Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > linux.kernel > #1252516
| From | kernel test robot <ying.huang@linux.intel.com> |
|---|---|
| Newsgroups | linux.kernel |
| Subject | [lkp] [sched/numa] b52da86e0a: -1.4% will-it-scale.per_thread_ops |
| Date | 2015-10-21 09:40 +0200 |
| Message-ID | <qlWa6-4E2-23@gated-at.bofh.it> (permalink) |
| Organization | linux.* mail to news gateway |
[Multipart message — attachments visible in raw view] - view raw
FYI, we noticed the below changes on
https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git master
commit b52da86e0ad58f096710977fcda856fd84da9233 ("sched/numa: Fix task_tick_fair() from disabling numa_balancing")
=========================================================================================
tbox_group/testcase/rootfs/kconfig/compiler/cpufreq_governor/test:
nhm4/will-it-scale/debian-x86_64-2015-02-07.cgz/x86_64-rhel/gcc-4.9/performance/readseek1
commit:
e2bf1c4b17aff25f07e0d2952d8c1c66643f33fe
b52da86e0ad58f096710977fcda856fd84da9233
e2bf1c4b17aff25f b52da86e0ad58f096710977fcd
---------------- --------------------------
%stddev %change %stddev
\ | \
1868374 ± 0% -1.4% 1843108 ± 0% will-it-scale.per_thread_ops
0.57 ± 0% +14.1% 0.65 ± 10% will-it-scale.scalability
14052 ± 0% -29.6% 9896 ± 0% will-it-scale.time.minor_page_faults
161.75 ± 45% -53.2% 75.75 ± 23% cpuidle.C1E-NHM.usage
14052 ± 0% -29.6% 9896 ± 0% time.minor_page_faults
6943 ± 0% -100.0% 0.00 ± -1% proc-vmstat.numa_hint_faults
6943 ± 0% -100.0% 0.00 ± -1% proc-vmstat.numa_hint_faults_local
7780 ± 0% -100.0% 0.00 ± -1% proc-vmstat.numa_pte_updates
1.19 ± 4% -14.1% 1.02 ± 4% perf-profile.cpu-cycles.__fget_light.sys_lseek.entry_SYSCALL_64_fastpath
5.47 ± 1% -11.8% 4.83 ± 2% perf-profile.cpu-cycles.entry_SYSCALL_64
1.17 ± 6% -16.7% 0.98 ± 12% perf-profile.cpu-cycles.fsnotify.vfs_read.sys_read.entry_SYSCALL_64_fastpath
1.70 ± 4% -15.0% 1.45 ± 4% perf-profile.cpu-cycles.shmem_file_llseek.sys_lseek.entry_SYSCALL_64_fastpath
5.39 ± 1% -12.8% 4.70 ± 5% perf-profile.cpu-cycles.sys_lseek.entry_SYSCALL_64_fastpath
116.50 ± 14% -35.4% 75.25 ± 13% sched_debug.cfs_rq[2]:/.load
95126 ± 8% +17.5% 111795 ± 10% sched_debug.cpu#0.nr_load_updates
2464 ± 6% -39.4% 1494 ± 29% sched_debug.cpu#2.curr->pid
116.50 ± 14% -35.4% 75.25 ± 13% sched_debug.cpu#2.load
1243 ± 2% +50.4% 1870 ± 22% sched_debug.cpu#3.curr->pid
17134 ± 29% +7834.3% 1359530 ±110% sched_debug.cpu#4.nr_switches
17204 ± 29% +7802.5% 1359602 ±110% sched_debug.cpu#4.sched_count
4538 ± 78% +12299.8% 562734 ±119% sched_debug.cpu#4.sched_goidle
950401 ± 4% -7.9% 875553 ± 4% sched_debug.cpu#7.avg_idle
nhm4: Nehalem
Memory: 4G
proc-vmstat.numa_pte_updates
8000 ++---------------------------------------------------*---*-*-----*-*-+
| * * *.* *
7000 ++ : |
| : |
| : |
6000 ++ : |
| : |
5000 ++ : |
| : |
4000 ++ : |
| : |
| : |
3000 ++ : |
*.*.*.*.*.*.*.*.*.*.*.*.*.*.*.*.*..*.*.*.*.*.*.*.* |
2000 ++-------------------------------------------------------------------+
proc-vmstat.numa_hint_faults
7000 ++-------------------------------------------------*-*-*-*-*-*-*-*-*-*
| : |
6000 ++ : |
| : |
| : |
5000 ++ : |
| : |
4000 ++ : |
| : |
3000 ++ : |
| : |
| : |
2000 ++ : |
*.*.*.*.*.*.*.*.*.*.*.*.*.*.*.*.*..*.*.*.*.*.*.*.* |
1000 ++-------------------------------------------------------------------+
proc-vmstat.numa_hint_faults_local
7000 ++-------------------------------------------------*-*-*-*-*-*-*-*-*-*
| : |
6000 ++ : |
| : |
| : |
5000 ++ : |
| : |
4000 ++ : |
| : |
3000 ++ : |
| : |
| : |
2000 ++ : |
*.*.*.*.*.*.*.*.*.*.*.*.*.*.*.*.*..*.*.*.*.*.*.*.* |
1000 ++-------------------------------------------------------------------+
[*] 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,
Ying Huang
Back to linux.kernel | Previous | Next | Find similar | Unroll thread
[lkp] [sched/numa] b52da86e0a: -1.4% will-it-scale.per_thread_ops kernel test robot <ying.huang@linux.intel.com> - 2015-10-21 09:40 +0200
csiph-web