Greeting, FYI, we noticed a 11.2% improvement of vm-scalability.throughput due to commit: commit: 170b04b7ae49634df103810dad67b22cf8a99aa6 ("mm/workingset: prepare the workingset detection infrastructure for anon LRU") https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git master in testcase: vm-scalability on test machine: 104 threads Skylake with 192G memory with following parameters: runtime: 300s size: 1T test: lru-shm cpufreq_governor: performance ucode: 0x2006906 test-description: The motivation behind this suite is to exercise functions and regions of the mm/ of the Linux kernel which are of interest to us. test-url: https://git.kernel.org/cgit/linux/kernel/git/wfg/vm-scalability.git/ Details are as below: --------------------------------------------------------------------------------------------------> To reproduce: git clone https://github.com/intel/lkp-tests.git cd lkp-tests bin/lkp install job.yaml # job file is attached in this email bin/lkp run job.yaml ========================================================================================= compiler/cpufreq_governor/kconfig/rootfs/runtime/size/tbox_group/test/testcase/ucode: gcc-9/performance/x86_64-rhel-8.3/debian-10.4-x86_64-20200603.cgz/300s/1T/lkp-skl-fpga01/lru-shm/vm-scalability/0x2006906 commit: b518154e59 ("mm/vmscan: protect the workingset on anonymous LRU") 170b04b7ae ("mm/workingset: prepare the workingset detection infrastructure for anon LRU") b518154e59aab3ad 170b04b7ae49634df103810dad6 ---------------- --------------------------- %stddev %change %stddev \ | \ 0.05 ± 5% -26.6% 0.03 ± 3% vm-scalability.free_time 417392 +12.3% 468921 vm-scalability.median 43811736 +11.2% 48709267 vm-scalability.throughput 243.55 -3.7% 234.62 vm-scalability.time.elapsed_time 243.55 -3.7% 234.62 vm-scalability.time.elapsed_time.max 57267 ± 10% -29.6% 40324 vm-scalability.time.involuntary_context_switches 2344 -7.0% 2180 vm-scalability.time.percent_of_cpu_this_job_got 3981 -20.7% 3155 ± 4% vm-scalability.time.system_time 1730 +13.3% 1960 ± 6% vm-scalability.time.user_time 75.25 +2.3% 77.00 vmstat.cpu.id 16.12 -2.8 13.32 ± 4% mpstat.cpu.all.sys% 6.96 +1.2 8.15 ± 6% mpstat.cpu.all.usr% 5511015 ± 5% -12.2% 4837482 ± 8% numa-meminfo.node0.Mapped 11416 ± 2% -13.5% 9871 ± 12% numa-meminfo.node0.PageTables 17772 ± 4% -10.2% 15957 meminfo.Active 17617 ± 4% -10.4% 15790 meminfo.Active(anon) 10864238 -9.0% 9886317 meminfo.Mapped 1386943 ± 4% -14.6% 1185103 ± 12% numa-vmstat.node0.nr_mapped 2842 ± 2% -14.4% 2433 ± 10% numa-vmstat.node0.nr_page_table_pages 17056 ± 27% +346.2% 76106 ± 63% numa-vmstat.node0.numa_other 1424454 ± 6% +19661.4% 2.815e+08 ±142% cpuidle.C1.time 84757 ± 4% +4598.2% 3982093 ±150% cpuidle.C1.usage 130831 ± 2% +230.4% 432228 ± 94% cpuidle.POLL.time 37049 ± 2% +90.7% 70640 ± 34% cpuidle.POLL.usage 1415 ± 7% +19.9% 1696 ± 6% slabinfo.dmaengine-unmap-16.active_objs 1415 ± 7% +19.9% 1696 ± 6% slabinfo.dmaengine-unmap-16.num_objs 3302 ± 4% -15.0% 2808 ± 5% slabinfo.fsnotify_mark_connector.active_objs 3302 ± 4% -15.0% 2808 ± 5% slabinfo.fsnotify_mark_connector.num_objs 4403 ± 5% -10.3% 3949 proc-vmstat.nr_active_anon 12556993 -0.7% 12471053 proc-vmstat.nr_inactive_anon 2697260 -9.9% 2430793 ± 5% proc-vmstat.nr_mapped 5545 -7.6% 5122 ± 5% proc-vmstat.nr_page_table_pages 12490172 -0.7% 12404206 proc-vmstat.nr_shmem 234365 +3.1% 241704 proc-vmstat.nr_unevictable 4403 ± 5% -10.3% 3949 proc-vmstat.nr_zone_active_anon 12556992 -0.7% 12471053 proc-vmstat.nr_zone_inactive_anon 234365 +3.1% 241704 proc-vmstat.nr_zone_unevictable 178.00 ± 4% +2661.2% 4915 ±156% interrupts.41:PCI-MSI.67633156-edge.eth0-TxRx-3 2366 ± 7% -14.0% 2036 ± 3% interrupts.CPU0.NMI:Non-maskable_interrupts 2366 ± 7% -14.0% 2036 ± 3% interrupts.CPU0.PMI:Performance_monitoring_interrupts 1108 ± 41% -51.8% 534.50 ± 5% interrupts.CPU1.CAL:Function_call_interrupts 588.25 ± 16% -16.5% 491.25 ± 3% interrupts.CPU17.CAL:Function_call_interrupts 159.50 ± 51% -37.9% 99.00 ± 9% interrupts.CPU22.RES:Rescheduling_interrupts 118.25 ± 12% -20.7% 93.75 ± 17% interrupts.CPU25.RES:Rescheduling_interrupts 2260 ± 2% -22.1% 1760 ± 24% interrupts.CPU26.NMI:Non-maskable_interrupts 2260 ± 2% -22.1% 1760 ± 24% interrupts.CPU26.PMI:Performance_monitoring_interrupts 178.00 ± 4% +2661.2% 4915 ±156% interrupts.CPU33.41:PCI-MSI.67633156-edge.eth0-TxRx-3 191.75 ± 58% -63.0% 71.00 ± 10% interrupts.CPU36.RES:Rescheduling_interrupts 140.75 ± 30% -41.9% 81.75 ± 27% interrupts.CPU40.RES:Rescheduling_interrupts 224.25 ± 72% -62.1% 85.00 ± 32% interrupts.CPU48.RES:Rescheduling_interrupts 726.75 ± 19% -29.9% 509.50 ± 5% interrupts.CPU52.CAL:Function_call_interrupts 734.25 ± 22% -22.9% 566.00 ± 19% interrupts.CPU53.CAL:Function_call_interrupts 939.00 ± 31% -37.3% 589.00 ± 4% interrupts.CPU54.CAL:Function_call_interrupts 710.25 ± 17% -24.7% 534.50 ± 11% interrupts.CPU56.CAL:Function_call_interrupts 648.75 ± 17% -22.5% 502.50 ± 4% interrupts.CPU61.CAL:Function_call_interrupts 790.75 ± 30% -30.7% 547.75 ± 16% interrupts.CPU65.CAL:Function_call_interrupts 2389 ± 11% -20.4% 1902 interrupts.CPU76.NMI:Non-maskable_interrupts 2389 ± 11% -20.4% 1902 interrupts.CPU76.PMI:Performance_monitoring_interrupts 2959 ± 42% -35.8% 1900 interrupts.CPU83.NMI:Non-maskable_interrupts 2959 ± 42% -35.8% 1900 interrupts.CPU83.PMI:Performance_monitoring_interrupts 879.50 ±137% -87.4% 110.75 ± 36% interrupts.CPU84.RES:Rescheduling_interrupts 2120 ± 17% -41.0% 1252 ± 42% interrupts.CPU85.NMI:Non-maskable_interrupts 2120 ± 17% -41.0% 1252 ± 42% interrupts.CPU85.PMI:Performance_monitoring_interrupts 94.00 ± 14% +42.8% 134.25 ± 20% interrupts.CPU85.RES:Rescheduling_interrupts 172.75 ± 46% -36.9% 109.00 ± 38% interrupts.CPU89.RES:Rescheduling_interrupts 1658 ± 87% -69.2% 510.50 ± 9% interrupts.CPU9.CAL:Function_call_interrupts 23465 ± 19% -45.1% 12876 ± 4% interrupts.RES:Rescheduling_interrupts 1.377e+10 +3.2% 1.422e+10 perf-stat.i.branch-instructions 48655466 +5.7% 51418236 perf-stat.i.cache-misses 6.967e+10 -7.5% 6.442e+10 perf-stat.i.cpu-cycles 1313 -5.1% 1247 perf-stat.i.cycles-between-cache-misses 1.402e+10 +3.1% 1.446e+10 perf-stat.i.dTLB-loads 2255710 +3.8% 2341501 perf-stat.i.dTLB-store-misses 3.814e+09 +3.6% 3.95e+09 perf-stat.i.dTLB-stores 3316652 +62.5% 5387931 ± 3% perf-stat.i.iTLB-load-misses 4.976e+10 +3.1% 5.132e+10 perf-stat.i.instructions 0.67 -7.7% 0.62 perf-stat.i.metric.GHz 305.99 +3.1% 315.37 perf-stat.i.metric.M/sec 2118563 +3.9% 2200504 perf-stat.i.minor-faults 7903787 +4.1% 8229701 perf-stat.i.node-stores 2118563 +3.9% 2200504 perf-stat.i.page-faults 1.40 -10.4% 1.26 perf-stat.overall.cpi 1432 -12.5% 1252 perf-stat.overall.cycles-between-cache-misses 39.88 ± 2% +12.5 52.34 ± 3% perf-stat.overall.iTLB-load-miss-rate% 15005 -36.5% 9534 ± 3% perf-stat.overall.instructions-per-iTLB-miss 0.71 +11.6% 0.80 perf-stat.overall.ipc 1.377e+10 +3.0% 1.419e+10 perf-stat.ps.branch-instructions 48656706 +5.5% 51312384 perf-stat.ps.cache-misses 6.968e+10 -7.7% 6.429e+10 perf-stat.ps.cpu-cycles 1.402e+10 +2.9% 1.442e+10 perf-stat.ps.dTLB-loads 2255925 +3.6% 2336219 perf-stat.ps.dTLB-store-misses 3.811e+09 +3.4% 3.94e+09 perf-stat.ps.dTLB-stores 3316265 +62.1% 5375280 ± 3% perf-stat.ps.iTLB-load-misses 4.975e+10 +2.9% 5.12e+10 perf-stat.ps.instructions 2119230 +3.6% 2195912 perf-stat.ps.minor-faults 7906342 +3.9% 8212206 perf-stat.ps.node-stores 2119230 +3.6% 2195912 perf-stat.ps.page-faults 10352 ± 2% +11.6% 11557 ± 6% softirqs.CPU100.RCU 10309 ± 2% +13.9% 11740 ± 10% softirqs.CPU101.RCU 10693 +18.0% 12621 ± 10% softirqs.CPU15.RCU 10974 ± 2% +9.5% 12021 ± 3% softirqs.CPU17.RCU 10715 ± 4% +11.6% 11958 ± 5% softirqs.CPU19.RCU 11394 ± 2% +6.8% 12171 ± 5% softirqs.CPU2.RCU 10500 ± 5% +18.0% 12387 ± 8% softirqs.CPU21.RCU 10393 ± 4% +13.8% 11830 ± 3% softirqs.CPU23.RCU 10377 ± 9% +14.0% 11832 ± 5% softirqs.CPU25.RCU 10270 ± 6% +17.1% 12023 ± 11% softirqs.CPU29.RCU 11134 ± 3% +14.1% 12700 ± 10% softirqs.CPU32.RCU 10837 ± 3% +10.2% 11940 ± 3% softirqs.CPU37.RCU 25056 ± 3% -10.7% 22376 ± 10% softirqs.CPU4.SCHED 10758 ± 2% +13.6% 12226 ± 4% softirqs.CPU41.RCU 10704 ± 2% +14.5% 12257 ± 9% softirqs.CPU45.RCU 10480 ± 6% +8.1% 11327 ± 4% softirqs.CPU47.RCU 10427 ± 3% +8.9% 11359 ± 6% softirqs.CPU48.RCU 10105 ± 4% +25.4% 12673 ± 14% softirqs.CPU49.RCU 10258 ± 4% +9.3% 11210 ± 4% softirqs.CPU52.RCU 11962 ± 18% -16.5% 9990 ± 5% softirqs.CPU53.RCU 10278 +11.4% 11454 ± 3% softirqs.CPU58.RCU 10115 ± 3% +11.6% 11288 ± 4% softirqs.CPU59.RCU 10227 ± 3% +13.7% 11624 ± 5% softirqs.CPU60.RCU 10524 ± 5% +8.5% 11423 ± 4% softirqs.CPU62.RCU 10546 ± 3% +11.8% 11790 ± 4% softirqs.CPU64.RCU 10005 ± 3% +13.7% 11378 ± 4% softirqs.CPU65.RCU 10201 ± 2% +17.3% 11969 ± 4% softirqs.CPU66.RCU 10367 ± 2% +12.3% 11637 ± 2% softirqs.CPU67.RCU 10233 ± 4% +16.5% 11920 ± 3% softirqs.CPU68.RCU 10701 ± 4% +8.6% 11623 ± 3% softirqs.CPU71.RCU 10005 ± 4% +13.7% 11374 ± 2% softirqs.CPU72.RCU 9034 ± 3% +14.8% 10368 softirqs.CPU75.RCU 9290 ± 3% +11.8% 10386 ± 3% softirqs.CPU76.RCU 10645 ± 5% +14.3% 12167 ± 2% softirqs.CPU81.RCU 10657 ± 3% +11.5% 11884 ± 3% softirqs.CPU82.RCU 10582 +35.4% 14324 ± 7% softirqs.CPU83.RCU 10474 ± 4% +24.0% 12983 ± 19% softirqs.CPU86.RCU 10093 ± 2% +16.8% 11788 ± 11% softirqs.CPU92.RCU 10275 +14.4% 11757 ± 8% softirqs.CPU95.RCU 10499 ± 4% +14.8% 12057 ± 9% softirqs.CPU96.RCU 10228 +16.6% 11931 ± 10% softirqs.CPU97.RCU 1120806 ± 2% +9.8% 1230893 ± 3% softirqs.RCU 27146 ± 10% -43.0% 15461 ± 17% sched_debug.cfs_rq:/.exec_clock.avg 35739 ± 9% -33.4% 23815 ± 13% sched_debug.cfs_rq:/.exec_clock.max 24614 ± 11% -43.1% 14012 ± 17% sched_debug.cfs_rq:/.exec_clock.min 2252 ± 14% -28.3% 1613 ± 8% sched_debug.cfs_rq:/.exec_clock.stddev 26221 ± 25% -48.2% 13579 ± 29% sched_debug.cfs_rq:/.load.avg 31.57 ± 21% -38.4% 19.45 ± 14% sched_debug.cfs_rq:/.load_avg.avg 594.11 ± 16% -18.6% 483.79 ± 5% sched_debug.cfs_rq:/.load_avg.max 102.33 ± 16% -29.0% 72.67 ± 11% sched_debug.cfs_rq:/.load_avg.stddev 2729295 ± 10% -43.0% 1556498 ± 17% sched_debug.cfs_rq:/.min_vruntime.avg 2841927 ± 9% -43.0% 1620688 ± 17% sched_debug.cfs_rq:/.min_vruntime.max 2545899 ± 11% -42.6% 1461699 ± 17% sched_debug.cfs_rq:/.min_vruntime.min 75990 ± 36% -56.0% 33447 ± 33% sched_debug.cfs_rq:/.min_vruntime.stddev 0.34 ± 10% -30.0% 0.24 ± 16% sched_debug.cfs_rq:/.nr_running.avg 20.09 ± 12% -43.1% 11.43 ± 19% sched_debug.cfs_rq:/.nr_spread_over.avg 164.81 ± 16% -37.1% 103.71 ± 20% sched_debug.cfs_rq:/.nr_spread_over.max 33.80 ± 15% -39.4% 20.50 ± 13% sched_debug.cfs_rq:/.nr_spread_over.stddev 375.78 ± 9% -24.0% 285.69 ± 12% sched_debug.cfs_rq:/.runnable_avg.avg -183105 -63.3% -67248 sched_debug.cfs_rq:/.spread0.min 76034 ± 37% -56.0% 33448 ± 33% sched_debug.cfs_rq:/.spread0.stddev 369.43 ± 9% -23.5% 282.61 ± 11% sched_debug.cfs_rq:/.util_avg.avg 696.88 ± 4% -20.5% 553.75 ± 9% sched_debug.cfs_rq:/.util_est_enqueued.max 93.11 ± 4% -17.5% 76.85 ± 9% sched_debug.cfs_rq:/.util_est_enqueued.stddev 131720 ± 7% +21.2% 159588 ± 5% sched_debug.cpu.avg_idle.stddev 153356 ± 8% -29.2% 108595 ± 12% sched_debug.cpu.clock.avg 153364 ± 8% -29.2% 108602 ± 12% sched_debug.cpu.clock.max 153348 ± 8% -29.2% 108588 ± 12% sched_debug.cpu.clock.min 151791 ± 8% -29.1% 107548 ± 12% sched_debug.cpu.clock_task.avg 152331 ± 8% -29.2% 107864 ± 12% sched_debug.cpu.clock_task.max 146177 ± 8% -30.1% 102228 ± 12% sched_debug.cpu.clock_task.min 10821 ± 8% -28.9% 7692 ± 12% sched_debug.cpu.curr->pid.max 0.32 ± 12% -27.4% 0.23 ± 19% sched_debug.cpu.nr_running.avg 5319 ± 9% -29.2% 3764 ± 11% sched_debug.cpu.nr_switches.avg 2044 ± 8% -27.5% 1481 ± 12% sched_debug.cpu.nr_switches.min 30.00 ± 28% +60.0% 48.00 ± 15% sched_debug.cpu.nr_uninterruptible.max 6.75 ± 14% +21.5% 8.19 ± 5% sched_debug.cpu.nr_uninterruptible.stddev 3776 ± 12% -41.9% 2194 ± 20% sched_debug.cpu.sched_count.avg 1236 ± 9% -41.2% 727.58 ± 17% sched_debug.cpu.sched_count.min 1477 ± 10% -36.7% 935.58 ± 20% sched_debug.cpu.sched_goidle.avg 386.73 ± 9% -37.5% 241.60 ± 22% sched_debug.cpu.sched_goidle.min 1705 ± 12% -42.8% 975.64 ± 21% sched_debug.cpu.ttwu_count.avg 24239 ± 22% -31.5% 16612 ± 15% sched_debug.cpu.ttwu_count.max 522.40 ± 13% -41.1% 307.79 ± 21% sched_debug.cpu.ttwu_count.min 2783 ± 15% -33.1% 1862 ± 12% sched_debug.cpu.ttwu_count.stddev 682.47 ± 15% -48.1% 354.52 ± 18% sched_debug.cpu.ttwu_local.avg 5220 ± 16% -59.3% 2122 ± 18% sched_debug.cpu.ttwu_local.max 282.79 ± 13% -39.4% 171.29 ± 21% sched_debug.cpu.ttwu_local.min 709.90 ± 15% -57.2% 304.11 ± 13% sched_debug.cpu.ttwu_local.stddev 153349 ± 8% -29.2% 108590 ± 12% sched_debug.cpu_clk 152855 ± 8% -29.3% 108094 ± 12% sched_debug.ktime 154041 ± 8% -29.1% 109253 ± 12% sched_debug.sched_clk 25.29 ±119% -25.3 0.00 perf-profile.calltrace.cycles-pp.asm_exc_page_fault 24.30 ±119% -24.3 0.00 perf-profile.calltrace.cycles-pp.exc_page_fault.asm_exc_page_fault 24.14 ±119% -24.1 0.00 perf-profile.calltrace.cycles-pp.do_user_addr_fault.exc_page_fault.asm_exc_page_fault 23.72 ±119% -23.7 0.00 perf-profile.calltrace.cycles-pp.handle_mm_fault.do_user_addr_fault.exc_page_fault.asm_exc_page_fault 26.28 ± 10% -14.8 11.48 ± 3% perf-profile.calltrace.cycles-pp.lru_cache_add.shmem_getpage_gfp.shmem_fault.__do_fault.do_fault 26.12 ± 10% -14.8 11.32 ± 3% perf-profile.calltrace.cycles-pp.pagevec_lru_move_fn.lru_cache_add.shmem_getpage_gfp.shmem_fault.__do_fault 23.96 ± 10% -14.5 9.45 ± 4% perf-profile.calltrace.cycles-pp.native_queued_spin_lock_slowpath._raw_spin_lock_irqsave.pagevec_lru_move_fn.lru_cache_add.shmem_getpage_gfp 24.02 ± 10% -14.5 9.51 ± 4% perf-profile.calltrace.cycles-pp._raw_spin_lock_irqsave.pagevec_lru_move_fn.lru_cache_add.shmem_getpage_gfp.shmem_fault 56.91 ± 12% -11.9 45.00 perf-profile.calltrace.cycles-pp.__handle_mm_fault.handle_mm_fault.do_user_addr_fault.exc_page_fault.asm_exc_page_fault 56.41 ± 12% -11.9 44.51 perf-profile.calltrace.cycles-pp.do_fault.__handle_mm_fault.handle_mm_fault.do_user_addr_fault.exc_page_fault 48.60 ± 12% -11.6 37.03 perf-profile.calltrace.cycles-pp.shmem_getpage_gfp.shmem_fault.__do_fault.do_fault.__handle_mm_fault 48.89 ± 12% -11.6 37.32 perf-profile.calltrace.cycles-pp.shmem_fault.__do_fault.do_fault.__handle_mm_fault.handle_mm_fault 48.96 ± 12% -11.6 37.39 perf-profile.calltrace.cycles-pp.__do_fault.do_fault.__handle_mm_fault.handle_mm_fault.do_user_addr_fault 2.75 ± 24% -2.5 0.30 ±101% perf-profile.calltrace.cycles-pp.entry_SYSCALL_64_after_hwframe.unlinkat 2.75 ± 24% -2.5 0.30 ±101% perf-profile.calltrace.cycles-pp.do_syscall_64.entry_SYSCALL_64_after_hwframe.unlinkat 2.75 ± 24% -2.5 0.30 ±101% perf-profile.calltrace.cycles-pp.do_unlinkat.do_syscall_64.entry_SYSCALL_64_after_hwframe.unlinkat 2.75 ± 24% -2.5 0.30 ±101% perf-profile.calltrace.cycles-pp.evict.do_unlinkat.do_syscall_64.entry_SYSCALL_64_after_hwframe.unlinkat 2.75 ± 24% -2.5 0.30 ±101% perf-profile.calltrace.cycles-pp.unlinkat 1.82 ± 18% -0.6 1.26 ± 2% perf-profile.calltrace.cycles-pp.__do_munmap.__vm_munmap.__x64_sys_munmap.do_syscall_64.entry_SYSCALL_64_after_hwframe 1.82 ± 18% -0.6 1.26 ± 2% perf-profile.calltrace.cycles-pp.unmap_region.__do_munmap.__vm_munmap.__x64_sys_munmap.do_syscall_64 1.80 ± 18% -0.6 1.24 ± 2% perf-profile.calltrace.cycles-pp.zap_pte_range.unmap_page_range.unmap_vmas.unmap_region.__do_munmap 1.81 ± 18% -0.6 1.25 ± 2% perf-profile.calltrace.cycles-pp.unmap_vmas.unmap_region.__do_munmap.__vm_munmap.__x64_sys_munmap 1.81 ± 18% -0.6 1.25 ± 2% perf-profile.calltrace.cycles-pp.unmap_page_range.unmap_vmas.unmap_region.__do_munmap.__vm_munmap 2.20 ± 11% -0.5 1.73 perf-profile.calltrace.cycles-pp.finish_fault.do_fault.__handle_mm_fault.handle_mm_fault.do_user_addr_fault 2.15 ± 11% -0.5 1.67 perf-profile.calltrace.cycles-pp.alloc_set_pte.finish_fault.do_fault.__handle_mm_fault.handle_mm_fault 1.81 ± 11% -0.5 1.33 perf-profile.calltrace.cycles-pp.page_add_file_rmap.alloc_set_pte.finish_fault.do_fault.__handle_mm_fault 1.35 ± 16% -0.3 1.02 ± 10% perf-profile.calltrace.cycles-pp.ret_from_fork 1.35 ± 16% -0.3 1.02 ± 10% perf-profile.calltrace.cycles-pp.kthread.ret_from_fork 1.33 ± 16% -0.3 1.00 ± 10% perf-profile.calltrace.cycles-pp.worker_thread.kthread.ret_from_fork 1.32 ± 16% -0.3 0.99 ± 10% perf-profile.calltrace.cycles-pp.drm_fb_helper_dirty_work.process_one_work.worker_thread.kthread.ret_from_fork 1.33 ± 17% -0.3 1.00 ± 10% perf-profile.calltrace.cycles-pp.process_one_work.worker_thread.kthread.ret_from_fork 1.28 ± 16% -0.3 0.96 ± 10% perf-profile.calltrace.cycles-pp.memcpy_erms.drm_fb_helper_dirty_work.process_one_work.worker_thread.kthread 1.68 ± 7% -0.2 1.48 perf-profile.calltrace.cycles-pp.__pagevec_lru_add_fn.pagevec_lru_move_fn.lru_cache_add.shmem_getpage_gfp.shmem_fault 0.62 ± 60% +0.4 1.05 ± 3% perf-profile.calltrace.cycles-pp.__irqentry_text_end.do_access 1.23 ± 61% +1.2 2.40 perf-profile.calltrace.cycles-pp.irqentry_exit_to_user_mode.asm_exc_page_fault.do_access 2.24 ± 21% +1.2 3.48 ± 5% perf-profile.calltrace.cycles-pp.get_mem_cgroup_from_mm.mem_cgroup_charge.shmem_add_to_page_cache.shmem_getpage_gfp.shmem_fault 0.87 ±114% +1.6 2.47 ± 3% perf-profile.calltrace.cycles-pp.do_syscall_64.entry_SYSCALL_64_after_hwframe 0.88 ±114% +1.6 2.47 ± 3% perf-profile.calltrace.cycles-pp.entry_SYSCALL_64_after_hwframe 9.19 ± 17% +2.4 11.55 ± 3% perf-profile.calltrace.cycles-pp.shmem_add_to_page_cache.shmem_getpage_gfp.shmem_fault.__do_fault.do_fault 0.00 +2.4 2.38 ± 3% perf-profile.calltrace.cycles-pp.do_unlinkat.do_syscall_64.entry_SYSCALL_64_after_hwframe 0.00 +2.4 2.38 ± 3% perf-profile.calltrace.cycles-pp.evict.do_unlinkat.do_syscall_64.entry_SYSCALL_64_after_hwframe 5.64 ± 21% +2.8 8.47 ± 4% perf-profile.calltrace.cycles-pp.mem_cgroup_charge.shmem_add_to_page_cache.shmem_getpage_gfp.shmem_fault.__do_fault 13.37 ± 61% +9.1 22.42 perf-profile.calltrace.cycles-pp.do_rw_once 26.17 ± 10% -14.8 11.35 ± 3% perf-profile.children.cycles-pp.pagevec_lru_move_fn 26.28 ± 10% -14.8 11.48 ± 3% perf-profile.children.cycles-pp.lru_cache_add 24.20 ± 10% -14.5 9.67 ± 4% perf-profile.children.cycles-pp._raw_spin_lock_irqsave 24.12 ± 10% -14.5 9.62 ± 4% perf-profile.children.cycles-pp.native_queued_spin_lock_slowpath 56.45 ± 12% -11.9 44.54 perf-profile.children.cycles-pp.do_fault 56.95 ± 12% -11.9 45.05 perf-profile.children.cycles-pp.__handle_mm_fault 59.12 ± 12% -11.8 47.29 perf-profile.children.cycles-pp.do_user_addr_fault 58.05 ± 12% -11.8 46.22 perf-profile.children.cycles-pp.handle_mm_fault 59.51 ± 12% -11.8 47.73 perf-profile.children.cycles-pp.exc_page_fault 48.62 ± 12% -11.6 37.05 perf-profile.children.cycles-pp.shmem_getpage_gfp 48.89 ± 12% -11.6 37.32 perf-profile.children.cycles-pp.shmem_fault 48.96 ± 12% -11.6 37.40 perf-profile.children.cycles-pp.__do_fault 63.35 ± 10% -10.5 52.82 perf-profile.children.cycles-pp.asm_exc_page_fault 2.75 ± 24% -2.2 0.51 ± 19% perf-profile.children.cycles-pp.unlinkat 1.83 ± 18% -0.6 1.27 ± 2% perf-profile.children.cycles-pp.unmap_region 1.84 ± 18% -0.6 1.27 ± 2% perf-profile.children.cycles-pp.__do_munmap 1.83 ± 18% -0.6 1.26 ± 2% perf-profile.children.cycles-pp.__vm_munmap 1.82 ± 18% -0.6 1.26 ± 2% perf-profile.children.cycles-pp.__x64_sys_munmap 1.82 ± 18% -0.6 1.26 ± 2% perf-profile.children.cycles-pp.zap_pte_range 1.82 ± 18% -0.6 1.27 ± 2% perf-profile.children.cycles-pp.unmap_page_range 1.82 ± 18% -0.6 1.27 ± 2% perf-profile.children.cycles-pp.unmap_vmas 1.16 ± 17% -0.5 0.66 ± 4% perf-profile.children.cycles-pp.page_remove_rmap 2.21 ± 11% -0.5 1.73 perf-profile.children.cycles-pp.finish_fault 1.82 ± 11% -0.5 1.34 perf-profile.children.cycles-pp.page_add_file_rmap 2.56 ± 11% -0.5 2.09 perf-profile.children.cycles-pp.alloc_set_pte 1.35 ± 16% -0.3 1.02 ± 10% perf-profile.children.cycles-pp.ret_from_fork 1.35 ± 16% -0.3 1.02 ± 10% perf-profile.children.cycles-pp.kthread 1.33 ± 16% -0.3 1.00 ± 10% perf-profile.children.cycles-pp.worker_thread 1.32 ± 16% -0.3 0.99 ± 10% perf-profile.children.cycles-pp.drm_fb_helper_dirty_work 1.33 ± 17% -0.3 1.00 ± 10% perf-profile.children.cycles-pp.process_one_work 1.32 ± 16% -0.3 0.99 ± 10% perf-profile.children.cycles-pp.memcpy_erms 1.69 ± 7% -0.2 1.50 perf-profile.children.cycles-pp.__pagevec_lru_add_fn 0.08 ± 11% -0.0 0.06 ± 9% perf-profile.children.cycles-pp.mem_cgroup_page_lruvec 0.27 ± 12% +0.1 0.32 ± 6% perf-profile.children.cycles-pp.xas_create_range 0.00 +0.1 0.06 ± 7% perf-profile.children.cycles-pp.get_vma_policy 0.00 +0.2 0.23 ± 6% perf-profile.children.cycles-pp._raw_spin_unlock_irqrestore 2.25 ± 21% +1.2 3.49 ± 5% perf-profile.children.cycles-pp.get_mem_cgroup_from_mm 9.22 ± 17% +2.4 11.59 ± 3% perf-profile.children.cycles-pp.shmem_add_to_page_cache 5.66 ± 21% +2.8 8.49 ± 4% perf-profile.children.cycles-pp.mem_cgroup_charge 12.88 ± 61% +9.7 22.58 perf-profile.children.cycles-pp.do_rw_once 24.12 ± 10% -14.5 9.62 ± 4% perf-profile.self.cycles-pp.native_queued_spin_lock_slowpath 1.36 ± 15% -0.6 0.80 perf-profile.self.cycles-pp.shmem_add_to_page_cache 0.65 ± 12% -0.4 0.26 ± 9% perf-profile.self.cycles-pp.page_add_file_rmap 1.31 ± 16% -0.3 0.98 ± 10% perf-profile.self.cycles-pp.memcpy_erms 0.48 ± 17% -0.3 0.17 ± 2% perf-profile.self.cycles-pp.page_remove_rmap 1.06 ± 11% -0.3 0.79 perf-profile.self.cycles-pp.__mod_memcg_lruvec_state 0.80 ± 7% -0.2 0.57 perf-profile.self.cycles-pp.__pagevec_lru_add_fn 0.07 ± 12% -0.0 0.04 ± 57% perf-profile.self.cycles-pp.mem_cgroup_page_lruvec 0.08 ± 15% -0.0 0.06 ± 6% perf-profile.self.cycles-pp.truncate_cleanup_page 0.08 ± 15% +0.0 0.12 ± 5% perf-profile.self.cycles-pp.xas_create_range 0.15 ± 19% +0.0 0.19 ± 4% perf-profile.self.cycles-pp.xas_find_conflict 0.00 +0.1 0.14 ± 3% perf-profile.self.cycles-pp._raw_spin_unlock_irqrestore 2.23 ± 21% +1.2 3.46 ± 5% perf-profile.self.cycles-pp.get_mem_cgroup_from_mm 2.15 ± 25% +1.5 3.66 ± 4% perf-profile.self.cycles-pp.mem_cgroup_charge 6.88 ± 61% +4.0 10.87 perf-profile.self.cycles-pp.do_access 10.47 ± 61% +8.7 19.20 perf-profile.self.cycles-pp.do_rw_once vm-scalability.throughput 5e+07 +-----------------------------------------------------------------+ | O O O O | 4.9e+07 |-+ O O O O O O | 4.8e+07 |-O O O O O O O O O O O O O | | O O O O | 4.7e+07 |-+ O | 4.6e+07 |.+. .+.+ | | + + | 4.5e+07 |-+ + +. .+.. .+ | 4.4e+07 |-+ + .. +. .+ .+ : .+. .+.| | +.+ +. .+ + + : .+ + | 4.3e+07 |-+ +.+.+.+ + + + + + | 4.2e+07 |-+ : : + + + | | : : + | 4.1e+07 +-----------------------------------------------------------------+ [*] bisect-good sample [O] bisect-bad sample 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. Thanks, Rong Chen