From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752462AbbCYDDx (ORCPT ); Tue, 24 Mar 2015 23:03:53 -0400 Received: from cantor2.suse.de ([195.135.220.15]:36964 "EHLO mx2.suse.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752243AbbCYDDu (ORCPT ); Tue, 24 Mar 2015 23:03:50 -0400 Date: Wed, 25 Mar 2015 14:03:59 +1100 From: NeilBrown To: Yuanahn Liu Cc: lkp@01.org, lkp@linux.intel.com, LKML , Chris Mason Subject: Re: performance changes on 4400755e: 200.0% fsmark.files_per_sec, -18.1% fsmark.files_per_sec, and few more Message-ID: <20150325140359.37763fd8@notabene.brown> In-Reply-To: <20150318050030.GC30894@yliu-dev.sh.intel.com> References: <20150318050030.GC30894@yliu-dev.sh.intel.com> X-Mailer: Claws Mail 3.10.1-162-g4d0ed6 (GTK+ 2.24.25; x86_64-suse-linux-gnu) MIME-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha1; boundary="Sig_/9QIveOZ39dGk4bkgR_Z=MYI"; protocol="application/pgp-signature" Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org --Sig_/9QIveOZ39dGk4bkgR_Z=MYI Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable On Wed, 18 Mar 2015 13:00:30 +0800 Yuanahn Liu wrote: > Hi, >=20 > FYI, we noticed performance changes on `fsmark.files_per_sec' by 4400755e= 356f9a2b0b7ceaa02f57b1c7546c3765: >=20 > > commit 4400755e356f9a2b0b7ceaa02f57b1c7546c3765 > > Author: NeilBrown > > AuthorDate: Thu Feb 26 12:47:56 2015 +1100 > > Commit: NeilBrown > > CommitDate: Wed Mar 4 13:40:19 2015 +1100 > >=20 > > md/raid5: allow the stripe_cache to grow and shrink. Thanks a lot for this testing!!! I was wondering how I could do some proper testing of this patch, and you've done it for me :-) The large number of improvements is very encouraging - that is what I was hoping for of course. The few regressions could be a concern. I note that are all NoSync. That seems to suggest that they could just be writing more data. i.e. the data is written a bit earlier (certainly possible) so it happen to introduce more delay .... I guess I'm not really sure how to interpret NoSync results, and suspect th= at poor NoSync result don't really reflect much on the underlying block device. Could that be right? Also, I'm a little confused by the fsmark.time.involuntary_context_switches statistic: > 1235 =C2=B1 2% -47.5% 649 =C2=B1 3% fsmark.time.perce= nt_of_cpu_this_job_got > 399 =C2=B1 4% -20.0% 319 =C2=B1 3% fsmark.time.perce= nt_of_cpu_this_job_got Does that means that the ext4 test changed from 12.4 cpus to 6.4, and that the btrfs test chnages from 4 cpus to 3.2 ??? Or does it just not mean anything? Thanks, NeilBrown >=20 > 26089f4902595a2f64c512066af07af6e82eb096 4400755e356f9a2b0b7ceaa02f57= b1c7546c3765 > ---------------------------------------- ----------------------------= ------------ > run time(m) metric_value =C2=B1stddev run time(m) metric_= value =C2=B1stddev change testbox/benchmark/sub-testcase > --- ------ ---------------------------- --- ------ ----------------= ------------ -------- ------------------------------ > 3 18.6 6.400 =C2=B10.0% 5 9.2 1= 9.200 =C2=B10.0% 200.0% ivb44/fsmark/1x-64t-3HDD-RAID5-ext4-4M-= 40G-fsyncBeforeClose > 3 24.7 6.400 =C2=B10.0% 3 13.7 1= 2.800 =C2=B10.0% 100.0% ivb44/fsmark/1x-64t-3HDD-RAID5-f2fs-4M-= 40G-fsyncBeforeClose > 3 17.5 28.267 =C2=B19.6% 3 12.3 4= 2.833 =C2=B16.5% 51.5% ivb44/fsmark/1x-64t-3HDD-RAID5-f2fs-4M-= 120G-NoSync > 3 16.7 30.700 =C2=B11.5% 3 12.6 4= 0.733 =C2=B12.4% 32.7% ivb44/fsmark/1x-1t-3HDD-RAID5-f2fs-4M-1= 20G-NoSync > 3 29.0 5.867 =C2=B10.8% 5 23.6 = 7.240 =C2=B10.7% 23.4% ivb44/fsmark/1x-1t-3HDD-RAID5-btrfs-4M-= 40G-fsyncBeforeClose > 3 28.5 6.000 =C2=B10.0% 3 23.2 = 7.367 =C2=B10.6% 22.8% ivb44/fsmark/1x-1t-3HDD-RAID5-f2fs-4M-4= 0G-fsyncBeforeClose > 5 11.7 14.600 =C2=B10.0% 5 9.7 1= 7.500 =C2=B10.4% 19.9% ivb44/fsmark/1x-1t-3HDD-RAID5-xfs-4M-40= G-fsyncBeforeClose > 3 22.4 25.600 =C2=B10.0% 5 17.9 3= 0.120 =C2=B14.1% 17.7% ivb44/fsmark/1x-64t-3HDD-RAID5-xfs-4M-1= 20G-NoSync > 5 10.8 47.320 =C2=B10.6% 5 9.3 5= 4.820 =C2=B10.2% 15.8% ivb44/fsmark/1x-1t-3HDD-RAID5-xfs-4M-12= 0G-NoSync > 1 0.5 252.400 =C2=B10.0% 1 0.5 26= 3.300 =C2=B10.0% 4.3% ivb44/fsmark/1x-1t-4BRD_12G-RAID5-ext4-= 4M-30G-NoSync >=20 > 3 0.5 273.100 =C2=B14.3% 3 0.6 22= 3.567 =C2=B16.5% -18.1% ivb44/fsmark/1x-64t-4BRD_12G-RAID5-btrf= s-4M-30G-NoSync > 3 8.1 63.133 =C2=B10.5% 3 9.2 5= 5.633 =C2=B10.2% -11.9% ivb44/fsmark/1x-1t-3HDD-RAID5-btrfs-4M-= 120G-NoSync > 3 8.2 64.000 =C2=B10.0% 3 9.2 5= 7.600 =C2=B10.0% -10.0% ivb44/fsmark/1x-64t-3HDD-RAID5-btrfs-4M= -120G-NoSync >=20 >=20 > NOTE: here are some more info about those test parameters for you to > understand the testcase better: >=20 > 1x: where 'x' means iterations or loop, corresponding to the 'L' op= tion of fsmark > 1t, 64t: where 't' means thread > 4M: means the single file size, corresponding to the '-s' option of= fsmark > 40G, 30G, 120G: means the total test size >=20 > 4BRD_12G: BRD is the ramdisk, where '4' means 4 ramdisk, and where = '12G' means > the size of one ramdisk. So, it would be 48G in total. An= d we made a > raid on those ramdisk. >=20 >=20 > As you can see from above data, interestingly, all performance > regressions come from btrfs testing. That's why Chris is also > in the cc list, with which just FYI. >=20 >=20 > FYI, here I listed more detailed changes for the maximal postive and negt= ive changes. >=20 > more detailed changes about ivb44/fsmark/1x-64t-3HDD-RAID5-ext4-4M-40G-fs= yncBeforeClose > --------- >=20 > 26089f4902595a2f 4400755e356f9a2b0b7ceaa02f =20 > ---------------- -------------------------- =20 > %stddev %change %stddev > \ | \ =20 > 6.40 =C2=B1 0% +200.0% 19.20 =C2=B1 0% fsmark.files_per_= sec > 1.015e+08 =C2=B1 1% -73.6% 26767355 =C2=B1 3% fsmark.time.volun= tary_context_switches > 13793 =C2=B1 1% -73.9% 3603 =C2=B1 5% fsmark.time.syste= m_time > 78473 =C2=B1 6% -64.3% 28016 =C2=B1 7% fsmark.time.invol= untary_context_switches > 15789555 =C2=B1 9% -54.7% 7159485 =C2=B1 13% fsmark.app_overhe= ad > 1115 =C2=B1 0% -50.3% 554 =C2=B1 1% fsmark.time.elaps= ed_time.max > 1115 =C2=B1 0% -50.3% 554 =C2=B1 1% fsmark.time.elaps= ed_time > 1235 =C2=B1 2% -47.5% 649 =C2=B1 3% fsmark.time.perce= nt_of_cpu_this_job_got > 456465 =C2=B1 1% -26.7% 334594 =C2=B1 4% fsmark.time.minor= _page_faults > 275 =C2=B1 0% +1257.7% 3733 =C2=B1 2% slabinfo.raid5-md= 0.num_objs > 275 =C2=B1 0% +1257.7% 3733 =C2=B1 2% slabinfo.raid5-md= 0.active_objs > 11 =C2=B1 0% +1250.9% 148 =C2=B1 2% slabinfo.raid5-md= 0.active_slabs > 11 =C2=B1 0% +1250.9% 148 =C2=B1 2% slabinfo.raid5-md= 0.num_slabs > 2407 =C2=B1 4% +293.4% 9471 =C2=B1 26% numa-meminfo.node= 0.Writeback > 600 =C2=B1 4% +294.9% 2372 =C2=B1 26% numa-vmstat.node0= .nr_writeback > 1114505 =C2=B1 0% -77.4% 251696 =C2=B1 2% softirqs.TASKLET > 1808027 =C2=B1 1% -77.7% 402378 =C2=B1 4% softirqs.RCU > 12158665 =C2=B1 1% -77.1% 2786069 =C2=B1 4% cpuidle.C3-IVT.us= age > 1119433 =C2=B1 0% -77.3% 254192 =C2=B1 2% softirqs.BLOCK > 37824202 =C2=B1 1% -75.1% 9405078 =C2=B1 4% cpuidle.C6-IVT.us= age > 1.015e+08 =C2=B1 1% -73.6% 26767355 =C2=B1 3% time.voluntary_co= ntext_switches > 13793 =C2=B1 1% -73.9% 3603 =C2=B1 5% time.system_time > 5971084 =C2=B1 1% -73.6% 1574912 =C2=B1 5% softirqs.SCHED > 10539492 =C2=B1 3% -72.0% 2956258 =C2=B1 6% cpuidle.C1E-IVT.u= sage > 2 =C2=B1 0% +230.0% 6 =C2=B1 12% vmstat.procs.b > 14064 =C2=B1 1% -71.2% 4049 =C2=B1 6% softirqs.HRTIMER > 7388306 =C2=B1 1% -71.2% 2129929 =C2=B1 4% softirqs.TIMER > 3.496e+09 =C2=B1 1% -70.3% 1.04e+09 =C2=B1 1% cpuidle.C3-IVT.ti= me > 0.88 =C2=B1 6% +224.9% 2.87 =C2=B1 11% turbostat.Pkg%pc6 > 19969464 =C2=B1 2% -66.2% 6750675 =C2=B1 5% cpuidle.C1-IVT.us= age > 78473 =C2=B1 6% -64.3% 28016 =C2=B1 7% time.involuntary_= context_switches > 4.23 =C2=B1 5% +181.4% 11.90 =C2=B1 3% turbostat.Pkg%pc2 > 2.551e+09 =C2=B1 1% -61.4% 9.837e+08 =C2=B1 3% cpuidle.C1E-IVT.t= ime > 8084 =C2=B1 3% +142.6% 19608 =C2=B1 3% meminfo.Writeback > 2026 =C2=B1 4% +141.6% 4895 =C2=B1 4% proc-vmstat.nr_wr= iteback > 165 =C2=B1 4% -56.9% 71 =C2=B1 14% numa-vmstat.node1= .nr_inactive_anon > 7.748e+09 =C2=B1 3% -50.3% 3.852e+09 =C2=B1 3% cpuidle.C1-IVT.ti= me > 175 =C2=B1 5% -53.2% 82 =C2=B1 13% numa-vmstat.node1= .nr_shmem > 1115 =C2=B1 0% -50.3% 554 =C2=B1 1% time.elapsed_time= .max > 1115 =C2=B1 0% -50.3% 554 =C2=B1 1% time.elapsed_time > 1147 =C2=B1 0% -49.0% 585 =C2=B1 1% uptime.boot > 2260889 =C2=B1 0% -48.8% 1157272 =C2=B1 1% proc-vmstat.pgfree > 16805 =C2=B1 2% -35.9% 10776 =C2=B1 23% numa-vmstat.node1= .nr_dirty > 1235 =C2=B1 2% -47.5% 649 =C2=B1 3% time.percent_of_c= pu_this_job_got > 67245 =C2=B1 2% -35.9% 43122 =C2=B1 23% numa-meminfo.node= 1.Dirty > 39041 =C2=B1 0% -45.7% 21212 =C2=B1 2% uptime.idle > 13 =C2=B1 9% -49.0% 6 =C2=B1 11% vmstat.procs.r > 3072 =C2=B1 10% -40.3% 1833 =C2=B1 9% cpuidle.POLL.usage > 3045115 =C2=B1 0% -46.1% 1642053 =C2=B1 1% proc-vmstat.pgfau= lt > 202 =C2=B1 1% -45.2% 110 =C2=B1 0% proc-vmstat.nr_in= active_anon > 4583079 =C2=B1 2% -31.4% 3143602 =C2=B1 16% numa-vmstat.node1= .numa_hit > 28.03 =C2=B1 0% +69.1% 47.39 =C2=B1 1% turbostat.CPU%c6 > 223 =C2=B1 1% -41.1% 131 =C2=B1 1% proc-vmstat.nr_sh= mem > 4518820 =C2=B1 3% -30.8% 3128304 =C2=B1 16% numa-vmstat.node1= .numa_local > 3363496 =C2=B1 3% -27.4% 2441619 =C2=B1 20% numa-vmstat.node1= .nr_dirtied > 3345346 =C2=B1 3% -27.4% 2428396 =C2=B1 20% numa-vmstat.node1= .nr_written > 0.18 =C2=B1 18% +105.6% 0.37 =C2=B1 36% turbostat.Pkg%pc3 > 3427913 =C2=B1 3% -27.3% 2492563 =C2=B1 20% numa-vmstat.node1= .nr_inactive_file > 13712431 =C2=B1 3% -27.3% 9971152 =C2=B1 20% numa-meminfo.node= 1.Inactive > 13711768 =C2=B1 3% -27.3% 9970866 =C2=B1 20% numa-meminfo.node= 1.Inactive(file) > 3444598 =C2=B1 3% -27.2% 2508920 =C2=B1 20% numa-vmstat.node1= .nr_file_pages > 13778510 =C2=B1 3% -27.2% 10036287 =C2=B1 20% numa-meminfo.node= 1.FilePages > 8819175 =C2=B1 1% -28.3% 6320188 =C2=B1 19% numa-numastat.nod= e1.numa_hit > 8819051 =C2=B1 1% -28.3% 6320152 =C2=B1 19% numa-numastat.nod= e1.local_node > 14350918 =C2=B1 3% -26.8% 10504070 =C2=B1 19% numa-meminfo.node= 1.MemUsed > 100892 =C2=B1 3% -26.0% 74623 =C2=B1 19% numa-vmstat.node1= .nr_slab_reclaimable > 403571 =C2=B1 3% -26.0% 298513 =C2=B1 19% numa-meminfo.node= 1.SReclaimable > 3525 =C2=B1 13% +36.6% 4817 =C2=B1 14% slabinfo.blkdev_r= equests.active_objs > 3552 =C2=B1 13% +36.3% 4841 =C2=B1 14% slabinfo.blkdev_r= equests.num_objs > 30779 =C2=B1 4% -34.7% 20084 =C2=B1 12% proc-vmstat.pgmig= rate_success > 30779 =C2=B1 4% -34.7% 20084 =C2=B1 12% proc-vmstat.numa_= pages_migrated > 447400 =C2=B1 2% -23.2% 343701 =C2=B1 16% numa-meminfo.node= 1.Slab > 2.532e+10 =C2=B1 0% -33.1% 1.694e+10 =C2=B1 1% cpuidle.C6-IVT.ti= me > 3081 =C2=B1 9% +28.0% 3945 =C2=B1 12% slabinfo.mnt_cach= e.num_objs > 3026 =C2=B1 9% +28.8% 3898 =C2=B1 12% slabinfo.mnt_cach= e.active_objs > 5822 =C2=B1 4% +77.8% 10350 =C2=B1 25% numa-meminfo.node= 1.Writeback > 1454 =C2=B1 4% +77.3% 2579 =C2=B1 25% numa-vmstat.node1= .nr_writeback > 424984 =C2=B1 1% -26.5% 312255 =C2=B1 3% proc-vmstat.numa_= pte_updates > 368001 =C2=B1 1% -26.8% 269440 =C2=B1 3% proc-vmstat.numa_= hint_faults > 456465 =C2=B1 1% -26.7% 334594 =C2=B1 4% time.minor_page_f= aults > 3.86 =C2=B1 3% -24.4% 2.92 =C2=B1 2% turbostat.CPU%c3 > 4661151 =C2=B1 2% +20.6% 5622999 =C2=B1 9% numa-vmstat.node1= .nr_free_pages > 18644452 =C2=B1 2% +20.6% 22491300 =C2=B1 9% numa-meminfo.node= 1.MemFree > 876 =C2=B1 2% +28.2% 1124 =C2=B1 5% slabinfo.kmalloc-= 4096.num_objs > 858 =C2=B1 3% +24.0% 1064 =C2=B1 5% slabinfo.kmalloc-= 4096.active_objs > 17767832 =C2=B1 8% -25.4% 13249545 =C2=B1 17% cpuidle.POLL.time > 285093 =C2=B1 1% -23.1% 219372 =C2=B1 5% proc-vmstat.numa_= hint_faults_local > 105423 =C2=B1 2% -16.1% 88498 =C2=B1 0% meminfo.Dirty > 26365 =C2=B1 1% -16.0% 22152 =C2=B1 1% proc-vmstat.nr_di= rty > 41.04 =C2=B1 1% -14.1% 35.26 =C2=B1 1% turbostat.CPU%c1 > 9385 =C2=B1 4% -14.3% 8043 =C2=B1 6% slabinfo.kmalloc-= 192.active_objs > 9574 =C2=B1 3% -13.9% 8241 =C2=B1 6% slabinfo.kmalloc-= 192.num_objs > 2411 =C2=B1 3% +17.0% 2820 =C2=B1 4% slabinfo.kmalloc-= 2048.active_objs > 12595574 =C2=B1 0% -10.0% 11338368 =C2=B1 1% proc-vmstat.pgall= oc_normal > 5262 =C2=B1 1% +13.3% 5962 =C2=B1 1% slabinfo.kmalloc-= 1024.num_objs > 5262 =C2=B1 1% +12.7% 5932 =C2=B1 1% slabinfo.kmalloc-= 1024.active_objs > 2538 =C2=B1 3% +13.7% 2885 =C2=B1 4% slabinfo.kmalloc-= 2048.num_objs > 5299546 =C2=B1 0% -9.9% 4776351 =C2=B1 0% slabinfo.buffer_h= ead.active_objs > 5299546 =C2=B1 0% -9.9% 4776351 =C2=B1 0% slabinfo.buffer_h= ead.num_objs > 135885 =C2=B1 0% -9.9% 122470 =C2=B1 0% slabinfo.buffer_h= ead.num_slabs > 135885 =C2=B1 0% -9.9% 122470 =C2=B1 0% slabinfo.buffer_h= ead.active_slabs > 28.04 =C2=B1 2% +715.6% 228.69 =C2=B1 3% iostat.sdb.avgrq-= sz > 28.05 =C2=B1 2% +708.1% 226.72 =C2=B1 2% iostat.sdc.avgrq-= sz > 2245 =C2=B1 3% -81.6% 413 =C2=B1 1% iostat.sda.w/s > 5.33 =C2=B1 1% +1008.2% 59.07 =C2=B1 1% iostat.sda.w_await > 5.85 =C2=B1 1% +1126.4% 71.69 =C2=B1 4% iostat.sda.r_await > 5.36 =C2=B1 1% +978.6% 57.79 =C2=B1 3% iostat.sdc.w_await > 1263 =C2=B1 4% -85.8% 179 =C2=B1 6% iostat.sdc.r/s > 2257 =C2=B1 3% -81.6% 414 =C2=B1 2% iostat.sdb.w/s > 1264 =C2=B1 4% -85.8% 179 =C2=B1 6% iostat.sdb.r/s > 5.55 =C2=B1 0% +1024.2% 62.37 =C2=B1 4% iostat.sdb.await > 5.89 =C2=B1 1% +1125.9% 72.16 =C2=B1 6% iostat.sdb.r_await > 5.36 =C2=B1 0% +1014.3% 59.75 =C2=B1 3% iostat.sdb.w_await > 5.57 =C2=B1 1% +987.9% 60.55 =C2=B1 3% iostat.sdc.await > 5.51 =C2=B1 0% +1017.3% 61.58 =C2=B1 1% iostat.sda.await > 1264 =C2=B1 4% -85.8% 179 =C2=B1 6% iostat.sda.r/s > 28.09 =C2=B1 2% +714.2% 228.73 =C2=B1 2% iostat.sda.avgrq-= sz > 5.95 =C2=B1 2% +1091.0% 70.82 =C2=B1 6% iostat.sdc.r_await > 2252 =C2=B1 3% -81.5% 417 =C2=B1 2% iostat.sdc.w/s > 4032 =C2=B1 2% +151.6% 10143 =C2=B1 1% iostat.sdb.wrqm/s > 4043 =C2=B1 2% +151.0% 10150 =C2=B1 1% iostat.sda.wrqm/s > 4035 =C2=B1 2% +151.2% 10138 =C2=B1 1% iostat.sdc.wrqm/s > 26252 =C2=B1 1% -54.0% 12077 =C2=B1 4% vmstat.system.in > 37813 =C2=B1 0% +101.0% 75998 =C2=B1 1% vmstat.io.bo > 37789 =C2=B1 0% +101.0% 75945 =C2=B1 1% iostat.md0.wkB/s > 205 =C2=B1 0% +96.1% 402 =C2=B1 1% iostat.md0.w/s > 164286 =C2=B1 1% -46.2% 88345 =C2=B1 2% vmstat.system.cs > 27.07 =C2=B1 2% -46.7% 14.42 =C2=B1 3% turbostat.%Busy > 810 =C2=B1 2% -46.7% 431 =C2=B1 3% turbostat.Avg_MHz > 15.56 =C2=B1 2% +71.7% 26.71 =C2=B1 1% iostat.sda.avgqu-= sz > 15.65 =C2=B1 2% +69.1% 26.46 =C2=B1 2% iostat.sdc.avgqu-= sz > 15.67 =C2=B1 2% +72.7% 27.06 =C2=B1 2% iostat.sdb.avgqu-= sz > 25151 =C2=B1 0% +68.3% 42328 =C2=B1 1% iostat.sda.wkB/s > 25153 =C2=B1 0% +68.2% 42305 =C2=B1 1% iostat.sdb.wkB/s > 25149 =C2=B1 0% +68.2% 42292 =C2=B1 1% iostat.sdc.wkB/s > 97.45 =C2=B1 0% -21.1% 76.90 =C2=B1 0% turbostat.CorWatt > 12517 =C2=B1 0% -20.2% 9994 =C2=B1 1% iostat.sdc.rkB/s > 12517 =C2=B1 0% -20.0% 10007 =C2=B1 1% iostat.sda.rkB/s > 12512 =C2=B1 0% -19.9% 10018 =C2=B1 1% iostat.sdb.rkB/s > 1863 =C2=B1 3% +24.7% 2325 =C2=B1 1% iostat.sdb.rrqm/s > 1865 =C2=B1 3% +24.3% 2319 =C2=B1 1% iostat.sdc.rrqm/s > 1864 =C2=B1 3% +24.6% 2322 =C2=B1 1% iostat.sda.rrqm/s > 128 =C2=B1 0% -16.4% 107 =C2=B1 0% turbostat.PkgWatt > 150569 =C2=B1 0% -8.7% 137525 =C2=B1 0% iostat.md0.avgqu-= sz > 4.29 =C2=B1 0% -5.1% 4.07 =C2=B1 0% turbostat.RAMWatt >=20 >=20 > more detailed changes about ivb44/fsmark/1x-64t-4BRD_12G-RAID5-btrfs-4M-3= 0G-NoSync > --------- >=20 > 26089f4902595a2f 4400755e356f9a2b0b7ceaa02f =20 > ---------------- -------------------------- =20 > %stddev %change %stddev > \ | \ =20 > 273 =C2=B1 4% -18.1% 223 =C2=B1 6% fsmark.files_per_= sec > 29.24 =C2=B1 1% +27.2% 37.20 =C2=B1 8% fsmark.time.elaps= ed_time.max > 29.24 =C2=B1 1% +27.2% 37.20 =C2=B1 8% fsmark.time.elaps= ed_time > 399 =C2=B1 4% -20.0% 319 =C2=B1 3% fsmark.time.perce= nt_of_cpu_this_job_got > 129891 =C2=B1 20% -28.9% 92334 =C2=B1 15% fsmark.time.volun= tary_context_switches > 266 =C2=B1 0% +413.4% 1365 =C2=B1 5% slabinfo.raid5-md= 0.num_objs > 266 =C2=B1 0% +413.4% 1365 =C2=B1 5% slabinfo.raid5-md= 0.active_objs > 0.23 =C2=B1 27% +98.6% 0.46 =C2=B1 35% turbostat.CPU%c3 > 56612063 =C2=B1 9% +36.7% 77369763 =C2=B1 20% cpuidle.C1-IVT.ti= me > 5579498 =C2=B1 14% -36.0% 3571516 =C2=B1 6% cpuidle.C1E-IVT.t= ime > 4668 =C2=B1 38% +64.7% 7690 =C2=B1 19% numa-vmstat.node0= .nr_unevictable > 18674 =C2=B1 38% +64.7% 30762 =C2=B1 19% numa-meminfo.node= 0.Unevictable > 9298 =C2=B1 37% +64.4% 15286 =C2=B1 19% proc-vmstat.nr_un= evictable > 4629 =C2=B1 37% +64.1% 7596 =C2=B1 19% numa-vmstat.node1= .nr_unevictable > 18535 =C2=B1 37% +63.9% 30385 =C2=B1 19% numa-meminfo.node= 1.Unevictable > 4270894 =C2=B1 19% +65.6% 7070923 =C2=B1 21% cpuidle.C3-IVT.ti= me > 38457 =C2=B1 37% +59.0% 61148 =C2=B1 19% meminfo.Unevictab= le > 3748226 =C2=B1 17% +26.6% 4743674 =C2=B1 16% numa-vmstat.node0= .numa_local > 4495283 =C2=B1 13% -24.8% 3382315 =C2=B1 17% numa-vmstat.node0= .nr_free_pages > 3818432 =C2=B1 16% +26.5% 4830938 =C2=B1 16% numa-vmstat.node0= .numa_hit > 17966826 =C2=B1 13% -24.7% 13537228 =C2=B1 17% numa-meminfo.node= 0.MemFree > 14901309 =C2=B1 15% +29.7% 19330906 =C2=B1 12% numa-meminfo.node= 0.MemUsed > 26 =C2=B1 21% -32.9% 17 =C2=B1 14% cpuidle.POLL.usage > 1.183e+09 =C2=B1 1% +29.6% 1.533e+09 =C2=B1 8% cpuidle.C6-IVT.ti= me > 29.24 =C2=B1 1% +27.2% 37.20 =C2=B1 8% time.elapsed_time > 29.24 =C2=B1 1% +27.2% 37.20 =C2=B1 8% time.elapsed_time= .max > 399 =C2=B1 4% -20.0% 319 =C2=B1 3% time.percent_of_c= pu_this_job_got > 850 =C2=B1 4% -8.6% 777 =C2=B1 5% slabinfo.blkdev_r= equests.num_objs > 850 =C2=B1 4% -8.6% 777 =C2=B1 5% slabinfo.blkdev_r= equests.active_objs > 14986 =C2=B1 9% +17.1% 17548 =C2=B1 8% numa-vmstat.node0= .nr_slab_reclaimable > 11943 =C2=B1 5% -12.6% 10441 =C2=B1 2% slabinfo.kmalloc-= 192.num_objs > 59986 =C2=B1 9% +17.0% 70186 =C2=B1 8% numa-meminfo.node= 0.SReclaimable > 3703 =C2=B1 6% +10.2% 4082 =C2=B1 7% slabinfo.btrfs_de= layed_data_ref.num_objs > 133551 =C2=B1 6% +16.1% 154995 =C2=B1 1% proc-vmstat.pgfau= lt > 129891 =C2=B1 20% -28.9% 92334 =C2=B1 15% time.voluntary_co= ntext_switches > 11823 =C2=B1 4% -12.0% 10409 =C2=B1 3% slabinfo.kmalloc-= 192.active_objs > 3703 =C2=B1 6% +9.7% 4061 =C2=B1 7% slabinfo.btrfs_de= layed_data_ref.active_objs > 19761 =C2=B1 2% -11.2% 17542 =C2=B1 6% slabinfo.anon_vma= .active_objs > 19761 =C2=B1 2% -11.2% 17544 =C2=B1 6% slabinfo.anon_vma= .num_objs > 13002 =C2=B1 3% +14.9% 14944 =C2=B1 5% slabinfo.kmalloc-= 256.num_objs > 12695 =C2=B1 3% +13.8% 14446 =C2=B1 7% slabinfo.kmalloc-= 256.active_objs > 1190 =C2=B1 1% -11.8% 1050 =C2=B1 3% slabinfo.mnt_cach= e.num_objs > 1190 =C2=B1 1% -11.8% 1050 =C2=B1 3% slabinfo.mnt_cach= e.active_objs > 136862 =C2=B1 1% -13.8% 117938 =C2=B1 7% cpuidle.C6-IVT.us= age > 1692630 =C2=B1 3% +12.3% 1900854 =C2=B1 0% numa-vmstat.node0= .nr_written > 1056 =C2=B1 2% +8.8% 1149 =C2=B1 3% slabinfo.mm_struc= t.active_objs > 1056 =C2=B1 2% +8.8% 1149 =C2=B1 3% slabinfo.mm_struc= t.num_objs > 24029 =C2=B1 11% -30.6% 16673 =C2=B1 8% vmstat.system.cs > 8859 =C2=B1 2% -15.0% 7530 =C2=B1 8% vmstat.system.in > 905630 =C2=B1 2% -16.8% 753097 =C2=B1 4% iostat.md0.wkB/s > 906433 =C2=B1 2% -16.9% 753482 =C2=B1 4% vmstat.io.bo > 3591 =C2=B1 2% -16.9% 2982 =C2=B1 4% iostat.md0.w/s > 13.22 =C2=B1 5% -16.3% 11.07 =C2=B1 1% turbostat.%Busy > 402 =C2=B1 4% -15.9% 338 =C2=B1 1% turbostat.Avg_MHz > 54236 =C2=B1 3% +10.4% 59889 =C2=B1 4% iostat.md0.avgqu-= sz > 7.67 =C2=B1 1% +4.5% 8.01 =C2=B1 1% turbostat.RAMWatt >=20 >=20 >=20 > --yliu > -- > To unsubscribe from this list: send the line "unsubscribe linux-kernel" in > the body of a message to majordomo@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html > Please read the FAQ at http://www.tux.org/lkml/ --Sig_/9QIveOZ39dGk4bkgR_Z=MYI Content-Type: application/pgp-signature Content-Description: OpenPGP digital signature -----BEGIN PGP SIGNATURE----- Version: GnuPG v2 iQIVAwUBVRIlnznsnt1WYoG5AQJbvg/9H/eAvXvyT76qVf8K5dqfBEpWb+rTKsdE vd55jGaT0mUc4mvkZxHt7hLSXCen8fgBDIw0q0MW0Hz1L3JeaEg6l01oA2oCf/4x kOAc7LZh9D2M8rPMGdCjWjfTrr7kfWfwyRPPftYPiXAXrWHkGclEXEek3QQHLyFX HqgyW633gzxcnJmLJhnh/dJJV1x2qAEm5gjIqh0q7+7Ym/2B/lw+yJcDTXjbjfiG BL+ZrrtYsQiuMw/k78tndCOG0zeqHp56VNsypzfpyJuVFJmUb795zN6qXJmjcW2z I5iCTFoszcLl5EEINRq/nauh7RD8gFmCDDRnBk4m2snFAyhzVOdKIm6AdNeJtLF3 x8vp3mcGQYvyFsm1fRZAhrFUcJWT2vTSJkgK+Q3C7wQ6bknhYDlQDMrMIzRP2oTr NPWdbAZq8CGt0HfbyvRQGhuVHCo4Krl9vy6R+P+N1ehNL7Iqp7/RsQ/YZ/vUI7XV Wb14HYPQsCAt9QNW/ArY4NYcqGvT/zzjpp93WoWWffnHDZQLnGABglA12rSUBdl2 6vy64Am1ksQ575f2uEAI7GfCgFrXJ30wdSdXNZK5DtoH1dabixrGplLyIJPt7a2u 5JgGsBFfGUxJYu5Lj7JiqjLqAI3nzAgMJ65H5RyIo8n9j6JxpaehQBNnJgCksDMp jd5sNB+pPPw= =m1Jo -----END PGP SIGNATURE----- --Sig_/9QIveOZ39dGk4bkgR_Z=MYI--