| 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
| ||
|
Message-ID: <20201230032753.GC22185@xsang-OptiPlex-9020>
Date: Wed, 30 Dec 2020 11:27:53 +0800
From: kernel test robot <oliver.sang@...el.com>
To: Kent Overstreet <kent.overstreet@...il.com>
Cc: Linus Torvalds <torvalds@...ux-foundation.org>,
Jens Axboe <axboe@...nel.dk>,
Matthew Wilcox <willy@...radead.org>,
kernel test robot <rong.a.chen@...el.com>,
Andrew Morton <akpm@...ux-foundation.org>,
LKML <linux-kernel@...r.kernel.org>, lkp@...ts.01.org,
lkp@...el.com, ying.huang@...el.com, feng.tang@...el.com,
zhengjun.xing@...el.com
Subject: [mm/filemap.c] 06c0444290: stress-ng.sendfile.ops_per_sec 26.7%
improvement
Greeting,
FYI, we noticed a 26.7% improvement of stress-ng.sendfile.ops_per_sec due to commit:
commit: 06c0444290cecf04c89c62e6d448b8461507d247 ("mm/filemap.c: generic_file_buffered_read() now uses find_get_pages_contig")
https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git master
in testcase: stress-ng
on test machine: 96 threads Intel(R) Xeon(R) Gold 6252 CPU @ 2.10GHz with 192G memory
with following parameters:
nr_threads: 100%
disk: 1HDD
testtime: 30s
class: pipe
cpufreq_governor: performance
ucode: 0x5003003
Details are as below:
-------------------------------------------------------------------------------------------------->
To reproduce:
git clone https://github.com/intel/lkp-tests.git
cd lkp-tests
bin/lkp install job.yaml # job file is attached in this email
bin/lkp run job.yaml
=========================================================================================
class/compiler/cpufreq_governor/disk/kconfig/nr_threads/rootfs/tbox_group/testcase/testtime/ucode:
pipe/gcc-9/performance/1HDD/x86_64-rhel-8.3/100%/debian-10.4-x86_64-20200603.cgz/lkp-csl-2sp5/stress-ng/30s/0x5003003
commit:
723ef24b9b ("mm/filemap/c: break generic_file_buffered_read up into multiple functions")
06c0444290 ("mm/filemap.c: generic_file_buffered_read() now uses find_get_pages_contig")
723ef24b9b379e59 06c0444290cecf04c89c62e6d44
---------------- ---------------------------
%stddev %change %stddev
\ | \
14865658 +26.7% 18839172 ± 2% stress-ng.sendfile.ops
495501 +26.7% 627957 ± 2% stress-ng.sendfile.ops_per_sec
17943 ± 12% +36.5% 24500 ± 5% proc-vmstat.numa_hint_faults
1585 ±170% +3019.7% 49447 ± 63% numa-numastat.node0.other_node
85104 ± 3% -56.2% 37244 ± 83% numa-numastat.node1.other_node
169349 ± 19% -21.8% 132475 meminfo.AnonHugePages
301479 ± 10% -11.8% 265842 meminfo.AnonPages
333754 ± 9% -10.6% 298511 meminfo.Inactive
333754 ± 9% -10.6% 298511 meminfo.Inactive(anon)
11540 ± 12% -15.2% 9785 ± 3% sched_debug.cfs_rq:/.load.avg
17531 ± 77% -77.9% 3878 ± 9% sched_debug.cfs_rq:/.load.stddev
28103 ± 22% +50.3% 42227 ± 30% sched_debug.cpu.avg_idle.min
6188 ± 20% +38.9% 8595 ± 4% sched_debug.cpu.curr->pid.min
495.48 ± 24% -46.2% 266.34 ± 18% sched_debug.cpu.curr->pid.stddev
3.336e+10 -5.6% 3.148e+10 ± 6% perf-stat.i.branch-instructions
0.03 ± 7% +0.0 0.04 ± 43% perf-stat.i.dTLB-load-miss-rate%
0.01 ± 18% +0.0 0.01 ± 10% perf-stat.i.dTLB-store-miss-rate%
6253 ± 3% -18.2% 5117 ± 9% perf-stat.i.instructions-per-iTLB-miss
0.64 +0.0 0.67 perf-stat.overall.branch-miss-rate%
3.264e+10 -5.5% 3.084e+10 ± 5% perf-stat.ps.branch-instructions
0.01 ± 5% +13.5% 0.01 ± 2% perf-sched.sch_delay.avg.ms.do_nanosleep.hrtimer_nanosleep.__x64_sys_nanosleep.do_syscall_64
0.05 ± 83% +1464.2% 0.85 ±157% perf-sched.sch_delay.max.ms.do_wait.kernel_wait4.__do_sys_wait4.do_syscall_64
494.00 ± 4% +734.5% 4122 ± 3% perf-sched.wait_and_delay.count.preempt_schedule_common._cond_resched.__splice_from_pipe.splice_from_pipe.direct_splice_actor
5392 ± 3% -98.4% 87.00 ± 10% perf-sched.wait_and_delay.count.preempt_schedule_common._cond_resched.generic_file_buffered_read.generic_file_splice_read.splice_direct_to_actor
18.80 ± 20% -68.8% 5.86 ± 73% perf-sched.wait_and_delay.max.ms.preempt_schedule_common._cond_resched.generic_file_buffered_read.generic_file_splice_read.splice_direct_to_actor
0.59 ± 24% -45.0% 0.32 ± 22% perf-sched.wait_time.avg.ms.wait_for_partner.fifo_open.do_dentry_open.path_openat
18.80 ± 20% -68.8% 5.86 ± 73% perf-sched.wait_time.max.ms.preempt_schedule_common._cond_resched.generic_file_buffered_read.generic_file_splice_read.splice_direct_to_actor
7626 -38.2% 4716 ± 33% interrupts.CPU30.NMI:Non-maskable_interrupts
7626 -38.2% 4716 ± 33% interrupts.CPU30.PMI:Performance_monitoring_interrupts
18521 ± 44% +241.7% 63289 ± 80% interrupts.CPU42.RES:Rescheduling_interrupts
30879 ± 63% +91.0% 58971 ± 31% interrupts.CPU49.RES:Rescheduling_interrupts
37970 ± 19% +115.4% 81806 ± 33% interrupts.CPU5.CAL:Function_call_interrupts
48131 ± 35% -55.7% 21307 ± 23% interrupts.CPU65.RES:Rescheduling_interrupts
33689 ± 39% +186.7% 96598 ± 88% interrupts.CPU7.CAL:Function_call_interrupts
37234 ± 52% +76.5% 65709 ± 45% interrupts.CPU71.CAL:Function_call_interrupts
22154 ± 18% +126.8% 50249 ± 70% interrupts.CPU82.RES:Rescheduling_interrupts
16632 ± 60% +310.7% 68311 ± 51% interrupts.CPU9.CAL:Function_call_interrupts
17920 ± 45% +264.1% 65238 ± 48% interrupts.CPU95.CAL:Function_call_interrupts
83.19 -3.7 79.44 ± 5% perf-profile.calltrace.cycles-pp.do_syscall_64.entry_SYSCALL_64_after_hwframe
83.28 -3.8 79.53 ± 5% perf-profile.children.cycles-pp.do_syscall_64
0.68 ± 3% +0.0 0.71 ± 2% perf-profile.children.cycles-pp.sched_clock
0.32 ± 12% +0.1 0.39 ± 6% perf-profile.children.cycles-pp.set_next_buddy
0.00 +0.1 0.07 ± 12% perf-profile.children.cycles-pp.perf_mux_hrtimer_handler
1.31 +0.1 1.39 ± 3% perf-profile.children.cycles-pp.__update_load_avg_se
3.41 +0.2 3.56 ± 2% perf-profile.children.cycles-pp.switch_mm_irqs_off
0.08 +0.0 0.10 ± 5% perf-profile.self.cycles-pp.perf_trace_run_bpf_submit
0.08 ± 10% +0.0 0.10 ± 4% perf-profile.self.cycles-pp.__bitmap_and
0.28 ± 9% +0.1 0.33 ± 9% perf-profile.self.cycles-pp.pipe_poll
0.93 +0.1 0.99 perf-profile.self.cycles-pp.switch_mm_irqs_off
0.29 ± 13% +0.1 0.36 ± 7% perf-profile.self.cycles-pp.set_next_buddy
1.28 +0.1 1.36 ± 3% perf-profile.self.cycles-pp.__update_load_avg_se
9207 ± 21% -40.3% 5500 ± 26% softirqs.CPU10.SCHED
8966 ± 22% -36.9% 5653 ± 18% softirqs.CPU11.SCHED
9107 ± 21% -41.4% 5341 ± 19% softirqs.CPU12.SCHED
8925 ± 23% -41.6% 5215 ± 14% softirqs.CPU13.SCHED
9040 ± 21% -41.5% 5285 ± 16% softirqs.CPU14.SCHED
8972 ± 23% -40.6% 5327 ± 16% softirqs.CPU15.SCHED
9021 ± 22% -41.5% 5279 ± 13% softirqs.CPU16.SCHED
8932 ± 20% -40.0% 5357 ± 15% softirqs.CPU17.SCHED
8870 ± 21% -39.0% 5409 ± 20% softirqs.CPU18.SCHED
8866 ± 23% -39.9% 5330 ± 19% softirqs.CPU19.SCHED
9042 ± 21% -37.1% 5683 ± 16% softirqs.CPU2.SCHED
8898 ± 22% -40.7% 5274 ± 20% softirqs.CPU20.SCHED
8989 ± 22% -39.8% 5412 ± 18% softirqs.CPU21.SCHED
8876 ± 22% -41.2% 5223 ± 19% softirqs.CPU22.SCHED
8892 ± 19% -38.7% 5455 ± 16% softirqs.CPU23.SCHED
6924 ± 31% +58.8% 10997 ± 7% softirqs.CPU24.SCHED
7098 ± 30% +64.3% 11663 ± 6% softirqs.CPU25.SCHED
7040 ± 32% +63.8% 11534 ± 7% softirqs.CPU26.SCHED
6977 ± 32% +63.6% 11416 ± 7% softirqs.CPU27.SCHED
7088 ± 30% +58.8% 11255 ± 8% softirqs.CPU28.SCHED
6857 ± 33% +65.5% 11352 ± 6% softirqs.CPU29.SCHED
9142 ± 25% -41.4% 5358 ± 19% softirqs.CPU3.SCHED
7061 ± 29% +62.5% 11472 ± 7% softirqs.CPU30.SCHED
6878 ± 30% +65.2% 11362 ± 8% softirqs.CPU31.SCHED
7173 ± 30% +64.9% 11828 ± 5% softirqs.CPU32.SCHED
7013 ± 31% +62.6% 11405 ± 8% softirqs.CPU33.SCHED
14139 ± 27% +30.8% 18492 ± 24% softirqs.CPU34.RCU
7033 ± 32% +58.8% 11166 ± 7% softirqs.CPU34.SCHED
6963 ± 29% +61.5% 11248 ± 7% softirqs.CPU35.SCHED
7012 ± 30% +61.6% 11332 ± 9% softirqs.CPU36.SCHED
6923 ± 29% +63.4% 11310 ± 8% softirqs.CPU37.SCHED
7070 ± 32% +59.8% 11298 ± 8% softirqs.CPU38.SCHED
6818 ± 31% +67.0% 11389 ± 7% softirqs.CPU39.SCHED
9088 ± 20% -42.2% 5250 ± 15% softirqs.CPU4.SCHED
7040 ± 29% +61.5% 11368 ± 7% softirqs.CPU40.SCHED
6980 ± 29% +62.2% 11321 ± 7% softirqs.CPU41.SCHED
6926 ± 29% +64.3% 11379 ± 8% softirqs.CPU42.SCHED
7062 ± 30% +57.4% 11114 ± 7% softirqs.CPU43.SCHED
6960 ± 31% +62.1% 11279 ± 9% softirqs.CPU44.SCHED
6854 ± 31% +65.0% 11310 ± 7% softirqs.CPU45.SCHED
7152 ± 29% +58.9% 11362 ± 9% softirqs.CPU46.SCHED
6828 ± 31% +64.2% 11210 ± 8% softirqs.CPU47.SCHED
8903 ± 21% -41.9% 5170 ± 17% softirqs.CPU48.SCHED
9048 ± 21% -41.0% 5336 ± 19% softirqs.CPU49.SCHED
8974 ± 23% -42.9% 5124 ± 18% softirqs.CPU5.SCHED
8742 ± 23% -40.8% 5177 ± 17% softirqs.CPU50.SCHED
8647 ± 23% -38.3% 5335 ± 19% softirqs.CPU51.SCHED
8783 ± 22% -41.7% 5118 ± 10% softirqs.CPU52.SCHED
8659 ± 26% -40.2% 5175 ± 18% softirqs.CPU53.SCHED
8852 ± 23% -40.2% 5292 ± 15% softirqs.CPU54.SCHED
9070 ± 20% -43.2% 5153 ± 16% softirqs.CPU55.SCHED
9191 ± 17% -42.7% 5266 ± 15% softirqs.CPU56.SCHED
8884 ± 24% -41.8% 5171 ± 16% softirqs.CPU57.SCHED
8986 ± 23% -40.5% 5344 ± 18% softirqs.CPU58.SCHED
9501 ± 24% -44.9% 5233 ± 20% softirqs.CPU59.SCHED
8897 ± 21% -38.5% 5467 ± 18% softirqs.CPU6.SCHED
9260 ± 18% -42.4% 5335 ± 22% softirqs.CPU60.SCHED
8966 ± 21% -42.3% 5170 ± 15% softirqs.CPU61.SCHED
8963 ± 22% -39.2% 5454 ± 16% softirqs.CPU62.SCHED
8948 ± 22% -39.0% 5462 ± 17% softirqs.CPU63.SCHED
8980 ± 22% -41.1% 5289 ± 16% softirqs.CPU64.SCHED
8969 ± 20% -41.4% 5260 ± 15% softirqs.CPU65.SCHED
8891 ± 23% -41.4% 5211 ± 20% softirqs.CPU66.SCHED
9193 ± 23% -40.0% 5520 ± 14% softirqs.CPU67.SCHED
8936 ± 24% -40.7% 5296 ± 16% softirqs.CPU68.SCHED
8871 ± 22% -40.0% 5320 ± 16% softirqs.CPU69.SCHED
8962 ± 19% -41.8% 5216 ± 17% softirqs.CPU7.SCHED
8671 ± 24% -38.7% 5315 ± 19% softirqs.CPU70.SCHED
7198 ± 29% +53.9% 11076 ± 4% softirqs.CPU72.SCHED
7133 ± 29% +61.0% 11488 ± 9% softirqs.CPU73.SCHED
6952 ± 30% +66.9% 11602 ± 8% softirqs.CPU74.SCHED
6975 ± 31% +60.8% 11214 ± 7% softirqs.CPU75.SCHED
6985 ± 31% +58.4% 11065 ± 10% softirqs.CPU76.SCHED
6811 ± 31% +63.6% 11146 ± 6% softirqs.CPU77.SCHED
7006 ± 29% +62.0% 11347 ± 7% softirqs.CPU78.SCHED
6827 ± 32% +65.8% 11316 ± 9% softirqs.CPU79.SCHED
8957 ± 19% -40.8% 5304 ± 18% softirqs.CPU8.SCHED
7102 ± 32% +59.7% 11345 ± 8% softirqs.CPU80.SCHED
7023 ± 30% +60.3% 11258 ± 8% softirqs.CPU81.SCHED
7046 ± 31% +57.8% 11121 ± 6% softirqs.CPU82.SCHED
6966 ± 30% +57.1% 10941 ± 8% softirqs.CPU83.SCHED
6953 ± 30% +62.0% 11261 ± 10% softirqs.CPU85.SCHED
6884 ± 31% +63.0% 11220 ± 9% softirqs.CPU86.SCHED
6765 ± 32% +66.3% 11249 ± 8% softirqs.CPU87.SCHED
6963 ± 29% +63.8% 11403 ± 7% softirqs.CPU88.SCHED
6869 ± 31% +63.6% 11241 ± 8% softirqs.CPU89.SCHED
9002 ± 21% -43.2% 5115 ± 18% softirqs.CPU9.SCHED
6759 ± 33% +69.4% 11450 ± 9% softirqs.CPU90.SCHED
7003 ± 30% +58.9% 11130 ± 9% softirqs.CPU91.SCHED
6994 ± 28% +64.9% 11533 ± 9% softirqs.CPU92.SCHED
6839 ± 32% +66.1% 11358 ± 6% softirqs.CPU93.SCHED
7121 ± 29% +59.2% 11339 ± 7% softirqs.CPU94.SCHED
6798 ± 32% +67.9% 11412 ± 7% softirqs.CPU95.SCHED
stress-ng.sendfile.ops
2e+07 +-----------------------------------------------------------------+
| O OO OO |
1.9e+07 |-+O OOOOOOO OO OOOO |
|OOOO O O OOO O |
| O O O O |
1.8e+07 |-+ |
| |
1.7e+07 |-+ |
| |
1.6e+07 |-+ |
| |
| |
1.5e+07 |+++++ + ++++++++++ +++++++++++ ++++++ +++++++ ++++++++++ + +++++|
|+ ++++ + +++ +++++ + ++ +++++ ++++++++ +++++ + + ++++ + |
1.4e+07 +-----------------------------------------------------------------+
stress-ng.sendfile.ops_per_sec
660000 +------------------------------------------------------------------+
640000 |-+ OO OOOOO O |
|OOOOOOOO O OOOOOOO |
620000 |-+OO O |
600000 |-+ O O O |
| |
580000 |-+ |
560000 |-+ |
540000 |-+ |
| |
520000 |-+ |
500000 |-+ + + + + + |
|++++++++++++++++++++++++++++++++++++++ ++++++++++++++++++++++++++|
480000 |-+ ++ + + ++ + + + + + +++++ ++ + + ++ + + |
460000 +------------------------------------------------------------------+
7000 +--------------------------------------------------------------------+
| |
6000 |-+ + + + + + + |
|++++++:++++++++++++:+++++++++++::++++:+ + ++++++:+++++++++++::++++++|
5000 |-+ ++++ + ++ + ++ + + + ++ + ++++ + + ++ + + + ++ + ++ |
| |
4000 |-+ |
| |
3000 |-+ |
| |
2000 |-+ |
| |
1000 |-+ |
| |
0 +--------------------------------------------------------------------+
[*] 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,
Oliver Sang
View attachment "config-5.10.0-g06c0444290ce" of type "text/plain" (171082 bytes)
View attachment "job-script" of type "text/plain" (7928 bytes)
View attachment "job.yaml" of type "text/plain" (5526 bytes)
View attachment "reproduce" of type "text/plain" (390 bytes)
Powered by blists - more mailing lists