Content-Type: multipart/mixed; boundary="===============5415114887085602684==" MIME-Version: 1.0 From: Huang Ying To: lkp@lists.01.org Subject: Re: [sched] a15b12ac36a: -46.9% time.voluntary_context_switches +1.5% will-it-scale.per_process_ops Date: Sun, 04 Jan 2015 08:39:34 +0800 Message-ID: <1420331974.8652.2.camel@intel.com> In-Reply-To: <4750651419325033@web14h.yandex.ru> List-Id: --===============5415114887085602684== Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Hi, Kirill, Sorry for late. On Tue, 2014-12-23 at 11:57 +0300, Kirill Tkhai wrote: > Hi, Huang, > = > what do these digits mean? What test does? > = > 23.12.2014, 08:16, "Huang Ying" : > > FYI, we noticed the below changes on > > > > commit a15b12ac36ad4e7b856a4ae54937ae26a51aebad ("sched: Do not stop cp= u in set_cpus_allowed_ptr() if task is not running") > > > > testbox/testcase/testparams: lkp-g5/will-it-scale/performance-lock1 > > > > 1ba93d42727c4400 a15b12ac36ad4e7b856a4ae549 > > ---------------- -------------------------- Above is the good commit and the bad commit. > > %stddev %change %stddev > > \ | \ > > 1517261 =C2=B1 0% +1.5% 1539994 =C2=B1 0% will-it-scale.p= er_process_ops We have basic description of data above, where %stddev is standard deviation. What's more do you want? Best Regards, Huang, Ying > > 247 =C2=B1 30% +131.8% 573 =C2=B1 49% sched_debug.cpu= #61.ttwu_count > > 225 =C2=B1 22% +142.8% 546 =C2=B1 34% sched_debug.cpu= #81.ttwu_local > > 15115 =C2=B1 44% +37.3% 20746 =C2=B1 40% numa-meminfo.no= de7.Active > > 1028 =C2=B1 38% +115.3% 2214 =C2=B1 36% sched_debug.cpu= #16.ttwu_local > > 2 =C2=B1 19% +133.3% 5 =C2=B1 43% sched_debug.cpu= #89.cpu_load[3] > > 21 =C2=B1 45% +88.2% 40 =C2=B1 23% sched_debug.cfs= _rq[99]:/.tg_load_contrib > > 414 =C2=B1 33% +98.6% 823 =C2=B1 28% sched_debug.cpu= #81.ttwu_count > > 4 =C2=B1 10% +88.2% 8 =C2=B1 12% sched_debug.cfs= _rq[33]:/.runnable_load_avg > > 22 =C2=B1 26% +80.9% 40 =C2=B1 24% sched_debug.cfs= _rq[103]:/.tg_load_contrib > > 7 =C2=B1 17% -41.4% 4 =C2=B1 25% sched_debug.cfs= _rq[41]:/.load > > 7 =C2=B1 17% -37.9% 4 =C2=B1 19% sched_debug.cpu= #41.load > > 3 =C2=B1 22% +106.7% 7 =C2=B1 10% sched_debug.cfs= _rq[36]:/.runnable_load_avg > > 174 =C2=B1 13% +48.7% 259 =C2=B1 31% sched_debug.cpu= #112.ttwu_count > > 4 =C2=B1 19% +88.9% 8 =C2=B1 5% sched_debug.cfs= _rq[35]:/.runnable_load_avg > > 260 =C2=B1 10% +55.6% 405 =C2=B1 26% numa-vmstat.nod= e3.nr_anon_pages > > 1042 =C2=B1 10% +56.0% 1626 =C2=B1 26% numa-meminfo.no= de3.AnonPages > > 26 =C2=B1 22% +74.3% 45 =C2=B1 16% sched_debug.cfs= _rq[65]:/.tg_load_contrib > > 21 =C2=B1 43% +71.3% 37 =C2=B1 26% sched_debug.cfs= _rq[100]:/.tg_load_contrib > > 3686 =C2=B1 21% +40.2% 5167 =C2=B1 19% sched_debug.cpu= #16.ttwu_count > > 142 =C2=B1 9% +34.4% 191 =C2=B1 24% sched_debug.cpu= #112.ttwu_local > > 5 =C2=B1 18% +69.6% 9 =C2=B1 15% sched_debug.cfs= _rq[35]:/.load > > 2 =C2=B1 30% +100.0% 5 =C2=B1 37% sched_debug.cpu= #106.cpu_load[1] > > 3 =C2=B1 23% +100.0% 6 =C2=B1 48% sched_debug.cpu= #106.cpu_load[2] > > 5 =C2=B1 18% +69.6% 9 =C2=B1 15% sched_debug.cpu= #35.load > > 9 =C2=B1 20% +48.6% 13 =C2=B1 16% sched_debug.cfs= _rq[7]:/.runnable_load_avg > > 1727 =C2=B1 15% +43.9% 2484 =C2=B1 30% sched_debug.cpu= #34.ttwu_local > > 10 =C2=B1 17% -40.5% 6 =C2=B1 13% sched_debug.cpu= #41.cpu_load[0] > > 10 =C2=B1 14% -29.3% 7 =C2=B1 5% sched_debug.cpu= #45.cpu_load[4] > > 10 =C2=B1 17% -33.3% 7 =C2=B1 10% sched_debug.cpu= #41.cpu_load[1] > > 6121 =C2=B1 8% +56.7% 9595 =C2=B1 30% sched_debug.cpu= #13.sched_goidle > > 13 =C2=B1 8% -25.9% 10 =C2=B1 17% sched_debug.cpu= #39.cpu_load[2] > > 12 =C2=B1 16% -24.0% 9 =C2=B1 15% sched_debug.cpu= #37.cpu_load[2] > > 492 =C2=B1 17% -21.3% 387 =C2=B1 24% sched_debug.cpu= #46.ttwu_count > > 3761 =C2=B1 11% -23.9% 2863 =C2=B1 15% sched_debug.cpu= #93.curr->pid > > 570 =C2=B1 19% +43.2% 816 =C2=B1 17% sched_debug.cpu= #86.ttwu_count > > 5279 =C2=B1 8% +63.5% 8631 =C2=B1 33% sched_debug.cpu= #13.ttwu_count > > 377 =C2=B1 22% -28.6% 269 =C2=B1 14% sched_debug.cpu= #46.ttwu_local > > 5396 =C2=B1 10% +29.9% 7007 =C2=B1 14% sched_debug.cpu= #16.sched_goidle > > 1959 =C2=B1 12% +36.9% 2683 =C2=B1 15% numa-vmstat.nod= e2.nr_slab_reclaimable > > 7839 =C2=B1 12% +37.0% 10736 =C2=B1 15% numa-meminfo.no= de2.SReclaimable > > 5 =C2=B1 15% +66.7% 8 =C2=B1 9% sched_debug.cfs= _rq[33]:/.load > > 5 =C2=B1 25% +47.8% 8 =C2=B1 10% sched_debug.cfs= _rq[37]:/.load > > 2 =C2=B1 0% +87.5% 3 =C2=B1 34% sched_debug.cpu= #89.cpu_load[4] > > 5 =C2=B1 15% +66.7% 8 =C2=B1 9% sched_debug.cpu= #33.load > > 6 =C2=B1 23% +41.7% 8 =C2=B1 10% sched_debug.cpu= #37.load > > 8 =C2=B1 10% -26.5% 6 =C2=B1 6% sched_debug.cpu= #51.cpu_load[1] > > 7300 =C2=B1 37% +63.6% 11943 =C2=B1 16% softirqs.TASKLET > > 2984 =C2=B1 6% +43.1% 4271 =C2=B1 23% sched_debug.cpu= #20.ttwu_count > > 328 =C2=B1 4% +40.5% 462 =C2=B1 25% sched_debug.cpu= #26.ttwu_local > > 10 =C2=B1 7% -27.5% 7 =C2=B1 5% sched_debug.cpu= #43.cpu_load[3] > > 9 =C2=B1 8% -30.8% 6 =C2=B1 6% sched_debug.cpu= #41.cpu_load[3] > > 9 =C2=B1 8% -27.0% 6 =C2=B1 6% sched_debug.cpu= #41.cpu_load[4] > > 10 =C2=B1 14% -32.5% 6 =C2=B1 6% sched_debug.cpu= #41.cpu_load[2] > > 16292 =C2=B1 6% +42.8% 23260 =C2=B1 25% sched_debug.cpu= #13.nr_switches > > 14 =C2=B1 28% +55.9% 23 =C2=B1 8% sched_debug.cpu= #99.cpu_load[0] > > 5 =C2=B1 8% +28.6% 6 =C2=B1 12% sched_debug.cpu= #17.load > > 13 =C2=B1 7% -23.1% 10 =C2=B1 12% sched_debug.cpu= #39.cpu_load[3] > > 7 =C2=B1 10% -35.7% 4 =C2=B1 11% sched_debug.cfs= _rq[45]:/.runnable_load_avg > > 5076 =C2=B1 13% -21.8% 3970 =C2=B1 11% numa-vmstat.nod= e0.nr_slab_unreclaimable > > 20306 =C2=B1 13% -21.8% 15886 =C2=B1 11% numa-meminfo.no= de0.SUnreclaim > > 10 =C2=B1 10% -28.6% 7 =C2=B1 6% sched_debug.cpu= #45.cpu_load[3] > > 11 =C2=B1 11% -29.5% 7 =C2=B1 14% sched_debug.cpu= #45.cpu_load[1] > > 10 =C2=B1 12% -26.8% 7 =C2=B1 6% sched_debug.cpu= #44.cpu_load[1] > > 10 =C2=B1 10% -28.6% 7 =C2=B1 6% sched_debug.cpu= #44.cpu_load[0] > > 7 =C2=B1 17% +48.3% 10 =C2=B1 7% sched_debug.cfs= _rq[11]:/.runnable_load_avg > > 11 =C2=B1 12% -34.1% 7 =C2=B1 11% sched_debug.cpu= #47.cpu_load[0] > > 10 =C2=B1 10% -27.9% 7 =C2=B1 5% sched_debug.cpu= #47.cpu_load[1] > > 10 =C2=B1 8% -26.8% 7 =C2=B1 11% sched_debug.cpu= #47.cpu_load[2] > > 10 =C2=B1 8% -28.6% 7 =C2=B1 14% sched_debug.cpu= #43.cpu_load[0] > > 10 =C2=B1 10% -27.9% 7 =C2=B1 10% sched_debug.cpu= #43.cpu_load[1] > > 10 =C2=B1 10% -28.6% 7 =C2=B1 6% sched_debug.cpu= #43.cpu_load[2] > > 12940 =C2=B1 3% +49.8% 19387 =C2=B1 35% numa-meminfo.no= de2.Active(anon) > > 3235 =C2=B1 2% +49.8% 4844 =C2=B1 35% numa-vmstat.nod= e2.nr_active_anon > > 17 =C2=B1 17% +36.6% 24 =C2=B1 9% sched_debug.cpu= #97.cpu_load[2] > > 14725 =C2=B1 8% +21.8% 17928 =C2=B1 11% sched_debug.cpu= #16.nr_switches > > 667 =C2=B1 10% +45.3% 969 =C2=B1 22% sched_debug.cpu= #17.ttwu_local > > 3257 =C2=B1 5% +22.4% 3988 =C2=B1 11% sched_debug.cpu= #118.curr->pid > > 3144 =C2=B1 15% -20.7% 2493 =C2=B1 8% sched_debug.cpu= #95.curr->pid > > 2192 =C2=B1 11% +50.9% 3308 =C2=B1 37% sched_debug.cpu= #18.ttwu_count > > 6 =C2=B1 11% +37.5% 8 =C2=B1 19% sched_debug.cfs= _rq[22]:/.load > > 12 =C2=B1 5% +27.1% 15 =C2=B1 8% sched_debug.cpu= #5.cpu_load[1] > > 11 =C2=B1 12% -23.4% 9 =C2=B1 13% sched_debug.cpu= #37.cpu_load[3] > > 6 =C2=B1 11% +37.5% 8 =C2=B1 19% sched_debug.cpu= #22.load > > 8 =C2=B1 8% -25.0% 6 =C2=B1 0% sched_debug.cpu= #51.cpu_load[2] > > 7 =C2=B1 6% -20.0% 6 =C2=B1 11% sched_debug.cpu= #55.cpu_load[3] > > 11 =C2=B1 9% -17.4% 9 =C2=B1 9% sched_debug.cpu= #39.cpu_load[4] > > 12 =C2=B1 5% -22.9% 9 =C2=B1 11% sched_debug.cpu= #38.cpu_load[3] > > 420 =C2=B1 13% +43.0% 601 =C2=B1 9% sched_debug.cpu= #30.ttwu_local > > 1682 =C2=B1 14% +38.5% 2329 =C2=B1 17% numa-meminfo.no= de7.AnonPages > > 423 =C2=B1 13% +37.0% 579 =C2=B1 16% numa-vmstat.nod= e7.nr_anon_pages > > 15 =C2=B1 13% +41.9% 22 =C2=B1 5% sched_debug.cpu= #99.cpu_load[1] > > 6 =C2=B1 20% +44.0% 9 =C2=B1 13% sched_debug.cfs= _rq[19]:/.runnable_load_avg > > 9 =C2=B1 4% -24.3% 7 =C2=B1 0% sched_debug.cpu= #43.cpu_load[4] > > 6341 =C2=B1 7% -19.6% 5100 =C2=B1 16% sched_debug.cpu= #43.curr->pid > > 2577 =C2=B1 11% -11.9% 2270 =C2=B1 10% sched_debug.cpu= #33.ttwu_count > > 13 =C2=B1 6% -18.5% 11 =C2=B1 12% sched_debug.cpu= #40.cpu_load[2] > > 4828 =C2=B1 6% +23.8% 5979 =C2=B1 6% sched_debug.cpu= #34.curr->pid > > 4351 =C2=B1 12% +33.9% 5824 =C2=B1 12% sched_debug.cpu= #36.curr->pid > > 10 =C2=B1 8% -23.8% 8 =C2=B1 8% sched_debug.cpu= #37.cpu_load[4] > > 10 =C2=B1 14% -28.6% 7 =C2=B1 6% sched_debug.cpu= #45.cpu_load[2] > > 17 =C2=B1 22% +40.6% 24 =C2=B1 7% sched_debug.cpu= #97.cpu_load[1] > > 11 =C2=B1 9% +21.3% 14 =C2=B1 5% sched_debug.cpu= #7.cpu_load[2] > > 10 =C2=B1 8% -26.2% 7 =C2=B1 10% sched_debug.cpu= #36.cpu_load[4] > > 12853 =C2=B1 2% +20.0% 15429 =C2=B1 11% numa-meminfo.no= de2.AnonPages > > 4744 =C2=B1 8% +30.8% 6204 =C2=B1 11% sched_debug.cpu= #35.curr->pid > > 3214 =C2=B1 2% +20.0% 3856 =C2=B1 11% numa-vmstat.nod= e2.nr_anon_pages > > 6181 =C2=B1 6% +24.9% 7718 =C2=B1 9% sched_debug.cpu= #13.curr->pid > > 6675 =C2=B1 23% +27.5% 8510 =C2=B1 10% sched_debug.cfs= _rq[91]:/.tg_load_avg > > 171261 =C2=B1 5% -22.2% 133177 =C2=B1 15% numa-numastat.n= ode0.local_node > > 6589 =C2=B1 21% +29.3% 8522 =C2=B1 11% sched_debug.cfs= _rq[89]:/.tg_load_avg > > 6508 =C2=B1 20% +28.0% 8331 =C2=B1 8% sched_debug.cfs= _rq[88]:/.tg_load_avg > > 6598 =C2=B1 22% +29.2% 8525 =C2=B1 11% sched_debug.cfs= _rq[90]:/.tg_load_avg > > 590 =C2=B1 13% -21.4% 464 =C2=B1 7% sched_debug.cpu= #105.ttwu_local > > 175392 =C2=B1 5% -21.7% 137308 =C2=B1 14% numa-numastat.n= ode0.numa_hit > > 11 =C2=B1 6% -18.2% 9 =C2=B1 7% sched_debug.cpu= #38.cpu_load[4] > > 6643 =C2=B1 23% +27.4% 8465 =C2=B1 10% sched_debug.cfs= _rq[94]:/.tg_load_avg > > 6764 =C2=B1 7% +13.8% 7695 =C2=B1 7% sched_debug.cpu= #12.curr->pid > > 29 =C2=B1 28% +34.5% 39 =C2=B1 5% sched_debug.cfs= _rq[98]:/.tg_load_contrib > > 1776 =C2=B1 7% +29.4% 2298 =C2=B1 13% sched_debug.cpu= #11.ttwu_local > > 13 =C2=B1 0% -19.2% 10 =C2=B1 8% sched_debug.cpu= #40.cpu_load[3] > > 7 =C2=B1 5% -17.2% 6 =C2=B1 0% sched_debug.cpu= #51.cpu_load[3] > > 7371 =C2=B1 20% -18.0% 6045 =C2=B1 3% sched_debug.cpu= #1.sched_goidle > > 26560 =C2=B1 2% +14.0% 30287 =C2=B1 7% numa-meminfo.no= de2.Slab > > 16161 =C2=B1 6% -9.4% 14646 =C2=B1 1% sched_debug.cfs= _rq[27]:/.avg->runnable_avg_sum > > 351 =C2=B1 6% -9.3% 318 =C2=B1 1% sched_debug.cfs= _rq[27]:/.tg_runnable_contrib > > 7753 =C2=B1 27% -22.9% 5976 =C2=B1 5% sched_debug.cpu= #2.sched_goidle > > 3828 =C2=B1 9% +17.3% 4490 =C2=B1 6% sched_debug.cpu= #23.sched_goidle > > 23925 =C2=B1 2% +23.0% 29419 =C2=B1 23% numa-meminfo.no= de2.Active > > 47 =C2=B1 6% -15.8% 40 =C2=B1 19% sched_debug.cpu= #42.cpu_load[1] > > 282 =C2=B1 5% -9.7% 254 =C2=B1 7% sched_debug.cfs= _rq[109]:/.tg_runnable_contrib > > 349 =C2=B1 5% -9.3% 317 =C2=B1 1% sched_debug.cfs= _rq[26]:/.tg_runnable_contrib > > 6941 =C2=B1 3% +8.9% 7558 =C2=B1 7% sched_debug.cpu= #61.nr_switches > > 16051 =C2=B1 5% -8.9% 14618 =C2=B1 1% sched_debug.cfs= _rq[26]:/.avg->runnable_avg_sum > > 238944 =C2=B1 3% +9.2% 260958 =C2=B1 5% numa-vmstat.nod= e2.numa_local > > 12966 =C2=B1 5% -9.5% 11732 =C2=B1 6% sched_debug.cfs= _rq[109]:/.avg->runnable_avg_sum > > 1004 =C2=B1 3% +8.2% 1086 =C2=B1 4% sched_debug.cpu= #118.sched_goidle > > 20746 =C2=B1 4% -8.4% 19000 =C2=B1 1% sched_debug.cfs= _rq[45]:/.avg->runnable_avg_sum > > 451 =C2=B1 4% -8.3% 413 =C2=B1 1% sched_debug.cfs= _rq[45]:/.tg_runnable_contrib > > 3538 =C2=B1 4% +17.2% 4147 =C2=B1 8% sched_debug.cpu= #26.ttwu_count > > 16 =C2=B1 9% +13.8% 18 =C2=B1 2% sched_debug.cpu= #99.cpu_load[3] > > 1531 =C2=B1 0% +11.3% 1704 =C2=B1 1% numa-meminfo.no= de7.KernelStack > > 3569 =C2=B1 3% +17.2% 4182 =C2=B1 10% sched_debug.cpu= #24.sched_goidle > > 1820 =C2=B1 3% -12.5% 1594 =C2=B1 8% slabinfo.taskst= ats.num_objs > > 1819 =C2=B1 3% -12.4% 1594 =C2=B1 8% slabinfo.taskst= ats.active_objs > > 4006 =C2=B1 5% +19.1% 4769 =C2=B1 8% sched_debug.cpu= #17.sched_goidle > > 21412 =C2=B1 19% -17.0% 17779 =C2=B1 3% sched_debug.cpu= #2.nr_switches > > 16 =C2=B1 9% +24.2% 20 =C2=B1 4% sched_debug.cpu= #99.cpu_load[2] > > 10493 =C2=B1 7% +13.3% 11890 =C2=B1 4% sched_debug.cpu= #23.nr_switches > > 1207 =C2=B1 2% -46.9% 640 =C2=B1 4% time.voluntary_= context_switches > > > > time.voluntary_context_switches > > > > 1300 ++-----------*--*--------------------*--------------------------= -----+ > > *..*.*..*.. + *.*..*..*.*..*..* .*..*..*. .*..*.*..*..= | > > 1200 ++ * * *. = *.*..* > > 1100 ++ = | > > | = | > > 1000 ++ = | > > | = | > > 900 ++ = | > > | = | > > 800 ++ = | > > 700 ++ = | > > O O O O O O O O O O O O O = | > > 600 ++ O O O O O O O O O = | > > | O = | > > 500 ++--------------------------------------------------------------= -----+ > > > > [*] 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 pr= ovided > > for informational purposes only. Any difference in system hardware or s= oftware > > design or configuration may affect actual performance. > > > = > Regards, > Kirill --===============5415114887085602684==--