Skip to contentSkip to conversation

d9b7e4f717a: +52.5% turbostat.%Busy

LKP Huang Ying 2 participants 2 messages Feb 11, 2015 – Feb 12, 2015

lore
Share & tools
lore

Huang Ying

author
#0 lore raw

FYI, we noticed the below changes on

commit d9b7e4f717a167610a49ceb9e5969e80146c89a8 ("ipmi: Periodically check to see if irqs and messages are set right")

testbox/testcase/testparams: avoton3/idle/performance-10m

6a11e5c67a397e9a d9b7e4f717a167610a49ceb9e5
---------------- --------------------------
%stddev %change %stddev
\ | \
0.10 ± 0% +52.5% 0.15 ± 2% turbostat.%Busy
2 ± 0% +50.0% 3 ± 0% turbostat.Avg_MHz

turbostat.%Busy

0.2 ++----O-O------------------------------------------------------------+

       |                                                                    |
  0.18 ++                                                                   |
       |                                                                    |
       |                                                                    |
  0.16 ++         O     O O  O  O       O  O  O O  O  O  O O          O     |
       O  O          O             O O                        O  O  O       |
  0.14 ++                                                                   |
       |                                                                    |
  0.12 ++    *.*                     *                                      |
       |   ..   +                   : +                                     |
       | .*      +                  :  +                                    |
   0.1 *+         *..  .*.*..  .*..*    *..*..*.*..*..*..*.*..*..*..*.*..*..*
       |             *.      *.                                             |

0.08 ++-------------------------------------------------------------------+

[*] bisect-good sample
[O] bisect-bad sample

To reproduce:

apt-get install ruby ruby-oj
git clone git://git.kernel.org/pub/scm/linux/kernel/git/wfg/lkp-tests.git
cd lkp-tests
bin/setup-local job.yaml # the job file attached in this email
bin/run-local job.yaml

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.

_______________________________________________
LKP mailing list
LKP(a)linux.intel.com

---
testcase: idle
default-monitors:
latency_stats:
interrupts:
softirqs:
cpuidle:
turbostat:
pmeter:
default_watchdogs:
watch-oom:
watchdog:
cpufreq_governor: performance
commit: f876f0cff6d460f03cd937104314f366dfa92261
model: Avoton
memory: 8G
hdd_partitions: "/dev/disk/by-id/ata-ST9250610NS_9XE0BSWH-part2"
swap_partitions:
rootfs_partition: "/dev/disk/by-id/ata-ST9250610NS_9XE0BSWH-part1"
user: wfg
rootfs: debian-x86_64-2015-02-07.cgz
pmeter_server: lkp-wildcat
pmeter_device: yokogawa-wt310
stop_daemons:
sleep: 10m
testbox: avoton3
tbox_group: avoton3
kconfig: x86_64-rhel
enqueue_time: 2015-02-07 15:27:31.085931347 +08:00
head_commit: f876f0cff6d460f03cd937104314f366dfa92261
base_commit: e36f014edff70fc02b3d3d79cead1d58f289332e
branch: linux-devel/devel-hourly-2015020522
kernel: "/kernel/x86_64-rhel/f876f0cff6d460f03cd937104314f366dfa92261/vmlinuz-3.19.0-rc7-gf876f0c"
queue: cyclic
result_root: "/result/avoton3/idle/performance-10m/debian-x86_64-2015-02-07.cgz/x86_64-rhel/f876f0cff6d460f03cd937104314f366dfa92261/0"
job_file: "/lkp/scheduled/avoton3/cyclic_idle-performance-10m-debian-x86_64.cgz-x86_64-rhel-HEAD-f876f0cff6d460f03cd937104314f366dfa92261-0-20150207-95893-1i1o2km.yaml"
dequeue_time: 2015-02-07 15:47:35.590497559 +08:00
nr_cpu: "$(nproc)"
job_state: finished
loadavg: 0.00 0.04 0.05 1/93 680
start_time: '1423295378'
end_time: '1423295978'
version: "/lkp/wfg/.src-20150207-125134"

#1 lore raw

I'm not quite sure how to interpret this data, but I'm assuming that with the periodic checks it's waking up a lot more than it used to.

I'm not sure why this is the case, but it depends on what userland software is doing with the driver. I'll try to reproduce and see what I can find.

-corey

Quoted 55 lines · Huang Ying

FYI, we noticed the below changes on

commit d9b7e4f717a167610a49ceb9e5969e80146c89a8 ("ipmi: Periodically check to see if irqs and messages are set right")

testbox/testcase/testparams: avoton3/idle/performance-10m

6a11e5c67a397e9a d9b7e4f717a167610a49ceb9e5
---------------- --------------------------
%stddev %change %stddev
\ | \
0.10 ± 0% +52.5% 0.15 ± 2% turbostat.%Busy
2 ± 0% +50.0% 3 ± 0% turbostat.Avg_MHz

turbostat.%Busy

   0.2 ++----O-O------------------------------------------------------------+
       |                                                                    |
  0.18 ++                                                                   |
       |                                                                    |
       |                                                                    |
  0.16 ++         O     O O  O  O       O  O  O O  O  O  O O          O     |
       O  O          O             O O                        O  O  O       |
  0.14 ++                                                                   |
       |                                                                    |
  0.12 ++    *.*                     *                                      |
       |   ..   +                   : +                                     |
       | .*      +                  :  +                                    |

0.1 *+ *.. .*.*.. .*..* *..*..*.*..*..*..*.*..*..*..*.*..*..*
| *. *. |
0.08 ++-------------------------------------------------------------------+

[*] bisect-good sample
[O] bisect-bad sample

To reproduce:

apt-get install ruby ruby-oj
git clone git://git.kernel.org/pub/scm/linux/kernel/git/wfg/lkp-tests.git
cd lkp-tests
bin/setup-local job.yaml # the job file attached in this email
bin/run-local job.yaml

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.

_______________________________________________
LKP mailing list
LKP(a)linux.intel.com