11:21 a.m.
FYI, we noticed the unixbench.score +2.7% improvement on:
git://bee.sh.intel.com/git/ydu19/tip flat_hierarchy_v1.2
commit 6e7cab53c9f783791705c900d12b5eccb4938440 ("sched/fair: Remove leftover period
when sched averages accrue")
=========================================================================================
compiler/kconfig/nr_task/rootfs/tbox_group/test/testcase:
gcc-4.9/x86_64-rhel/100%/debian-x86_64-2015-02-07.cgz/nhm-white2/shell8/unixbench
commit:
8bc776c99c82240cf91a1b2a42136bdf9004ae16
6e7cab53c9f783791705c900d12b5eccb4938440
8bc776c99c82240c 6e7cab53c9f783791705c900d1
---------------- --------------------------
fail:runs %reproduction fail:runs
| | |
:4 25% 1:4 last_state.is_incomplete_run
%stddev %change %stddev
\ | \
10400 ± 0% +2.7% 10677 ± 0% unixbench.score
1838780 ± 0% -57.6% 778800 ± 0% unixbench.time.involuntary_context_switches
1.245e+08 ± 0% +2.7% 1.279e+08 ± 0% unixbench.time.minor_page_faults
736.25 ± 0% +1.9% 750.00 ± 0% unixbench.time.percent_of_cpu_this_job_got
558.03 ± 0% +3.6% 577.91 ± 0% unixbench.time.user_time
5358 ± 1% +15.9% 6209 ± 0% meminfo.KernelStack
332.25 ± 1% +16.1% 385.67 ± 0% proc-vmstat.nr_kernel_stack
1660 ± 1% +13.7% 1887 ± 0% slabinfo.cred_jar.active_objs
137725 ± 0% -39.3% 83558 ± 0% softirqs.SCHED
1838780 ± 0% -57.6% 778800 ± 0% time.involuntary_context_switches
235.05 ± 0% -11.9% 207.06 ± 0% uptime.idle
37202 ± 0% -30.3% 25924 ± 0% vmstat.system.cs
10784 ± 0% -7.5% 9978 ± 0% vmstat.system.in
93.30 ± 0% +1.5% 94.72 ± 0% turbostat.%Busy
2729 ± 0% +1.5% 2771 ± 0% turbostat.Avg_MHz
1.48 ± 1% -89.0% 0.16 ± 2% turbostat.CPU%c1
5384304 ± 1% -91.9% 434216 ± 2% cpuidle.C1-NHM.time
89950 ± 1% -87.2% 11554 ± 7% cpuidle.C1-NHM.usage
9082096 ± 1% -94.4% 506892 ± 2% cpuidle.C1E-NHM.time
137950 ± 1% -94.7% 7341 ± 5% cpuidle.C1E-NHM.usage
7193562 ± 4% -82.5% 1257977 ± 13% cpuidle.C3-NHM.time
89945 ± 4% -94.9% 4622 ± 6% cpuidle.C3-NHM.usage
23531 ± 8% -55.1% 10566 ± 8% cpuidle.C6-NHM.usage
1709 ± 15% -92.7% 124.00 ± 15% cpuidle.POLL.time
83.00 ± 8% -88.8% 9.33 ± 5% cpuidle.POLL.usage
63.74 ± 19% -73.3% 17.03 ± 9% sched_debug.cfs_rq:/.exec_clock.stddev
73.09 ± 18% -97.8% 1.62 ± 0% sched_debug.cfs_rq:/.load.avg
202.38 ± 46% -99.0% 2.00 ± 0% sched_debug.cfs_rq:/.load.max
12.44 ± 32% -87.9% 1.50 ± 0% sched_debug.cfs_rq:/.load.min
63.17 ± 46% -99.7% 0.22 ± 0% sched_debug.cfs_rq:/.load.stddev
86.61 ± 5% +3.9e+07% 33743136 ± 0% sched_debug.cfs_rq:/.load_avg.avg
151.50 ± 14% +2.9e+07% 43843754 ± 2% sched_debug.cfs_rq:/.load_avg.max
58.88 ± 1% +4.2e+07% 24874922 ± 11% sched_debug.cfs_rq:/.load_avg.min
32.02 ± 23% +1.9e+07% 6050630 ± 21% sched_debug.cfs_rq:/.load_avg.stddev
1968797 ± 0% +2082.6% 42971060 ± 0% sched_debug.cfs_rq:/.min_vruntime.avg
1986695 ± 0% +2070.9% 43128549 ± 0% sched_debug.cfs_rq:/.min_vruntime.max
1949176 ± 0% +2096.8% 42820332 ± 0% sched_debug.cfs_rq:/.min_vruntime.min
10884 ± 17% +830.1% 101238 ± 40% sched_debug.cfs_rq:/.min_vruntime.stddev
0.08 ± 72% +22086.7% 17.33 ± 3% sched_debug.cfs_rq:/.nr_spread_over.avg
0.62 ± 72% +5420.0% 34.50 ± 4% sched_debug.cfs_rq:/.nr_spread_over.max
0.21 ± 72% +5297.0% 11.16 ± 8% sched_debug.cfs_rq:/.nr_spread_over.stddev
59.09 ± 2% +1307.9% 832.00 ± 0% sched_debug.cfs_rq:/.runnable_load_avg.avg
67.94 ± 2% +1407.3% 1024 ± 0% sched_debug.cfs_rq:/.runnable_load_avg.max
47.56 ± 14% +1514.7% 768.00 ± 0% sched_debug.cfs_rq:/.runnable_load_avg.min
6.27 ± 30% +1668.2% 110.85 ± 0%
sched_debug.cfs_rq:/.runnable_load_avg.stddev
7796 ± 68% -1935.4% -143099 ±-64% sched_debug.cfs_rq:/.spread0.avg
-11809 ±-60% +2386.8% -293674 ±-40% sched_debug.cfs_rq:/.spread0.min
10877 ± 17% +831.2% 101294 ± 40% sched_debug.cfs_rq:/.spread0.stddev
807.66 ± 1% +45873.1% 371304 ± 2% sched_debug.cfs_rq:/.util_avg.avg
867.56 ± 1% +1.9e+05% 1637161 ± 6% sched_debug.cfs_rq:/.util_avg.max
730.44 ± 0% +508.4% 4443 ± 64% sched_debug.cfs_rq:/.util_avg.min
41.66 ± 9% +1.4e+06% 573825 ± 6% sched_debug.cfs_rq:/.util_avg.stddev
469251 ± 7% +26.0% 591177 ± 5% sched_debug.cpu.avg_idle.avg
211409 ± 16% +105.7% 434763 ± 12% sched_debug.cpu.avg_idle.min
192961 ± 12% -39.9% 116038 ± 14% sched_debug.cpu.avg_idle.stddev
2.00 ± 28% -42.3% 1.15 ± 30% sched_debug.cpu.clock.stddev
1.99 ± 28% -42.2% 1.15 ± 30% sched_debug.cpu.clock_task.stddev
58.80 ± 2% +1314.9% 832.00 ± 3% sched_debug.cpu.cpu_load[0].avg
68.00 ± 2% +1405.9% 1024 ± 0% sched_debug.cpu.cpu_load[0].max
48.38 ± 15% +1311.2% 682.67 ± 17% sched_debug.cpu.cpu_load[0].min
6.18 ± 39% +2221.5% 143.43 ± 25% sched_debug.cpu.cpu_load[0].stddev
58.53 ± 2% +1309.2% 824.80 ± 1% sched_debug.cpu.cpu_load[1].avg
67.25 ± 2% +1369.1% 988.00 ± 3% sched_debug.cpu.cpu_load[1].max
49.50 ± 6% +1347.8% 716.67 ± 10% sched_debug.cpu.cpu_load[1].min
5.49 ± 21% +1648.7% 96.08 ± 40% sched_debug.cpu.cpu_load[1].stddev
58.20 ± 1% +1315.3% 823.77 ± 1% sched_debug.cpu.cpu_load[2].avg
66.50 ± 2% +1295.0% 927.67 ± 2% sched_debug.cpu.cpu_load[2].max
49.81 ± 6% +1383.9% 739.17 ± 7% sched_debug.cpu.cpu_load[2].min
5.19 ± 25% +1070.4% 60.75 ± 47% sched_debug.cpu.cpu_load[2].stddev
57.96 ± 1% +1330.1% 828.93 ± 0% sched_debug.cpu.cpu_load[3].avg
65.50 ± 3% +1270.6% 897.75 ± 1% sched_debug.cpu.cpu_load[3].max
50.31 ± 5% +1411.9% 760.67 ± 5% sched_debug.cpu.cpu_load[3].min
4.60 ± 27% +803.4% 41.52 ± 40% sched_debug.cpu.cpu_load[3].stddev
58.10 ± 1% +1340.0% 836.69 ± 0% sched_debug.cpu.cpu_load[4].avg
64.56 ± 4% +1272.7% 886.25 ± 2% sched_debug.cpu.cpu_load[4].max
51.50 ± 3% +1422.8% 784.25 ± 3% sched_debug.cpu.cpu_load[4].min
4.13 ± 31% +688.6% 32.53 ± 33% sched_debug.cpu.cpu_load[4].stddev
64.74 ± 4% -81.0% 12.29 ±122% sched_debug.cpu.load.avg
12.38 ± 34% -87.9% 1.50 ± 0% sched_debug.cpu.load.min
0.00 ± 13% -18.1% 0.00 ± 8% sched_debug.cpu.next_balance.stddev
3.36 ± 6% -14.2% 2.88 ± 14% sched_debug.cpu.nr_running.stddev
424371 ± 0% -29.9% 297353 ± 0% sched_debug.cpu.nr_switches.avg
443398 ± 0% -28.8% 315529 ± 0% sched_debug.cpu.nr_switches.max
410840 ± 0% -31.2% 282757 ± 2% sched_debug.cpu.nr_switches.min
90.50 ± 7% +93.3% 174.92 ± 22% sched_debug.cpu.nr_uninterruptible.max
-81.81 ±-31% +94.0% -158.75 ±-24% sched_debug.cpu.nr_uninterruptible.min
52.60 ± 14% +94.7% 102.40 ± 22% sched_debug.cpu.nr_uninterruptible.stddev
420259 ± 0% -30.2% 293195 ± 0% sched_debug.cpu.sched_count.avg
437167 ± 0% -29.6% 307604 ± 0% sched_debug.cpu.sched_count.max
407813 ± 0% -31.3% 280108 ± 2% sched_debug.cpu.sched_count.min
19966 ± 1% -91.5% 1703 ± 6% sched_debug.cpu.sched_goidle.avg
21891 ± 3% -83.4% 3642 ± 11% sched_debug.cpu.sched_goidle.max
18268 ± 2% -94.3% 1042 ± 15% sched_debug.cpu.sched_goidle.min
1151 ± 11% -31.0% 794.30 ± 13% sched_debug.cpu.sched_goidle.stddev
209338 ± 0% -22.5% 162153 ± 0% sched_debug.cpu.ttwu_count.avg
222545 ± 0% -21.0% 175766 ± 1% sched_debug.cpu.ttwu_count.max
198999 ± 0% -24.2% 150879 ± 1% sched_debug.cpu.ttwu_count.min
157836 ± 0% -20.1% 126120 ± 0% sched_debug.cpu.ttwu_local.avg
163936 ± 0% -20.9% 129672 ± 0% sched_debug.cpu.ttwu_local.max
152229 ± 0% -19.6% 122365 ± 0% sched_debug.cpu.ttwu_local.min
4082 ± 15% -42.9% 2332 ± 22% sched_debug.cpu.ttwu_local.stddev
nhm-white2: 8 threads Nehalem with 4G memory
unixbench.time.user_time
600 ++-----------------------------------O--------------------------------+
OO.OO.OO.O*.OO O.OO.OO.OO.OO.OO.O*.O*.O*.**.**.**.**.**.**.**.**.**.**
500 ++ : : |
| : : |
| : : |
400 ++ : : |
| : : |
300 ++ : : |
| :: |
200 ++ :: |
| :: |
| : |
100 ++ : |
| : |
0 ++--------O----O------------------O-----------------------------------+
unixbench.time.minor_page_faults
1.4e+08 ++----------------------------------------------------------------+
OO OO OOO O O O OO OOO OO OOO O OO O |
1.2e+08 **.**.***.**.* *.**.***.**.***.**.**.***.**.***.**.***.**.***.**.**
| : : |
1e+08 ++ : : |
| : : |
8e+07 ++ : : |
| : : |
6e+07 ++ :: |
| : |
4e+07 ++ : |
| : |
2e+07 ++ : |
| : |
0 ++--------O---O-----------------O---------------------------------+
unixbench.time.involuntary_context_switches
2e+06 ++----------------------------------------------------------------+
1.8e+06 **.**.***.**.* *.**.***.**.***.**.**.***.**.***.**.***.**.***.**.**
| : : |
1.6e+06 ++ : : |
1.4e+06 ++ : : |
| : : |
1.2e+06 ++ : : |
1e+06 ++ : : |
800000 ++ : OO O O |
OO OO OOO O O:O OO OOO OO OO O |
600000 ++ : |
400000 ++ : |
| : |
200000 ++ : |
0 ++--------O---O-----------------O---------------------------------+
[*] bisect-good sample
[O] bisect-bad sample
To reproduce:
git clone git://git.kernel.org/pub/scm/linux/kernel/git/wfg/lkp-tests.git
cd lkp-tests
bin/lkp install job.yaml # job file is attached in this email
bin/lkp run job.yaml
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 Ye
1775
days inactive
1775
days old
0 comments
1 participants
participants (1)
-
kernel test robot