Re: performance changes on 4400755e: 200.0% fsmark.files_per_sec, -18.1% fsmark.files_per_sec, and few more

From: NeilBrown
Date: Tue Mar 24 2015 - 23:04:03 EST


On Wed, 18 Mar 2015 13:00:30 +0800 Yuanahn Liu <yuanhan.liu@xxxxxxxxxxxxxxx>
wrote:

> Hi,
>
> FYI, we noticed performance changes on `fsmark.files_per_sec' by 4400755e356f9a2b0b7ceaa02f57b1c7546c3765:
>
> > commit 4400755e356f9a2b0b7ceaa02f57b1c7546c3765
> > Author: NeilBrown <neilb@xxxxxxx>
> > AuthorDate: Thu Feb 26 12:47:56 2015 +1100
> > Commit: NeilBrown <neilb@xxxxxxx>
> > CommitDate: Wed Mar 4 13:40:19 2015 +1100
> >
> > 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 that
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 Â 2% -47.5% 649 Â 3% fsmark.time.percent_of_cpu_this_job_got

> 399 Â 4% -20.0% 319 Â 3% fsmark.time.percent_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




>
> 26089f4902595a2f64c512066af07af6e82eb096 4400755e356f9a2b0b7ceaa02f57b1c7546c3765
> ---------------------------------------- ----------------------------------------
> run time(m) metric_value Âstddev run time(m) metric_value Âstddev change testbox/benchmark/sub-testcase
> --- ------ ---------------------------- --- ------ ---------------------------- -------- ------------------------------
> 3 18.6 6.400 Â0.0% 5 9.2 19.200 Â0.0% 200.0% ivb44/fsmark/1x-64t-3HDD-RAID5-ext4-4M-40G-fsyncBeforeClose
> 3 24.7 6.400 Â0.0% 3 13.7 12.800 Â0.0% 100.0% ivb44/fsmark/1x-64t-3HDD-RAID5-f2fs-4M-40G-fsyncBeforeClose
> 3 17.5 28.267 Â9.6% 3 12.3 42.833 Â6.5% 51.5% ivb44/fsmark/1x-64t-3HDD-RAID5-f2fs-4M-120G-NoSync
> 3 16.7 30.700 Â1.5% 3 12.6 40.733 Â2.4% 32.7% ivb44/fsmark/1x-1t-3HDD-RAID5-f2fs-4M-120G-NoSync
> 3 29.0 5.867 Â0.8% 5 23.6 7.240 Â0.7% 23.4% ivb44/fsmark/1x-1t-3HDD-RAID5-btrfs-4M-40G-fsyncBeforeClose
> 3 28.5 6.000 Â0.0% 3 23.2 7.367 Â0.6% 22.8% ivb44/fsmark/1x-1t-3HDD-RAID5-f2fs-4M-40G-fsyncBeforeClose
> 5 11.7 14.600 Â0.0% 5 9.7 17.500 Â0.4% 19.9% ivb44/fsmark/1x-1t-3HDD-RAID5-xfs-4M-40G-fsyncBeforeClose
> 3 22.4 25.600 Â0.0% 5 17.9 30.120 Â4.1% 17.7% ivb44/fsmark/1x-64t-3HDD-RAID5-xfs-4M-120G-NoSync
> 5 10.8 47.320 Â0.6% 5 9.3 54.820 Â0.2% 15.8% ivb44/fsmark/1x-1t-3HDD-RAID5-xfs-4M-120G-NoSync
> 1 0.5 252.400 Â0.0% 1 0.5 263.300 Â0.0% 4.3% ivb44/fsmark/1x-1t-4BRD_12G-RAID5-ext4-4M-30G-NoSync
>
> 3 0.5 273.100 Â4.3% 3 0.6 223.567 Â6.5% -18.1% ivb44/fsmark/1x-64t-4BRD_12G-RAID5-btrfs-4M-30G-NoSync
> 3 8.1 63.133 Â0.5% 3 9.2 55.633 Â0.2% -11.9% ivb44/fsmark/1x-1t-3HDD-RAID5-btrfs-4M-120G-NoSync
> 3 8.2 64.000 Â0.0% 3 9.2 57.600 Â0.0% -10.0% ivb44/fsmark/1x-64t-3HDD-RAID5-btrfs-4M-120G-NoSync
>
>
> NOTE: here are some more info about those test parameters for you to
> understand the testcase better:
>
> 1x: where 'x' means iterations or loop, corresponding to the 'L' option 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
>
> 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. And we made a
> raid on those ramdisk.
>
>
> 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.
>
>
> FYI, here I listed more detailed changes for the maximal postive and negtive changes.
>
> more detailed changes about ivb44/fsmark/1x-64t-3HDD-RAID5-ext4-4M-40G-fsyncBeforeClose
> ---------
>
> 26089f4902595a2f 4400755e356f9a2b0b7ceaa02f
> ---------------- --------------------------
> %stddev %change %stddev
> \ | \
> 6.40 Â 0% +200.0% 19.20 Â 0% fsmark.files_per_sec
> 1.015e+08 Â 1% -73.6% 26767355 Â 3% fsmark.time.voluntary_context_switches
> 13793 Â 1% -73.9% 3603 Â 5% fsmark.time.system_time
> 78473 Â 6% -64.3% 28016 Â 7% fsmark.time.involuntary_context_switches
> 15789555 Â 9% -54.7% 7159485 Â 13% fsmark.app_overhead
> 1115 Â 0% -50.3% 554 Â 1% fsmark.time.elapsed_time.max
> 1115 Â 0% -50.3% 554 Â 1% fsmark.time.elapsed_time
> 1235 Â 2% -47.5% 649 Â 3% fsmark.time.percent_of_cpu_this_job_got
> 456465 Â 1% -26.7% 334594 Â 4% fsmark.time.minor_page_faults
> 275 Â 0% +1257.7% 3733 Â 2% slabinfo.raid5-md0.num_objs
> 275 Â 0% +1257.7% 3733 Â 2% slabinfo.raid5-md0.active_objs
> 11 Â 0% +1250.9% 148 Â 2% slabinfo.raid5-md0.active_slabs
> 11 Â 0% +1250.9% 148 Â 2% slabinfo.raid5-md0.num_slabs
> 2407 Â 4% +293.4% 9471 Â 26% numa-meminfo.node0.Writeback
> 600 Â 4% +294.9% 2372 Â 26% numa-vmstat.node0.nr_writeback
> 1114505 Â 0% -77.4% 251696 Â 2% softirqs.TASKLET
> 1808027 Â 1% -77.7% 402378 Â 4% softirqs.RCU
> 12158665 Â 1% -77.1% 2786069 Â 4% cpuidle.C3-IVT.usage
> 1119433 Â 0% -77.3% 254192 Â 2% softirqs.BLOCK
> 37824202 Â 1% -75.1% 9405078 Â 4% cpuidle.C6-IVT.usage
> 1.015e+08 Â 1% -73.6% 26767355 Â 3% time.voluntary_context_switches
> 13793 Â 1% -73.9% 3603 Â 5% time.system_time
> 5971084 Â 1% -73.6% 1574912 Â 5% softirqs.SCHED
> 10539492 Â 3% -72.0% 2956258 Â 6% cpuidle.C1E-IVT.usage
> 2 Â 0% +230.0% 6 Â 12% vmstat.procs.b
> 14064 Â 1% -71.2% 4049 Â 6% softirqs.HRTIMER
> 7388306 Â 1% -71.2% 2129929 Â 4% softirqs.TIMER
> 3.496e+09 Â 1% -70.3% 1.04e+09 Â 1% cpuidle.C3-IVT.time
> 0.88 Â 6% +224.9% 2.87 Â 11% turbostat.Pkg%pc6
> 19969464 Â 2% -66.2% 6750675 Â 5% cpuidle.C1-IVT.usage
> 78473 Â 6% -64.3% 28016 Â 7% time.involuntary_context_switches
> 4.23 Â 5% +181.4% 11.90 Â 3% turbostat.Pkg%pc2
> 2.551e+09 Â 1% -61.4% 9.837e+08 Â 3% cpuidle.C1E-IVT.time
> 8084 Â 3% +142.6% 19608 Â 3% meminfo.Writeback
> 2026 Â 4% +141.6% 4895 Â 4% proc-vmstat.nr_writeback
> 165 Â 4% -56.9% 71 Â 14% numa-vmstat.node1.nr_inactive_anon
> 7.748e+09 Â 3% -50.3% 3.852e+09 Â 3% cpuidle.C1-IVT.time
> 175 Â 5% -53.2% 82 Â 13% numa-vmstat.node1.nr_shmem
> 1115 Â 0% -50.3% 554 Â 1% time.elapsed_time.max
> 1115 Â 0% -50.3% 554 Â 1% time.elapsed_time
> 1147 Â 0% -49.0% 585 Â 1% uptime.boot
> 2260889 Â 0% -48.8% 1157272 Â 1% proc-vmstat.pgfree
> 16805 Â 2% -35.9% 10776 Â 23% numa-vmstat.node1.nr_dirty
> 1235 Â 2% -47.5% 649 Â 3% time.percent_of_cpu_this_job_got
> 67245 Â 2% -35.9% 43122 Â 23% numa-meminfo.node1.Dirty
> 39041 Â 0% -45.7% 21212 Â 2% uptime.idle
> 13 Â 9% -49.0% 6 Â 11% vmstat.procs.r
> 3072 Â 10% -40.3% 1833 Â 9% cpuidle.POLL.usage
> 3045115 Â 0% -46.1% 1642053 Â 1% proc-vmstat.pgfault
> 202 Â 1% -45.2% 110 Â 0% proc-vmstat.nr_inactive_anon
> 4583079 Â 2% -31.4% 3143602 Â 16% numa-vmstat.node1.numa_hit
> 28.03 Â 0% +69.1% 47.39 Â 1% turbostat.CPU%c6
> 223 Â 1% -41.1% 131 Â 1% proc-vmstat.nr_shmem
> 4518820 Â 3% -30.8% 3128304 Â 16% numa-vmstat.node1.numa_local
> 3363496 Â 3% -27.4% 2441619 Â 20% numa-vmstat.node1.nr_dirtied
> 3345346 Â 3% -27.4% 2428396 Â 20% numa-vmstat.node1.nr_written
> 0.18 Â 18% +105.6% 0.37 Â 36% turbostat.Pkg%pc3
> 3427913 Â 3% -27.3% 2492563 Â 20% numa-vmstat.node1.nr_inactive_file
> 13712431 Â 3% -27.3% 9971152 Â 20% numa-meminfo.node1.Inactive
> 13711768 Â 3% -27.3% 9970866 Â 20% numa-meminfo.node1.Inactive(file)
> 3444598 Â 3% -27.2% 2508920 Â 20% numa-vmstat.node1.nr_file_pages
> 13778510 Â 3% -27.2% 10036287 Â 20% numa-meminfo.node1.FilePages
> 8819175 Â 1% -28.3% 6320188 Â 19% numa-numastat.node1.numa_hit
> 8819051 Â 1% -28.3% 6320152 Â 19% numa-numastat.node1.local_node
> 14350918 Â 3% -26.8% 10504070 Â 19% numa-meminfo.node1.MemUsed
> 100892 Â 3% -26.0% 74623 Â 19% numa-vmstat.node1.nr_slab_reclaimable
> 403571 Â 3% -26.0% 298513 Â 19% numa-meminfo.node1.SReclaimable
> 3525 Â 13% +36.6% 4817 Â 14% slabinfo.blkdev_requests.active_objs
> 3552 Â 13% +36.3% 4841 Â 14% slabinfo.blkdev_requests.num_objs
> 30779 Â 4% -34.7% 20084 Â 12% proc-vmstat.pgmigrate_success
> 30779 Â 4% -34.7% 20084 Â 12% proc-vmstat.numa_pages_migrated
> 447400 Â 2% -23.2% 343701 Â 16% numa-meminfo.node1.Slab
> 2.532e+10 Â 0% -33.1% 1.694e+10 Â 1% cpuidle.C6-IVT.time
> 3081 Â 9% +28.0% 3945 Â 12% slabinfo.mnt_cache.num_objs
> 3026 Â 9% +28.8% 3898 Â 12% slabinfo.mnt_cache.active_objs
> 5822 Â 4% +77.8% 10350 Â 25% numa-meminfo.node1.Writeback
> 1454 Â 4% +77.3% 2579 Â 25% numa-vmstat.node1.nr_writeback
> 424984 Â 1% -26.5% 312255 Â 3% proc-vmstat.numa_pte_updates
> 368001 Â 1% -26.8% 269440 Â 3% proc-vmstat.numa_hint_faults
> 456465 Â 1% -26.7% 334594 Â 4% time.minor_page_faults
> 3.86 Â 3% -24.4% 2.92 Â 2% turbostat.CPU%c3
> 4661151 Â 2% +20.6% 5622999 Â 9% numa-vmstat.node1.nr_free_pages
> 18644452 Â 2% +20.6% 22491300 Â 9% numa-meminfo.node1.MemFree
> 876 Â 2% +28.2% 1124 Â 5% slabinfo.kmalloc-4096.num_objs
> 858 Â 3% +24.0% 1064 Â 5% slabinfo.kmalloc-4096.active_objs
> 17767832 Â 8% -25.4% 13249545 Â 17% cpuidle.POLL.time
> 285093 Â 1% -23.1% 219372 Â 5% proc-vmstat.numa_hint_faults_local
> 105423 Â 2% -16.1% 88498 Â 0% meminfo.Dirty
> 26365 Â 1% -16.0% 22152 Â 1% proc-vmstat.nr_dirty
> 41.04 Â 1% -14.1% 35.26 Â 1% turbostat.CPU%c1
> 9385 Â 4% -14.3% 8043 Â 6% slabinfo.kmalloc-192.active_objs
> 9574 Â 3% -13.9% 8241 Â 6% slabinfo.kmalloc-192.num_objs
> 2411 Â 3% +17.0% 2820 Â 4% slabinfo.kmalloc-2048.active_objs
> 12595574 Â 0% -10.0% 11338368 Â 1% proc-vmstat.pgalloc_normal
> 5262 Â 1% +13.3% 5962 Â 1% slabinfo.kmalloc-1024.num_objs
> 5262 Â 1% +12.7% 5932 Â 1% slabinfo.kmalloc-1024.active_objs
> 2538 Â 3% +13.7% 2885 Â 4% slabinfo.kmalloc-2048.num_objs
> 5299546 Â 0% -9.9% 4776351 Â 0% slabinfo.buffer_head.active_objs
> 5299546 Â 0% -9.9% 4776351 Â 0% slabinfo.buffer_head.num_objs
> 135885 Â 0% -9.9% 122470 Â 0% slabinfo.buffer_head.num_slabs
> 135885 Â 0% -9.9% 122470 Â 0% slabinfo.buffer_head.active_slabs
> 28.04 Â 2% +715.6% 228.69 Â 3% iostat.sdb.avgrq-sz
> 28.05 Â 2% +708.1% 226.72 Â 2% iostat.sdc.avgrq-sz
> 2245 Â 3% -81.6% 413 Â 1% iostat.sda.w/s
> 5.33 Â 1% +1008.2% 59.07 Â 1% iostat.sda.w_await
> 5.85 Â 1% +1126.4% 71.69 Â 4% iostat.sda.r_await
> 5.36 Â 1% +978.6% 57.79 Â 3% iostat.sdc.w_await
> 1263 Â 4% -85.8% 179 Â 6% iostat.sdc.r/s
> 2257 Â 3% -81.6% 414 Â 2% iostat.sdb.w/s
> 1264 Â 4% -85.8% 179 Â 6% iostat.sdb.r/s
> 5.55 Â 0% +1024.2% 62.37 Â 4% iostat.sdb.await
> 5.89 Â 1% +1125.9% 72.16 Â 6% iostat.sdb.r_await
> 5.36 Â 0% +1014.3% 59.75 Â 3% iostat.sdb.w_await
> 5.57 Â 1% +987.9% 60.55 Â 3% iostat.sdc.await
> 5.51 Â 0% +1017.3% 61.58 Â 1% iostat.sda.await
> 1264 Â 4% -85.8% 179 Â 6% iostat.sda.r/s
> 28.09 Â 2% +714.2% 228.73 Â 2% iostat.sda.avgrq-sz
> 5.95 Â 2% +1091.0% 70.82 Â 6% iostat.sdc.r_await
> 2252 Â 3% -81.5% 417 Â 2% iostat.sdc.w/s
> 4032 Â 2% +151.6% 10143 Â 1% iostat.sdb.wrqm/s
> 4043 Â 2% +151.0% 10150 Â 1% iostat.sda.wrqm/s
> 4035 Â 2% +151.2% 10138 Â 1% iostat.sdc.wrqm/s
> 26252 Â 1% -54.0% 12077 Â 4% vmstat.system.in
> 37813 Â 0% +101.0% 75998 Â 1% vmstat.io.bo
> 37789 Â 0% +101.0% 75945 Â 1% iostat.md0.wkB/s
> 205 Â 0% +96.1% 402 Â 1% iostat.md0.w/s
> 164286 Â 1% -46.2% 88345 Â 2% vmstat.system.cs
> 27.07 Â 2% -46.7% 14.42 Â 3% turbostat.%Busy
> 810 Â 2% -46.7% 431 Â 3% turbostat.Avg_MHz
> 15.56 Â 2% +71.7% 26.71 Â 1% iostat.sda.avgqu-sz
> 15.65 Â 2% +69.1% 26.46 Â 2% iostat.sdc.avgqu-sz
> 15.67 Â 2% +72.7% 27.06 Â 2% iostat.sdb.avgqu-sz
> 25151 Â 0% +68.3% 42328 Â 1% iostat.sda.wkB/s
> 25153 Â 0% +68.2% 42305 Â 1% iostat.sdb.wkB/s
> 25149 Â 0% +68.2% 42292 Â 1% iostat.sdc.wkB/s
> 97.45 Â 0% -21.1% 76.90 Â 0% turbostat.CorWatt
> 12517 Â 0% -20.2% 9994 Â 1% iostat.sdc.rkB/s
> 12517 Â 0% -20.0% 10007 Â 1% iostat.sda.rkB/s
> 12512 Â 0% -19.9% 10018 Â 1% iostat.sdb.rkB/s
> 1863 Â 3% +24.7% 2325 Â 1% iostat.sdb.rrqm/s
> 1865 Â 3% +24.3% 2319 Â 1% iostat.sdc.rrqm/s
> 1864 Â 3% +24.6% 2322 Â 1% iostat.sda.rrqm/s
> 128 Â 0% -16.4% 107 Â 0% turbostat.PkgWatt
> 150569 Â 0% -8.7% 137525 Â 0% iostat.md0.avgqu-sz
> 4.29 Â 0% -5.1% 4.07 Â 0% turbostat.RAMWatt
>
>
> more detailed changes about ivb44/fsmark/1x-64t-4BRD_12G-RAID5-btrfs-4M-30G-NoSync
> ---------
>
> 26089f4902595a2f 4400755e356f9a2b0b7ceaa02f
> ---------------- --------------------------
> %stddev %change %stddev
> \ | \
> 273 Â 4% -18.1% 223 Â 6% fsmark.files_per_sec
> 29.24 Â 1% +27.2% 37.20 Â 8% fsmark.time.elapsed_time.max
> 29.24 Â 1% +27.2% 37.20 Â 8% fsmark.time.elapsed_time
> 399 Â 4% -20.0% 319 Â 3% fsmark.time.percent_of_cpu_this_job_got
> 129891 Â 20% -28.9% 92334 Â 15% fsmark.time.voluntary_context_switches
> 266 Â 0% +413.4% 1365 Â 5% slabinfo.raid5-md0.num_objs
> 266 Â 0% +413.4% 1365 Â 5% slabinfo.raid5-md0.active_objs
> 0.23 Â 27% +98.6% 0.46 Â 35% turbostat.CPU%c3
> 56612063 Â 9% +36.7% 77369763 Â 20% cpuidle.C1-IVT.time
> 5579498 Â 14% -36.0% 3571516 Â 6% cpuidle.C1E-IVT.time
> 4668 Â 38% +64.7% 7690 Â 19% numa-vmstat.node0.nr_unevictable
> 18674 Â 38% +64.7% 30762 Â 19% numa-meminfo.node0.Unevictable
> 9298 Â 37% +64.4% 15286 Â 19% proc-vmstat.nr_unevictable
> 4629 Â 37% +64.1% 7596 Â 19% numa-vmstat.node1.nr_unevictable
> 18535 Â 37% +63.9% 30385 Â 19% numa-meminfo.node1.Unevictable
> 4270894 Â 19% +65.6% 7070923 Â 21% cpuidle.C3-IVT.time
> 38457 Â 37% +59.0% 61148 Â 19% meminfo.Unevictable
> 3748226 Â 17% +26.6% 4743674 Â 16% numa-vmstat.node0.numa_local
> 4495283 Â 13% -24.8% 3382315 Â 17% numa-vmstat.node0.nr_free_pages
> 3818432 Â 16% +26.5% 4830938 Â 16% numa-vmstat.node0.numa_hit
> 17966826 Â 13% -24.7% 13537228 Â 17% numa-meminfo.node0.MemFree
> 14901309 Â 15% +29.7% 19330906 Â 12% numa-meminfo.node0.MemUsed
> 26 Â 21% -32.9% 17 Â 14% cpuidle.POLL.usage
> 1.183e+09 Â 1% +29.6% 1.533e+09 Â 8% cpuidle.C6-IVT.time
> 29.24 Â 1% +27.2% 37.20 Â 8% time.elapsed_time
> 29.24 Â 1% +27.2% 37.20 Â 8% time.elapsed_time.max
> 399 Â 4% -20.0% 319 Â 3% time.percent_of_cpu_this_job_got
> 850 Â 4% -8.6% 777 Â 5% slabinfo.blkdev_requests.num_objs
> 850 Â 4% -8.6% 777 Â 5% slabinfo.blkdev_requests.active_objs
> 14986 Â 9% +17.1% 17548 Â 8% numa-vmstat.node0.nr_slab_reclaimable
> 11943 Â 5% -12.6% 10441 Â 2% slabinfo.kmalloc-192.num_objs
> 59986 Â 9% +17.0% 70186 Â 8% numa-meminfo.node0.SReclaimable
> 3703 Â 6% +10.2% 4082 Â 7% slabinfo.btrfs_delayed_data_ref.num_objs
> 133551 Â 6% +16.1% 154995 Â 1% proc-vmstat.pgfault
> 129891 Â 20% -28.9% 92334 Â 15% time.voluntary_context_switches
> 11823 Â 4% -12.0% 10409 Â 3% slabinfo.kmalloc-192.active_objs
> 3703 Â 6% +9.7% 4061 Â 7% slabinfo.btrfs_delayed_data_ref.active_objs
> 19761 Â 2% -11.2% 17542 Â 6% slabinfo.anon_vma.active_objs
> 19761 Â 2% -11.2% 17544 Â 6% slabinfo.anon_vma.num_objs
> 13002 Â 3% +14.9% 14944 Â 5% slabinfo.kmalloc-256.num_objs
> 12695 Â 3% +13.8% 14446 Â 7% slabinfo.kmalloc-256.active_objs
> 1190 Â 1% -11.8% 1050 Â 3% slabinfo.mnt_cache.num_objs
> 1190 Â 1% -11.8% 1050 Â 3% slabinfo.mnt_cache.active_objs
> 136862 Â 1% -13.8% 117938 Â 7% cpuidle.C6-IVT.usage
> 1692630 Â 3% +12.3% 1900854 Â 0% numa-vmstat.node0.nr_written
> 1056 Â 2% +8.8% 1149 Â 3% slabinfo.mm_struct.active_objs
> 1056 Â 2% +8.8% 1149 Â 3% slabinfo.mm_struct.num_objs
> 24029 Â 11% -30.6% 16673 Â 8% vmstat.system.cs
> 8859 Â 2% -15.0% 7530 Â 8% vmstat.system.in
> 905630 Â 2% -16.8% 753097 Â 4% iostat.md0.wkB/s
> 906433 Â 2% -16.9% 753482 Â 4% vmstat.io.bo
> 3591 Â 2% -16.9% 2982 Â 4% iostat.md0.w/s
> 13.22 Â 5% -16.3% 11.07 Â 1% turbostat.%Busy
> 402 Â 4% -15.9% 338 Â 1% turbostat.Avg_MHz
> 54236 Â 3% +10.4% 59889 Â 4% iostat.md0.avgqu-sz
> 7.67 Â 1% +4.5% 8.01 Â 1% turbostat.RAMWatt
>
>
>
> --yliu
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@xxxxxxxxxxxxxxx
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/

Attachment: pgpeLU7LEcZqV.pgp
Description: OpenPGP digital signature