* [linus:master] [mm] ab04945f91: stress-ng.fd-abuse.ops_per_sec 3.4% improvement
@ 2025-12-19 6:21 kernel test robot
0 siblings, 0 replies; only message in thread
From: kernel test robot @ 2025-12-19 6:21 UTC (permalink / raw)
To: Lorenzo Stoakes
Cc: oe-lkp, lkp, linux-kernel, Andrew Morton, Jason Gunthorpe,
Alexander Gordeev, Al Viro, Andreas Larsson, Andrey Konovalov,
Arnd Bergmann, Baolin Wang, Baoquan He, Chatre, Reinette,
Christian Borntraeger, Christian Brauner, Dan Williams,
Dave Jiang, Dave Martin, Dave Young, David Hildenbrand,
David S. Miller, Dmitriy Vyukov, Greg Kroah-Hartman, Guo Ren,
Heiko Carstens, Hugh Dickins, James Morse, Jan Kara, Jann Horn,
Jonathan Corbet, Kevin Tian, Konstantin Komarov, Liam Howlett,
Luck, Tony, Matthew Wilcox, Michal Hocko, Mike Rapoport,
Muchun Song, Nicolas Pitre, Oscar Salvador, Pedro Falcato,
Robin Murohy, Sumanth Korikkar, Suren Baghdasaryan, Sven Schnelle,
Thomas Bogendoerfer, Uladzislau Rezki (Sony), Vasily Gorbik,
Vishal Verma, Vivek Goyal, Vlastimil Babka, Will Deacon,
oliver.sang
Hello,
kernel test robot noticed a 3.4% improvement of stress-ng.fd-abuse.ops_per_sec on:
commit: ab04945f91bcad1668af57bbb575771e794aea8d ("mm: update mem char driver to use mmap_prepare")
https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git master
testcase: stress-ng
config: x86_64-rhel-9.4
compiler: gcc-14
test machine: 256 threads 4 sockets INTEL(R) XEON(R) PLATINUM 8592+ (Emerald Rapids) with 256G memory
parameters:
nr_threads: 100%
testtime: 60s
test: fd-abuse
cpufreq_governor: performance
Details are as below:
-------------------------------------------------------------------------------------------------->
The kernel config and materials to reproduce are available at:
https://download.01.org/0day-ci/archive/20251218/202512181616.16b76cde-lkp@intel.com
=========================================================================================
compiler/cpufreq_governor/kconfig/nr_threads/rootfs/tbox_group/test/testcase/testtime:
gcc-14/performance/x86_64-rhel-9.4/100%/debian-13-x86_64-20250902.cgz/lkp-emr-2sp1/fd-abuse/stress-ng/60s
commit:
89646d9c74 ("mm: add shmem_zero_setup_desc()")
ab04945f91 ("mm: update mem char driver to use mmap_prepare")
89646d9c748c0902 ab04945f91bcad1668af57bbb57
---------------- ---------------------------
%stddev %change %stddev
\ | \
19.28 +2.8% 19.82 turbostat.RAMWatt
2438603 ± 3% -13.5% 2109471 ± 7% numa-numastat.node1.local_node
2517328 ± 4% -11.9% 2217162 ± 5% numa-numastat.node1.numa_hit
72858 ± 6% +91.5% 139519 ± 3% proc-vmstat.nr_slab_reclaimable
339884 +107.0% 703398 proc-vmstat.nr_slab_unreclaimable
20261925 ± 2% -25.3% 15142663 proc-vmstat.pgalloc_normal
18968240 ± 2% -31.6% 12978914 ± 2% proc-vmstat.pgfree
3.706e+09 +3.4% 3.833e+09 stress-ng.fd-abuse.ops
61777966 +3.4% 63895149 stress-ng.fd-abuse.ops_per_sec
285.34 +4.4% 297.95 stress-ng.time.user_time
16701 +2.7% 17144 stress-ng.time.voluntary_context_switches
8804549 +61.4% 14208934 meminfo.Committed_AS
291574 ± 6% +91.1% 557114 ± 3% meminfo.KReclaimable
12991379 ± 3% +16.6% 15149678 ± 2% meminfo.Memused
291574 ± 6% +91.1% 557114 ± 3% meminfo.SReclaimable
1353499 +107.7% 2811130 meminfo.SUnreclaim
1645073 +104.7% 3368244 meminfo.Slab
13335448 ± 3% +15.9% 15450436 ± 2% meminfo.max_used_kB
17516 ± 34% +97.4% 34572 ± 6% numa-vmstat.node0.nr_slab_reclaimable
88462 ± 6% +102.5% 179171 ± 2% numa-vmstat.node0.nr_slab_unreclaimable
66.79 ± 54% +106.9% 138.19 ± 33% numa-vmstat.node1.nr_anon_transparent_hugepages
16982 ± 18% +94.3% 32997 ± 19% numa-vmstat.node1.nr_slab_reclaimable
81860 ± 3% +113.2% 174536 ± 2% numa-vmstat.node1.nr_slab_unreclaimable
2521117 ± 4% -12.0% 2218931 ± 5% numa-vmstat.node1.numa_hit
2442392 ± 3% -13.6% 2111240 ± 7% numa-vmstat.node1.numa_local
19931 ± 28% +90.9% 38056 ± 7% numa-vmstat.node2.nr_slab_reclaimable
86088 ± 5% +99.7% 171897 numa-vmstat.node2.nr_slab_unreclaimable
18146 ± 12% +82.8% 33181 ± 13% numa-vmstat.node3.nr_slab_reclaimable
83060 ± 4% +111.5% 175660 ± 2% numa-vmstat.node3.nr_slab_unreclaimable
2.679e+10 +1.5% 2.718e+10 perf-stat.i.branch-instructions
43.06 +0.9 44.01 perf-stat.i.cache-miss-rate%
88883420 +5.0% 93333404 perf-stat.i.cache-misses
6.01 -1.2% 5.94 perf-stat.i.cpi
8654 -5.3% 8192 perf-stat.i.cycles-between-cache-misses
1.211e+11 +1.6% 1.23e+11 perf-stat.i.instructions
0.73 +3.5% 0.76 perf-stat.overall.MPKI
42.85 +0.9 43.75 perf-stat.overall.cache-miss-rate%
6.03 -1.3% 5.95 perf-stat.overall.cpi
8266 -4.7% 7879 perf-stat.overall.cycles-between-cache-misses
0.17 +1.3% 0.17 perf-stat.overall.ipc
2.635e+10 +1.4% 2.673e+10 perf-stat.ps.branch-instructions
86986582 +5.1% 91435274 perf-stat.ps.cache-misses
2.03e+08 +2.9% 2.09e+08 perf-stat.ps.cache-references
1.191e+11 +1.6% 1.21e+11 perf-stat.ps.instructions
7.248e+12 +1.6% 7.363e+12 perf-stat.total.instructions
70689 ± 34% +96.6% 138955 ± 6% numa-meminfo.node0.KReclaimable
70689 ± 34% +96.6% 138955 ± 6% numa-meminfo.node0.SReclaimable
356946 ± 6% +101.3% 718428 ± 3% numa-meminfo.node0.SUnreclaim
427635 ± 10% +100.5% 857383 ± 2% numa-meminfo.node0.Slab
139096 ± 54% +104.6% 284523 ± 33% numa-meminfo.node1.AnonHugePages
68643 ± 18% +93.3% 132663 ± 19% numa-meminfo.node1.KReclaimable
2076019 ± 6% +57.0% 3258377 ± 43% numa-meminfo.node1.MemUsed
68643 ± 18% +93.3% 132663 ± 19% numa-meminfo.node1.SReclaimable
330581 ± 3% +111.7% 699874 ± 2% numa-meminfo.node1.SUnreclaim
399224 ± 3% +108.5% 832538 ± 4% numa-meminfo.node1.Slab
80254 ± 27% +90.6% 152941 ± 7% numa-meminfo.node2.KReclaimable
80254 ± 27% +90.6% 152941 ± 7% numa-meminfo.node2.SReclaimable
347619 ± 4% +98.2% 689115 numa-meminfo.node2.SUnreclaim
427873 ± 7% +96.8% 842056 numa-meminfo.node2.Slab
73141 ± 11% +82.0% 133137 ± 13% numa-meminfo.node3.KReclaimable
3108062 ± 11% +41.5% 4398558 ± 26% numa-meminfo.node3.MemUsed
73141 ± 11% +82.0% 133137 ± 13% numa-meminfo.node3.SReclaimable
335184 ± 5% +110.2% 704492 ± 2% numa-meminfo.node3.SUnreclaim
408326 ± 5% +105.1% 837630 ± 3% numa-meminfo.node3.Slab
5.23 -1.9 3.31 perf-profile.calltrace.cycles-pp.stress_fd_lseek
46.32 -1.3 45.05 perf-profile.calltrace.cycles-pp.inode_sb_list_add.new_inode.__shmem_get_inode.__shmem_file_setup.shmem_zero_setup
46.38 -1.3 45.12 perf-profile.calltrace.cycles-pp.task_work_run.exit_to_user_mode_loop.do_syscall_64.entry_SYSCALL_64_after_hwframe.__munmap
46.32 -1.3 45.06 perf-profile.calltrace.cycles-pp.dput.__fput.task_work_run.exit_to_user_mode_loop.do_syscall_64
46.34 -1.3 45.08 perf-profile.calltrace.cycles-pp.__fput.task_work_run.exit_to_user_mode_loop.do_syscall_64.entry_SYSCALL_64_after_hwframe
46.44 -1.3 45.18 perf-profile.calltrace.cycles-pp.new_inode.__shmem_get_inode.__shmem_file_setup.shmem_zero_setup.__mmap_new_vma
46.54 -1.3 45.28 perf-profile.calltrace.cycles-pp.__shmem_get_inode.__shmem_file_setup.shmem_zero_setup.__mmap_new_vma.__mmap_region
46.30 -1.3 45.04 perf-profile.calltrace.cycles-pp.__dentry_kill.dput.__fput.task_work_run.exit_to_user_mode_loop
46.14 -1.3 44.88 perf-profile.calltrace.cycles-pp.evict.__dentry_kill.dput.__fput.task_work_run
45.84 -1.2 44.60 perf-profile.calltrace.cycles-pp.native_queued_spin_lock_slowpath._raw_spin_lock.evict.__dentry_kill.dput
45.99 -1.2 44.74 perf-profile.calltrace.cycles-pp._raw_spin_lock.evict.__dentry_kill.dput.__fput
0.67 +0.0 0.70 perf-profile.calltrace.cycles-pp.llseek
2.40 +0.0 2.44 perf-profile.calltrace.cycles-pp.mmap_region.do_mmap.vm_mmap_pgoff.ksys_mmap_pgoff.do_syscall_64
2.38 +0.0 2.43 perf-profile.calltrace.cycles-pp.__mmap_region.mmap_region.do_mmap.vm_mmap_pgoff.ksys_mmap_pgoff
2.54 +0.1 2.60 perf-profile.calltrace.cycles-pp.do_mmap.vm_mmap_pgoff.ksys_mmap_pgoff.do_syscall_64.entry_SYSCALL_64_after_hwframe
0.75 ± 2% +0.1 0.81 perf-profile.calltrace.cycles-pp.do_syscall_64.entry_SYSCALL_64_after_hwframe
2.58 +0.1 2.64 perf-profile.calltrace.cycles-pp.ksys_mmap_pgoff.do_syscall_64.entry_SYSCALL_64_after_hwframe.__mmap
0.77 ± 2% +0.1 0.83 perf-profile.calltrace.cycles-pp.entry_SYSCALL_64_after_hwframe
2.56 +0.1 2.62 perf-profile.calltrace.cycles-pp.vm_mmap_pgoff.ksys_mmap_pgoff.do_syscall_64.entry_SYSCALL_64_after_hwframe.__mmap
1.02 +0.1 1.09 perf-profile.calltrace.cycles-pp.do_syscall_64.entry_SYSCALL_64_after_hwframe.stress_fd_lseek
1.06 +0.1 1.14 perf-profile.calltrace.cycles-pp.entry_SYSCALL_64_after_hwframe.stress_fd_lseek
44.33 +0.8 45.15 perf-profile.calltrace.cycles-pp.exit_to_user_mode_loop.do_syscall_64.entry_SYSCALL_64_after_hwframe.__munmap
46.28 +0.8 47.11 perf-profile.calltrace.cycles-pp._raw_spin_lock.inode_sb_list_add.new_inode.__shmem_get_inode.__shmem_file_setup
46.20 +0.8 47.04 perf-profile.calltrace.cycles-pp.native_queued_spin_lock_slowpath._raw_spin_lock.inode_sb_list_add.new_inode.__shmem_get_inode
44.63 +0.8 45.47 perf-profile.calltrace.cycles-pp.shmem_zero_setup.__mmap_new_vma.__mmap_region.do_mmap.vm_mmap_pgoff
44.62 +0.8 45.47 perf-profile.calltrace.cycles-pp.__shmem_file_setup.shmem_zero_setup.__mmap_new_vma.__mmap_region.do_mmap
44.66 +0.8 45.51 perf-profile.calltrace.cycles-pp.do_syscall_64.entry_SYSCALL_64_after_hwframe.__munmap
44.66 +0.8 45.51 perf-profile.calltrace.cycles-pp.entry_SYSCALL_64_after_hwframe.__munmap
44.69 +0.8 45.53 perf-profile.calltrace.cycles-pp.__munmap
44.78 +0.9 45.63 perf-profile.calltrace.cycles-pp.__mmap_new_vma.__mmap_region.do_mmap.vm_mmap_pgoff.do_syscall_64
44.98 +0.9 45.84 perf-profile.calltrace.cycles-pp.__mmap_region.do_mmap.vm_mmap_pgoff.do_syscall_64.entry_SYSCALL_64_after_hwframe
45.11 +0.9 45.97 perf-profile.calltrace.cycles-pp.vm_mmap_pgoff.do_syscall_64.entry_SYSCALL_64_after_hwframe.__mmap
45.09 +0.9 45.96 perf-profile.calltrace.cycles-pp.do_mmap.vm_mmap_pgoff.do_syscall_64.entry_SYSCALL_64_after_hwframe.__mmap
47.74 +0.9 48.66 perf-profile.calltrace.cycles-pp.do_syscall_64.entry_SYSCALL_64_after_hwframe.__mmap
47.74 +0.9 48.67 perf-profile.calltrace.cycles-pp.entry_SYSCALL_64_after_hwframe.__mmap
47.80 +0.9 48.73 perf-profile.calltrace.cycles-pp.__mmap
0.00 +2.1 2.11 perf-profile.calltrace.cycles-pp.inode_sb_list_add.new_inode.__shmem_get_inode.__shmem_file_setup.shmem_zero_setup_desc
0.00 +2.1 2.12 perf-profile.calltrace.cycles-pp.new_inode.__shmem_get_inode.__shmem_file_setup.shmem_zero_setup_desc.__mmap_region
0.00 +2.1 2.12 perf-profile.calltrace.cycles-pp.__shmem_get_inode.__shmem_file_setup.shmem_zero_setup_desc.__mmap_region.mmap_region
0.00 +2.1 2.13 perf-profile.calltrace.cycles-pp.__shmem_file_setup.shmem_zero_setup_desc.__mmap_region.mmap_region.do_mmap
0.00 +2.1 2.13 perf-profile.calltrace.cycles-pp.shmem_zero_setup_desc.__mmap_region.mmap_region.do_mmap.vm_mmap_pgoff
5.24 -1.9 3.32 perf-profile.children.cycles-pp.stress_fd_lseek
46.39 -1.3 45.13 perf-profile.children.cycles-pp.task_work_run
46.73 -1.3 45.47 perf-profile.children.cycles-pp.shmem_zero_setup
46.30 -1.3 45.04 perf-profile.children.cycles-pp.__dentry_kill
46.48 -1.3 45.22 perf-profile.children.cycles-pp.exit_to_user_mode_loop
46.36 -1.3 45.11 perf-profile.children.cycles-pp.dput
46.14 -1.3 44.88 perf-profile.children.cycles-pp.evict
46.44 -1.2 45.20 perf-profile.children.cycles-pp.__fput
47.02 -1.2 45.78 perf-profile.children.cycles-pp.__mmap_new_vma
46.96 -1.2 45.72 perf-profile.children.cycles-pp.__munmap
92.11 -0.4 91.71 perf-profile.children.cycles-pp.native_queued_spin_lock_slowpath
92.39 -0.4 91.99 perf-profile.children.cycles-pp._raw_spin_lock
97.69 -0.1 97.58 perf-profile.children.cycles-pp.do_syscall_64
97.82 -0.1 97.71 perf-profile.children.cycles-pp.entry_SYSCALL_64_after_hwframe
0.08 -0.0 0.07 perf-profile.children.cycles-pp.__destroy_inode
0.11 +0.0 0.12 perf-profile.children.cycles-pp.kthread
0.11 +0.0 0.12 perf-profile.children.cycles-pp.ret_from_fork
0.11 +0.0 0.12 perf-profile.children.cycles-pp.ret_from_fork_asm
0.13 +0.0 0.14 perf-profile.children.cycles-pp.__fstat64
0.17 +0.0 0.18 perf-profile.children.cycles-pp.__x64_sys_pselect6
0.16 +0.0 0.17 perf-profile.children.cycles-pp.entry_SYSCALL_64_safe_stack
0.08 +0.0 0.09 perf-profile.children.cycles-pp.generic_perform_write
0.13 +0.0 0.14 perf-profile.children.cycles-pp.inet_listen
0.13 +0.0 0.14 perf-profile.children.cycles-pp.mas_store_gfp
0.08 +0.0 0.09 perf-profile.children.cycles-pp.mas_store_prealloc
0.11 ± 3% +0.0 0.12 perf-profile.children.cycles-pp.generic_file_write_iter
0.14 ± 2% +0.0 0.15 perf-profile.children.cycles-pp.mas_wr_node_store
0.15 +0.0 0.16 ± 2% perf-profile.children.cycles-pp.perf_event_mmap_event
0.38 +0.0 0.39 perf-profile.children.cycles-pp.arch_exit_to_user_mode_prepare
0.21 +0.0 0.22 perf-profile.children.cycles-pp.__x64_sys_fcntl
0.08 ± 5% +0.0 0.10 ± 4% perf-profile.children.cycles-pp.do_iter_readv_writev
0.10 +0.0 0.12 ± 4% perf-profile.children.cycles-pp.run_ksoftirqd
0.21 +0.0 0.23 ± 2% perf-profile.children.cycles-pp.kmem_cache_alloc_noprof
0.10 +0.0 0.12 ± 4% perf-profile.children.cycles-pp.smpboot_thread_fn
0.13 +0.0 0.15 ± 3% perf-profile.children.cycles-pp.handle_softirqs
0.12 ± 3% +0.0 0.14 ± 2% perf-profile.children.cycles-pp.rcu_core
0.12 +0.0 0.14 ± 3% perf-profile.children.cycles-pp.alloc_inode
0.12 +0.0 0.14 ± 3% perf-profile.children.cycles-pp.rcu_do_batch
0.67 +0.0 0.69 perf-profile.children.cycles-pp.entry_SYSCALL_64
0.46 +0.0 0.48 perf-profile.children.cycles-pp.do_vmi_munmap
0.44 +0.0 0.46 perf-profile.children.cycles-pp.do_vmi_align_munmap
0.48 +0.0 0.51 perf-profile.children.cycles-pp.entry_SYSRETQ_unsafe_stack
0.47 +0.0 0.50 perf-profile.children.cycles-pp.__vm_munmap
0.47 +0.0 0.50 perf-profile.children.cycles-pp.__x64_sys_munmap
0.05 +0.0 0.08 perf-profile.children.cycles-pp.inode_init_always_gfp
0.31 ± 4% +0.0 0.34 ± 4% perf-profile.children.cycles-pp.do_filp_open
0.31 ± 5% +0.0 0.34 ± 4% perf-profile.children.cycles-pp.path_openat
0.37 ± 4% +0.0 0.40 ± 3% perf-profile.children.cycles-pp.__x64_sys_openat
0.37 ± 4% +0.0 0.40 ± 3% perf-profile.children.cycles-pp.do_sys_openat2
0.85 +0.0 0.89 perf-profile.children.cycles-pp.llseek
2.40 +0.0 2.44 perf-profile.children.cycles-pp.mmap_region
2.59 +0.1 2.64 perf-profile.children.cycles-pp.ksys_mmap_pgoff
46.32 +0.8 47.16 perf-profile.children.cycles-pp.inode_sb_list_add
46.44 +0.9 47.30 perf-profile.children.cycles-pp.new_inode
46.54 +0.9 47.40 perf-profile.children.cycles-pp.__shmem_get_inode
46.73 +0.9 47.60 perf-profile.children.cycles-pp.__shmem_file_setup
47.37 +0.9 48.28 perf-profile.children.cycles-pp.__mmap_region
47.64 +0.9 48.56 perf-profile.children.cycles-pp.do_mmap
47.68 +0.9 48.60 perf-profile.children.cycles-pp.vm_mmap_pgoff
47.83 +0.9 48.76 perf-profile.children.cycles-pp.__mmap
0.00 +2.1 2.13 perf-profile.children.cycles-pp.shmem_zero_setup_desc
91.71 -0.4 91.31 perf-profile.self.cycles-pp.native_queued_spin_lock_slowpath
0.22 +0.0 0.23 perf-profile.self.cycles-pp.entry_SYSCALL_64
0.30 +0.0 0.31 perf-profile.self.cycles-pp.do_syscall_64
0.35 +0.0 0.36 perf-profile.self.cycles-pp.arch_exit_to_user_mode_prepare
0.47 +0.0 0.50 perf-profile.self.cycles-pp.entry_SYSRETQ_unsafe_stack
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:[~2025-12-19 6:22 UTC | newest]
Thread overview: (only message) (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2025-12-19 6:21 [linus:master] [mm] ab04945f91: stress-ng.fd-abuse.ops_per_sec 3.4% improvement kernel test robot
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.