All of lore.kernel.org
 help / color / mirror / Atom feed
From: kernel test robot <oliver.sang@intel.com>
To: Peter Zijlstra <peterz@infradead.org>
Cc: <oe-lkp@lists.linux.dev>, <lkp@intel.com>,
	<linux-kernel@vger.kernel.org>, <x86@kernel.org>,
	Joseph Salisbury <joseph.salisbury@oracle.com>,
	Adam Li <adamli@os.amperecomputing.com>,
	Dietmar Eggemann <dietmar.eggemann@arm.com>,
	Hazem Mohamed Abuelfotoh <abuehaze@amazon.com>,
	Chris Mason <clm@meta.com>, <aubrey.li@linux.intel.com>,
	<yu.c.chen@intel.com>, <oliver.sang@intel.com>
Subject: [tip:sched/core] [sched/fair]  d206fbad93:  hackbench.throughput 6.5% regression
Date: Fri, 28 Nov 2025 14:56:32 +0800	[thread overview]
Message-ID: <202511281428.bfcbd5a8-lkp@intel.com> (raw)



Hello,


for the change reverted by this commit, we reported several regressions before
[1][2][3] majorly for unixbench tests.

[1] https://lore.kernel.org/all/202507150846.538fc133-lkp@intel.com/
[2] https://lore.kernel.org/all/202507281046.e71b853a-lkp@intel.com/
[3] https://lore.kernel.org/all/202508150416.d7153a48-lkp@intel.com/

but now, for d206fbad93, we also capture a regression in hackbench tests.

below report is just FYI what's the possible performance impact of this revert
commit.


kernel test robot noticed a 6.5% regression of hackbench.throughput on:


commit: d206fbad9328ddb68ebabd7cf7413392acd38081 ("sched/fair: Revert max_newidle_lb_cost bump")
https://git.kernel.org/cgit/linux/kernel/git/tip/tip.git sched/core

[still regression on linux-next/master ef68bf704646690aba5e81c2f7be8d6ef13d7ad8]

testcase: hackbench
config: x86_64-rhel-9.4
compiler: gcc-14
test machine: 64 threads 2 sockets Intel(R) Xeon(R) Gold 6346 CPU @ 3.10GHz (Ice Lake) with 256G memory
parameters:

	nr_threads: 50%
	iterations: 4
	mode: threads
	ipc: pipe
	cpufreq_governor: performance




If you fix the issue in a separate patch/commit (i.e. not just a new version of
the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <oliver.sang@intel.com>
| Closes: https://lore.kernel.org/oe-lkp/202511281428.bfcbd5a8-lkp@intel.com


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


The kernel config and materials to reproduce are available at:
https://download.01.org/0day-ci/archive/20251128/202511281428.bfcbd5a8-lkp@intel.com

=========================================================================================
compiler/cpufreq_governor/ipc/iterations/kconfig/mode/nr_threads/rootfs/tbox_group/testcase:
  gcc-14/performance/pipe/4/x86_64-rhel-9.4/threads/50%/debian-13-x86_64-20250902.cgz/lkp-icl-2sp7/hackbench

commit: 
  e837456fdc ("sched/fair: Reimplement NEXT_BUDDY to align with EEVDF goals")
  d206fbad93 ("sched/fair: Revert max_newidle_lb_cost bump")

e837456fdca81899 d206fbad9328ddb68ebabd7cf74 
---------------- --------------------------- 
         %stddev     %change         %stddev
             \          |                \  
 2.603e+08 ±  6%     -33.6%   1.73e+08 ±  4%  cpuidle..time
   5881942 ±  3%     -62.9%    2180816 ± 13%  cpuidle..usage
     97892 ±  5%     +17.9%     115437 ±  2%  perf-c2c.HITM.local
     98482 ±  5%     +17.6%     115850 ±  2%  perf-c2c.HITM.total
      3.99 ±  8%      -1.4        2.62 ±  3%  mpstat.cpu.all.idle%
      0.71 ±  2%      -0.1        0.64 ±  2%  mpstat.cpu.all.irq%
      0.05            -0.0        0.03 ±  2%  mpstat.cpu.all.soft%
   1305956 ±  8%     +19.7%    1562588 ± 11%  numa-meminfo.node1.Active
   1305948 ±  8%     +19.7%    1562580 ± 11%  numa-meminfo.node1.Active(anon)
    863383           +20.4%    1039636 ± 11%  numa-meminfo.node1.Shmem
    326485 ±  8%     +19.7%     390946 ± 10%  numa-vmstat.node1.nr_active_anon
    215818           +20.6%     260214 ± 11%  numa-vmstat.node1.nr_shmem
    326485 ±  8%     +19.7%     390945 ± 10%  numa-vmstat.node1.nr_zone_active_anon
   1631510           +10.4%    1801907 ±  6%  meminfo.Active
   1631494           +10.4%    1801890 ±  6%  meminfo.Active(anon)
   1772671            +9.5%    1940770 ±  5%  meminfo.Committed_AS
    909474           +19.7%    1088760 ± 10%  meminfo.Shmem
      5.36 ±107%      -3.2        2.11 ±223%  perf-profile.calltrace.cycles-pp.process_simple.reader__read_event.perf_session__process_events.record__finish_output.cmd_record
      6.59 ±100%      -2.2        4.41 ±141%  perf-profile.calltrace.cycles-pp.__ordered_events__flush.perf_session__process_user_event.reader__read_event.perf_session__process_events.record__finish_output
      6.59 ±100%      -2.2        4.41 ±141%  perf-profile.calltrace.cycles-pp.perf_session__process_user_event.reader__read_event.perf_session__process_events.record__finish_output.cmd_record
      5.36 ±107%      -3.2        2.11 ±223%  perf-profile.children.cycles-pp.process_simple
      0.34 ±  4%      -0.2        0.11 ± 14%  turbostat.C1%
      1.10            -0.8        0.28 ± 16%  turbostat.C1E%
      2.96 ±  8%      -0.6        2.38 ±  2%  turbostat.C6%
      1.25 ±  5%     -15.5%       1.05 ±  4%  turbostat.CPU%c1
    795.01 ±  4%     -10.1%     714.37 ±  2%  sched_debug.cfs_rq:/.util_est.avg
      4969 ±  9%    +122.0%      11030 ± 47%  sched_debug.cpu.avg_idle.min
      1716 ±141%    +111.4%       3628 ± 61%  sched_debug.cpu.curr->pid.min
   1063335 ±  3%     -47.4%     559053 ± 23%  sched_debug.cpu.max_idle_balance_cost.avg
    712039 ±  2%     -29.8%     500000        sched_debug.cpu.max_idle_balance_cost.min
    407716           +10.5%     450503 ±  6%  proc-vmstat.nr_active_anon
   1139851            +3.9%    1184811 ±  2%  proc-vmstat.nr_file_pages
     33054            +0.8%      33323        proc-vmstat.nr_kernel_stack
     88427 ±  2%      +6.0%      93718 ±  2%  proc-vmstat.nr_mapped
    227232           +19.8%     272194 ± 10%  proc-vmstat.nr_shmem
    407716           +10.5%     450503 ±  6%  proc-vmstat.nr_zone_active_anon
  44509495            -2.8%   43284126        proc-vmstat.numa_hit
  44442837            -2.8%   43217032        proc-vmstat.numa_local
  44571379            -2.7%   43348289        proc-vmstat.pgalloc_normal
  43993792            -2.9%   42727500        proc-vmstat.pgfree
    342580            -6.5%     320227        hackbench.throughput
    334468            -6.2%     313595        hackbench.throughput_avg
    342580            -6.5%     320227        hackbench.throughput_best
    319665            -6.4%     299259        hackbench.throughput_worst
     90.24            +6.6%      96.22        hackbench.time.elapsed_time
     90.24            +6.6%      96.22        hackbench.time.elapsed_time.max
  1.47e+08           +10.0%  1.617e+08        hackbench.time.involuntary_context_switches
     37736 ±  3%      +8.1%      40787        hackbench.time.minor_page_faults
      6097            +1.8%       6206        hackbench.time.percent_of_cpu_this_job_got
      5041            +9.1%       5501        hackbench.time.system_time
    461.18            +2.2%     471.26        hackbench.time.user_time
 4.238e+08            +4.5%  4.429e+08        hackbench.time.voluntary_context_switches
 2.525e+10            -4.4%  2.415e+10        perf-stat.i.branch-instructions
      5.10            -0.6        4.55 ±  4%  perf-stat.i.cache-miss-rate%
 3.459e+08           +12.9%  3.906e+08        perf-stat.i.cache-references
      1.83            +5.5%       1.93        perf-stat.i.cpi
  1.92e+11            +1.3%  1.946e+11        perf-stat.i.cpu-cycles
    423018           +19.6%     506120        perf-stat.i.cpu-migrations
 1.074e+11            -4.1%   1.03e+11        perf-stat.i.instructions
      0.56            -5.4%       0.53        perf-stat.i.ipc
      4.79 ±  2%      -0.7        4.14 ±  5%  perf-stat.overall.cache-miss-rate%
      1.79            +5.7%       1.89        perf-stat.overall.cpi
      0.56            -5.4%       0.53        perf-stat.overall.ipc
 2.499e+10            -4.3%  2.392e+10        perf-stat.ps.branch-instructions
  3.42e+08           +13.0%  3.865e+08        perf-stat.ps.cache-references
   1.9e+11            +1.4%  1.926e+11        perf-stat.ps.cpu-cycles
    417843           +19.8%     500456        perf-stat.ps.cpu-migrations
 1.063e+11            -4.1%   1.02e+11        perf-stat.ps.instructions
 9.709e+12            +2.4%  9.941e+12        perf-stat.total.instructions




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


                 reply	other threads:[~2025-11-28  6:56 UTC|newest]

Thread overview: [no followups] expand[flat|nested]  mbox.gz  Atom feed

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=202511281428.bfcbd5a8-lkp@intel.com \
    --to=oliver.sang@intel.com \
    --cc=abuehaze@amazon.com \
    --cc=adamli@os.amperecomputing.com \
    --cc=aubrey.li@linux.intel.com \
    --cc=clm@meta.com \
    --cc=dietmar.eggemann@arm.com \
    --cc=joseph.salisbury@oracle.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=lkp@intel.com \
    --cc=oe-lkp@lists.linux.dev \
    --cc=peterz@infradead.org \
    --cc=x86@kernel.org \
    --cc=yu.c.chen@intel.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.