From: Fengguang Wu <fengguang.wu@intel.com>
To: lkp@lists.01.org
Subject: [sched] 177672afd93: -40.9% dbench.throughput-MB/sec
Date: Sat, 05 Apr 2014 22:16:42 +0800 [thread overview]
Message-ID: <20140405141642.GB13106@localhost> (raw)
[-- Attachment #1: Type: text/plain, Size: 7225 bytes --]
Hi Tim,
Up to now, the system catches one regression in
git://bee.sh.intel.com/git/wfg/linux-devel.git tim-idle-load-balance
commit 177672afd93461854abe26c05bd549647efc09db ("sched: Reduce the rate of idle load balancing")
test case: nhm8/micro/dbench/100%
v3.14 177672afd93461854abe26c05
--------------- -------------------------
8214 ~ 0% -40.9% 4854 ~ 0% TOTAL dbench.throughput-MB/sec
443685 ~ 0% +156.5% 1138160 ~20% TOTAL softirqs.RCU
3.38 ~ 0% -52.1% 1.62 ~ 0% TOTAL turbostat.GHz
81260 ~ 1% -42.7% 46534 ~ 1% TOTAL slabinfo.kmalloc-256.active_objs
280 ~23% +85.4% 520 ~12% TOTAL cpuidle.C3-NHM.usage
2556 ~ 1% -42.6% 1467 ~ 1% TOTAL slabinfo.kmalloc-256.num_slabs
2556 ~ 1% -42.6% 1467 ~ 1% TOTAL slabinfo.kmalloc-256.active_slabs
81822 ~ 1% -42.6% 46982 ~ 1% TOTAL slabinfo.kmalloc-256.num_objs
8.34 ~ 0% -42.2% 4.82 ~ 1% TOTAL perf-profile.cpu-cycles.copy_user_generic_string.generic_file_buffered_write.__generic_file_aio_write.generic_file_aio_write.do_sync_write
17.87 ~ 0% -42.6% 10.26 ~ 0% TOTAL perf-profile.cpu-cycles.copy_user_generic_string.generic_file_aio_read.do_sync_read.vfs_read.sys_pread64
7029777 ~ 0% -40.8% 4164671 ~ 0% TOTAL proc-vmstat.pgactivate
0.78 ~ 1% +41.2% 1.10 ~ 1% TOTAL perf-profile.cpu-cycles.filldir.dcache_readdir.iterate_dir.sys_getdents.system_call_fastpath
0.75 ~ 2% +42.9% 1.07 ~ 1% TOTAL perf-profile.cpu-cycles.strchr(a)plt
91820 ~ 0% -25.1% 68743 ~ 0% TOTAL slabinfo.dentry.active_objs
2234 ~ 0% -25.0% 1675 ~ 0% TOTAL slabinfo.dentry.num_slabs
2234 ~ 0% -25.0% 1675 ~ 0% TOTAL slabinfo.dentry.active_slabs
93857 ~ 0% -25.0% 70409 ~ 0% TOTAL slabinfo.dentry.num_objs
1.48 ~ 1% +34.2% 1.98 ~ 0% TOTAL perf-profile.cpu-cycles.copy_user_generic_string.dcache_readdir.iterate_dir.sys_getdents.system_call_fastpath
0.82 ~ 5% +31.9% 1.08 ~ 4% TOTAL perf-profile.cpu-cycles.__d_lookup_rcu.lookup_fast.link_path_walk.path_lookupat.filename_lookup
9908 ~ 1% -21.3% 7798 ~ 0% TOTAL proc-vmstat.nr_slab_unreclaimable
39507 ~ 0% -21.4% 31047 ~ 0% TOTAL meminfo.SUnreclaim
61 ~ 1% -20.1% 49 ~ 2% TOTAL turbostat.CTMP
0.85 ~ 2% +22.9% 1.04 ~ 1% TOTAL perf-profile.cpu-cycles.close
1.31 ~ 1% +22.9% 1.61 ~ 1% TOTAL perf-profile.cpu-cycles.link_path_walk.path_lookupat.filename_lookup.user_path_at_empty.user_path_at
92159 ~ 0% -16.6% 76833 ~ 0% TOTAL meminfo.Slab
14061 ~ 4% +20.9% 16994 ~ 1% TOTAL interrupts.RES
1.45 ~ 1% +16.9% 1.69 ~ 0% TOTAL perf-profile.cpu-cycles._raw_spin_lock.dcache_readdir.iterate_dir.sys_getdents.system_call_fastpath
13180 ~ 0% -13.1% 11451 ~ 0% TOTAL proc-vmstat.nr_slab_reclaimable
52651 ~ 0% -13.0% 45786 ~ 0% TOTAL meminfo.SReclaimable
3096 ~ 0% +15.6% 3581 ~ 0% TOTAL time.user_time
5514 ~ 0% -8.9% 5023 ~ 0% TOTAL time.system_time
225997 ~ 0% +3.1% 233003 ~ 1% TOTAL time.involuntary_context_switches
Legend:
~XX% - stddev percent
[+-]XX% - change percent
time.user_time
3600 ++-------------O-------O------O-------------------------------------+
O O |
3500 ++ |
| |
| |
3400 ++ |
| |
3300 ++ |
| |
3200 ++ |
| |
*....... ....*...... ....*
3100 ++ *......*.......*......*... *.......*......*... |
| |
3000 ++------------------------------------------------------------------+
time.system_time
5600 ++------------------------------------------------------------------+
| |
5500 ++ ....*......*.......*......*.......*......*.......*......*.......*
*... |
| |
5400 ++ |
| |
5300 ++ |
| |
5200 ++ |
| |
| |
5100 ++ |
O O O |
5000 ++-------------O-------O--------------------------------------------+
dbench.throughput-MB/sec
8500 ++------------------------------------------------------------------+
*.......*......*.......*......*.......*......*.......*......*.......*
8000 ++ |
7500 ++ |
| |
7000 ++ |
| |
6500 ++ |
| |
6000 ++ |
5500 ++ |
| |
5000 ++ |
O O O O O |
4500 ++------------------------------------------------------------------+
[*] bisect-good sample
[O] bisect-bad sample
Thanks,
Fengguang
[-- Attachment #2: reproduce.ksh --]
[-- Type: text/plain, Size: 42 bytes --]
dbench 12 -c /usr/share/dbench/client.txt
next reply other threads:[~2014-04-05 14:16 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-04-05 14:16 Fengguang Wu [this message]
2014-04-08 2:08 ` [sched] 177672afd93: -40.9% dbench.throughput-MB/sec Fengguang Wu
2014-04-12 2:51 ` Fengguang Wu
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=20140405141642.GB13106@localhost \
--to=fengguang.wu@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.