* Re: [lkp-robot] [fs] 5c6de586e8: vm-scalability.throughput +12.4% improvement (from reorganizing struct inode?)
[not found] <CAOQ4uxgRXzq+ZSmwRB_nyGijFJsiuWXaM_F74W-3JOyTvAEz-Q@mail.gmail.com>
@ 2018-07-02 18:59 ` Linus Torvalds
2018-07-04 23:34 ` Al Viro
2018-07-05 6:01 ` Amir Goldstein
1 sibling, 1 reply; 5+ messages in thread
From: Linus Torvalds @ 2018-07-02 18:59 UTC (permalink / raw)
To: lkp
[-- Attachment #1: Type: text/plain, Size: 568 bytes --]
On Sun, Jul 1, 2018 at 11:27 PM Amir Goldstein <amir73il@gmail.com> wrote:
>
> This may be a test fluctuation or as a result of moving
> i_blkbits closer to i_bytes and i_lock.
Hey, I certainly am not against shrinking the inode (and in general,
removing 'enum's from internal structures), although that benchmark
improvement looks suspiciously large. The kernel test robot is lovely,
but some of the performance fluctuations may be more noise than
others.
I was hoping the patch would go through the regular vfs tree, though. Al?
Linus
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [lkp-robot] [fs] 5c6de586e8: vm-scalability.throughput +12.4% improvement (from reorganizing struct inode?)
2018-07-02 18:59 ` [lkp-robot] [fs] 5c6de586e8: vm-scalability.throughput +12.4% improvement (from reorganizing struct inode?) Linus Torvalds
@ 2018-07-04 23:34 ` Al Viro
2018-07-05 1:37 ` Linus Torvalds
0 siblings, 1 reply; 5+ messages in thread
From: Al Viro @ 2018-07-04 23:34 UTC (permalink / raw)
To: lkp
[-- Attachment #1: Type: text/plain, Size: 1059 bytes --]
On Mon, Jul 02, 2018 at 11:59:30AM -0700, Linus Torvalds wrote:
> On Sun, Jul 1, 2018 at 11:27 PM Amir Goldstein <amir73il@gmail.com> wrote:
> >
> > This may be a test fluctuation or as a result of moving
> > i_blkbits closer to i_bytes and i_lock.
>
> Hey, I certainly am not against shrinking the inode (and in general,
> removing 'enum's from internal structures), although that benchmark
> improvement looks suspiciously large. The kernel test robot is lovely,
> but some of the performance fluctuations may be more noise than
> others.
>
> I was hoping the patch would go through the regular vfs tree, though. Al?
Sure, but
* please, repost it. github's webshite is atrocious ;-/
* I would like more details about the variation of timing -
what's the dispersion from boot to boot, for starters?
I don't hate that patch, but there are immediate followup questions - e.g.
how sensitive is relative position of i_lock/i_hash/i_sb? Those are *not*
close to each other. E.g. what happens if one moves i_hash right after
i_ino?
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [lkp-robot] [fs] 5c6de586e8: vm-scalability.throughput +12.4% improvement (from reorganizing struct inode?)
2018-07-04 23:34 ` Al Viro
@ 2018-07-05 1:37 ` Linus Torvalds
2018-07-05 5:50 ` Amir Goldstein
0 siblings, 1 reply; 5+ messages in thread
From: Linus Torvalds @ 2018-07-05 1:37 UTC (permalink / raw)
To: lkp
[-- Attachment #1: Type: text/plain, Size: 909 bytes --]
On Wed, Jul 4, 2018 at 4:34 PM Al Viro <viro@zeniv.linux.org.uk> wrote:
>
> I don't hate that patch, but there are immediate followup questions - e.g.
> how sensitive is relative position of i_lock/i_hash/i_sb? Those are *not*
> close to each other. E.g. what happens if one moves i_hash right after
> i_ino?
I think that's a separate issue from just packing things better.
The original patch is at
https://marc.info/?l=linux-fsdevel&m=152880180732361&w=2
although I certainly agree with the "just re-send as a proper patch,
now with a note on potential performance improvements".
I do agree that we could also see if there is something we could do
about sorting the inode by commonly used fields. We already do to some
degree (thge first fields up to and including i_security are certainly
hot), but the field ordering is old and maybe things have changed.
Linus
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [lkp-robot] [fs] 5c6de586e8: vm-scalability.throughput +12.4% improvement (from reorganizing struct inode?)
2018-07-05 1:37 ` Linus Torvalds
@ 2018-07-05 5:50 ` Amir Goldstein
0 siblings, 0 replies; 5+ messages in thread
From: Amir Goldstein @ 2018-07-05 5:50 UTC (permalink / raw)
To: Linus Torvalds
Cc: Al Viro, Jan Kara, kernel test robot, LKP, linux-kernel, linux-fsdevel
On Thu, Jul 5, 2018 at 4:37 AM, Linus Torvalds
<torvalds@linux-foundation.org> wrote:
> On Wed, Jul 4, 2018 at 4:34 PM Al Viro <viro@zeniv.linux.org.uk> wrote:
>>
>> I don't hate that patch, but there are immediate followup questions - e.g.
>> how sensitive is relative position of i_lock/i_hash/i_sb? Those are *not*
>> close to each other. E.g. what happens if one moves i_hash right after
>> i_ino?
>
> I think that's a separate issue from just packing things better.
>
> The original patch is at
>
> https://marc.info/?l=linux-fsdevel&m=152880180732361&w=2
>
> although I certainly agree with the "just re-send as a proper patch,
> now with a note on potential performance improvements".
>
Will do. But first, let me CC this discussion to some public lists,
so I have somewhere to refer to in commit message...
Thanks,
Amir.
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [lkp-robot] [fs] 5c6de586e8: vm-scalability.throughput +12.4% improvement (from reorganizing struct inode?)
[not found] <CAOQ4uxgRXzq+ZSmwRB_nyGijFJsiuWXaM_F74W-3JOyTvAEz-Q@mail.gmail.com>
2018-07-02 18:59 ` [lkp-robot] [fs] 5c6de586e8: vm-scalability.throughput +12.4% improvement (from reorganizing struct inode?) Linus Torvalds
@ 2018-07-05 6:01 ` Amir Goldstein
1 sibling, 0 replies; 5+ messages in thread
From: Amir Goldstein @ 2018-07-05 6:01 UTC (permalink / raw)
To: Linus Torvalds
Cc: Al Viro, Jan Kara, kernel test robot, LKP, linux-kernel, linux-fsdevel
On Mon, Jul 2, 2018 at 9:27 AM, Amir Goldstein <amir73il@gmail.com> wrote:
> Linus,
>
> This may be a test fluctuation or as a result of moving
> i_blkbits closer to i_bytes and i_lock.
>
> In any case, ping for:
> https://marc.info/?l=linux-fsdevel&m=152882624707975&w=2
>
Linus,
Per your request, I will re-post the origin patch with a link to this
discussion (which has now been made public).
Thanks,
Amir.
>
> ---------- Forwarded message ----------
> From: kernel test robot <xiaolong.ye@intel.com>
> Date: Mon, Jul 2, 2018 at 8:14 AM
> Subject: [lkp-robot] [fs] 5c6de586e8: vm-scalability.throughput
> +12.4% improvement
> To: Amir Goldstein <amir73il@gmail.com>
> Cc: lkp@01.org
>
>
>
> Greeting,
>
> FYI, we noticed a +12.4% improvement of vm-scalability.throughput due to commit:
>
>
> commit: 5c6de586e899a4a80a0ffa26468639f43dee1009 ("[PATCH] fs: shave 8
> bytes off of struct inode")
> url: https://github.com/0day-ci/linux/commits/Amir-Goldstein/fs-shave-8-bytes-off-of-struct-inode/20180612-192311
>
>
> in testcase: vm-scalability
> on test machine: 56 threads Intel(R) Xeon(R) CPU E5-2695 v3 @ 2.30GHz
> with 256G memory
> with following parameters:
>
> runtime: 300s
> test: small-allocs
> cpufreq_governor: performance
>
> test-description: The motivation behind this suite is to exercise
> functions and regions of the mm/ of the Linux kernel which are of
> interest to us.
> test-url: https://git.kernel.org/cgit/linux/kernel/git/wfg/vm-scalability.git/
>
>
>
> Details are as below:
> -------------------------------------------------------------------------------------------------->
>
>
> To reproduce:
>
> git clone https://github.com/intel/lkp-tests.git
> cd lkp-tests
> bin/lkp install job.yaml # job file is attached in this email
> bin/lkp run job.yaml
>
> =========================================================================================
> compiler/cpufreq_governor/kconfig/rootfs/runtime/tbox_group/test/testcase:
> gcc-7/performance/x86_64-rhel-7.2/debian-x86_64-2018-04-03.cgz/300s/lkp-hsw-ep5/small-allocs/vm-scalability
>
> commit:
> 8efcf34a26 (" ARM: SoC: late updates")
> 5c6de586e8 ("fs: shave 8 bytes off of struct inode")
>
> 8efcf34a263965e4 5c6de586e899a4a80a0ffa2646
> ---------------- --------------------------
> %stddev %change %stddev
> \ | \
> 19335952 +12.4% 21729332 vm-scalability.throughput
> 693688 +11.9% 775935 vm-scalability.median
> 0.56 ą 55% -43.6% 0.32 ą 62% vm-scalability.stddev
> 288.16 -7.0% 267.96 vm-scalability.time.elapsed_time
> 288.16 -7.0% 267.96
> vm-scalability.time.elapsed_time.max
> 48921 ą 6% -3.3% 47314 ą 5%
> vm-scalability.time.involuntary_context_switches
> 3777 -4.4% 3610
> vm-scalability.time.maximum_resident_set_size
> 1.074e+09 +0.0% 1.074e+09
> vm-scalability.time.minor_page_faults
> 4096 +0.0% 4096 vm-scalability.time.page_size
> 2672 +0.6% 2689
> vm-scalability.time.percent_of_cpu_this_job_got
> 5457 -9.4% 4942 vm-scalability.time.system_time
> 2244 +0.9% 2263 vm-scalability.time.user_time
> 5529533 ą 6% -65.2% 1923014 ą 6%
> vm-scalability.time.voluntary_context_switches
> 4.832e+09 +0.0% 4.832e+09 vm-scalability.workload
> 93827 ą 3% -7.0% 87299 ą 3%
> interrupts.CAL:Function_call_interrupts
> 26.50 -2.0% 25.98 ą 3% boot-time.boot
> 16.69 -3.2% 16.16 ą 5% boot-time.dhcp
> 674.18 -1.1% 666.43 ą 2% boot-time.idle
> 17.61 -3.5% 17.00 ą 5% boot-time.kernel_boot
> 15034 ą 62% -30.0% 10528 ą 79% softirqs.NET_RX
> 453251 ą 9% -2.9% 440306 ą 14% softirqs.RCU
> 46795 -14.9% 39806 ą 2% softirqs.SCHED
> 3565160 ą 8% +6.1% 3784023 softirqs.TIMER
> 4.87 ą 4% -0.6 4.25 ą 3% mpstat.cpu.idle%
> 0.00 ą 13% -0.0 0.00 ą 14% mpstat.cpu.iowait%
> 0.00 ą 37% +0.0 0.00 ą 37% mpstat.cpu.soft%
> 67.35 -1.8 65.60 mpstat.cpu.sys%
> 27.78 +2.4 30.14 mpstat.cpu.usr%
> 1038 -0.5% 1033 vmstat.memory.buff
> 1117006 -0.3% 1113807 vmstat.memory.cache
> 2.463e+08 -0.2% 2.457e+08 vmstat.memory.free
> 26.00 +1.9% 26.50 vmstat.procs.r
> 42239 ą 6% -55.9% 18619 ą 5% vmstat.system.cs
> 31359 -0.7% 31132 vmstat.system.in
> 0.00 -100.0% 0.00 numa-numastat.node0.interleave_hit
> 2713006 -1.0% 2685108 numa-numastat.node0.local_node
> 2716657 -1.0% 2689682 numa-numastat.node0.numa_hit
> 3651 ą 36% +25.3% 4576 ą 34% numa-numastat.node0.other_node
> 0.00 -100.0% 0.00 numa-numastat.node1.interleave_hit
> 2713025 -0.5% 2699801 numa-numastat.node1.local_node
> 2714924 -0.5% 2700769 numa-numastat.node1.numa_hit
> 1900 ą 68% -49.1% 968.00 ą162% numa-numastat.node1.other_node
> 21859882 ą 6% -56.1% 9599175 ą 2% cpuidle.C1.time
> 5231991 ą 6% -65.3% 1814200 ą 8% cpuidle.C1.usage
> 620147 ą 9% -22.4% 481528 ą 10% cpuidle.C1E.time
> 7829 ą 6% -34.5% 5126 ą 17% cpuidle.C1E.usage
> 5343219 ą 5% -58.8% 2202020 ą 4% cpuidle.C3.time
> 22942 ą 5% -54.9% 10349 ą 4% cpuidle.C3.usage
> 3.345e+08 ą 3% -15.1% 2.839e+08 ą 3% cpuidle.C6.time
> 355754 ą 3% -16.3% 297683 ą 3% cpuidle.C6.usage
> 248800 ą 6% -74.5% 63413 ą 5% cpuidle.POLL.time
> 90897 ą 6% -76.4% 21409 ą 7% cpuidle.POLL.usage
> 2631 +0.5% 2644 turbostat.Avg_MHz
> 95.35 +0.5 95.88 turbostat.Busy%
> 2759 -0.1% 2757 turbostat.Bzy_MHz
> 5227940 ą 6% -65.4% 1809983 ą 8% turbostat.C1
> 0.27 ą 5% -0.1 0.13 ą 3% turbostat.C1%
> 7646 ą 5% -36.0% 4894 ą 17% turbostat.C1E
> 0.01 +0.0 0.01 turbostat.C1E%
> 22705 ą 5% -56.0% 9995 ą 3% turbostat.C3
> 0.07 ą 7% -0.0 0.03 turbostat.C3%
> 354625 ą 3% -16.3% 296732 ą 3% turbostat.C6
> 4.11 ą 3% -0.4 3.75 ą 2% turbostat.C6%
> 1.68 ą 3% -15.6% 1.42 ą 2% turbostat.CPU%c1
> 0.04 -62.5% 0.01 ą 33% turbostat.CPU%c3
> 2.93 ą 3% -8.1% 2.69 ą 2% turbostat.CPU%c6
> 64.50 -1.6% 63.50 ą 2% turbostat.CoreTmp
> 9095020 -7.6% 8400668 turbostat.IRQ
> 11.78 ą 5% -2.3 9.45 ą 6% turbostat.PKG_%
> 0.10 ą 27% -4.9% 0.10 ą 22% turbostat.Pkg%pc2
> 0.00 ą173% -100.0% 0.00 turbostat.Pkg%pc6
> 68.50 ą 2% +0.0% 68.50 turbostat.PkgTmp
> 230.97 +0.1% 231.10 turbostat.PkgWatt
> 22.45 -0.8% 22.27 turbostat.RAMWatt
> 10304 -8.6% 9415 ą 2% turbostat.SMI
> 2300 +0.0% 2300 turbostat.TSC_MHz
> 269189 -0.6% 267505 meminfo.Active
> 269118 -0.6% 267429 meminfo.Active(anon)
> 167369 -1.7% 164478 meminfo.AnonHugePages
> 245606 +0.2% 246183 meminfo.AnonPages
> 1042 -0.2% 1039 meminfo.Buffers
> 1066883 -0.2% 1064908 meminfo.Cached
> 203421 -0.0% 203421 meminfo.CmaFree
> 204800 +0.0% 204800 meminfo.CmaTotal
> 1.32e+08 -0.0% 1.32e+08 meminfo.CommitLimit
> 485388 ą 15% -5.3% 459546 ą 9% meminfo.Committed_AS
> 2.65e+08 +0.0% 2.65e+08 meminfo.DirectMap1G
> 5240434 ą 8% +0.3% 5257281 ą 16% meminfo.DirectMap2M
> 169088 ą 6% -10.0% 152241 ą 5% meminfo.DirectMap4k
> 2048 +0.0% 2048 meminfo.Hugepagesize
> 150295 +0.1% 150389 meminfo.Inactive
> 149164 +0.1% 149264 meminfo.Inactive(anon)
> 1130 -0.6% 1124 meminfo.Inactive(file)
> 7375 -0.4% 7345 meminfo.KernelStack
> 28041 -0.5% 27895 meminfo.Mapped
> 2.451e+08 -0.2% 2.445e+08 meminfo.MemAvailable
> 2.462e+08 -0.2% 2.457e+08 meminfo.MemFree
> 2.64e+08 -0.0% 2.64e+08 meminfo.MemTotal
> 1179 ą 57% -35.2% 764.50 ą100% meminfo.Mlocked
> 4507108 +3.4% 4660163 meminfo.PageTables
> 50580 -1.9% 49633 meminfo.SReclaimable
> 11620141 +3.3% 12007592 meminfo.SUnreclaim
> 172885 -1.4% 170543 meminfo.Shmem
> 11670722 +3.3% 12057225 meminfo.Slab
> 894024 +0.0% 894327 meminfo.Unevictable
> 3.436e+10 +0.0% 3.436e+10 meminfo.VmallocTotal
> 6.873e+12 -2.1% 6.727e+12 perf-stat.branch-instructions
> 0.08 ą 2% -0.0 0.08 perf-stat.branch-miss-rate%
> 5.269e+09 ą 2% -3.7% 5.072e+09 perf-stat.branch-misses
> 37.34 +1.1 38.44 perf-stat.cache-miss-rate%
> 8.136e+09 ą 2% -7.0% 7.568e+09 ą 3% perf-stat.cache-misses
> 2.179e+10 ą 2% -9.6% 1.969e+10 ą 3% perf-stat.cache-references
> 12287807 ą 6% -59.1% 5020001 ą 5% perf-stat.context-switches
> 0.87 -3.5% 0.84 perf-stat.cpi
> 2.116e+13 -6.5% 1.978e+13 perf-stat.cpu-cycles
> 24160 ą 5% -13.8% 20819 ą 3% perf-stat.cpu-migrations
> 0.16 ą 6% +0.0 0.16 ą 4% perf-stat.dTLB-load-miss-rate%
> 1.057e+10 ą 6% +0.4% 1.061e+10 ą 5% perf-stat.dTLB-load-misses
> 6.792e+12 -3.5% 6.557e+12 perf-stat.dTLB-loads
> 0.00 ą 9% +0.0 0.00 ą 20% perf-stat.dTLB-store-miss-rate%
> 22950816 ą 9% +6.2% 24373827 ą 20% perf-stat.dTLB-store-misses
> 9.067e+11 -0.7% 9.005e+11 perf-stat.dTLB-stores
> 95.10 +2.7 97.81 perf-stat.iTLB-load-miss-rate%
> 2.437e+09 +6.8% 2.604e+09 ą 4% perf-stat.iTLB-load-misses
> 1.257e+08 ą 8% -53.7% 58211601 ą 6% perf-stat.iTLB-loads
> 2.44e+13 -3.1% 2.364e+13 perf-stat.instructions
> 10011 -9.1% 9100 ą 4%
> perf-stat.instructions-per-iTLB-miss
> 1.15 +3.6% 1.20 perf-stat.ipc
> 1.074e+09 -0.0% 1.074e+09 perf-stat.minor-faults
> 64.40 ą 4% -4.2 60.22 ą 5% perf-stat.node-load-miss-rate%
> 4.039e+09 ą 2% -10.9% 3.599e+09 ą 3% perf-stat.node-load-misses
> 2.241e+09 ą 10% +6.7% 2.39e+09 ą 12% perf-stat.node-loads
> 49.24 -1.7 47.53 perf-stat.node-store-miss-rate%
> 9.005e+08 -18.3% 7.357e+08 ą 2% perf-stat.node-store-misses
> 9.282e+08 -12.5% 8.123e+08 ą 2% perf-stat.node-stores
> 1.074e+09 -0.0% 1.074e+09 perf-stat.page-faults
> 5049 -3.1% 4893 perf-stat.path-length
> 67282 -0.6% 66860 proc-vmstat.nr_active_anon
> 61404 +0.2% 61545 proc-vmstat.nr_anon_pages
> 6117769 -0.2% 6104302
> proc-vmstat.nr_dirty_background_threshold
> 12250497 -0.2% 12223531 proc-vmstat.nr_dirty_threshold
> 266952 -0.2% 266464 proc-vmstat.nr_file_pages
> 50855 -0.0% 50855 proc-vmstat.nr_free_cma
> 61552505 -0.2% 61417644 proc-vmstat.nr_free_pages
> 37259 +0.1% 37286 proc-vmstat.nr_inactive_anon
> 282.25 -0.6% 280.50 proc-vmstat.nr_inactive_file
> 7375 -0.4% 7344 proc-vmstat.nr_kernel_stack
> 7124 -0.5% 7085 proc-vmstat.nr_mapped
> 295.00 ą 57% -35.3% 190.75 ą100% proc-vmstat.nr_mlock
> 1125531 +3.4% 1163908 proc-vmstat.nr_page_table_pages
> 43193 -1.3% 42613 proc-vmstat.nr_shmem
> 12644 -1.9% 12407 proc-vmstat.nr_slab_reclaimable
> 2901812 +3.3% 2998814 proc-vmstat.nr_slab_unreclaimable
> 223506 +0.0% 223581 proc-vmstat.nr_unevictable
> 67282 -0.6% 66860 proc-vmstat.nr_zone_active_anon
> 37259 +0.1% 37286 proc-vmstat.nr_zone_inactive_anon
> 282.25 -0.6% 280.50 proc-vmstat.nr_zone_inactive_file
> 223506 +0.0% 223581 proc-vmstat.nr_zone_unevictable
> 2685 ą104% +108.2% 5591 ą 84% proc-vmstat.numa_hint_faults
> 1757 ą140% +77.9% 3125 ą 87% proc-vmstat.numa_hint_faults_local
> 5456982 -0.7% 5417832 proc-vmstat.numa_hit
> 5451431 -0.7% 5412285 proc-vmstat.numa_local
> 5551 -0.1% 5547 proc-vmstat.numa_other
> 977.50 ą 40% +193.7% 2871 ą101% proc-vmstat.numa_pages_migrated
> 10519 ą113% +140.4% 25286 ą 97% proc-vmstat.numa_pte_updates
> 10726 ą 8% -3.8% 10315 ą 7% proc-vmstat.pgactivate
> 8191987 -0.5% 8150192 proc-vmstat.pgalloc_normal
> 1.074e+09 -0.0% 1.074e+09 proc-vmstat.pgfault
> 8143430 -2.7% 7926613 proc-vmstat.pgfree
> 977.50 ą 40% +193.7% 2871 ą101% proc-vmstat.pgmigrate_success
> 2155 -0.4% 2147 proc-vmstat.pgpgin
> 2049 -0.0% 2048 proc-vmstat.pgpgout
> 67375 -0.2% 67232
> slabinfo.Acpi-Namespace.active_objs
> 67375 -0.2% 67232 slabinfo.Acpi-Namespace.num_objs
> 604.00 ą 19% -0.1% 603.50 ą 9% slabinfo.Acpi-ParseExt.active_objs
> 604.00 ą 19% -0.1% 603.50 ą 9% slabinfo.Acpi-ParseExt.num_objs
> 7972 ą 4% -0.3% 7949 ą 2% slabinfo.anon_vma.active_objs
> 7972 ą 4% -0.3% 7949 ą 2% slabinfo.anon_vma.num_objs
> 1697 ą 12% -10.0% 1528 ą 7% slabinfo.avtab_node.active_objs
> 1697 ą 12% -10.0% 1528 ą 7% slabinfo.avtab_node.num_objs
> 58071 -0.1% 58034 slabinfo.dentry.active_objs
> 1351 ą 25% -31.2% 930.50 ą 30%
> slabinfo.dmaengine-unmap-16.active_objs
> 1351 ą 25% -31.2% 930.50 ą 30%
> slabinfo.dmaengine-unmap-16.num_objs
> 1354 ą 6% -7.1% 1258 ą 9% slabinfo.eventpoll_pwq.active_objs
> 1354 ą 6% -7.1% 1258 ą 9% slabinfo.eventpoll_pwq.num_objs
> 8880 ą 5% +0.0% 8883 ą 6% slabinfo.filp.num_objs
> 2715 ą 7% -9.0% 2470 ą 5% slabinfo.kmalloc-1024.active_objs
> 2831 ą 7% -10.2% 2543 ą 3% slabinfo.kmalloc-1024.num_objs
> 12480 -2.7% 12140 ą 2% slabinfo.kmalloc-16.active_objs
> 12480 -2.7% 12140 ą 2% slabinfo.kmalloc-16.num_objs
> 39520 ą 3% +1.0% 39922 slabinfo.kmalloc-32.active_objs
> 37573 +0.1% 37616 slabinfo.kmalloc-64.active_objs
> 37590 +0.2% 37673 slabinfo.kmalloc-64.num_objs
> 17182 +2.1% 17550 ą 2% slabinfo.kmalloc-8.active_objs
> 17663 +2.2% 18045 ą 2% slabinfo.kmalloc-8.num_objs
> 4428 ą 7% -2.9% 4298 ą 4% slabinfo.kmalloc-96.active_objs
> 956.50 ą 10% +4.5% 999.50 ą 16% slabinfo.nsproxy.active_objs
> 956.50 ą 10% +4.5% 999.50 ą 16% slabinfo.nsproxy.num_objs
> 19094 ą 3% -3.3% 18463 ą 5% slabinfo.pid.active_objs
> 19094 ą 3% -3.0% 18523 ą 5% slabinfo.pid.num_objs
> 2088 ą 14% -21.1% 1648 ą 8%
> slabinfo.skbuff_head_cache.active_objs
> 2136 ą 16% -19.5% 1720 ą 7%
> slabinfo.skbuff_head_cache.num_objs
> 872.00 ą 16% -9.3% 791.00 ą 15% slabinfo.task_group.active_objs
> 872.00 ą 16% -9.3% 791.00 ą 15% slabinfo.task_group.num_objs
> 57677936 +3.4% 59655533
> slabinfo.vm_area_struct.active_objs
> 1442050 +3.4% 1491482
> slabinfo.vm_area_struct.active_slabs
> 57682039 +3.4% 59659325 slabinfo.vm_area_struct.num_objs
> 1442050 +3.4% 1491482 slabinfo.vm_area_struct.num_slabs
> 132527 ą 15% +1.9% 135006 ą 2% numa-meminfo.node0.Active
> 132475 ą 15% +1.9% 134985 ą 2% numa-meminfo.node0.Active(anon)
> 86736 ą 32% +6.5% 92413 ą 15% numa-meminfo.node0.AnonHugePages
> 124613 ą 19% +3.4% 128788 ą 4% numa-meminfo.node0.AnonPages
> 556225 ą 11% -5.1% 527584 ą 12% numa-meminfo.node0.FilePages
> 100973 ą 57% -27.7% 73011 ą 89% numa-meminfo.node0.Inactive
> 100128 ą 57% -27.4% 72686 ą 90% numa-meminfo.node0.Inactive(anon)
> 843.75 ą 57% -61.5% 324.75 ą115% numa-meminfo.node0.Inactive(file)
> 4058 ą 3% +3.5% 4200 ą 4% numa-meminfo.node0.KernelStack
> 12846 ą 26% +1.6% 13053 ą 26% numa-meminfo.node0.Mapped
> 1.228e+08 -0.1% 1.226e+08 numa-meminfo.node0.MemFree
> 1.32e+08 +0.0% 1.32e+08 numa-meminfo.node0.MemTotal
> 9191573 +1.7% 9347737 numa-meminfo.node0.MemUsed
> 2321138 +2.2% 2372650 numa-meminfo.node0.PageTables
> 24816 ą 13% -5.8% 23377 ą 16% numa-meminfo.node0.SReclaimable
> 5985628 +2.2% 6116406 numa-meminfo.node0.SUnreclaim
> 108046 ą 57% -26.9% 78955 ą 80% numa-meminfo.node0.Shmem
> 6010444 +2.2% 6139784 numa-meminfo.node0.Slab
> 447344 +0.2% 448404 numa-meminfo.node0.Unevictable
> 136675 ą 13% -3.0% 132521 ą 2% numa-meminfo.node1.Active
> 136655 ą 13% -3.1% 132467 ą 2% numa-meminfo.node1.Active(anon)
> 80581 ą 35% -10.5% 72102 ą 18% numa-meminfo.node1.AnonHugePages
> 120993 ą 20% -3.0% 117407 ą 4% numa-meminfo.node1.AnonPages
> 511901 ą 12% +5.2% 538396 ą 11% numa-meminfo.node1.FilePages
> 49526 ą116% +56.3% 77402 ą 84% numa-meminfo.node1.Inactive
> 49238 ą115% +55.6% 76603 ą 85% numa-meminfo.node1.Inactive(anon)
> 287.75 ą168% +177.6% 798.75 ą 47% numa-meminfo.node1.Inactive(file)
> 3316 ą 3% -5.2% 3143 ą 5% numa-meminfo.node1.KernelStack
> 15225 ą 23% -2.0% 14918 ą 22% numa-meminfo.node1.Mapped
> 1.234e+08 -0.3% 1.23e+08 numa-meminfo.node1.MemFree
> 1.321e+08 -0.0% 1.321e+08 numa-meminfo.node1.MemTotal
> 8664187 +4.4% 9041729 numa-meminfo.node1.MemUsed
> 2185511 +4.6% 2286356 numa-meminfo.node1.PageTables
> 25762 ą 12% +1.9% 26255 ą 15% numa-meminfo.node1.SReclaimable
> 5634774 +4.5% 5887602 numa-meminfo.node1.SUnreclaim
> 65037 ą 92% +40.9% 91621 ą 68% numa-meminfo.node1.Shmem
> 5660536 +4.5% 5913858 numa-meminfo.node1.Slab
> 446680 -0.2% 445922 numa-meminfo.node1.Unevictable
> 15553 ą 18% -14.1% 13366 ą 11% numa-vmstat.node0
> 33116 ą 15% +1.9% 33742 ą 2% numa-vmstat.node0.nr_active_anon
> 31157 ą 19% +3.3% 32196 ą 4% numa-vmstat.node0.nr_anon_pages
> 139001 ą 11% -5.1% 131864 ą 12% numa-vmstat.node0.nr_file_pages
> 30692447 -0.1% 30654328 numa-vmstat.node0.nr_free_pages
> 24983 ą 57% -27.4% 18142 ą 90% numa-vmstat.node0.nr_inactive_anon
> 210.25 ą 57% -61.6% 80.75 ą116% numa-vmstat.node0.nr_inactive_file
> 4058 ą 3% +3.5% 4199 ą 4% numa-vmstat.node0.nr_kernel_stack
> 3304 ą 26% +1.2% 3344 ą 25% numa-vmstat.node0.nr_mapped
> 139.00 ą 60% -20.3% 110.75 ą100% numa-vmstat.node0.nr_mlock
> 579931 +2.1% 592262
> numa-vmstat.node0.nr_page_table_pages
> 26956 ą 57% -26.9% 19707 ą 80% numa-vmstat.node0.nr_shmem
> 6203 ą 13% -5.8% 5844 ą 16%
> numa-vmstat.node0.nr_slab_reclaimable
> 1495541 +2.2% 1527781
> numa-vmstat.node0.nr_slab_unreclaimable
> 111835 +0.2% 112100 numa-vmstat.node0.nr_unevictable
> 33116 ą 15% +1.9% 33742 ą 2%
> numa-vmstat.node0.nr_zone_active_anon
> 24983 ą 57% -27.4% 18142 ą 90%
> numa-vmstat.node0.nr_zone_inactive_anon
> 210.25 ą 57% -61.6% 80.75 ą116%
> numa-vmstat.node0.nr_zone_inactive_file
> 111835 +0.2% 112100
> numa-vmstat.node0.nr_zone_unevictable
> 1840693 ą 2% +1.6% 1869501 numa-vmstat.node0.numa_hit
> 144048 +0.2% 144385 numa-vmstat.node0.numa_interleave
> 1836656 ą 2% +1.5% 1864579 numa-vmstat.node0.numa_local
> 4036 ą 33% +21.9% 4921 ą 31% numa-vmstat.node0.numa_other
> 11577 ą 24% +17.8% 13635 ą 11% numa-vmstat.node1
> 34171 ą 13% -3.1% 33126 ą 2% numa-vmstat.node1.nr_active_anon
> 30247 ą 20% -3.0% 29352 ą 4% numa-vmstat.node1.nr_anon_pages
> 127979 ą 12% +5.2% 134601 ą 11% numa-vmstat.node1.nr_file_pages
> 50855 -0.0% 50855 numa-vmstat.node1.nr_free_cma
> 30858027 -0.3% 30763794 numa-vmstat.node1.nr_free_pages
> 12305 ą116% +55.6% 19145 ą 85% numa-vmstat.node1.nr_inactive_anon
> 71.75 ą168% +177.7% 199.25 ą 47% numa-vmstat.node1.nr_inactive_file
> 3315 ą 3% -5.2% 3144 ą 5% numa-vmstat.node1.nr_kernel_stack
> 3823 ą 23% -1.8% 3754 ą 22% numa-vmstat.node1.nr_mapped
> 155.00 ą 60% -48.2% 80.25 ą100% numa-vmstat.node1.nr_mlock
> 545973 +4.6% 571069
> numa-vmstat.node1.nr_page_table_pages
> 16263 ą 92% +40.9% 22907 ą 68% numa-vmstat.node1.nr_shmem
> 6440 ą 12% +1.9% 6563 ą 15%
> numa-vmstat.node1.nr_slab_reclaimable
> 1407904 +4.5% 1471030
> numa-vmstat.node1.nr_slab_unreclaimable
> 111669 -0.2% 111480 numa-vmstat.node1.nr_unevictable
> 34171 ą 13% -3.1% 33126 ą 2%
> numa-vmstat.node1.nr_zone_active_anon
> 12305 ą116% +55.6% 19145 ą 85%
> numa-vmstat.node1.nr_zone_inactive_anon
> 71.75 ą168% +177.7% 199.25 ą 47%
> numa-vmstat.node1.nr_zone_inactive_file
> 111669 -0.2% 111480
> numa-vmstat.node1.nr_zone_unevictable
> 1846889 ą 2% +1.4% 1872108 numa-vmstat.node1.numa_hit
> 144151 -0.2% 143830 numa-vmstat.node1.numa_interleave
> 1699975 ą 2% +1.6% 1726462 numa-vmstat.node1.numa_local
> 146913 -0.9% 145645 numa-vmstat.node1.numa_other
> 0.00 +1.2e+12% 12083 ą100%
> sched_debug.cfs_rq:/.MIN_vruntime.avg
> 0.00 +3.4e+13% 338347 ą100%
> sched_debug.cfs_rq:/.MIN_vruntime.max
> 0.00 +0.0% 0.00
> sched_debug.cfs_rq:/.MIN_vruntime.min
> 0.00 +1.5e+28% 62789 ą100%
> sched_debug.cfs_rq:/.MIN_vruntime.stddev
> 118226 +0.4% 118681
> sched_debug.cfs_rq:/.exec_clock.avg
> 119425 +0.3% 119724
> sched_debug.cfs_rq:/.exec_clock.max
> 117183 +0.1% 117244
> sched_debug.cfs_rq:/.exec_clock.min
> 395.73 ą 14% +11.0% 439.18 ą 23%
> sched_debug.cfs_rq:/.exec_clock.stddev
> 32398 +14.1% 36980 ą 9% sched_debug.cfs_rq:/.load.avg
> 73141 ą 5% +128.0% 166780 ą 57% sched_debug.cfs_rq:/.load.max
> 17867 ą 19% +30.4% 23301 ą 13% sched_debug.cfs_rq:/.load.min
> 11142 ą 3% +146.4% 27458 ą 62% sched_debug.cfs_rq:/.load.stddev
> 59.52 ą 2% -2.3% 58.13 ą 6% sched_debug.cfs_rq:/.load_avg.avg
> 305.15 ą 10% -8.8% 278.35 ą 3% sched_debug.cfs_rq:/.load_avg.max
> 27.20 ą 6% +8.6% 29.55 sched_debug.cfs_rq:/.load_avg.min
> 71.12 ą 9% -8.1% 65.38 ą 5%
> sched_debug.cfs_rq:/.load_avg.stddev
> 0.00 +1.2e+12% 12083 ą100%
> sched_debug.cfs_rq:/.max_vruntime.avg
> 0.00 +3.4e+13% 338347 ą100%
> sched_debug.cfs_rq:/.max_vruntime.max
> 0.00 +0.0% 0.00
> sched_debug.cfs_rq:/.max_vruntime.min
> 0.00 +1.5e+28% 62789 ą100%
> sched_debug.cfs_rq:/.max_vruntime.stddev
> 3364874 +0.8% 3391671
> sched_debug.cfs_rq:/.min_vruntime.avg
> 3399316 +0.7% 3423051
> sched_debug.cfs_rq:/.min_vruntime.max
> 3303899 +0.9% 3335083
> sched_debug.cfs_rq:/.min_vruntime.min
> 20061 ą 14% -2.5% 19552 ą 16%
> sched_debug.cfs_rq:/.min_vruntime.stddev
> 0.87 +3.1% 0.89 ą 2%
> sched_debug.cfs_rq:/.nr_running.avg
> 1.00 +5.0% 1.05 ą 8%
> sched_debug.cfs_rq:/.nr_running.max
> 0.50 ą 19% +20.0% 0.60
> sched_debug.cfs_rq:/.nr_running.min
> 0.16 ą 14% -6.0% 0.15 ą 11%
> sched_debug.cfs_rq:/.nr_running.stddev
> 4.14 ą 6% -6.5% 3.87 ą 10%
> sched_debug.cfs_rq:/.nr_spread_over.avg
> 15.10 ą 7% +29.8% 19.60 ą 12%
> sched_debug.cfs_rq:/.nr_spread_over.max
> 1.50 ą 11% -16.7% 1.25 ą 20%
> sched_debug.cfs_rq:/.nr_spread_over.min
> 2.82 ą 8% +23.0% 3.47 ą 13%
> sched_debug.cfs_rq:/.nr_spread_over.stddev
> 7.31 -6.2% 6.86 ą 69%
> sched_debug.cfs_rq:/.removed.load_avg.avg
> 204.80 -28.2% 147.10 ą 57%
> sched_debug.cfs_rq:/.removed.load_avg.max
> 38.01 -20.3% 30.29 ą 60%
> sched_debug.cfs_rq:/.removed.load_avg.stddev
> 337.44 -6.0% 317.11 ą 69%
> sched_debug.cfs_rq:/.removed.runnable_sum.avg
> 9448 -27.8% 6819 ą 57%
> sched_debug.cfs_rq:/.removed.runnable_sum.max
> 1753 -20.2% 1399 ą 60%
> sched_debug.cfs_rq:/.removed.runnable_sum.stddev
> 2.16 ą 56% -25.7% 1.60 ą 57%
> sched_debug.cfs_rq:/.removed.util_avg.avg
> 60.40 ą 56% -37.3% 37.90 ą 61%
> sched_debug.cfs_rq:/.removed.util_avg.max
> 11.21 ą 56% -33.8% 7.42 ą 58%
> sched_debug.cfs_rq:/.removed.util_avg.stddev
> 30.32 ą 2% +0.6% 30.52 ą 3%
> sched_debug.cfs_rq:/.runnable_load_avg.avg
> 80.90 ą 14% -19.2% 65.35 ą 34%
> sched_debug.cfs_rq:/.runnable_load_avg.max
> 14.95 ą 24% +47.2% 22.00 ą 12%
> sched_debug.cfs_rq:/.runnable_load_avg.min
> 12.10 ą 19% -27.7% 8.74 ą 42%
> sched_debug.cfs_rq:/.runnable_load_avg.stddev
> 30853 +13.2% 34914 ą 10%
> sched_debug.cfs_rq:/.runnable_weight.avg
> 61794 +152.3% 155911 ą 64%
> sched_debug.cfs_rq:/.runnable_weight.max
> 17867 ą 19% +30.4% 23300 ą 13%
> sched_debug.cfs_rq:/.runnable_weight.min
> 8534 ą 7% +193.7% 25066 ą 72%
> sched_debug.cfs_rq:/.runnable_weight.stddev
> 45914 ą 45% -50.7% 22619 ą 62% sched_debug.cfs_rq:/.spread0.avg
> 80385 ą 26% -32.8% 53990 ą 34% sched_debug.cfs_rq:/.spread0.max
> -15013 +126.3% -33973 sched_debug.cfs_rq:/.spread0.min
> 20053 ą 14% -2.5% 19541 ą 16%
> sched_debug.cfs_rq:/.spread0.stddev
> 964.19 -0.2% 962.12 sched_debug.cfs_rq:/.util_avg.avg
> 1499 ą 14% -13.2% 1301 ą 4% sched_debug.cfs_rq:/.util_avg.max
> 510.75 ą 12% +35.6% 692.65 ą 13% sched_debug.cfs_rq:/.util_avg.min
> 177.84 ą 21% -34.4% 116.72 ą 23%
> sched_debug.cfs_rq:/.util_avg.stddev
> 768.04 +5.1% 807.40
> sched_debug.cfs_rq:/.util_est_enqueued.avg
> 1192 ą 14% -22.5% 924.20
> sched_debug.cfs_rq:/.util_est_enqueued.max
> 170.90 ą 99% +89.6% 324.05 ą 21%
> sched_debug.cfs_rq:/.util_est_enqueued.min
> 201.22 ą 17% -36.7% 127.44 ą 15%
> sched_debug.cfs_rq:/.util_est_enqueued.stddev
> 111567 ą 4% +7.6% 120067 ą 7% sched_debug.cpu.avg_idle.avg
> 549432 ą 16% -16.6% 458264 ą 5% sched_debug.cpu.avg_idle.max
> 4419 ą 79% +87.7% 8293 ą 31% sched_debug.cpu.avg_idle.min
> 123967 ą 13% -8.9% 112928 ą 15% sched_debug.cpu.avg_idle.stddev
> 147256 -0.4% 146724 sched_debug.cpu.clock.avg
> 147258 -0.4% 146728 sched_debug.cpu.clock.max
> 147252 -0.4% 146720 sched_debug.cpu.clock.min
> 1.70 ą 11% +39.5% 2.37 ą 29% sched_debug.cpu.clock.stddev
> 147256 -0.4% 146724 sched_debug.cpu.clock_task.avg
> 147258 -0.4% 146728 sched_debug.cpu.clock_task.max
> 147252 -0.4% 146720 sched_debug.cpu.clock_task.min
> 1.70 ą 11% +39.4% 2.37 ą 29% sched_debug.cpu.clock_task.stddev
> 30.85 +0.4% 30.97 ą 3% sched_debug.cpu.cpu_load[0].avg
> 84.15 ą 13% -13.8% 72.55 ą 21% sched_debug.cpu.cpu_load[0].max
> 17.35 ą 24% +26.5% 21.95 ą 24% sched_debug.cpu.cpu_load[0].min
> 12.75 ą 18% -19.9% 10.22 ą 22% sched_debug.cpu.cpu_load[0].stddev
> 30.88 +1.1% 31.22 ą 2% sched_debug.cpu.cpu_load[1].avg
> 77.35 ą 19% -9.8% 69.75 ą 18% sched_debug.cpu.cpu_load[1].max
> 18.60 ą 19% +31.2% 24.40 ą 10% sched_debug.cpu.cpu_load[1].min
> 11.10 ą 24% -20.5% 8.82 ą 25% sched_debug.cpu.cpu_load[1].stddev
> 31.13 +2.3% 31.84 ą 2% sched_debug.cpu.cpu_load[2].avg
> 71.40 ą 26% +0.4% 71.70 ą 20% sched_debug.cpu.cpu_load[2].max
> 19.45 ą 21% +32.9% 25.85 ą 5% sched_debug.cpu.cpu_load[2].min
> 9.61 ą 36% -8.3% 8.81 ą 27% sched_debug.cpu.cpu_load[2].stddev
> 31.79 +2.9% 32.71 ą 3% sched_debug.cpu.cpu_load[3].avg
> 76.75 ą 19% +8.1% 82.95 ą 37% sched_debug.cpu.cpu_load[3].max
> 20.25 ą 14% +33.3% 27.00 ą 3% sched_debug.cpu.cpu_load[3].min
> 9.88 ą 28% +6.7% 10.54 ą 52% sched_debug.cpu.cpu_load[3].stddev
> 32.95 +2.3% 33.71 ą 3% sched_debug.cpu.cpu_load[4].avg
> 107.65 ą 9% +3.9% 111.90 ą 33% sched_debug.cpu.cpu_load[4].max
> 20.35 ą 8% +30.5% 26.55 sched_debug.cpu.cpu_load[4].min
> 15.33 ą 12% +1.2% 15.51 ą 46% sched_debug.cpu.cpu_load[4].stddev
> 1244 +2.0% 1269 sched_debug.cpu.curr->pid.avg
> 4194 -0.7% 4165 sched_debug.cpu.curr->pid.max
> 655.90 ą 22% +16.7% 765.75 ą 3% sched_debug.cpu.curr->pid.min
> 656.54 -0.8% 651.46 sched_debug.cpu.curr->pid.stddev
> 32481 +13.8% 36973 ą 9% sched_debug.cpu.load.avg
> 73132 ą 5% +130.1% 168294 ą 56% sched_debug.cpu.load.max
> 17867 ą 19% +20.3% 21497 sched_debug.cpu.load.min
> 11194 ą 3% +148.6% 27826 ą 61% sched_debug.cpu.load.stddev
> 500000 +0.0% 500000
> sched_debug.cpu.max_idle_balance_cost.avg
> 500000 +0.0% 500000
> sched_debug.cpu.max_idle_balance_cost.max
> 500000 +0.0% 500000
> sched_debug.cpu.max_idle_balance_cost.min
> 4294 -0.0% 4294 sched_debug.cpu.next_balance.avg
> 4294 -0.0% 4294 sched_debug.cpu.next_balance.max
> 4294 -0.0% 4294 sched_debug.cpu.next_balance.min
> 0.00 ą 4% -3.6% 0.00 ą 5%
> sched_debug.cpu.next_balance.stddev
> 126191 -0.0% 126167
> sched_debug.cpu.nr_load_updates.avg
> 133303 -0.6% 132467
> sched_debug.cpu.nr_load_updates.max
> 124404 +0.4% 124852
> sched_debug.cpu.nr_load_updates.min
> 1826 ą 13% -8.5% 1672 ą 5%
> sched_debug.cpu.nr_load_updates.stddev
> 0.90 +2.2% 0.92 ą 2% sched_debug.cpu.nr_running.avg
> 1.80 ą 7% -5.6% 1.70 ą 5% sched_debug.cpu.nr_running.max
> 0.50 ą 19% +20.0% 0.60 sched_debug.cpu.nr_running.min
> 0.29 ą 7% -13.6% 0.25 ą 4% sched_debug.cpu.nr_running.stddev
> 204123 ą 8% -56.8% 88239 ą 5% sched_debug.cpu.nr_switches.avg
> 457439 ą 15% -60.4% 181234 ą 8% sched_debug.cpu.nr_switches.max
> 116531 ą 16% -62.8% 43365 ą 13% sched_debug.cpu.nr_switches.min
> 72910 ą 22% -50.5% 36095 ą 13% sched_debug.cpu.nr_switches.stddev
> 0.03 ą 19% -52.9% 0.01 ą 35%
> sched_debug.cpu.nr_uninterruptible.avg
> 16.50 ą 14% -11.5% 14.60 ą 16%
> sched_debug.cpu.nr_uninterruptible.max
> -16.90 -26.3% -12.45
> sched_debug.cpu.nr_uninterruptible.min
> 7.97 ą 12% -19.4% 6.42 ą 15%
> sched_debug.cpu.nr_uninterruptible.stddev
> 210655 ą 8% -56.7% 91289 ą 5% sched_debug.cpu.sched_count.avg
> 467518 ą 15% -60.4% 185362 ą 9% sched_debug.cpu.sched_count.max
> 120764 ą 16% -62.5% 45233 ą 14% sched_debug.cpu.sched_count.min
> 74259 ą 23% -51.0% 36403 ą 14% sched_debug.cpu.sched_count.stddev
> 89621 ą 8% -63.5% 32750 ą 7% sched_debug.cpu.sched_goidle.avg
> 189668 ą 8% -67.5% 61630 ą 18% sched_debug.cpu.sched_goidle.max
> 54342 ą 16% -68.0% 17412 ą 15% sched_debug.cpu.sched_goidle.min
> 28685 ą 15% -58.7% 11834 ą 16%
> sched_debug.cpu.sched_goidle.stddev
> 109820 ą 8% -56.0% 48303 ą 5% sched_debug.cpu.ttwu_count.avg
> 144975 ą 13% -41.1% 85424 ą 7% sched_debug.cpu.ttwu_count.max
> 96409 ą 8% -61.1% 37542 ą 6% sched_debug.cpu.ttwu_count.min
> 12310 ą 20% -2.4% 12009 ą 14% sched_debug.cpu.ttwu_count.stddev
> 9749 ą 10% -5.0% 9257 ą 6% sched_debug.cpu.ttwu_local.avg
> 45094 ą 30% +4.8% 47270 ą 13% sched_debug.cpu.ttwu_local.max
> 1447 ą 6% +1.2% 1465 ą 7% sched_debug.cpu.ttwu_local.min
> 12231 ą 24% -6.1% 11487 ą 13% sched_debug.cpu.ttwu_local.stddev
> 147253 -0.4% 146720 sched_debug.cpu_clk
> 996147 +0.0% 996147 sched_debug.dl_rq:.dl_bw->bw.avg
> 996147 +0.0% 996147 sched_debug.dl_rq:.dl_bw->bw.max
> 996147 +0.0% 996147 sched_debug.dl_rq:.dl_bw->bw.min
> 4.295e+09 -0.0% 4.295e+09 sched_debug.jiffies
> 147253 -0.4% 146720 sched_debug.ktime
> 950.00 +0.0% 950.00 sched_debug.rt_rq:/.rt_runtime.avg
> 950.00 +0.0% 950.00 sched_debug.rt_rq:/.rt_runtime.max
> 950.00 +0.0% 950.00 sched_debug.rt_rq:/.rt_runtime.min
> 0.00 ą146% -69.8% 0.00 ą100% sched_debug.rt_rq:/.rt_time.avg
> 0.04 ą146% -69.8% 0.01 ą100% sched_debug.rt_rq:/.rt_time.max
> 0.01 ą146% -69.8% 0.00 ą100% sched_debug.rt_rq:/.rt_time.stddev
> 147626 -0.3% 147114 sched_debug.sched_clk
> 1.00 +0.0% 1.00 sched_debug.sched_clock_stable()
> 4118331 +0.0% 4118331
> sched_debug.sysctl_sched.sysctl_sched_features
> 24.00 +0.0% 24.00
> sched_debug.sysctl_sched.sysctl_sched_latency
> 3.00 +0.0% 3.00
> sched_debug.sysctl_sched.sysctl_sched_min_granularity
> 1.00 +0.0% 1.00
> sched_debug.sysctl_sched.sysctl_sched_tunable_scaling
> 4.00 +0.0% 4.00
> sched_debug.sysctl_sched.sysctl_sched_wakeup_granularity
> 68.63 -2.2 66.43
> perf-profile.calltrace.cycles-pp.osq_lock.rwsem_down_write_failed.call_rwsem_down_write_failed.down_write.vma_link
> 73.63 -1.9 71.69
> perf-profile.calltrace.cycles-pp.call_rwsem_down_write_failed.down_write.vma_link.mmap_region.do_mmap
> 73.63 -1.9 71.69
> perf-profile.calltrace.cycles-pp.rwsem_down_write_failed.call_rwsem_down_write_failed.down_write.vma_link.mmap_region
> 73.99 -1.9 72.05
> perf-profile.calltrace.cycles-pp.down_write.vma_link.mmap_region.do_mmap.vm_mmap_pgoff
> 78.36 -1.7 76.67
> perf-profile.calltrace.cycles-pp.vma_link.mmap_region.do_mmap.vm_mmap_pgoff.ksys_mmap_pgoff
> 79.68 -1.5 78.19
> perf-profile.calltrace.cycles-pp.mmap_region.do_mmap.vm_mmap_pgoff.ksys_mmap_pgoff.do_syscall_64
> 81.34 -1.2 80.10
> perf-profile.calltrace.cycles-pp.do_mmap.vm_mmap_pgoff.ksys_mmap_pgoff.do_syscall_64.entry_SYSCALL_64_after_hwframe
> 83.36 -1.2 82.15
> perf-profile.calltrace.cycles-pp.do_syscall_64.entry_SYSCALL_64_after_hwframe
> 81.59 -1.2 80.39
> perf-profile.calltrace.cycles-pp.vm_mmap_pgoff.ksys_mmap_pgoff.do_syscall_64.entry_SYSCALL_64_after_hwframe
> 83.38 -1.2 82.18
> perf-profile.calltrace.cycles-pp.entry_SYSCALL_64_after_hwframe
> 82.00 -1.2 80.83
> perf-profile.calltrace.cycles-pp.ksys_mmap_pgoff.do_syscall_64.entry_SYSCALL_64_after_hwframe
> 2.44 ą 5% -0.7 1.79
> perf-profile.calltrace.cycles-pp.__handle_mm_fault.handle_mm_fault.__do_page_fault.do_page_fault.page_fault
> 3.17 ą 4% -0.6 2.62
> perf-profile.calltrace.cycles-pp.handle_mm_fault.__do_page_fault.do_page_fault.page_fault
> 1.15 ą 2% -0.2 0.97 ą 2%
> perf-profile.calltrace.cycles-pp.up_write.vma_link.mmap_region.do_mmap.vm_mmap_pgoff
> 1.27 ą 18% -0.0 1.24 ą 16%
> perf-profile.calltrace.cycles-pp.task_numa_work.task_work_run.exit_to_usermode_loop.do_syscall_64.entry_SYSCALL_64_after_hwframe
> 1.27 ą 18% -0.0 1.24 ą 16%
> perf-profile.calltrace.cycles-pp.exit_to_usermode_loop.do_syscall_64.entry_SYSCALL_64_after_hwframe
> 1.27 ą 18% -0.0 1.24 ą 16%
> perf-profile.calltrace.cycles-pp.task_work_run.exit_to_usermode_loop.do_syscall_64.entry_SYSCALL_64_after_hwframe
> 0.55 ą 4% +0.0 0.58 ą 3%
> perf-profile.calltrace.cycles-pp.__rb_insert_augmented.vma_link.mmap_region.do_mmap.vm_mmap_pgoff
> 0.94 ą 14% +0.1 1.02 ą 12%
> perf-profile.calltrace.cycles-pp.task_numa_work.task_work_run.exit_to_usermode_loop.prepare_exit_to_usermode.swapgs_restore_regs_and_return_to_usermode
> 0.94 ą 14% +0.1 1.02 ą 12%
> perf-profile.calltrace.cycles-pp.exit_to_usermode_loop.prepare_exit_to_usermode.swapgs_restore_regs_and_return_to_usermode
> 0.94 ą 14% +0.1 1.02 ą 12%
> perf-profile.calltrace.cycles-pp.task_work_run.exit_to_usermode_loop.prepare_exit_to_usermode.swapgs_restore_regs_and_return_to_usermode
> 6.67 +0.1 6.76
> perf-profile.calltrace.cycles-pp.__do_page_fault.do_page_fault.page_fault
> 0.95 ą 14% +0.1 1.04 ą 11%
> perf-profile.calltrace.cycles-pp.prepare_exit_to_usermode.swapgs_restore_regs_and_return_to_usermode
> 6.70 +0.1 6.79
> perf-profile.calltrace.cycles-pp.do_page_fault.page_fault
> 6.75 +0.1 6.86
> perf-profile.calltrace.cycles-pp.page_fault
> 0.61 ą 6% +0.1 0.75 ą 3%
> perf-profile.calltrace.cycles-pp.___perf_sw_event.__perf_sw_event.__do_page_fault.do_page_fault.page_fault
> 1.09 ą 2% +0.1 1.23 ą 3%
> perf-profile.calltrace.cycles-pp.vmacache_find.find_vma.__do_page_fault.do_page_fault.page_fault
> 0.74 ą 4% +0.2 0.91 ą 3%
> perf-profile.calltrace.cycles-pp.__perf_sw_event.__do_page_fault.do_page_fault.page_fault
> 1.35 +0.2 1.56
> perf-profile.calltrace.cycles-pp.unmapped_area_topdown.arch_get_unmapped_area_topdown.get_unmapped_area.do_mmap.vm_mmap_pgoff
> 1.39 +0.2 1.61
> perf-profile.calltrace.cycles-pp.arch_get_unmapped_area_topdown.get_unmapped_area.do_mmap.vm_mmap_pgoff.ksys_mmap_pgoff
> 1.56 +0.2 1.78 ą 2%
> perf-profile.calltrace.cycles-pp.find_vma.__do_page_fault.do_page_fault.page_fault
> 1.50 +0.2 1.72
> perf-profile.calltrace.cycles-pp.get_unmapped_area.do_mmap.vm_mmap_pgoff.ksys_mmap_pgoff.do_syscall_64
> 2.71 +0.3 3.05
> perf-profile.calltrace.cycles-pp.native_irq_return_iret
> 2.15 +0.4 2.50 ą 3%
> perf-profile.calltrace.cycles-pp.vma_interval_tree_insert.vma_link.mmap_region.do_mmap.vm_mmap_pgoff
> 2.70 ą 5% +0.4 3.07 ą 3%
> perf-profile.calltrace.cycles-pp.swapgs_restore_regs_and_return_to_usermode
> 3.83 +0.4 4.26
> perf-profile.calltrace.cycles-pp.rwsem_spin_on_owner.rwsem_down_write_failed.call_rwsem_down_write_failed.down_write.vma_link
> 0.00 +0.5 0.52
> perf-profile.calltrace.cycles-pp.perf_event_mmap.mmap_region.do_mmap.vm_mmap_pgoff.ksys_mmap_pgoff
> 68.65 -2.2 66.45
> perf-profile.children.cycles-pp.osq_lock
> 73.63 -1.9 71.69
> perf-profile.children.cycles-pp.call_rwsem_down_write_failed
> 73.63 -1.9 71.69
> perf-profile.children.cycles-pp.rwsem_down_write_failed
> 73.99 -1.9 72.05
> perf-profile.children.cycles-pp.down_write
> 78.36 -1.7 76.68
> perf-profile.children.cycles-pp.vma_link
> 79.70 -1.5 78.21
> perf-profile.children.cycles-pp.mmap_region
> 81.35 -1.2 80.12
> perf-profile.children.cycles-pp.do_mmap
> 81.61 -1.2 80.41
> perf-profile.children.cycles-pp.vm_mmap_pgoff
> 83.48 -1.2 82.28
> perf-profile.children.cycles-pp.entry_SYSCALL_64_after_hwframe
> 83.46 -1.2 82.26
> perf-profile.children.cycles-pp.do_syscall_64
> 82.01 -1.2 80.85
> perf-profile.children.cycles-pp.ksys_mmap_pgoff
> 2.51 ą 5% -0.6 1.87
> perf-profile.children.cycles-pp.__handle_mm_fault
> 3.24 ą 4% -0.5 2.70
> perf-profile.children.cycles-pp.handle_mm_fault
> 1.23 ą 2% -0.2 1.06 ą 2%
> perf-profile.children.cycles-pp.up_write
> 0.20 ą 12% -0.1 0.10 ą 7%
> perf-profile.children.cycles-pp.do_idle
> 0.20 ą 12% -0.1 0.10 ą 7%
> perf-profile.children.cycles-pp.secondary_startup_64
> 0.20 ą 12% -0.1 0.10 ą 7%
> perf-profile.children.cycles-pp.cpu_startup_entry
> 0.18 ą 13% -0.1 0.09 ą 13%
> perf-profile.children.cycles-pp.start_secondary
> 0.26 ą 11% -0.1 0.18 ą 8%
> perf-profile.children.cycles-pp.rwsem_wake
> 0.26 ą 11% -0.1 0.18 ą 6%
> perf-profile.children.cycles-pp.call_rwsem_wake
> 0.07 ą 17% -0.1 0.00
> perf-profile.children.cycles-pp.intel_idle
> 0.08 ą 14% -0.1 0.01 ą173%
> perf-profile.children.cycles-pp.cpuidle_enter_state
> 0.06 ą 13% -0.1 0.00
> perf-profile.children.cycles-pp.schedule
> 0.06 ą 6% -0.1 0.00
> perf-profile.children.cycles-pp.save_stack_trace_tsk
> 0.18 ą 9% -0.1 0.12
> perf-profile.children.cycles-pp.wake_up_q
> 0.06 ą 7% -0.1 0.00
> perf-profile.children.cycles-pp.sched_ttwu_pending
> 0.06 ą 7% -0.1 0.00
> perf-profile.children.cycles-pp.__save_stack_trace
> 0.18 ą 8% -0.1 0.13
> perf-profile.children.cycles-pp.try_to_wake_up
> 0.05 -0.1 0.00
> perf-profile.children.cycles-pp.unwind_next_frame
> 0.11 ą 6% -0.0 0.06 ą 13%
> perf-profile.children.cycles-pp.enqueue_task_fair
> 0.12 ą 9% -0.0 0.07 ą 10%
> perf-profile.children.cycles-pp.ttwu_do_activate
> 0.08 ą 5% -0.0 0.04 ą 57%
> perf-profile.children.cycles-pp.__account_scheduler_latency
> 0.11 ą 10% -0.0 0.06 ą 13%
> perf-profile.children.cycles-pp.enqueue_entity
> 0.10 ą 17% -0.0 0.06 ą 7%
> perf-profile.children.cycles-pp.__schedule
> 0.36 ą 5% -0.0 0.32 ą 7%
> perf-profile.children.cycles-pp.osq_unlock
> 0.03 ą100% -0.0 0.00
> perf-profile.children.cycles-pp.serial8250_console_write
> 0.03 ą100% -0.0 0.00
> perf-profile.children.cycles-pp.uart_console_write
> 0.03 ą100% -0.0 0.00
> perf-profile.children.cycles-pp.wait_for_xmitr
> 0.03 ą100% -0.0 0.00
> perf-profile.children.cycles-pp.serial8250_console_putchar
> 0.03 ą100% -0.0 0.00
> perf-profile.children.cycles-pp.__softirqentry_text_start
> 0.08 ą 11% -0.0 0.05 ą 9%
> perf-profile.children.cycles-pp._raw_spin_lock_irqsave
> 0.03 ą100% -0.0 0.01 ą173%
> perf-profile.children.cycles-pp.console_unlock
> 0.05 ą 9% -0.0 0.04 ą 58%
> perf-profile.children.cycles-pp.update_load_avg
> 0.03 ą100% -0.0 0.01 ą173%
> perf-profile.children.cycles-pp.irq_work_run_list
> 0.03 ą100% -0.0 0.01 ą173%
> perf-profile.children.cycles-pp.perf_mux_hrtimer_handler
> 0.03 ą100% -0.0 0.01 ą173%
> perf-profile.children.cycles-pp.irq_exit
> 0.01 ą173% -0.0 0.00
> perf-profile.children.cycles-pp.process_one_work
> 0.01 ą173% -0.0 0.00
> perf-profile.children.cycles-pp.ktime_get
> 0.01 ą173% -0.0 0.00
> perf-profile.children.cycles-pp.__vma_link_file
> 0.37 ą 3% -0.0 0.36 ą 5%
> perf-profile.children.cycles-pp.apic_timer_interrupt
> 0.24 ą 6% -0.0 0.23 ą 6%
> perf-profile.children.cycles-pp.__hrtimer_run_queues
> 0.37 ą 4% -0.0 0.36 ą 5%
> perf-profile.children.cycles-pp.smp_apic_timer_interrupt
> 0.06 ą 14% -0.0 0.05 ą 9%
> perf-profile.children.cycles-pp.file_has_perm
> 0.31 ą 5% -0.0 0.31 ą 7%
> perf-profile.children.cycles-pp.hrtimer_interrupt
> 0.16 ą 10% -0.0 0.15 ą 7%
> perf-profile.children.cycles-pp.update_process_times
> 0.04 ą 57% -0.0 0.04 ą 58%
> perf-profile.children.cycles-pp.write
> 0.06 ą 7% -0.0 0.06 ą 9%
> perf-profile.children.cycles-pp.native_iret
> 0.62 ą 5% +0.0 0.62 ą 2%
> perf-profile.children.cycles-pp.__rb_insert_augmented
> 0.16 ą 10% +0.0 0.16 ą 9%
> perf-profile.children.cycles-pp.tick_sched_handle
> 0.01 ą173% +0.0 0.01 ą173%
> perf-profile.children.cycles-pp.ksys_write
> 0.01 ą173% +0.0 0.01 ą173%
> perf-profile.children.cycles-pp.worker_thread
> 0.18 ą 9% +0.0 0.18 ą 8%
> perf-profile.children.cycles-pp.tick_sched_timer
> 0.05 ą 8% +0.0 0.06 ą 9%
> perf-profile.children.cycles-pp._cond_resched
> 0.11 ą 4% +0.0 0.11 ą 3%
> perf-profile.children.cycles-pp.scheduler_tick
> 0.08 ą 10% +0.0 0.09 ą 13%
> perf-profile.children.cycles-pp.mem_cgroup_from_task
> 0.08 ą 5% +0.0 0.09 ą 7%
> perf-profile.children.cycles-pp.task_tick_fair
> 0.06 +0.0 0.07 ą 7%
> perf-profile.children.cycles-pp.security_mmap_addr
> 0.07 ą 7% +0.0 0.07 ą 5%
> perf-profile.children.cycles-pp.pmd_devmap_trans_unstable
> 0.09 ą 5% +0.0 0.10 ą 5%
> perf-profile.children.cycles-pp.vma_gap_callbacks_rotate
> 0.01 ą173% +0.0 0.03 ą100%
> perf-profile.children.cycles-pp.ret_from_fork
> 0.01 ą173% +0.0 0.03 ą100%
> perf-profile.children.cycles-pp.kthread
> 0.06 ą 13% +0.0 0.07 ą 5%
> perf-profile.children.cycles-pp.fput
> 0.07 ą 7% +0.0 0.08 ą 6%
> perf-profile.children.cycles-pp.__slab_alloc
> 0.05 +0.0 0.06 ą 6%
> perf-profile.children.cycles-pp.prepend_path
> 0.05 +0.0 0.06 ą 6%
> perf-profile.children.cycles-pp.new_slab
> 0.00 +0.0 0.01 ą173%
> perf-profile.children.cycles-pp.ktime_get_update_offsets_now
> 0.00 +0.0 0.01 ą173%
> perf-profile.children.cycles-pp.vfs_write
> 0.00 +0.0 0.01 ą173%
> perf-profile.children.cycles-pp.down_write_killable
> 0.06 ą 6% +0.0 0.08 ą 6%
> perf-profile.children.cycles-pp.___slab_alloc
> 0.07 +0.0 0.08 ą 5%
> perf-profile.children.cycles-pp.perf_exclude_event
> 0.04 ą 58% +0.0 0.06 ą 9%
> perf-profile.children.cycles-pp.get_page_from_freelist
> 0.04 ą 58% +0.0 0.06 ą 7%
> perf-profile.children.cycles-pp.__alloc_pages_nodemask
> 0.03 ą100% +0.0 0.04 ą 58%
> perf-profile.children.cycles-pp.__pte_alloc
> 0.21 ą 3% +0.0 0.23 ą 6%
> perf-profile.children.cycles-pp.__entry_trampoline_start
> 0.21 ą 4% +0.0 0.23 ą 2%
> perf-profile.children.cycles-pp.vma_interval_tree_augment_rotate
> 0.11 ą 7% +0.0 0.14 ą 8%
> perf-profile.children.cycles-pp.selinux_mmap_file
> 0.04 ą 57% +0.0 0.06
> perf-profile.children.cycles-pp.kfree
> 0.04 ą 57% +0.0 0.06 ą 11%
> perf-profile.children.cycles-pp.avc_has_perm
> 0.17 ą 4% +0.0 0.19 ą 4%
> perf-profile.children.cycles-pp.d_path
> 0.00 +0.0 0.03 ą100%
> perf-profile.children.cycles-pp.pte_alloc_one
> 0.01 ą173% +0.0 0.04 ą 57%
> perf-profile.children.cycles-pp.perf_swevent_event
> 0.12 ą 10% +0.0 0.15 ą 7%
> perf-profile.children.cycles-pp.security_mmap_file
> 0.18 ą 3% +0.0 0.21 ą 3%
> perf-profile.children.cycles-pp.up_read
> 0.14 ą 3% +0.0 0.17 ą 3%
> perf-profile.children.cycles-pp.__might_sleep
> 0.32 ą 3% +0.0 0.35 ą 4%
> perf-profile.children.cycles-pp.__fget
> 0.12 ą 3% +0.0 0.15 ą 2%
> perf-profile.children.cycles-pp.kmem_cache_alloc_trace
> 0.16 ą 4% +0.0 0.19 ą 5%
> perf-profile.children.cycles-pp.kmem_cache_alloc
> 0.21 ą 3% +0.0 0.25 ą 4%
> perf-profile.children.cycles-pp.down_read_trylock
> 0.25 ą 5% +0.0 0.29 ą 4%
> perf-profile.children.cycles-pp.__vma_link_rb
> 0.17 ą 6% +0.0 0.21 ą 7%
> perf-profile.children.cycles-pp.vma_compute_subtree_gap
> 2.22 ą 15% +0.0 2.26 ą 14%
> perf-profile.children.cycles-pp.exit_to_usermode_loop
> 0.21 ą 7% +0.0 0.26 ą 4%
> perf-profile.children.cycles-pp._raw_spin_lock
> 0.00 +0.0 0.04 ą 58%
> perf-profile.children.cycles-pp.perf_iterate_sb
> 2.22 ą 15% +0.0 2.27 ą 14%
> perf-profile.children.cycles-pp.task_numa_work
> 0.16 ą 5% +0.0 0.20 ą 4%
> perf-profile.children.cycles-pp.___might_sleep
> 0.24 ą 16% +0.0 0.29 ą 13%
> perf-profile.children.cycles-pp.vma_policy_mof
> 2.21 ą 15% +0.0 2.26 ą 14%
> perf-profile.children.cycles-pp.task_work_run
> 0.37 ą 3% +0.0 0.42 ą 5%
> perf-profile.children.cycles-pp.sync_regs
> 0.08 ą 23% +0.0 0.13 ą 14%
> perf-profile.children.cycles-pp.get_task_policy
> 0.39 ą 2% +0.1 0.45
> perf-profile.children.cycles-pp.syscall_return_via_sysret
> 0.46 +0.1 0.53 ą 2%
> perf-profile.children.cycles-pp.perf_event_mmap
> 0.97 ą 14% +0.1 1.05 ą 11%
> perf-profile.children.cycles-pp.prepare_exit_to_usermode
> 6.75 +0.1 6.85
> perf-profile.children.cycles-pp.do_page_fault
> 6.76 +0.1 6.86
> perf-profile.children.cycles-pp.page_fault
> 6.77 +0.1 6.87
> perf-profile.children.cycles-pp.__do_page_fault
> 1.10 ą 2% +0.1 1.25 ą 3%
> perf-profile.children.cycles-pp.vmacache_find
> 0.63 ą 6% +0.2 0.78 ą 3%
> perf-profile.children.cycles-pp.___perf_sw_event
> 0.75 ą 4% +0.2 0.93 ą 3%
> perf-profile.children.cycles-pp.__perf_sw_event
> 1.35 +0.2 1.56
> perf-profile.children.cycles-pp.unmapped_area_topdown
> 1.42 +0.2 1.63
> perf-profile.children.cycles-pp.arch_get_unmapped_area_topdown
> 1.58 +0.2 1.81 ą 2%
> perf-profile.children.cycles-pp.find_vma
> 1.50 +0.2 1.73
> perf-profile.children.cycles-pp.get_unmapped_area
> 2.72 +0.3 3.06
> perf-profile.children.cycles-pp.native_irq_return_iret
> 2.15 +0.4 2.50 ą 3%
> perf-profile.children.cycles-pp.vma_interval_tree_insert
> 2.70 ą 5% +0.4 3.07 ą 3%
> perf-profile.children.cycles-pp.swapgs_restore_regs_and_return_to_usermode
> 3.83 +0.4 4.26
> perf-profile.children.cycles-pp.rwsem_spin_on_owner
> 68.47 -2.2 66.29
> perf-profile.self.cycles-pp.osq_lock
> 2.16 ą 7% -0.7 1.45
> perf-profile.self.cycles-pp.__handle_mm_fault
> 0.74 ą 3% -0.1 0.64 ą 2%
> perf-profile.self.cycles-pp.rwsem_down_write_failed
> 0.97 -0.1 0.89
> perf-profile.self.cycles-pp.up_write
> 0.07 ą 17% -0.1 0.00
> perf-profile.self.cycles-pp.intel_idle
> 0.04 ą 58% -0.0 0.00
> perf-profile.self.cycles-pp._raw_spin_lock_irqsave
> 0.36 ą 5% -0.0 0.32 ą 7%
> perf-profile.self.cycles-pp.osq_unlock
> 2.00 ą 14% -0.0 1.99 ą 14%
> perf-profile.self.cycles-pp.task_numa_work
> 0.01 ą173% -0.0 0.00
> perf-profile.self.cycles-pp.__vma_link_file
> 0.62 ą 5% -0.0 0.61 ą 2%
> perf-profile.self.cycles-pp.__rb_insert_augmented
> 0.06 ą 7% -0.0 0.06 ą 9%
> perf-profile.self.cycles-pp.native_iret
> 0.01 ą173% +0.0 0.01 ą173%
> perf-profile.self.cycles-pp.ksys_mmap_pgoff
> 0.01 ą173% +0.0 0.01 ą173%
> perf-profile.self.cycles-pp._cond_resched
> 0.06 ą 9% +0.0 0.06 ą 14%
> perf-profile.self.cycles-pp.arch_get_unmapped_area_topdown
> 0.34 ą 5% +0.0 0.35 ą 3%
> perf-profile.self.cycles-pp.down_write
> 0.08 ą 5% +0.0 0.09 ą 4%
> perf-profile.self.cycles-pp.perf_event_mmap
> 0.11 ą 7% +0.0 0.11 ą 3%
> perf-profile.self.cycles-pp.__vma_link_rb
> 0.08 ą 10% +0.0 0.09 ą 13%
> perf-profile.self.cycles-pp.mem_cgroup_from_task
> 0.07 +0.0 0.08 ą 5%
> perf-profile.self.cycles-pp.do_page_fault
> 0.06 ą 6% +0.0 0.07 ą 10%
> perf-profile.self.cycles-pp.pmd_devmap_trans_unstable
> 0.07 ą 7% +0.0 0.07 ą 5%
> perf-profile.self.cycles-pp.vma_gap_callbacks_rotate
> 0.12 ą 8% +0.0 0.13 ą 6% perf-profile.self.cycles-pp.d_path
> 0.06 +0.0 0.07 ą 10%
> perf-profile.self.cycles-pp.kmem_cache_alloc
> 0.00 +0.0 0.01 ą173%
> perf-profile.self.cycles-pp.kmem_cache_alloc_trace
> 0.00 +0.0 0.01 ą173%
> perf-profile.self.cycles-pp.prepend_path
> 0.06 ą 11% +0.0 0.07 ą 5% perf-profile.self.cycles-pp.fput
> 0.07 +0.0 0.08 ą 5%
> perf-profile.self.cycles-pp.perf_exclude_event
> 0.21 ą 4% +0.0 0.22
> perf-profile.self.cycles-pp.vma_interval_tree_augment_rotate
> 0.21 ą 3% +0.0 0.23 ą 6%
> perf-profile.self.cycles-pp.__entry_trampoline_start
> 0.04 ą 57% +0.0 0.06 ą 7% perf-profile.self.cycles-pp.kfree
> 0.04 ą 57% +0.0 0.06 ą 11%
> perf-profile.self.cycles-pp.avc_has_perm
> 0.16 ą 13% +0.0 0.19 ą 13%
> perf-profile.self.cycles-pp.vma_policy_mof
> 0.01 ą173% +0.0 0.04 ą 57%
> perf-profile.self.cycles-pp.perf_swevent_event
> 0.14 ą 3% +0.0 0.16 ą 2%
> perf-profile.self.cycles-pp.__might_sleep
> 0.32 ą 3% +0.0 0.34 ą 5% perf-profile.self.cycles-pp.__fget
> 0.18 ą 3% +0.0 0.21 ą 3%
> perf-profile.self.cycles-pp.up_read
> 0.15 ą 7% +0.0 0.17 ą 2%
> perf-profile.self.cycles-pp.__perf_sw_event
> 0.14 ą 10% +0.0 0.18 ą 4%
> perf-profile.self.cycles-pp.do_mmap
> 0.21 ą 3% +0.0 0.25 ą 4%
> perf-profile.self.cycles-pp.down_read_trylock
> 0.17 ą 6% +0.0 0.21 ą 7%
> perf-profile.self.cycles-pp.vma_compute_subtree_gap
> 0.21 ą 7% +0.0 0.25 ą 3%
> perf-profile.self.cycles-pp._raw_spin_lock
> 0.00 +0.0 0.04 ą 58%
> perf-profile.self.cycles-pp.perf_iterate_sb
> 0.16 ą 5% +0.0 0.20 ą 4%
> perf-profile.self.cycles-pp.___might_sleep
> 0.08 ą 23% +0.0 0.12 ą 14%
> perf-profile.self.cycles-pp.get_task_policy
> 0.37 ą 3% +0.0 0.42 ą 5%
> perf-profile.self.cycles-pp.sync_regs
> 0.00 +0.1 0.05
> perf-profile.self.cycles-pp.do_syscall_64
> 0.39 ą 2% +0.1 0.45
> perf-profile.self.cycles-pp.syscall_return_via_sysret
> 0.48 +0.1 0.55
> perf-profile.self.cycles-pp.find_vma
> 0.69 ą 2% +0.1 0.77 ą 2%
> perf-profile.self.cycles-pp.mmap_region
> 0.70 ą 3% +0.1 0.81 ą 2%
> perf-profile.self.cycles-pp.handle_mm_fault
> 0.54 ą 7% +0.1 0.69 ą 3%
> perf-profile.self.cycles-pp.___perf_sw_event
> 1.10 ą 2% +0.1 1.25 ą 3%
> perf-profile.self.cycles-pp.vmacache_find
> 0.73 +0.1 0.88 ą 5%
> perf-profile.self.cycles-pp.__do_page_fault
> 1.35 +0.2 1.56
> perf-profile.self.cycles-pp.unmapped_area_topdown
> 1.75 +0.3 2.03
> perf-profile.self.cycles-pp.swapgs_restore_regs_and_return_to_usermode
> 2.72 +0.3 3.06
> perf-profile.self.cycles-pp.native_irq_return_iret
> 2.15 +0.3 2.49 ą 3%
> perf-profile.self.cycles-pp.vma_interval_tree_insert
> 3.82 +0.4 4.25
> perf-profile.self.cycles-pp.rwsem_spin_on_owner
>
>
>
> vm-scalability.throughput
>
> 2.25e+07 +-+--------------------------------------------------------------+
> | O O |
> 2.2e+07 O-+O O OO O O OO O O O O O |
> 2.15e+07 +-+ O O O |
> | O O |
> 2.1e+07 +-+ |
> | |
> 2.05e+07 +-+ |
> | |
> 2e+07 +-+ +. +. .++.+ ++ +. + + +.+ |
> 1.95e+07 +-+ .+ : + + + + + .+ +.++.+ + .+ + : +.+ .|
> | ++ +: ++ + + +.+ + +.+ + + |
> 1.9e+07 +-+ + + + :+ : |
> | + + |
> 1.85e+07 +-+--------------------------------------------------------------+
>
>
> vm-scalability.median
>
> 800000 +-O----------------------------------------------------------------+
> O O O O O O O O O |
> 780000 +-+O O O O O O O |
> | O O O O |
> 760000 +-+ |
> | |
> 740000 +-+ |
> | |
> 720000 +-+ .+ .+ |
> | + +.+.++.+ +. + + .+. .+ .+.+ +.+ ++. .+ |
> 700000 +-+ + : : +. : ++ + + :+ +. + + : .+ + +.|
> | ++ :: + + + :.+ :+ + : |
> 680000 +-+ + + + + |
> | |
> 660000 +-+----------------------------------------------------------------+
>
>
> vm-scalability.time.voluntary_context_switches
>
> 8e+06 +-+-----------------------------------------------------------------+
> | +. +. |
> 7e+06 +-+ : + + +. + +. + + |
> | : :+ + : +. + + : :: .+ |
> 6e+06 +-++.+.+. + + + + : : : .+.++.+ +.+.++. .++.+. |
> |+ : + + +.+ +.+ + .|
> 5e+06 +-+ + |
> | |
> 4e+06 +-+ |
> | |
> 3e+06 +-O O O O OO O |
> O O O O O O |
> 2e+06 +-+ O O OO O OO O |
> | |
> 1e+06 +-+-----------------------------------------------------------------+
>
>
> [*] bisect-good sample
> [O] bisect-bad sample
>
>
>
> Disclaimer:
> Results have been estimated based on internal Intel analysis and are provided
> for informational purposes only. Any difference in system hardware or software
> design or configuration may affect actual performance.
>
>
> Thanks,
> Xiaolong
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2018-07-05 6:01 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
[not found] <CAOQ4uxgRXzq+ZSmwRB_nyGijFJsiuWXaM_F74W-3JOyTvAEz-Q@mail.gmail.com>
2018-07-02 18:59 ` [lkp-robot] [fs] 5c6de586e8: vm-scalability.throughput +12.4% improvement (from reorganizing struct inode?) Linus Torvalds
2018-07-04 23:34 ` Al Viro
2018-07-05 1:37 ` Linus Torvalds
2018-07-05 5:50 ` Amir Goldstein
2018-07-05 6:01 ` Amir Goldstein
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.