linux-mm.kvack.org archive mirror
 help / color / mirror / Atom feed
* [linus:master] [mm]  6bef4c2f97:  stress-ng.mlockmany.ops_per_sec 5.2% improvement
@ 2025-06-10 14:38 kernel test robot
  0 siblings, 0 replies; only message in thread
From: kernel test robot @ 2025-06-10 14:38 UTC (permalink / raw)
  To: Suren Baghdasaryan
  Cc: oe-lkp, lkp, linux-kernel, Andrew Morton, Lorenzo Stoakes,
	Shivank Garg, Vlastimil Babka, Christian Brauner,
	David Hildenbrand, David Howells, Davidlohr Bueso, Hugh Dickins,
	Jann Horn, Johannes Weiner, Jonathan Corbet, Klara Modin,
	Liam R. Howlett, Lokesh Gidra, Mateusz Guzik, Matthew Wilcox,
	Mel Gorman, Michal Hocko, Minchan Kim, Oleg Nesterov,
	Pasha Tatashin, Paul E . McKenney, Peter Xu, Peter Zijlstra,
	Shakeel Butt, Sourav Panda, Wei Yang, Will Deacon, Heiko Carstens,
	Stephen Rothwell, linux-mm, oliver.sang



Hello,

kernel test robot noticed a 5.2% improvement of stress-ng.mlockmany.ops_per_sec on:


commit: 6bef4c2f97221f3b595d08c8656eb5845ef80fe9 ("mm: move lesser used vma_area_struct members into the last cacheline")
https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git master


testcase: stress-ng
config: x86_64-rhel-9.4
compiler: gcc-12
test machine: 64 threads 2 sockets Intel(R) Xeon(R) Gold 6346 CPU @ 3.10GHz (Ice Lake) with 256G memory
parameters:

	nr_threads: 100%
	testtime: 60s
	test: mlockmany
	cpufreq_governor: performance



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


The kernel config and materials to reproduce are available at:
https://download.01.org/0day-ci/archive/20250610/202506102254.13cda0af-lkp@intel.com

=========================================================================================
compiler/cpufreq_governor/kconfig/nr_threads/rootfs/tbox_group/test/testcase/testtime:
  gcc-12/performance/x86_64-rhel-9.4/100%/debian-12-x86_64-20240206.cgz/lkp-icl-2sp7/mlockmany/stress-ng/60s

commit: 
  f35ab95ca0 ("mm: replace vm_lock and detached flag with a reference count")
  6bef4c2f97 ("mm: move lesser used vma_area_struct members into the last cacheline")

f35ab95ca0af7a27 6bef4c2f97221f3b595d08c8656 
---------------- --------------------------- 
         %stddev     %change         %stddev
             \          |                \  
      0.66 ±  5%      -0.1        0.57 ±  9%  mpstat.cpu.all.soft%
     27183            +1.9%      27708        vmstat.system.cs
    264643            +5.2%     278326        stress-ng.mlockmany.ops
      4406            +5.2%       4634        stress-ng.mlockmany.ops_per_sec
    314509            +4.9%     329874        stress-ng.time.voluntary_context_switches
    343582            -3.7%     330742 ±  2%  proc-vmstat.nr_active_anon
    454064            -2.7%     441886        proc-vmstat.nr_anon_pages
     54743            -3.5%      52828        proc-vmstat.nr_slab_unreclaimable
    343583            -3.7%     330741 ±  2%  proc-vmstat.nr_zone_active_anon
      1.99 ±  8%     -14.0%       1.72 ± 12%  sched_debug.cfs_rq:/.h_nr_queued.stddev
      1.98 ±  8%     -13.9%       1.71 ± 12%  sched_debug.cfs_rq:/.h_nr_runnable.stddev
      0.00 ± 18%     -24.8%       0.00 ± 20%  sched_debug.cpu.next_balance.stddev
      1.99 ±  8%     -13.8%       1.72 ± 12%  sched_debug.cpu.nr_running.stddev
      0.25            +0.0        0.25        perf-stat.i.branch-miss-rate%
  21663531            +1.7%   22033919        perf-stat.i.branch-misses
     27855            +1.8%      28352        perf-stat.i.context-switches
      0.25            +0.0        0.25        perf-stat.overall.branch-miss-rate%
  21319615            +1.7%   21691011        perf-stat.ps.branch-misses
     27388            +1.7%      27866        perf-stat.ps.context-switches
     19.64 ±  7%     -18.7%      15.97 ± 11%  perf-sched.sch_delay.avg.ms.__cond_resched.__vmalloc_area_node.__vmalloc_node_range_noprof.alloc_thread_stack_node.dup_task_struct
     11.34 ±  8%     -13.5%       9.80 ±  6%  perf-sched.sch_delay.avg.ms.__cond_resched.down_read.__mm_populate.do_mlock.__x64_sys_mlock
     17.11 ±  4%      -8.2%      15.70 ±  5%  perf-sched.sch_delay.avg.ms.__cond_resched.mlock_pte_range.walk_pmd_range.isra.0
     10.51 ± 10%     +35.6%      14.26 ± 15%  perf-sched.sch_delay.avg.ms.__cond_resched.wp_page_copy.__handle_mm_fault.handle_mm_fault.do_user_addr_fault
     52.76 ± 22%     -31.2%      36.28 ± 18%  perf-sched.sch_delay.max.ms.irqentry_exit_to_user_mode.asm_exc_page_fault.[unknown].[unknown]
     50.19 ±  7%     -26.9%      36.68 ± 45%  perf-sched.wait_and_delay.avg.ms.__cond_resched.__vmalloc_area_node.__vmalloc_node_range_noprof.alloc_thread_stack_node.dup_task_struct
     23.36 ±  9%     -14.2%      20.03 ±  6%  perf-sched.wait_and_delay.avg.ms.__cond_resched.down_read.__mm_populate.do_mlock.__x64_sys_mlock
     51.05 ± 10%     -34.3%      33.53 ± 45%  perf-sched.wait_and_delay.avg.ms.__cond_resched.kmem_cache_alloc_noprof.prepare_creds.copy_creds.copy_process
    245.67 ±  6%     -47.6%     128.83 ±  4%  perf-sched.wait_and_delay.count.__cond_resched.copy_page_range.dup_mmap.dup_mm.constprop
    286.83 ±  7%     -21.0%     226.67 ±  5%  perf-sched.wait_and_delay.count.__cond_resched.down_write.anon_vma_clone.anon_vma_fork.dup_mmap
    120.67 ±  9%     +32.6%     160.00 ±  8%  perf-sched.wait_and_delay.count.schedule_preempt_disabled.rwsem_down_write_slowpath.down_write.anon_vma_fork
    225.41 ± 31%     -33.7%     149.44 ±  7%  perf-sched.wait_and_delay.max.ms.__cond_resched.copy_page_range.dup_mmap.dup_mm.constprop
     77.77 ± 73%     +79.0%     139.22 ± 15%  perf-sched.wait_and_delay.max.ms.__cond_resched.uprobe_start_dup_mmap.dup_mm.constprop.0
     12.02 ± 11%     -14.9%      10.23 ±  6%  perf-sched.wait_time.avg.ms.__cond_resched.down_read.__mm_populate.do_mlock.__x64_sys_mlock
     31.78 ± 18%     -31.9%      21.63 ± 11%  perf-sched.wait_time.avg.ms.__cond_resched.kmem_cache_alloc_noprof.prepare_creds.copy_creds.copy_process
     16.57 ±  5%      -9.3%      15.03 ±  5%  perf-sched.wait_time.avg.ms.__cond_resched.mlock_pte_range.walk_pmd_range.isra.0
     25.21 ±  7%     +12.4%      28.34 ±  6%  perf-sched.wait_time.avg.ms.__cond_resched.mutex_lock_killable.pcpu_alloc_noprof.mm_init.dup_mm
     24.68 ± 29%     +39.0%      34.31 ± 15%  perf-sched.wait_time.avg.ms.__cond_resched.uprobe_start_dup_mmap.dup_mm.constprop.0
    207.48 ± 35%     -32.5%     140.02 ±  6%  perf-sched.wait_time.max.ms.__cond_resched.copy_page_range.dup_mmap.dup_mm.constprop
     70.62 ± 41%     +75.6%     124.03 ± 15%  perf-sched.wait_time.max.ms.__cond_resched.uprobe_start_dup_mmap.dup_mm.constprop.0




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



^ permalink raw reply	[flat|nested] only message in thread

only message in thread, other threads:[~2025-06-10 14:39 UTC | newest]

Thread overview: (only message) (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2025-06-10 14:38 [linus:master] [mm] 6bef4c2f97: stress-ng.mlockmany.ops_per_sec 5.2% improvement kernel test robot

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).