench_client_72_directio.works/sec
3.054e+08 +3.3% 3.154e+08 fxmark.time.file_system_outputs
1032 ? 2% -8.7% 942.00 fxmark.time.percent_of_cpu_this_job_got
6151 ? 2% -9.2% 5584 fxmark.time.system_time
154.57 +3.8% 160.38 fxmark.time.user_time
8290165 +5.5% 8747536 fxmark.time.voluntary_context_switches
6242070 ? 2% +9.6% 6844007 ? 2% numa-numastat.node1.numa_hit
33.42 +3.5% 34.59 iostat.cpu.iowait
36.41 -3.1% 35.27 iostat.cpu.system
4347817 -24.5% 3281458 vmstat.memory.cache
45757 +4.7% 47892 vmstat.system.cs
38.17 -9.1% 34.69 boot-time.boot
23.34 -15.2% 19.79 boot-time.dhcp
4374 -10.1% 3934 boot-time.idle
0.29 ? 10% +16.3% 0.33 ? 6% sched_debug.cfs_rq:/.h_nr_running.stddev
3361937 ? 4% -14.6% 2870581 ? 3% sched_debug.cfs_rq:/.min_vruntime.avg
3453968 ? 4% -13.8% 2976035 ? 2% sched_debug.cfs_rq:/.min_vruntime.max
3074025 ? 4% -16.7% 2559836 ? 6% sched_debug.cfs_rq:/.min_vruntime.min
4234791 -25.1% 3170271 meminfo.Cached
13280609 ? 8% -23.2% 10193251 ? 9% meminfo.DirectMap2M
5623315 -19.1% 4546975 meminfo.Memused
3825966 -27.8% 2760633 meminfo.Unevictable
7683659 -13.9% 6614519 meminfo.max_used_kB
45394 +4.8% 47577 perf-stat.i.context-switches
878.12 ? 2% +7.8% 946.23 ? 3% perf-stat.i.cpu-migrations
0.00 ?116% +0.0 0.00 ? 66% perf-stat.i.dTLB-load-miss-rate%
45558 +4.8% 47762 perf-stat.ps.context-switches
886.77 ? 2% +7.5% 953.70 ? 3% perf-stat.ps.cpu-migrations
0.49 ? 19% -0.3 0.18 ? 52% perf-profile.children.cycles-pp.get_cpu_device
0.34 ? 24% -0.1 0.20 ? 20% perf-profile.children.cycles-pp.s_next
0.06 ?114% +0.1 0.14 ? 34% perf-profile.children.cycles-pp.account_process_tick
0.48 ? 20% -0.3 0.16 ? 54% perf-profile.self.cycles-pp.get_cpu_device
0.05 ?141% +0.1 0.14 ? 41% perf-profile.self.cycles-pp.account_process_tick
0.40 ? 9% +0.1 0.50 ? 13% perf-profile.self.cycles-pp.tick_nohz_next_event
109190 ? 30% -60.3% 43316 ? 57% numa-meminfo.node0.AnonHugePages
154521 ? 26% -43.2% 87783 ? 27% numa-meminfo.node0.AnonPages
437517 ? 8% -16.8% 364128 ? 6% numa-meminfo.node0.Inactive
160530 ? 25% -41.4% 94026 ? 25% numa-meminfo.node0.Inactive(anon)
87985 ? 39% +73.3% 152466 ? 16% numa-meminfo.node1.AnonHugePages
157511 ? 26% +41.4% 222774 ? 10% numa-meminfo.node1.AnonPages
207719 ? 17% +34.9% 280314 ? 9% numa-meminfo.node1.Inactive
161988 ? 25% +40.3% 227216 ? 10% numa-meminfo.node1.Inactive(anon)
38635 ? 26% -43.2% 21947 ? 27% numa-vmstat.node0.nr_anon_pages
40136 ? 25% -41.4% 23505 ? 25% numa-vmstat.node0.nr_inactive_anon
40136 ? 25% -41.4% 23505 ? 25% numa-vmstat.node0.nr_zone_inactive_anon
39376 ? 26% +41.5% 55699 ? 10% numa-vmstat.node1.nr_anon_pages
6131790 ? 2% +10.0% 6744942 ? 2% numa-vmstat.node1.nr_dirtied
40495 ? 25% +40.3% 56809 ? 10% numa-vmstat.node1.nr_inactive_anon
40495 ? 25% +40.3% 56809 ? 10% numa-vmstat.node1.nr_zone_inactive_anon
6241898 ? 2% +9.6% 6843576 ? 2% numa-vmstat.node1.numa_hit
38173261 +3.3% 39428621 proc-vmstat.nr_dirtied
1059388 -25.1% 793340 proc-vmstat.nr_file_pages
53799 -1.9% 52794 proc-vmstat.nr_slab_unreclaimable
956491 -27.8% 690158 proc-vmstat.nr_unevictable
9850651 +3.5% 10194926 proc-vmstat.nr_written
956491 -27.8% 690158 proc-vmstat.nr_zone_unevictable
38481863 +3.2% 39707334 proc-vmstat.numa_hit
38362225 +3.2% 39589589 proc-vmstat.numa_local
1333657 +2.1% 1361550 proc-vmstat.pgactivate
38472842 +3.2% 39705274 proc-vmstat.pgalloc_normal
38505715 +3.2% 39737694 proc-vmstat.pgfree
66251245 +2.3% 67801130 proc-vmstat.pgpgout
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