Date:   Fri, 30 Dec 2022 15:43:19 +0800
From:   kernel test robot <yujie.liu@...el.com>
To:     Marco Elver <elver@...gle.com>
CC:     <oe-lkp@...ts.linux.dev>, <lkp@...el.com>,
        Peter Zijlstra <peterz@...radead.org>,
        Dmitry Vyukov <dvyukov@...gle.com>,
        Ian Rogers <irogers@...gle.com>,
        <linux-kernel@...r.kernel.org>, <linux-perf-users@...r.kernel.org>,
        <ying.huang@...el.com>, <feng.tang@...el.com>,
        <zhengjun.xing@...ux.intel.com>, <fengwei.yin@...el.com>
Subject: [linus:master] [perf/hw_breakpoint] 0370dc314d:
 stress-ng.kill.ops_per_sec 14.1% improvement

Greeting,

FYI, we noticed a 14.1% improvement of stress-ng.kill.ops_per_sec due to commit:

commit: 0370dc314df35579b751d1b77c9169f071444962 ("perf/hw_breakpoint: Optimize list of per-task breakpoints")
https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git master

in testcase: stress-ng
on test machine: 96 threads 2 sockets Intel(R) Xeon(R) Gold 6252 CPU @ 2.10GHz (Cascade Lake) with 512G memory
with following parameters:

	nr_threads: 10%
	disk: 1HDD
	testtime: 60s
	fs: ext4
	class: os
	test: kill
	cpufreq_governor: performance


Details are as below:

=========================================================================================
class/compiler/cpufreq_governor/disk/fs/kconfig/nr_threads/rootfs/tbox_group/test/testcase/testtime:
  os/gcc-11/performance/1HDD/ext4/x86_64-rhel-8.3/10%/debian-11.1-x86_64-20220510.cgz/lkp-csl-2sp7/kill/stress-ng/60s

commit: 
  089cdcb0cd ("perf/hw_breakpoint: Clean up headers")
  0370dc314d ("perf/hw_breakpoint: Optimize list of per-task breakpoints")

089cdcb0cd1c2534 0370dc314df35579b751d1b77c9 
---------------- --------------------------- 
         %stddev     %change         %stddev
             \          |                \  
      2025           +15.3%       2334        stress-ng.kill.kill_calls_per_sec
    369933           +14.1%     422083        stress-ng.kill.ops
      6165           +14.1%       7034        stress-ng.kill.ops_per_sec
      2235 ±  6%     +32.7%       2966 ±  3%  stress-ng.time.involuntary_context_switches
    722685 ±  2%      +7.8%     779391 ±  2%  stress-ng.time.voluntary_context_switches
    846890 ± 12%     +19.8%    1014212 ±  6%  meminfo.DirectMap4k
     24473 ±  2%      +7.1%      26207 ±  2%  vmstat.system.cs
      3.44 ±  4%      -1.9        1.59 ±  6%  perf-profile.calltrace.cycles-pp.aa_may_signal.apparmor_task_kill.security_task_kill.kill_something_info.__x64_sys_kill
      0.20 ±122%      +0.4        0.55 ±  5%  perf-profile.calltrace.cycles-pp.check_kill_permission.kill_something_info.__x64_sys_kill.do_syscall_64.entry_SYSCALL_64_after_hwframe
      3.48 ±  4%      -1.9        1.62 ±  6%  perf-profile.children.cycles-pp.aa_may_signal
      0.08 ±  9%      +0.0        0.10 ± 10%  perf-profile.children.cycles-pp.audit_signal_info
      0.13 ± 11%      +0.0        0.15 ±  4%  perf-profile.children.cycles-pp.apparmor_capable
      0.09 ± 37%      +0.1        0.14 ± 22%  perf-profile.children.cycles-pp.tick_sched_do_timer
      0.51 ±  9%      +0.1        0.60 ±  5%  perf-profile.children.cycles-pp.check_kill_permission
      3.46 ±  5%      -1.9        1.60 ±  6%  perf-profile.self.cycles-pp.aa_may_signal
      0.13 ± 10%      +0.0        0.15 ±  4%  perf-profile.self.cycles-pp.apparmor_capable
      0.16 ± 12%      +0.0        0.19 ±  9%  perf-profile.self.cycles-pp.check_kill_permission
      0.12 ± 11%      +0.0        0.14 ±  8%  perf-profile.self.cycles-pp.kill_something_info
     31.01           -33.4%      20.65 ±  2%  perf-stat.i.MPKI
 6.476e+08            +9.3%  7.079e+08        perf-stat.i.branch-instructions
     25.80 ±  3%      -8.3       17.52        perf-stat.i.cache-miss-rate%
  24950550 ±  3%     -50.4%   12381137 ±  2%  perf-stat.i.cache-misses
  95492313           -26.5%   70215337        perf-stat.i.cache-references
     25172 ±  2%      +7.4%      27042        perf-stat.i.context-switches
      8.90            -9.1%       8.09        perf-stat.i.cpi
      1137 ±  2%     +98.4%       2257 ±  3%  perf-stat.i.cycles-between-cache-misses
 9.911e+08            +9.9%  1.089e+09        perf-stat.i.dTLB-loads
 5.059e+08           +10.3%  5.578e+08        perf-stat.i.dTLB-stores
     35.29            +1.1       36.42        perf-stat.i.iTLB-load-miss-rate%
   1172159            +6.0%    1242747        perf-stat.i.iTLB-load-misses
 3.273e+09            +9.4%  3.582e+09        perf-stat.i.instructions
      2839            +3.3%       2933        perf-stat.i.instructions-per-iTLB-miss
      0.14            +8.2%       0.15        perf-stat.i.ipc
    298.28 ± 71%    +176.8%     825.53        perf-stat.i.metric.K/sec
     23.13 ±  2%      +6.0%      24.52        perf-stat.i.metric.M/sec
   3366282 ±  3%     -15.4%    2848063        perf-stat.i.node-load-misses
    429795 ±  7%     -20.9%     340164 ±  6%  perf-stat.i.node-loads
   3305245           +11.6%    3687084        perf-stat.i.node-store-misses
     29.20           -32.8%      19.62        perf-stat.overall.MPKI
      1.11 ±  2%      -0.1        1.04 ±  2%  perf-stat.overall.branch-miss-rate%
     26.13 ±  2%      -8.5       17.64        perf-stat.overall.cache-miss-rate%
      8.38            -8.2%       7.69        perf-stat.overall.cpi
      1099 ±  3%    +102.2%       2223 ±  2%  perf-stat.overall.cycles-between-cache-misses
     35.32            +1.1       36.44        perf-stat.overall.iTLB-load-miss-rate%
      2791 ±  2%      +3.2%       2880        perf-stat.overall.instructions-per-iTLB-miss
      0.12            +9.0%       0.13        perf-stat.overall.ipc
 6.369e+08            +9.3%  6.962e+08        perf-stat.ps.branch-instructions
  24566232 ±  3%     -50.4%   12190440 ±  2%  perf-stat.ps.cache-misses
  94005574           -26.5%   69121490        perf-stat.ps.cache-references
     24783 ±  2%      +7.4%      26622        perf-stat.ps.context-switches
 9.751e+08            +9.9%  1.071e+09        perf-stat.ps.dTLB-loads
 4.979e+08           +10.2%  5.489e+08        perf-stat.ps.dTLB-stores
   1153650            +6.0%    1222951        perf-stat.ps.iTLB-load-misses
 3.219e+09            +9.4%  3.523e+09        perf-stat.ps.instructions
   3314459 ±  3%     -15.4%    2804239        perf-stat.ps.node-load-misses
    423139 ±  7%     -20.8%     334925 ±  6%  perf-stat.ps.node-loads
   3254493           +11.5%    3630306        perf-stat.ps.node-store-misses
  2.05e+11            +9.1%  2.236e+11        perf-stat.total.instructions



To reproduce:

        git clone https://github.com/intel/lkp-tests.git
        cd lkp-tests
        sudo bin/lkp install job.yaml           # job file is attached in this email
        bin/lkp split-job --compatible job.yaml # generate the yaml file for lkp run
        sudo bin/lkp run generated-yaml-file

        # if come across any failure that blocks the test,
        # please remove ~/.lkp and /lkp dir to run from a clean state.


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.


-- 
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests

View attachment "config-6.0.0-rc2-00021-g0370dc314df3" of type "text/plain" (164374 bytes)

View attachment "job-script" of type "text/plain" (8442 bytes)

View attachment "job.yaml" of type "text/plain" (5711 bytes)

View attachment "reproduce" of type "text/plain" (532 bytes)

Powered by blists - more mailing lists