Message-ID: <202306071452.887afdea-oliver.sang@intel.com>
Date:   Wed, 7 Jun 2023 14:32:17 +0800
From:   kernel test robot <oliver.sang@...el.com>
To:     Linus Torvalds <torvalds@...ux-foundation.org>
CC:     <oe-lkp@...ts.linux.dev>, <lkp@...el.com>,
        <linux-kernel@...r.kernel.org>, <ying.huang@...el.com>,
        <feng.tang@...el.com>, <fengwei.yin@...el.com>,
        <oliver.sang@...el.com>
Subject: [linus:master] [x86]  20f3337d35:  stress-ng.lockofd.ops_per_sec
 3.4% improvement



Hello,

kernel test robot noticed a 3.4% improvement of stress-ng.lockofd.ops_per_sec on:


commit: 20f3337d350c4e1b4ac66d731fd4e98565bf6cc0 ("x86: don't use REP_GOOD or ERMS for small memory clearing")
https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git master

testcase: stress-ng
test machine: 64 threads 2 sockets Intel(R) Xeon(R) Gold 6346 CPU @ 3.10GHz (Ice Lake) with 256G memory
parameters:

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






Details are as below:
-------------------------------------------------------------------------------------------------->


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.

=========================================================================================
class/compiler/cpufreq_governor/disk/fs/kconfig/nr_threads/rootfs/tbox_group/test/testcase/testtime:
  os/gcc-12/performance/1HDD/ext4/x86_64-rhel-8.3/10%/debian-11.1-x86_64-20220510.cgz/lkp-icl-2sp7/lockofd/stress-ng/60s

commit: 
  68674f94ff ("x86: don't use REP_GOOD or ERMS for small memory copies")
  20f3337d35 ("x86: don't use REP_GOOD or ERMS for small memory clearing")

68674f94ffc9dddc 20f3337d350c4e1b4ac66d731fd 
---------------- --------------------------- 
         %stddev     %change         %stddev
             \          |                \  
     39.25            +2.2%      40.10        iostat.cpu.user
      0.05            -0.0        0.03 ± 16%  mpstat.cpu.all.soft%
 1.212e+08            +3.4%  1.253e+08        stress-ng.lockofd.ops
   2019682            +3.4%    2087883        stress-ng.lockofd.ops_per_sec
 9.311e+08            +6.8%  9.944e+08        perf-stat.i.branch-instructions
      0.67            -7.1%       0.63        perf-stat.i.cpi
 1.172e+09            +2.5%  1.202e+09        perf-stat.i.dTLB-loads
      0.00 ±  3%      -0.0        0.00 ±  2%  perf-stat.i.dTLB-store-miss-rate%
 8.309e+08           +15.2%  9.575e+08        perf-stat.i.dTLB-stores
 4.579e+09            +7.8%  4.934e+09        perf-stat.i.instructions
      1.49            +7.8%       1.60        perf-stat.i.ipc
      1.16           +62.7%       1.89 ±  2%  perf-stat.i.metric.G/sec
      1771           -28.8%       1261 ±  4%  perf-stat.i.metric.M/sec
     19.22 ±  2%      -1.8       17.39        perf-profile.calltrace.cycles-pp.fcntl_setlk.do_fcntl.__x64_sys_fcntl.do_syscall_64.entry_SYSCALL_64_after_hwframe
     53.24            -1.8       51.44        perf-profile.calltrace.cycles-pp.do_fcntl.__x64_sys_fcntl.do_syscall_64.entry_SYSCALL_64_after_hwframe
     60.31            -1.5       58.83        perf-profile.calltrace.cycles-pp.__x64_sys_fcntl.do_syscall_64.entry_SYSCALL_64_after_hwframe
      6.48 ±  4%      -0.6        5.87 ±  4%  perf-profile.calltrace.cycles-pp.kmem_cache_alloc.fcntl_setlk.do_fcntl.__x64_sys_fcntl.do_syscall_64
      1.03 ±  7%      -0.2        0.88 ± 10%  perf-profile.calltrace.cycles-pp.flock64_to_posix_lock.fcntl_getlk.do_fcntl.__x64_sys_fcntl.do_syscall_64
      5.76 ±  3%      +0.4        6.12 ±  2%  perf-profile.calltrace.cycles-pp.stress_mwc64modn
      0.00            +1.5        1.51 ±  4%  perf-profile.calltrace.cycles-pp.memset_orig.kmem_cache_alloc.fcntl_setlk.do_fcntl.__x64_sys_fcntl
      0.00            +1.6        1.60 ±  7%  perf-profile.calltrace.cycles-pp.memset_orig.kmem_cache_alloc.fcntl_getlk.do_fcntl.__x64_sys_fcntl
     53.94            -1.7       52.24        perf-profile.children.cycles-pp.do_fcntl
     19.65            -1.6       18.07        perf-profile.children.cycles-pp.fcntl_setlk
     13.82            -1.0       12.83 ±  3%  perf-profile.children.cycles-pp.kmem_cache_alloc
      5.96 ±  3%      +0.4        6.37 ±  3%  perf-profile.children.cycles-pp.stress_mwc64modn
      5.88 ±  2%      +0.5        6.39 ±  3%  perf-profile.children.cycles-pp.entry_SYSRETQ_unsafe_stack
      0.00            +3.3        3.29 ±  5%  perf-profile.children.cycles-pp.memset_orig
      0.84 ±  9%      +0.2        0.99 ±  9%  perf-profile.self.cycles-pp.exit_to_user_mode_prepare
      5.60 ±  2%      +0.4        6.04 ±  3%  perf-profile.self.cycles-pp.stress_mwc64modn
      5.58 ±  2%      +0.5        6.09 ±  2%  perf-profile.self.cycles-pp.entry_SYSRETQ_unsafe_stack
      0.00            +3.1        3.11 ±  5%  perf-profile.self.cycles-pp.memset_orig




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/wiki



View attachment "config-6.3.0-rc7-00002-g20f3337d350c" of type "text/plain" (158337 bytes)

View attachment "job-script" of type "text/plain" (9238 bytes)

View attachment "job.yaml" of type "text/plain" (6216 bytes)

View attachment "reproduce" of type "text/plain" (535 bytes)

Powered by blists - more mailing lists