[linux-next:master] [mm/page_alloc] 6698721815: vm-scalability.throughput 16.9% improvement
kernel test robot <[email protected]>
| Newsgroups | dev.linux.lists.oe-lkp,org.kvack.linux-mm |
|---|---|
| Message-ID | <[email protected]> |
Hello,
kernel test robot noticed a 16.9% improvement of vm-scalability.throughput on:
commit: 66987218154918a6341a3e3eeeee58110a69e0bb ("mm/page_alloc: ignore the exact initial compaction result")
https://git.kernel.org/cgit/linux/kernel/git/next/linux-next.git master
testcase: vm-scalability
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:
runtime: 300s
test: lru-file-readonce
cpufreq_governor: performance
Details are as below:
-------------------------------------------------------------------------------------------------->
The kernel config and materials to reproduce are available at:
https://download.01.org/0day-ci/archive/20260206/[email protected]
=========================================================================================
compiler/cpufreq_governor/kconfig/rootfs/runtime/tbox_group/test/testcase:
gcc-14/performance/x86_64-rhel-9.4/debian-13-x86_64-20250902.cgz/300s/lkp-emr-2sp1/lru-file-readonce/vm-scalability
commit:
d17f02417a ("mm/rmap: separate out fork-only logic on anon_vma_clone()")
6698721815 ("mm/page_alloc: ignore the exact initial compaction result")
d17f02417a337de0 66987218154918a6341a3e3eeee
---------------- ---------------------------
%stddev %change %stddev
\ | \
2.07 ±201% +213.6% 6.50 ± 96% vmstat.memory.buff
30.09 ± 3% +2.8 32.89 ± 3% mpstat.cpu.all.idle%
0.15 ± 3% +0.0 0.16 ± 3% mpstat.cpu.all.irq%
1.10 +0.1 1.21 mpstat.cpu.all.usr%
29.13 ± 4% +2.7 31.84 ± 2% turbostat.C6%
10961264 -12.1% 9635977 turbostat.IRQ
39.55 +6.9% 42.27 turbostat.RAMWatt
881807 ± 13% -19.1% 713350 ± 4% numa-meminfo.node0.KReclaimable
881807 ± 13% -19.1% 713350 ± 4% numa-meminfo.node0.SReclaimable
1055803 ± 11% -15.9% 888061 ± 4% numa-meminfo.node0.Slab
702201 ± 7% +18.8% 833965 ± 11% numa-meminfo.node2.KReclaimable
702201 ± 7% +18.8% 833965 ± 11% numa-meminfo.node2.SReclaimable
837079 ± 6% +16.2% 973054 ± 10% numa-meminfo.node2.Slab
426746 ± 2% +17.3% 500442 ± 2% vm-scalability.median
1166 ± 28% +1053.6 2219 ± 15% vm-scalability.stddev%
1.11e+08 ± 2% +16.9% 1.297e+08 vm-scalability.throughput
54.43 -10.4% 48.78 vm-scalability.time.elapsed_time
54.43 -10.4% 48.78 vm-scalability.time.elapsed_time.max
103326 ± 4% -15.5% 87260 ± 3% vm-scalability.time.involuntary_context_switches
17909 -3.9% 17205 vm-scalability.time.percent_of_cpu_this_job_got
9605 ± 2% -14.1% 8252 vm-scalability.time.system_time
1.17 ± 24% -0.6 0.58 ± 82% perf-profile.calltrace.cycles-pp.ksys_mmap_pgoff.do_syscall_64.entry_SYSCALL_64_after_hwframe
1.17 ± 24% -0.6 0.58 ± 82% perf-profile.calltrace.cycles-pp.vm_mmap_pgoff.ksys_mmap_pgoff.do_syscall_64.entry_SYSCALL_64_after_hwframe
0.97 ± 37% -0.6 0.40 ± 57% perf-profile.children.cycles-pp.rcu_pending
0.03 ±223% +0.2 0.24 ± 33% perf-profile.children.cycles-pp.__get_user_pages
0.03 ±223% +0.2 0.24 ± 33% perf-profile.children.cycles-pp.get_user_pages_remote
0.27 ± 52% +0.2 0.50 ± 20% perf-profile.children.cycles-pp.load_elf_interp
0.08 ±101% +0.2 0.33 ± 46% perf-profile.children.cycles-pp.get_arg_page
0.59 ± 37% -0.4 0.21 ±102% perf-profile.self.cycles-pp.__percpu_counter_sum
0.03 ±223% +0.2 0.21 ± 30% perf-profile.self.cycles-pp.sched_setaffinity
7.31 ± 2% +19.9% 8.77 ± 4% perf-stat.i.MPKI
58.30 +2.2 60.49 perf-stat.i.cache-miss-rate%
7.426e+08 +17.5% 8.722e+08 perf-stat.i.cache-misses
1.228e+09 +12.0% 1.376e+09 perf-stat.i.cache-references
580.71 ± 7% -16.7% 483.74 ± 3% perf-stat.i.cycles-between-cache-misses
7.41 +18.3% 8.76 ± 5% perf-stat.overall.MPKI
60.61 +3.0 63.61 perf-stat.overall.cache-miss-rate%
704.31 -17.1% 583.78 perf-stat.overall.cycles-between-cache-misses
1296 ± 2% -11.5% 1147 ± 5% perf-stat.overall.path-length
7.497e+08 +16.2% 8.715e+08 perf-stat.ps.cache-misses
1.237e+09 +10.8% 1.37e+09 perf-stat.ps.cache-references
5.281e+11 ± 2% -3.7% 5.086e+11 perf-stat.ps.cpu-cycles
5.569e+12 ± 2% -11.5% 4.928e+12 ± 5% perf-stat.total.instructions
3008740 ± 2% +33.6% 4020245 ± 3% proc-vmstat.allocstall_movable
482140 ± 81% +81.6% 875667 ± 51% proc-vmstat.compact_stall
313239 ± 60% +108.2% 652063 ± 45% proc-vmstat.compact_success
8065 -2.8% 7836 proc-vmstat.nr_page_table_pages
39936876 ± 2% -59.1% 16330199 proc-vmstat.numa_foreign
95893347 -76.7% 22371986 proc-vmstat.numa_hit
95488037 -77.0% 21972039 proc-vmstat.numa_local
39916786 ± 2% -59.1% 16315472 proc-vmstat.numa_miss
40314192 ± 2% -58.5% 16714902 proc-vmstat.numa_other
687855 -2.2% 672475 proc-vmstat.pgfault
37269 -2.4% 36389 proc-vmstat.pgreuse
7.849e+08 -8.8% 7.158e+08 ± 3% proc-vmstat.pgscan_direct
2.277e+08 ± 5% +29.8% 2.956e+08 ± 9% proc-vmstat.pgscan_kswapd
7.849e+08 -8.8% 7.158e+08 ± 3% proc-vmstat.pgsteal_direct
2.277e+08 ± 5% +29.8% 2.956e+08 ± 9% proc-vmstat.pgsteal_kswapd
23897099 -77.3% 5427738 ± 5% numa-numastat.node0.local_node
9808674 ± 3% -65.9% 3340632 ± 3% numa-numastat.node0.numa_foreign
23996306 -76.9% 5539552 ± 4% numa-numastat.node0.numa_hit
9935480 ± 3% -48.6% 5104083 ± 2% numa-numastat.node0.numa_miss
10033238 ± 3% -48.0% 5215298 ± 3% numa-numastat.node0.other_node
23818598 ± 2% -77.6% 5336433 ± 3% numa-numastat.node1.local_node
10422599 ± 4% -60.6% 4107508 ± 5% numa-numastat.node1.numa_foreign
23903690 ± 2% -77.3% 5425282 ± 3% numa-numastat.node1.numa_hit
9568102 ± 3% -62.3% 3606134 ± 2% numa-numastat.node1.numa_miss
9650828 ± 3% -61.7% 3694897 ± 2% numa-numastat.node1.other_node
23244424 -76.6% 5433322 ± 3% numa-numastat.node2.local_node
10157128 ± 2% -58.5% 4212315 ± 3% numa-numastat.node2.numa_foreign
23367486 -76.3% 5531118 ± 3% numa-numastat.node2.numa_hit
9655573 ± 4% -59.5% 3909187 ± 3% numa-numastat.node2.numa_miss
9778660 ± 4% -59.0% 4007790 ± 2% numa-numastat.node2.other_node
24517225 ± 2% -76.5% 5765127 ± 3% numa-numastat.node3.local_node
9548474 ± 4% -51.1% 4669743 ± 4% numa-numastat.node3.numa_foreign
24615123 ± 2% -76.2% 5866615 ± 3% numa-numastat.node3.numa_hit
10757630 ± 2% -65.6% 3696067 ± 3% numa-numastat.node3.numa_miss
10851465 -65.0% 3796917 ± 2% numa-numastat.node3.other_node
1363284 ± 6% +12.5% 1533491 ± 7% numa-vmstat.node0.nr_free_pages_blocks
219953 ± 13% -19.0% 178205 ± 4% numa-vmstat.node0.nr_slab_reclaimable
9808674 ± 3% -65.9% 3340632 ± 3% numa-vmstat.node0.numa_foreign
23996706 -76.9% 5539212 ± 4% numa-vmstat.node0.numa_hit
23897500 -77.3% 5427399 ± 5% numa-vmstat.node0.numa_local
9935480 ± 3% -48.6% 5104083 ± 2% numa-vmstat.node0.numa_miss
10033238 ± 3% -48.0% 5215298 ± 3% numa-vmstat.node0.numa_other
1470383 ± 20% -82.6% 255546 ± 73% numa-vmstat.node0.workingset_nodereclaim
10422599 ± 4% -60.6% 4107508 ± 5% numa-vmstat.node1.numa_foreign
23903980 ± 2% -77.3% 5425155 ± 3% numa-vmstat.node1.numa_hit
23818888 ± 2% -77.6% 5336306 ± 3% numa-vmstat.node1.numa_local
9568102 ± 3% -62.3% 3606134 ± 2% numa-vmstat.node1.numa_miss
9650828 ± 3% -61.7% 3694897 ± 2% numa-vmstat.node1.numa_other
1342994 ± 6% +14.2% 1533478 ± 9% numa-vmstat.node2.nr_free_pages_blocks
175084 ± 7% +19.0% 208437 ± 10% numa-vmstat.node2.nr_slab_reclaimable
10157128 ± 2% -58.5% 4212315 ± 3% numa-vmstat.node2.numa_foreign
23367487 -76.3% 5530880 ± 3% numa-vmstat.node2.numa_hit
23244424 -76.6% 5433084 ± 3% numa-vmstat.node2.numa_local
9655573 ± 4% -59.5% 3909187 ± 3% numa-vmstat.node2.numa_miss
9778660 ± 4% -59.0% 4007790 ± 2% numa-vmstat.node2.numa_other
9548474 ± 4% -51.1% 4669743 ± 4% numa-vmstat.node3.numa_foreign
24615201 ± 2% -76.2% 5866947 ± 3% numa-vmstat.node3.numa_hit
24517303 ± 2% -76.5% 5765459 ± 3% numa-vmstat.node3.numa_local
10757630 ± 2% -65.6% 3696067 ± 3% numa-vmstat.node3.numa_miss
10851465 -65.0% 3796917 ± 2% numa-vmstat.node3.numa_other
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