| 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: <20170411014608.GU17682@yexl-desktop>
Date: Tue, 11 Apr 2017 09:46:09 +0800
From: kernel test robot <xiaolong.ye@...el.com>
To: Filipe Manana <fdmanana@...e.com>
Cc: LKML <linux-kernel@...r.kernel.org>,
Stephen Rothwell <sfr@...b.auug.org.au>, lkp@...org
Subject: [lkp-robot] [Btrfs] 4bcecb33b4: fsmark.files_per_sec -65%
regression
Greeting,
FYI, we noticed a -65% regression of fsmark.files_per_sec due to commit:
commit: 4bcecb33b4bc919e0a0383d97d9d4508c8cf78b8 ("Btrfs: fix reported number of inode blocks")
https://git.kernel.org/cgit/linux/kernel/git/next/linux-next.git master
in testcase: fsmark
on test machine: 144 threads Intel(R) Xeon(R) CPU E7-8890 v3 @ 2.50GHz with 128G memory
with following parameters:
iterations: 1x
nr_threads: 64t
disk: 8BRD_12G
md: RAID0
fs: btrfs
filesize: 4M
test_size: 60G
sync_method: fsyncBeforeClose
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 https://github.com/01org/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/1x-64t-8BRD_12G-RAID0-btrfs-4M-60G-fsyncBeforeClose-performance/lkp-hsx02
2a2bb87c2c221d7d 4bcecb33b4bc919e0a0383d97d
---------------- --------------------------
%stddev change %stddev
\ | \
474957 -6% 444428 fsmark.app_overhead
1061 -65% 370 ± 4% fsmark.files_per_sec
558421 781% 4919293 ± 4% fsmark.time.voluntary_context_switches
14.70 186% 42.10 ± 4% fsmark.time.elapsed_time
14.70 186% 42.10 ± 4% fsmark.time.elapsed_time.max
1210 ± 5% 153% 3067 ± 5% fsmark.time.involuntary_context_switches
564 120% 1243 ± 3% fsmark.time.system_time
3839 -23% 2955 fsmark.time.percent_of_cpu_this_job_got
26843 ± 8% 138% 63940 ± 5% interrupts.CAL:Function_call_interrupts
15703 -61% 6157 ± 4% iostat.md0.w/s
3758783 -61% 1460865 ± 4% iostat.md0.wkB/s
3675077 -61% 1430247 ± 4% vmstat.io.bo
109228 264% 397596 vmstat.system.cs
57567 -4% 55284 vmstat.system.in
779 12% 870 turbostat.Avg_MHz
27.29 11% 30.17 turbostat.%Busy
413 4% 430 turbostat.PkgWatt
59.62 -4% 57.16 turbostat.RAMWatt
1856908 852% 17672157 ± 5% perf-stat.context-switches
1179 ± 7% 571% 7907 ± 5% perf-stat.cpu-migrations
18478436 ± 3% 494% 1.098e+08 ± 5% perf-stat.iTLB-loads
1.8e+11 240% 6.126e+11 ± 5% perf-stat.branch-instructions
1.771e+12 ± 3% 205% 5.395e+12 ± 5% perf-stat.cpu-cycles
1.889e+08 174% 5.17e+08 ± 3% perf-stat.node-load-misses
7.372e+11 154% 1.87e+12 ± 4% perf-stat.instructions
1.752e+11 ± 12% 137% 4.15e+11 ± 7% perf-stat.dTLB-loads
1.195e+08 134% 2.796e+08 perf-stat.node-store-misses
1.91e+08 ± 3% 121% 4.225e+08 ± 3% perf-stat.branch-misses
77270 113% 164378 ± 3% perf-stat.page-faults
77270 113% 164378 ± 3% perf-stat.minor-faults
395924 ± 38% 111% 836529 ± 16% perf-stat.instructions-per-iTLB-miss
5.534e+08 91% 1.057e+09 ± 3% perf-stat.cache-misses
2.066e+09 60% 3.299e+09 perf-stat.cache-references
4.201e+08 ± 13% 47% 6.178e+08 ± 15% perf-stat.dTLB-load-misses
42780444 ± 10% 42% 60856921 ± 3% perf-stat.dTLB-store-misses
4.081e+10 ± 6% 38% 5.63e+10 perf-stat.dTLB-stores
1.468e+08 33% 1.951e+08 ± 3% perf-stat.node-stores
44.86 31% 58.91 perf-stat.node-store-miss-rate%
70.58 30% 91.96 perf-stat.node-load-miss-rate%
26.80 20% 32.05 perf-stat.cache-miss-rate%
0.42 -17% 0.35 perf-stat.ipc
0.11 -35% 0.07 perf-stat.branch-miss-rate%
0.24 ± 6% -38% 0.15 ± 10% perf-stat.dTLB-load-miss-rate%
78732161 ± 4% -43% 45125639 perf-stat.node-loads
fsmark.files_per_sec
1200 ++-------------------------------------------------------------------+
| * * * * * |
1100 **** ** * ****** **** **** **.** * ****** ******** * ** *** * ***
1000 ++ * ***** ** * * * * * * |
| |
900 ++ |
800 ++ |
| |
700 ++ |
600 ++ |
| |
500 ++ |
400 ++ |
|OOOOOOOOOOOOOOOO O OOOOOOOOOOO |
300 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
View attachment "config-4.10.0-rc8-00201-g4bcecb3" of type "text/plain" (102826 bytes)
View attachment "job-script" of type "text/plain" (7224 bytes)
View attachment "job.yaml" of type "text/plain" (4721 bytes)
View attachment "reproduce" of type "text/plain" (1607 bytes)
Powered by blists - more mailing lists