From: Li, Aubrey <aubrey.li@linux.intel.com>
To: lkp@lists.01.org
Subject: Re: [lkp-robot] [sched/idle] bb400b924e: fio.write_clat_95%_us -8% improvement
Date: Thu, 06 Apr 2017 13:56:38 +0800 [thread overview]
Message-ID: <a68cc4b0-e794-2016-5419-cd61ce8f2402@linux.intel.com> (raw)
In-Reply-To: <20170405071210.GD20286@yexl-desktop>
[-- Attachment #1: Type: text/plain, Size: 23590 bytes --]
Hi Xiaolong,
Thanks to report this. Several questions on my side:
(1) how to make a report between <linux-stable tree tag:4.8.8> and
<fast_idle tree HEAD>
(2) Where can I get the raw data of the report
(3) I need to dig into the test machine bios to confirm if it's
compliant to my expected criteria, how to do that?
Thanks,
-Aubrey
On 2017/4/5 15:12, kernel test robot wrote:
>
> Greeting,
>
> FYI, we noticed a -8% improvement of fio.write_clat_95%_us due to commit:
>
>
> commit: bb400b924e06e94eb12047621bd70d61564fca4c ("sched/idle: make the fast idle path for short idle periods")
> git://bee.sh.intel.com/git/aubrey/fast_idle.git 4.8.x
>
> in testcase: fio-basic
> on test machine: 88 threads Intel(R) Xeon(R) CPU E5-2699 v4 @ 2.20GHz with 128G memory
> with following parameters:
>
> disk: 1SSD
> fs: btrfs
> runtime: 300s
> nr_task: 8
> rw: randwrite
> bs: 4k
> ioengine: sync
> test_size: 512g
> cpufreq_governor: performance
>
> test-description: Fio is a tool that will spawn a number of threads or processes doing a particular type of I/O action as specified by the user.
> test-url: https://github.com/axboe/fio
>
> 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: fio-basic/1SSD-btrfs-300s-8-randwrite-4k-sync-512g-performance/lkp-bdw-ep2
>
> 4cb0c8c00bded81a bb400b924e06e94eb12047621b
> ---------------- --------------------------
> 0.17 ± 20% -58% 0.07 ± 17% fio.latency_4us%
> 237 233 fio.write_clat_mean_us
> 332 -8% 307 fio.write_clat_95%_us
> 304 -8% 279 fio.write_clat_90%_us
> 22.22 ± 4% -10% 19.96 ± 9% fio.latency_500us%
> 0.12 ± 14% -45% 0.07 ± 24% fio.latency_100us%
> 130 -17% 108 fio.time.system_time
> 50 -18% 41 fio.time.percent_of_cpu_this_job_got
> 237 21% 287 pmeter.Average_Active_Power
> 141 -16% 119 pmeter.performance_per_watt
> 128251 -3% 123873 vmstat.system.in
> 128 24% 159 turbostat.PkgWatt
> 53.98 11% 60.15 turbostat.RAMWatt
> 7.54 -41% 4.42 turbostat.%Busy
> 201 -39% 123 turbostat.Avg_MHz
> 10549 ±100% 2e+04 32779 ±173% latency_stats.avg.cgroup_kn_lock_live.__cgroup_procs_write.cgroup_procs_write.cgroup_file_write.kernfs_fop_write.__vfs_write.vfs_write.SyS_write.entry_SYSCALL_64_fastpath
> 22588 ± 7% 1e+04 36323 ±150% latency_stats.avg.max
> 10549 ±100% 2e+04 32779 ±173% latency_stats.max.cgroup_kn_lock_live.__cgroup_procs_write.cgroup_procs_write.cgroup_file_write.kernfs_fop_write.__vfs_write.vfs_write.SyS_write.entry_SYSCALL_64_fastpath
> 10549 ±100% 2e+04 32779 ±173% latency_stats.sum.cgroup_kn_lock_live.__cgroup_procs_write.cgroup_procs_write.cgroup_file_write.kernfs_fop_write.__vfs_write.vfs_write.SyS_write.entry_SYSCALL_64_fastpath
> 6644 250% 23269 ± 5% perf-stat.instructions-per-iTLB-miss
> 0.63 43% 0.90 perf-stat.ipc
> 5.33 37% 7.29 perf-stat.cache-miss-rate%
> 2.145e+09 ± 3% 15% 2.469e+09 perf-stat.iTLB-loads
> 788668 -26% 583795 perf-stat.cpu-migrations
> 3.019e+09 3.091e+09 perf-stat.cache-misses
> 8.685e+11 -15% 7.377e+11 perf-stat.dTLB-loads
> 3.254e+12 -16% 2.742e+12 perf-stat.instructions
> 5.527e+11 -15% 4.696e+11 perf-stat.dTLB-stores
> 6.852e+11 -18% 5.637e+11 perf-stat.branch-instructions
> 5.661e+10 -25% 4.239e+10 perf-stat.cache-references
> 18.12 ± 30% -40% 10.89 ± 13% perf-stat.node-store-miss-rate%
> 49893336 ± 26% -42% 28851077 ± 12% perf-stat.node-store-misses
> 5.195e+12 -41% 3.052e+12 perf-stat.cpu-cycles
> 1.08 -22% 0.85 perf-stat.branch-miss-rate%
> 7.393e+09 -35% 4.775e+09 perf-stat.branch-misses
> 0.21 -41% 0.12 ± 7% perf-stat.dTLB-load-miss-rate%
> 1.8e+09 -50% 9.054e+08 ± 8% perf-stat.dTLB-load-misses
> 18.60 -75% 4.57 ± 4% perf-stat.iTLB-load-miss-rate%
> 4.898e+08 -76% 1.182e+08 ± 5% perf-stat.iTLB-load-misses
> 0.04 -77% 0.01 ± 6% perf-stat.dTLB-store-miss-rate%
> 1.982e+08 -80% 38854803 ± 7% perf-stat.dTLB-store-misses
>
>
>
> perf-stat.cache-references
>
> 6e+10 ++----------------------------------------------------------------+
> 5.8e+10 ++ .*... |
> | ..*.. .*..*...*..*..*...*..*. *..*..*.. ..*..*
> 5.6e+10 *+.*...*..*..*. *. *. |
> 5.4e+10 ++ |
> | |
> 5.2e+10 ++ |
> 5e+10 ++ |
> 4.8e+10 ++ |
> | |
> 4.6e+10 O+ |
> 4.4e+10 ++ O O O O O |
> | O O O O |
> 4.2e+10 ++ O |
> 4e+10 ++----------------------------------------------------------------+
>
>
> perf-stat.branch-misses
>
> 8e+09 ++----------------------------------------------------------------+
> | .*... |
> 7.5e+09 ++.*... .*...*.. .*. *..*..*.. .*
> *. *..*. *..*..*...*..*..*...*. *...*. |
> 7e+09 ++ |
> | |
> 6.5e+09 ++ |
> | |
> 6e+09 ++ |
> | |
> 5.5e+09 ++ |
> O O O O O O |
> 5e+09 ++ O |
> | O O O O |
> 4.5e+09 ++----------------------------------------------------------------+
>
>
> perf-stat.dTLB-load-misses
>
> 2e+09 ++----------------------------------------------------------------+
> *..*... ..*.. .*... |
> 1.8e+09 ++ *..*..*. *..*. *..*..*...*..*..*...*..*..*..*...*..*
> | |
> 1.6e+09 ++ |
> | |
> 1.4e+09 ++ |
> | |
> 1.2e+09 ++ |
> | O |
> 1e+09 O+ O O O O |
> | O O O O |
> 8e+08 ++ O |
> | |
> 6e+08 ++----------------------------------------------------------------+
>
>
> perf-stat.dTLB-stores
>
> 5.6e+11 ++-----*----------------------*---------------*------*--*---------*
> 5.5e+11 *+.*. *..*...*.. .*..*. *..*...*..*. *. ..|
> | *. *...* |
> 5.4e+11 ++ |
> 5.3e+11 ++ |
> 5.2e+11 ++ |
> 5.1e+11 ++ |
> | |
> 5e+11 ++ |
> 4.9e+11 O+ O |
> 4.8e+11 ++ O O O O |
> 4.7e+11 ++ O O O O |
> | |
> 4.6e+11 ++ O |
> 4.5e+11 ++----------------------------------------------------------------+
>
>
> perf-stat.dTLB-store-misses
>
> 2.2e+08 ++----------------------------------------------------------------+
> 2e+08 *+.*... ..*..*.. .*...*..*..*...*..*..*... .*..*... |
> | *..*..*. *. *..*. *..*
> 1.8e+08 ++ |
> 1.6e+08 ++ |
> | |
> 1.4e+08 ++ |
> 1.2e+08 ++ |
> 1e+08 ++ |
> | |
> 8e+07 ++ |
> 6e+07 O+ O O O O O |
> | O O |
> 4e+07 ++ O O O |
> 2e+07 ++----------------------------------------------------------------+
>
>
> perf-stat.iTLB-load-misses
>
> 7e+08 ++------------------------------------------------------------------+
> | |
> 6e+08 ++.*... |
> *. *..*...* |
> | + *.. |
> 5e+08 ++ + .. *...*..*...*..*..*...*..*...*..*
> | *..*...*..*...* |
> 4e+08 ++ |
> | |
> 3e+08 ++ |
> | |
> | |
> 2e+08 ++ O |
> O O O O O O O |
> 1e+08 ++----------------------O------O--O---------------------------------+
>
>
> perf-stat.cpu-migrations
>
> 850000 ++-----------------------------------------------------------------+
> | *.. *.. |
> 800000 ++.*.. + *... .. |
> *. . + *..*.. .*... .*..*..*...* *... .*
> | * *...*. *..*.. .. *. |
> 750000 ++ * |
> | |
> 700000 ++ |
> | |
> 650000 ++ |
> | O |
> O O |
> 600000 ++ O O O O O |
> | O O O |
> 550000 ++-----------------------------------------------------------------+
>
>
> perf-stat.cache-miss-rate_
>
> 8 ++--------------------------------------------------------------------+
> | O O |
> 7.5 ++ |
> | O O O O O |
> | O O |
> 7 O+ O |
> | |
> 6.5 ++ |
> | |
> 6 ++ |
> | |
> | |
> 5.5 ++..*.. .*... .*...*.. ..*.. ..*..*...*..|
> *. *...*. *. *...*..*...*..*. *...*..*. *
> 5 ++--------------------------------------------------------------------+
>
>
> perf-stat.dTLB-load-miss-rate_
>
> 0.22 ++--------------------------*----------------------------------------+
> *..*... .*... .. .*.. ..*.. ..*..*... |
> 0.2 ++ *..*...*. *..* *...*. *. *...*..*. *..*
> | |
> | |
> 0.18 ++ |
> | |
> 0.16 ++ |
> | |
> 0.14 ++ O |
> O O O O |
> | O O O |
> 0.12 ++ O |
> | O O |
> 0.1 ++-------------------------------------------------------------------+
>
>
> perf-stat.dTLB-store-miss-rate_
>
> 0.04 ++------------------------------------------------------------------+
> | .*... .*..*... .*... .*...*.. ..*.. .*... |
> 0.035 *+ *..*...*. *. *..*. *. *..*...*. *..*
> | |
> 0.03 ++ |
> | |
> 0.025 ++ |
> | |
> 0.02 ++ |
> | |
> 0.015 ++ O |
> O O O O O |
> 0.01 ++ O O |
> | O O O |
> 0.005 ++------------------------------------------------------------------+
>
>
> perf-stat.iTLB-load-miss-rate_
>
> 24 ++---------------------------------------------------------------------+
> 22 ++..*.. |
> *. ..*..*. |
> 20 ++ *. .. .*.. ..*.. .*...|
> 18 ++ .. *...*..*. *...*..*...*. *
> | *..*...*..*...* |
> 16 ++ |
> 14 ++ |
> 12 ++ |
> | |
> 10 ++ |
> 8 ++ |
> | O O |
> 6 O+ O O O O O |
> 4 ++-----------------------O------O---O----------------------------------+
>
>
> perf-stat.instructions-per-iTLB-miss
>
> 26000 ++------------------------------------------------------------------+
> 24000 ++ O O |
> | O O |
> 22000 ++ O O |
> 20000 ++ O |
> 18000 O+ O |
> 16000 ++ O |
> | O |
> 14000 ++ |
> 12000 ++ |
> 10000 ++ |
> 8000 ++ |
> | .*..*...*..*...*.. .*...*..*...*..*..*...*..*...*..*
> 6000 *+.*...*..*...*. *. |
> 4000 ++------------------------------------------------------------------+
>
>
> fio.write_clat_90__us
>
> 310 ++------------*------*-------------*-------------*--------------------+
> *...*.. .*. .. + *.. + .. .*. *.. *...*..|
> 305 ++ .. + + .. .. |
> 300 ++ * * *...* *..* *...* *
> | |
> 295 ++ |
> | |
> 290 ++ |
> | O |
> 285 ++ |
> 280 ++ O O O O |
> | O O O O |
> 275 ++ |
> | O |
> 270 O+--------------------------------------------------------------------+
>
>
> fio.write_clat_95__us
>
> 345 ++--------------------------------------------------------------------+
> 340 ++ *... |
> *...*.. .*..*... .. *.. .*.. .*... |
> 335 ++ .. * ..*. . ..*. *.. .*...*..|
> 330 ++ * *. *..*. *...*. *
> | |
> 325 ++ |
> 320 ++ |
> 315 ++ |
> | O |
> 310 ++ O O O O |
> 305 ++ O O |
> | O O O |
> 300 O+ |
> 295 ++--------------------------------------------------------------------+
>
> [*] 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
>
next prev parent reply other threads:[~2017-04-06 5:56 UTC|newest] Thread overview: 33+ messages / expand[flat|nested] mbox.gz Atom feed top 2017-04-05 7:12 [lkp-robot] [sched/idle] bb400b924e: fio.write_clat_95%_us -8% improvement kernel test robot 2017-04-06 5:56 ` Li, Aubrey [this message] 2017-04-06 6:26 ` Ye Xiaolong 2017-04-06 13:54 ` Li, Aubrey 2017-04-07 1:33 ` Huang, Ying 2017-04-07 1:47 ` Li, Aubrey 2017-04-07 2:21 ` Huang, Ying 2017-04-07 3:10 ` Li, Aubrey 2017-04-07 1:46 ` Ye Xiaolong 2017-04-07 2:12 ` Li, Aubrey 2017-04-07 2:24 ` Huang, Ying 2017-04-07 5:26 ` Li, Aubrey 2017-04-07 6:02 ` Huang, Ying 2017-04-07 6:06 ` Li, Aubrey 2017-04-07 6:11 ` Ye Xiaolong 2017-04-07 6:16 ` Li, Aubrey 2017-04-07 6:25 ` Huang, Ying 2017-04-07 6:41 ` Li, Aubrey 2017-04-07 6:44 ` Huang, Ying 2017-04-07 6:51 ` Li, Aubrey 2017-04-07 6:58 ` Huang, Ying 2017-04-07 6:12 ` Huang, Ying 2017-04-07 3:34 ` Li, Aubrey 2017-04-07 3:44 ` Ye Xiaolong 2017-04-07 6:58 ` Li, Aubrey 2017-04-07 7:03 ` Huang, Ying 2017-04-07 7:12 ` Li, Aubrey 2017-04-07 7:18 ` Huang, Ying 2017-04-07 8:32 ` Du, Julie 2017-04-07 8:39 ` Li, Aubrey 2017-04-07 8:43 ` Huang, Ying 2017-04-07 8:48 ` Li, Aubrey 2017-04-07 8:53 ` Li, Aubrey
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=a68cc4b0-e794-2016-5419-cd61ce8f2402@linux.intel.com \
--to=aubrey.li@linux.intel.com \
--cc=lkp@lists.01.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.