Date:   Mon, 25 Dec 2017 14:15:46 +0800
From:   kernel test robot <xiaolong.ye@...el.com>
To:     Jeff Layton <jlayton@...hat.com>
Cc:     LKML <linux-kernel@...r.kernel.org>,
        Jeff Layton <jlayton@...hat.com>, lkp@...org
Subject: [lkp-robot] [fs]  7717d6821a:  aim7.jobs-per-min 2.7% improvement


Greeting,

FYI, we noticed a 2.7% improvement of aim7.jobs-per-min due to commit:


commit: 7717d6821a068194fc70ed98e921dad49acae257 ("fs: handle inode->i_version more efficiently")
https://git.kernel.org/cgit/linux/kernel/git/jlayton/linux.git iversion

in testcase: aim7
on test machine: 40 threads Intel(R) Xeon(R) CPU E5-2690 v2 @ 3.00GHz with 384G memory
with following parameters:

	disk: 1BRD_48G
	fs: xfs
	test: sync_disk_rw
	load: 600
	cpufreq_governor: performance

test-description: AIM7 is a traditional UNIX system level benchmark suite which is used to test and measure the performance of multiuser system.
test-url: https://sourceforge.net/projects/aimbench/files/aim-suite7/


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/disk/fs/kconfig/load/rootfs/tbox_group/test/testcase:
  gcc-7/performance/1BRD_48G/xfs/x86_64-rhel-7.2/600/debian-x86_64-2016-08-31.cgz/lkp-ivb-ep01/sync_disk_rw/aim7

commit: 
  95751ce80f ("btrfs: only dirty the inode in btrfs_update_time if something was changed")
  7717d6821a ("fs: handle inode->i_version more efficiently")

95751ce80f794e00 7717d6821a068194fc70ed98e9 
---------------- -------------------------- 
         %stddev     %change         %stddev
             \          |                \  
     12123            +2.7%      12451        aim7.jobs-per-min
    297.12            -2.6%     289.30        aim7.time.elapsed_time
    297.12            -2.6%     289.30        aim7.time.elapsed_time.max
   2346204           -17.0%    1948233        aim7.time.involuntary_context_switches
      7364            -2.5%       7180        aim7.time.system_time
  82044965            -1.7%   80644825        aim7.time.voluntary_context_switches
    242040            -4.7%     230740 ±  2%  interrupts.CAL:Function_call_interrupts
    448043            +2.7%     460207        vmstat.io.bo
      1094 ±  5%     +76.5%       1931 ± 66%  sched_debug.cfs_rq:/.load_avg.stddev
      1.93 ±  7%     -14.7%       1.65 ±  6%  sched_debug.cfs_rq:/.nr_spread_over.avg
      5.70 ± 20%     -28.0%       4.10 ± 16%  sched_debug.cfs_rq:/.nr_spread_over.max
      1.12 ± 11%     -23.7%       0.86 ± 12%  sched_debug.cfs_rq:/.nr_spread_over.stddev
    315.45 ±  8%     +22.2%     385.57 ±  9%  sched_debug.cfs_rq:/.util_avg.min
 7.101e+11            -2.8%  6.898e+11        perf-stat.branch-instructions
 5.528e+09            -4.5%   5.28e+09        perf-stat.branch-misses
 8.804e+09            -4.7%  8.389e+09        perf-stat.cache-misses
 4.922e+10            -4.1%  4.718e+10        perf-stat.cache-references
 1.226e+08            -2.5%  1.194e+08        perf-stat.context-switches
 1.004e+13            -2.4%  9.791e+12        perf-stat.cpu-cycles
  15409312            -3.5%   14869259        perf-stat.cpu-migrations
  8.12e+11            -3.6%   7.83e+11        perf-stat.dTLB-loads
     3e+11            -6.9%  2.792e+11        perf-stat.dTLB-stores
 2.873e+09            -2.8%  2.793e+09        perf-stat.iTLB-loads
 3.171e+12            -3.4%  3.063e+12        perf-stat.instructions
      0.32            -1.0%       0.31        perf-stat.ipc
 3.654e+09            -7.3%  3.387e+09        perf-stat.node-load-misses
 4.442e+09            -6.2%  4.165e+09        perf-stat.node-loads
 2.942e+09            -4.5%  2.809e+09        perf-stat.node-store-misses
 4.318e+09            -3.0%  4.188e+09        perf-stat.node-stores

                                                                                

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

View attachment "config-4.15.0-rc3-00022-g7717d68" of type "text/plain" (163807 bytes)

View attachment "job-script" of type "text/plain" (7544 bytes)

View attachment "job.yaml" of type "text/plain" (5201 bytes)

View attachment "reproduce" of type "text/plain" (719 bytes)

Powered by blists - more mailing lists