From mboxrd@z Thu Jan 1 00:00:00 1970 Content-Type: multipart/mixed; boundary="===============7451010360236667752==" MIME-Version: 1.0 From: Paul E. McKenney To: lkp@lists.01.org Subject: Re: [rcu] 7c66b15f870: +1.8% turbostat.Pkg_W Date: Wed, 25 Jun 2014 19:10:23 -0700 Message-ID: <20140626021023.GO4603@linux.vnet.ibm.com> In-Reply-To: <20140626013308.GA12239@localhost> List-Id: --===============7451010360236667752== Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable On Thu, Jun 26, 2014 at 09:33:08AM +0800, Fengguang Wu wrote: > Hi Paul, > = > We noticed increased power consumption in our internal merge commit. > It's not a direct evidence, however in the hope you can see any > obvious clues. :) Hello, Fengguang, This particular branch is obsolete, and has been replaced by commit 4a81e8328d37 (Reduce overhead of cond_resched() checks for RCU). Just out of curiosity, how do you measure the power consumption? Thanx, Paul > git://internal_merge_and_test_tree devel-hourly-2014062315 > commit 7c66b15f8704703f6861faa246387342f7a05108 ("Merge 'rcu/rcu_cond_res= ched.2014.06.20c' into devel-hourly-2014062315") > = > Merge sequence is: > = > 7c66b15 Merge 'rcu/rcu_cond_resched.2014.06.20c' into devel-hourly-201406= 2315 > db47b74 Merge 'ipvs-next/master' into devel-hourly-2014062315 > ca8d737 Merge 'regulator/topic/ab8500' into devel-hourly-2014062315 > aee9255 Merge 'renesas/devel' into devel-hourly-2014062315 > af18816 Merge 'vireshk/tick/lowres-go-tickless' into devel-hourly-2014062= 315 > 3cf94bc1 Merge 'vireshk/tick/ONESHOT-STOPPED' into devel-hourly-2014062315 > 94ae897 Merge 'asoc/topic/tlv320aic32x4' into devel-hourly-2014062315 > 28ef1ff Merge 'tianyu/dep_support' into devel-hourly-2014062315 > f04bc00 Merge 'robclark/msm-next' into devel-hourly-2014062315 > 03f1b1e Merge 'net/master' into devel-hourly-2014062315 > a1a9b33 Merge 'spi/fix/qup' into devel-hourly-2014062315 > 45bf5db Merge 'renesas/next' into devel-hourly-2014062315 > fb20fce Merge 'asoc/for-next' into devel-hourly-2014062315 > a385be7 Merge 'asoc/topic/samsung' into devel-hourly-2014062315 > 1a9f804 0day base guard for 'devel-hourly-2014062315' > a497c3b Linux 3.16-rc2 > = > test case: brickland3/vm-scalability/300s-anon-rx-seq-mt > = > db47b74b78e1623 7c66b15f8704703f6861faa24 = > --------------- ------------------------- = > 0.73 ~ 7% -13.4% 0.64 ~ 6% TOTAL turbostat.%c1 > 1499 ~ 5% +11.6% 1672 ~ 2% TOTAL slabinfo.sock_inode_cac= he.num_objs > 1499 ~ 5% +11.6% 1672 ~ 2% TOTAL slabinfo.sock_inode_cac= he.active_objs > 4.36 ~ 1% -8.1% 4.01 ~ 1% TOTAL turbostat.%c6 > 11782 ~ 1% +8.0% 12726 ~ 2% TOTAL time.involuntary_contex= t_switches > 2848 ~ 0% +4.8% 2983 ~ 0% TOTAL time.user_time > 24.26 ~ 0% +4.4% 25.33 ~ 0% TOTAL time.elapsed_time > 402 ~ 0% +2.0% 410 ~ 0% TOTAL turbostat.Cor_W > 469 ~ 0% +1.8% 477 ~ 0% TOTAL turbostat.Pkg_W > = > c8bb7487275a9b7 7c66b15f8704703f6861faa24 > --------------- ------------------------- > 2.047e+08 ~ 0% -5.2% 1.942e+08 ~ 0% TOTAL vm-scalability.throughp= ut > 58629 ~11% +25.1% 73348 ~13% TOTAL numa-numastat.node3.loc= al_node > 58670 ~11% +25.1% 73387 ~13% TOTAL numa-numastat.node3.num= a_hit > 41 ~ 6% -12.6% 36 ~ 1% TOTAL numa-numastat.node2.oth= er_node > 493777 ~ 2% -12.7% 431077 ~ 3% TOTAL proc-vmstat.pgfault > 386476 ~ 1% -10.1% 347456 ~ 2% TOTAL proc-vmstat.pgalloc_nor= mal > 366616 ~ 2% -9.6% 331587 ~ 2% TOTAL proc-vmstat.numa_hit > 366489 ~ 2% -9.6% 331475 ~ 2% TOTAL proc-vmstat.numa_local > 11970 ~ 2% +6.3% 12726 ~ 2% TOTAL time.involuntary_contex= t_switches > 2829 ~ 0% +5.5% 2983 ~ 0% TOTAL time.user_time > 24.02 ~ 0% +5.5% 25.33 ~ 0% TOTAL time.elapsed_time > 402 ~ 0% +2.0% 410 ~ 0% TOTAL turbostat.Cor_W > 469 ~ 0% +1.7% 477 ~ 0% TOTAL turbostat.Pkg_W > = > Legend: > ~XX% - stddev percent > [+-]XX% - change percent > = > = > time.user_time > = > 3000 ++----------------------------------------------------------------= ---+ > 2980 ++O O O O O O O O O O O O O = | > | O O O O O = | > 2960 O+ = | > 2940 ++ = | > | = | > 2920 ++ = | > 2900 ++ = | > 2880 ++ = | > | = | > 2860 ++ *. .*. .*. *. .*. = | > 2840 ++ + * *.* *. + *.* *= . | > *. .*.* *.*.*.*.*..*.*.*. .*.*.*.*.*.*.* = *.* > 2820 ++*.* * = | > 2800 ++----------------------------------------------------------------= ---+ > = > = > time.elapsed_time > = > 29 ++----------------------------------------------------------------= ---+ > 28.5 O+ = | > | = | > 28 ++ = | > 27.5 ++ = | > 27 ++ = | > 26.5 ++ = | > | = | > 26 ++ = | > 25.5 ++O O O O O O O O O O O O O O O O O = | > 25 ++ O = | > 24.5 ++ = | > | .*. .*.*.*.*.*.*.*. .*.*.*.*.*= . | > 24 *+*.* * *.*.*.*.*..*.*.*.*.*.*.*.*.*.*.* = *.* > 23.5 ++----------------------------------------------------------------= ---+ > = > = > vm-scalability.throughput > = > 2.06e+08 ++*-*---------------------------------*-----------------------= ---+ > * + *.*.*.*.*.**.* *.*.*.*.*.*.* = *.| > 2.04e+08 ++ *.*. + + = + * > | **.*.*.*.*.* **.*.*.*= | > 2.02e+08 ++ = | > | = | > 2e+08 ++ = | > | = | > 1.98e+08 ++ = | > | = | > 1.96e+08 O+ = | > | O OO O = | > 1.94e+08 ++ O O O O O O O O O O O O OO = | > | = | > 1.92e+08 ++------------------------------------------------------------= ---+ > = > = > [*] bisect-good sample > [O] bisect-bad sample > = > = > Disclaimer: > Results have been estimated based on internal Intel analysis and are prov= ided > for informational purposes only. Any difference in system hardware or sof= tware > design or configuration may affect actual performance. > = > Thanks, > Fengguang > = --===============7451010360236667752==--