* [linux-next:master] [pipe] f017b0a495: hackbench.throughput 10.5% improvement
@ 2025-02-14 7:53 kernel test robot
0 siblings, 0 replies; only message in thread
From: kernel test robot @ 2025-02-14 7:53 UTC (permalink / raw)
To: Oleg Nesterov
Cc: oe-lkp, lkp, Christian Brauner, K Prateek Nayak, linux-fsdevel,
oliver.sang
Hello,
kernel test robot noticed a 10.5% improvement of hackbench.throughput on:
commit: f017b0a4951fac8f150232661b2cc0b67e0c57f0 ("pipe: don't update {a,c,m}time for anonymous pipes")
https://git.kernel.org/cgit/linux/kernel/git/next/linux-next.git master
testcase: hackbench
config: x86_64-rhel-9.4
compiler: gcc-12
test machine: 128 threads 2 sockets Intel(R) Xeon(R) Gold 6338 CPU @ 2.00GHz (Ice Lake) with 256G memory
parameters:
nr_threads: 800%
iterations: 4
mode: threads
ipc: pipe
cpufreq_governor: performance
Details are as below:
-------------------------------------------------------------------------------------------------->
The kernel config and materials to reproduce are available at:
https://download.01.org/0day-ci/archive/20250214/202502141548.9fa68773-lkp@intel.com
=========================================================================================
compiler/cpufreq_governor/ipc/iterations/kconfig/mode/nr_threads/rootfs/tbox_group/testcase:
gcc-12/performance/pipe/4/x86_64-rhel-9.4/threads/800%/debian-12-x86_64-20240206.cgz/lkp-icl-2sp2/hackbench
commit:
262b2fa99c ("pipe: introduce struct file_operations pipeanon_fops")
f017b0a495 ("pipe: don't update {a,c,m}time for anonymous pipes")
262b2fa99cbe02a7 f017b0a4951fac8f150232661b2
---------------- ---------------------------
%stddev %change %stddev
\ | \
319054 -2.8% 310139 proc-vmstat.nr_active_anon
319054 -2.8% 310139 proc-vmstat.nr_zone_active_anon
549457 ± 92% -94.4% 30640 ± 30% sched_debug.cfs_rq:/.load.max
49885 ± 87% -88.9% 5535 ± 17% sched_debug.cfs_rq:/.load.stddev
1266298 +10.5% 1399088 hackbench.throughput
1237971 ± 2% +10.0% 1361485 ± 2% hackbench.throughput_avg
1266298 +10.5% 1399088 hackbench.throughput_best
4837 ± 2% -11.3% 4289 ± 2% hackbench.time.system_time
6.114e+10 -4.2% 5.86e+10 perf-stat.i.branch-instructions
2.74e+11 -2.0% 2.686e+11 perf-stat.i.cpu-cycles
1167 ± 3% -7.4% 1080 ± 3% perf-stat.i.cycles-between-cache-misses
2.527e+11 -6.0% 2.376e+11 perf-stat.i.instructions
0.87 ± 3% +15.0% 1.00 ± 4% perf-stat.overall.MPKI
1.07 +4.2% 1.12 perf-stat.overall.cpi
1233 ± 3% -9.3% 1118 ± 4% perf-stat.overall.cycles-between-cache-misses
0.93 -4.0% 0.89 perf-stat.overall.ipc
6.45e+10 -4.5% 6.161e+10 perf-stat.ps.branch-instructions
2.318e+08 ± 2% +7.7% 2.496e+08 ± 4% perf-stat.ps.cache-misses
2.856e+11 -2.4% 2.788e+11 perf-stat.ps.cpu-cycles
2.662e+11 -6.3% 2.494e+11 perf-stat.ps.instructions
10565 ± 3% +8.0% 11409 ± 2% perf-stat.ps.minor-faults
10565 ± 3% +8.0% 11409 ± 2% perf-stat.ps.page-faults
1.435e+13 -14.2% 1.232e+13 perf-stat.total.instructions
299.84 ± 47% -100.0% 0.00 perf-sched.sch_delay.avg.ms.__cond_resched.__alloc_frozen_pages_noprof.alloc_pages_mpol.alloc_pages_noprof.pipe_write
35.32 ± 24% -46.6% 18.84 ± 30% perf-sched.sch_delay.avg.ms.__cond_resched.__wait_for_common.affine_move_task.__set_cpus_allowed_ptr.__sched_setaffinity
169.52 ± 79% -100.0% 0.00 perf-sched.sch_delay.avg.ms.__cond_resched.mutex_lock.pipe_read.vfs_read.ksys_read
308.81 ± 34% -100.0% 0.00 perf-sched.sch_delay.avg.ms.__cond_resched.mutex_lock.pipe_write.vfs_write.ksys_write
308.90 ± 30% -47.0% 163.58 ± 19% perf-sched.sch_delay.avg.ms.irqentry_exit_to_user_mode.asm_sysvec_apic_timer_interrupt.[unknown].[unknown]
10.16 ±210% -99.7% 0.03 ±115% perf-sched.sch_delay.avg.ms.irqentry_exit_to_user_mode.asm_sysvec_reschedule_ipi.[unknown]
85.33 ± 25% -100.0% 0.00 perf-sched.sch_delay.avg.ms.pipe_read.vfs_read.ksys_read.do_syscall_64
209.12 ± 31% -100.0% 0.00 perf-sched.sch_delay.avg.ms.pipe_write.vfs_write.ksys_write.do_syscall_64
85.21 ± 62% -100.0% 0.00 perf-sched.sch_delay.avg.ms.schedule_preempt_disabled.__mutex_lock.constprop.0.pipe_read
374.84 ± 38% -100.0% 0.00 perf-sched.sch_delay.avg.ms.schedule_preempt_disabled.__mutex_lock.constprop.0.pipe_write
39.29 ± 55% -55.1% 17.63 ± 13% perf-sched.sch_delay.avg.ms.syscall_exit_to_user_mode.do_syscall_64.entry_SYSCALL_64_after_hwframe.[unknown]
5455 ± 49% -100.0% 0.00 perf-sched.sch_delay.max.ms.__cond_resched.__alloc_frozen_pages_noprof.alloc_pages_mpol.alloc_pages_noprof.pipe_write
6980 ± 12% -100.0% 0.00 perf-sched.sch_delay.max.ms.__cond_resched.mutex_lock.pipe_read.vfs_read.ksys_read
8278 ± 8% -100.0% 0.00 perf-sched.sch_delay.max.ms.__cond_resched.mutex_lock.pipe_write.vfs_write.ksys_write
8110 ± 9% -36.9% 5114 ± 16% perf-sched.sch_delay.max.ms.irqentry_exit_to_user_mode.asm_sysvec_apic_timer_interrupt.[unknown].[unknown]
8143 ± 12% -100.0% 0.00 perf-sched.sch_delay.max.ms.pipe_read.vfs_read.ksys_read.do_syscall_64
8560 ± 9% -100.0% 0.00 perf-sched.sch_delay.max.ms.pipe_write.vfs_write.ksys_write.do_syscall_64
2455 ±109% -100.0% 0.00 perf-sched.sch_delay.max.ms.schedule_preempt_disabled.__mutex_lock.constprop.0.pipe_read
7556 ± 13% -100.0% 0.00 perf-sched.sch_delay.max.ms.schedule_preempt_disabled.__mutex_lock.constprop.0.pipe_write
8543 ± 11% -37.6% 5332 ± 16% perf-sched.sch_delay.max.ms.syscall_exit_to_user_mode.do_syscall_64.entry_SYSCALL_64_after_hwframe.[unknown]
96.49 ± 28% -44.7% 53.38 ± 12% perf-sched.total_sch_delay.average.ms
8719 ± 10% -37.3% 5462 ± 15% perf-sched.total_sch_delay.max.ms
261.40 ± 29% -46.0% 141.08 ± 12% perf-sched.total_wait_and_delay.average.ms
17438 ± 10% -37.9% 10828 ± 16% perf-sched.total_wait_and_delay.max.ms
164.90 ± 30% -46.8% 87.70 ± 13% perf-sched.total_wait_time.average.ms
8862 ± 11% -35.6% 5710 ± 15% perf-sched.total_wait_time.max.ms
846.91 ± 36% -100.0% 0.00 perf-sched.wait_and_delay.avg.ms.__cond_resched.__alloc_frozen_pages_noprof.alloc_pages_mpol.alloc_pages_noprof.pipe_write
846.15 ± 37% -100.0% 0.00 perf-sched.wait_and_delay.avg.ms.__cond_resched.mutex_lock.pipe_write.vfs_write.ksys_write
858.41 ± 34% -50.4% 426.01 ± 19% perf-sched.wait_and_delay.avg.ms.irqentry_exit_to_user_mode.asm_sysvec_apic_timer_interrupt.[unknown].[unknown]
227.96 ± 27% -100.0% 0.00 perf-sched.wait_and_delay.avg.ms.pipe_read.vfs_read.ksys_read.do_syscall_64
576.02 ± 31% -100.0% 0.00 perf-sched.wait_and_delay.avg.ms.pipe_write.vfs_write.ksys_write.do_syscall_64
983.53 ± 40% -100.0% 0.00 perf-sched.wait_and_delay.avg.ms.schedule_preempt_disabled.__mutex_lock.constprop.0.pipe_write
67.17 ± 10% -100.0% 0.00 perf-sched.wait_and_delay.count.__cond_resched.__alloc_frozen_pages_noprof.alloc_pages_mpol.alloc_pages_noprof.pipe_write
7320 ± 6% -100.0% 0.00 perf-sched.wait_and_delay.count.__cond_resched.mutex_lock.pipe_write.vfs_write.ksys_write
752867 ± 2% -100.0% 0.00 perf-sched.wait_and_delay.count.pipe_read.vfs_read.ksys_read.do_syscall_64
96327 ± 3% -100.0% 0.00 perf-sched.wait_and_delay.count.pipe_write.vfs_write.ksys_write.do_syscall_64
1106 ± 10% -100.0% 0.00 perf-sched.wait_and_delay.count.schedule_preempt_disabled.__mutex_lock.constprop.0.pipe_write
11731 ± 36% -100.0% 0.00 perf-sched.wait_and_delay.max.ms.__cond_resched.__alloc_frozen_pages_noprof.alloc_pages_mpol.alloc_pages_noprof.pipe_write
16557 ± 8% -100.0% 0.00 perf-sched.wait_and_delay.max.ms.__cond_resched.mutex_lock.pipe_write.vfs_write.ksys_write
16104 ± 9% -36.4% 10235 ± 16% perf-sched.wait_and_delay.max.ms.irqentry_exit_to_user_mode.asm_sysvec_apic_timer_interrupt.[unknown].[unknown]
16318 ± 12% -100.0% 0.00 perf-sched.wait_and_delay.max.ms.pipe_read.vfs_read.ksys_read.do_syscall_64
17121 ± 9% -100.0% 0.00 perf-sched.wait_and_delay.max.ms.pipe_write.vfs_write.ksys_write.do_syscall_64
15123 ± 13% -100.0% 0.00 perf-sched.wait_and_delay.max.ms.schedule_preempt_disabled.__mutex_lock.constprop.0.pipe_write
8275 ± 15% -33.0% 5544 ± 15% perf-sched.wait_and_delay.max.ms.smpboot_thread_fn.kthread.ret_from_fork.ret_from_fork_asm
17047 ± 11% -37.3% 10687 ± 16% perf-sched.wait_and_delay.max.ms.syscall_exit_to_user_mode.do_syscall_64.entry_SYSCALL_64_after_hwframe.[unknown]
547.07 ± 33% -100.0% 0.00 perf-sched.wait_time.avg.ms.__cond_resched.__alloc_frozen_pages_noprof.alloc_pages_mpol.alloc_pages_noprof.pipe_write
374.49 ± 48% -100.0% 0.00 perf-sched.wait_time.avg.ms.__cond_resched.__mutex_lock.constprop.0.pipe_write
36.27 ± 19% -44.7% 20.06 ± 25% perf-sched.wait_time.avg.ms.__cond_resched.__wait_for_common.affine_move_task.__set_cpus_allowed_ptr.__sched_setaffinity
280.11 ± 85% -100.0% 0.00 perf-sched.wait_time.avg.ms.__cond_resched.mutex_lock.pipe_read.vfs_read.ksys_read
537.34 ± 38% -100.0% 0.00 perf-sched.wait_time.avg.ms.__cond_resched.mutex_lock.pipe_write.vfs_write.ksys_write
0.35 ±138% -100.0% 0.00 perf-sched.wait_time.avg.ms.__cond_resched.pipe_read.vfs_read.ksys_read.do_syscall_64
549.51 ± 37% -52.2% 262.43 ± 20% perf-sched.wait_time.avg.ms.irqentry_exit_to_user_mode.asm_sysvec_apic_timer_interrupt.[unknown].[unknown]
10.16 ±210% -99.6% 0.04 ±134% perf-sched.wait_time.avg.ms.irqentry_exit_to_user_mode.asm_sysvec_reschedule_ipi.[unknown]
142.63 ± 28% -100.0% 0.00 perf-sched.wait_time.avg.ms.pipe_read.vfs_read.ksys_read.do_syscall_64
366.89 ± 31% -100.0% 0.00 perf-sched.wait_time.avg.ms.pipe_write.vfs_write.ksys_write.do_syscall_64
39.52 ± 95% -73.6% 10.44 ± 53% perf-sched.wait_time.avg.ms.rcu_gp_kthread.kthread.ret_from_fork.ret_from_fork_asm
149.09 ± 38% -100.0% 0.00 perf-sched.wait_time.avg.ms.schedule_preempt_disabled.__mutex_lock.constprop.0.pipe_read
608.70 ± 42% -100.0% 0.00 perf-sched.wait_time.avg.ms.schedule_preempt_disabled.__mutex_lock.constprop.0.pipe_write
11.89 ±178% +22112.3% 2641 ± 61% perf-sched.wait_time.avg.ms.schedule_preempt_disabled.rwsem_down_write_slowpath.down_write_killable.vm_mmap_pgoff
43.32 ± 27% -61.1% 16.86 ± 29% perf-sched.wait_time.avg.ms.schedule_timeout.rcu_gp_fqs_loop.rcu_gp_kthread.kthread
6944 ± 17% -100.0% 0.00 perf-sched.wait_time.max.ms.__cond_resched.__alloc_frozen_pages_noprof.alloc_pages_mpol.alloc_pages_noprof.pipe_write
1676 ±126% -100.0% 0.00 perf-sched.wait_time.max.ms.__cond_resched.__mutex_lock.constprop.0.pipe_write
7277 ± 10% -100.0% 0.00 perf-sched.wait_time.max.ms.__cond_resched.mutex_lock.pipe_read.vfs_read.ksys_read
8328 ± 8% -100.0% 0.00 perf-sched.wait_time.max.ms.__cond_resched.mutex_lock.pipe_write.vfs_write.ksys_write
3.54 ±175% -100.0% 0.00 perf-sched.wait_time.max.ms.__cond_resched.pipe_read.vfs_read.ksys_read.do_syscall_64
8192 ± 9% -37.5% 5122 ± 16% perf-sched.wait_time.max.ms.irqentry_exit_to_user_mode.asm_sysvec_apic_timer_interrupt.[unknown].[unknown]
7035 ± 5% -68.5% 2216 ± 81% perf-sched.wait_time.max.ms.irqentry_exit_to_user_mode.asm_sysvec_call_function_single.[unknown].[unknown]
8490 ± 12% -100.0% 0.00 perf-sched.wait_time.max.ms.pipe_read.vfs_read.ksys_read.do_syscall_64
8581 ± 9% -100.0% 0.00 perf-sched.wait_time.max.ms.pipe_write.vfs_write.ksys_write.do_syscall_64
915.16 ±118% -86.0% 127.99 ± 79% perf-sched.wait_time.max.ms.rcu_gp_kthread.kthread.ret_from_fork.ret_from_fork_asm
4449 ± 64% -100.0% 0.00 perf-sched.wait_time.max.ms.schedule_preempt_disabled.__mutex_lock.constprop.0.pipe_read
8142 ± 13% -100.0% 0.00 perf-sched.wait_time.max.ms.schedule_preempt_disabled.__mutex_lock.constprop.0.pipe_write
11.89 ±178% +27553.2% 3288 ± 58% perf-sched.wait_time.max.ms.schedule_preempt_disabled.rwsem_down_write_slowpath.down_write_killable.vm_mmap_pgoff
8275 ± 15% -33.0% 5544 ± 15% perf-sched.wait_time.max.ms.smpboot_thread_fn.kthread.ret_from_fork.ret_from_fork_asm
8719 ± 11% -36.0% 5584 ± 16% perf-sched.wait_time.max.ms.syscall_exit_to_user_mode.do_syscall_64.entry_SYSCALL_64_after_hwframe.[unknown]
8432 ± 10% -36.3% 5373 ± 19% perf-sched.wait_time.max.ms.worker_thread.kthread.ret_from_fork.ret_from_fork_asm
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:[~2025-02-14 7:55 UTC | newest]
Thread overview: (only message) (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2025-02-14 7:53 [linux-next:master] [pipe] f017b0a495: hackbench.throughput 10.5% 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.