Greeting, FYI, we noticed a 2.2% improvement of will-it-scale.per_process_ops due to commit: commit: 95f755a4ef7b9ccbedf6012b4112a990120b6a6c ("[PATCH v2 16/17] driver core: Refactor fw_devlink feature") url: https://github.com/0day-ci/linux/commits/Saravana-Kannan/Refactor-fw_devlink-to-significantly-improve-boot-time/20201121-100850 base: https://git.kernel.org/cgit/linux/kernel/git/rafael/linux-pm.git linux-next in testcase: will-it-scale on test machine: 192 threads Intel(R) Xeon(R) Platinum 9242 CPU @ 2.30GHz with 192G memory with following parameters: nr_task: 50% mode: process test: futex3 cpufreq_governor: performance ucode: 0x5003003 test-description: Will It Scale takes a testcase and runs it from 1 through to n parallel copies to see if the testcase will scale. It builds both a process and threads based test in order to see any differences between the two. test-url: https://github.com/antonblanchard/will-it-scale In addition to that, the commit also has significant impact on the following tests: +------------------+---------------------------------------------------------------------------+ | testcase: change | will-it-scale: will-it-scale.per_process_ops 5.2% improvement | | test machine | 192 threads Intel(R) Xeon(R) Platinum 9242 CPU @ 2.30GHz with 192G memory | | test parameters | cpufreq_governor=performance | | | mode=process | | | nr_task=50% | | | test=dup1 | | | ucode=0x5003003 | +------------------+---------------------------------------------------------------------------+ 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/mode/nr_task/rootfs/tbox_group/test/testcase/ucode: gcc-9/performance/x86_64-rhel-8.3/process/50%/debian-10.4-x86_64-20200603.cgz/lkp-csl-2ap2/futex3/will-it-scale/0x5003003 commit: 03b7843215 ("efi: Update implementation of add_links() to create fwnode links") 95f755a4ef ("driver core: Refactor fw_devlink feature") 03b7843215575338 95f755a4ef7b9ccbedf6012b411 ---------------- --------------------------- %stddev %change %stddev \ | \ 10181734 +2.2% 10402895 will-it-scale.per_process_ops 9.774e+08 +2.2% 9.987e+08 will-it-scale.workload 13832427 ± 6% -10.8% 12342804 ± 4% meminfo.DirectMap2M 2454 -5.0% 2332 ± 2% vmstat.system.cs 16233 ± 29% +172.3% 44205 ± 33% numa-vmstat.node0.nr_anon_pages 17076 ± 33% +162.3% 44793 ± 35% numa-vmstat.node0.nr_inactive_anon 17076 ± 33% +162.3% 44793 ± 35% numa-vmstat.node0.nr_zone_inactive_anon 23545 ± 69% +388.7% 115074 ± 36% numa-meminfo.node0.AnonHugePages 64841 ± 29% +172.6% 176742 ± 33% numa-meminfo.node0.AnonPages 68213 ± 33% +162.6% 179095 ± 35% numa-meminfo.node0.Inactive 68213 ± 33% +162.6% 179095 ± 35% numa-meminfo.node0.Inactive(anon) 4.38 ± 10% -2.5 1.87 ± 8% perf-profile.calltrace.cycles-pp.syscall_enter_from_user_mode.do_syscall_64.entry_SYSCALL_64_after_hwframe.syscall 4.61 ± 9% -2.7 1.88 ± 8% perf-profile.children.cycles-pp.syscall_enter_from_user_mode 0.46 ± 11% -0.2 0.24 ± 9% perf-profile.children.cycles-pp.__x86_indirect_thunk_rax 4.44 ± 10% -2.7 1.71 ± 8% perf-profile.self.cycles-pp.syscall_enter_from_user_mode 2.00 ± 9% +0.6 2.63 ± 8% perf-profile.self.cycles-pp.do_futex 2.52 ± 10% +0.8 3.37 ± 9% perf-profile.self.cycles-pp.__x64_sys_futex 70868 ± 13% +21.5% 86124 sched_debug.cfs_rq:/.exec_clock.avg 125411 ± 10% +17.9% 147892 sched_debug.cfs_rq:/.exec_clock.max 7286186 ± 12% +21.3% 8840402 sched_debug.cfs_rq:/.min_vruntime.avg 12873032 ± 10% +17.7% 15149207 sched_debug.cfs_rq:/.min_vruntime.max 0.80 ± 13% +29.9% 1.04 ± 9% sched_debug.cfs_rq:/.nr_spread_over.avg -828241 +84.1% -1524980 sched_debug.cfs_rq:/.spread0.min 7478 ± 5% +9.6% 8199 sched_debug.cpu.curr->pid.max 21364 ± 19% +41.9% 30320 ± 23% sched_debug.cpu.ttwu_count.max 1997 ± 14% +27.8% 2552 ± 15% sched_debug.cpu.ttwu_count.stddev 6.451e+10 +2.3% 6.597e+10 perf-stat.i.branch-instructions 4.942e+08 +2.3% 5.054e+08 perf-stat.i.branch-misses 13.97 ± 2% -1.0 13.00 ± 3% perf-stat.i.cache-miss-rate% 1429864 ± 3% -8.8% 1303388 ± 4% perf-stat.i.cache-misses 2410 -5.2% 2285 ± 3% perf-stat.i.context-switches 0.71 -2.6% 0.69 perf-stat.i.cpi 229727 ± 2% +9.8% 252139 ± 4% perf-stat.i.cycles-between-cache-misses 1.062e+11 +2.3% 1.087e+11 perf-stat.i.dTLB-loads 127892 +1.6% 129886 perf-stat.i.dTLB-store-misses 8.09e+10 +2.2% 8.272e+10 perf-stat.i.dTLB-stores 4259891 ± 2% -4.0% 4089092 perf-stat.i.iTLB-loads 4.208e+11 +2.3% 4.303e+11 perf-stat.i.instructions 1.42 +2.2% 1.45 perf-stat.i.ipc 0.65 +3.1% 0.67 perf-stat.i.major-faults 1.33 +3.6% 1.38 perf-stat.i.metric.K/sec 1310 +2.3% 1340 perf-stat.i.metric.M/sec 14.08 ± 2% -1.2 12.92 ± 3% perf-stat.overall.cache-miss-rate% 0.70 -2.1% 0.69 perf-stat.overall.cpi 205608 ± 3% +10.0% 226168 ± 4% perf-stat.overall.cycles-between-cache-misses 1.42 +2.2% 1.45 perf-stat.overall.ipc 6.429e+10 +2.3% 6.574e+10 perf-stat.ps.branch-instructions 4.925e+08 +2.3% 5.037e+08 perf-stat.ps.branch-misses 1439249 ± 3% -8.9% 1310558 ± 4% perf-stat.ps.cache-misses 2401 -5.3% 2274 ± 3% perf-stat.ps.context-switches 1.059e+11 +2.3% 1.083e+11 perf-stat.ps.dTLB-loads 127806 +1.7% 129943 perf-stat.ps.dTLB-store-misses 8.062e+10 +2.2% 8.243e+10 perf-stat.ps.dTLB-stores 4246454 ± 2% -4.1% 4074221 perf-stat.ps.iTLB-loads 4.193e+11 +2.3% 4.288e+11 perf-stat.ps.instructions 32095 ± 2% -6.6% 29971 ± 5% perf-stat.ps.node-loads 1.27e+14 +2.1% 1.297e+14 perf-stat.total.instructions 2251 ± 13% -25.6% 1674 ± 16% interrupts.CPU100.CAL:Function_call_interrupts 4823 ± 46% +34.3% 6475 ± 35% interrupts.CPU105.NMI:Non-maskable_interrupts 4823 ± 46% +34.3% 6475 ± 35% interrupts.CPU105.PMI:Performance_monitoring_interrupts 7063 ± 15% -47.6% 3704 ± 47% interrupts.CPU11.NMI:Non-maskable_interrupts 7063 ± 15% -47.6% 3704 ± 47% interrupts.CPU11.PMI:Performance_monitoring_interrupts 1743 ± 6% +29.0% 2249 ± 22% interrupts.CPU126.CAL:Function_call_interrupts 6249 ± 6% -41.3% 3670 ± 14% interrupts.CPU130.NMI:Non-maskable_interrupts 6249 ± 6% -41.3% 3670 ± 14% interrupts.CPU130.PMI:Performance_monitoring_interrupts 7215 ± 17% -42.5% 4148 ± 60% interrupts.CPU137.NMI:Non-maskable_interrupts 7215 ± 17% -42.5% 4148 ± 60% interrupts.CPU137.PMI:Performance_monitoring_interrupts 235.50 ± 19% -77.8% 52.25 ± 91% interrupts.CPU144.RES:Rescheduling_interrupts 5653 ± 37% -50.4% 2803 ± 33% interrupts.CPU146.NMI:Non-maskable_interrupts 5653 ± 37% -50.4% 2803 ± 33% interrupts.CPU146.PMI:Performance_monitoring_interrupts 6894 ± 16% -50.3% 3426 ± 22% interrupts.CPU147.NMI:Non-maskable_interrupts 6894 ± 16% -50.3% 3426 ± 22% interrupts.CPU147.PMI:Performance_monitoring_interrupts 6768 ± 20% -57.4% 2886 ± 36% interrupts.CPU149.NMI:Non-maskable_interrupts 6768 ± 20% -57.4% 2886 ± 36% interrupts.CPU149.PMI:Performance_monitoring_interrupts 6750 ± 19% -36.7% 4274 ± 24% interrupts.CPU15.NMI:Non-maskable_interrupts 6750 ± 19% -36.7% 4274 ± 24% interrupts.CPU15.PMI:Performance_monitoring_interrupts 74.75 ± 95% +181.6% 210.50 ± 32% interrupts.CPU157.RES:Rescheduling_interrupts 3955 ± 41% +114.3% 8474 ± 5% interrupts.CPU159.NMI:Non-maskable_interrupts 3955 ± 41% +114.3% 8474 ± 5% interrupts.CPU159.PMI:Performance_monitoring_interrupts 157.00 ± 50% -82.0% 28.25 ± 83% interrupts.CPU161.RES:Rescheduling_interrupts 6259 ± 11% -42.8% 3579 ± 34% interrupts.CPU183.NMI:Non-maskable_interrupts 6259 ± 11% -42.8% 3579 ± 34% interrupts.CPU183.PMI:Performance_monitoring_interrupts 354.50 ± 77% -87.2% 45.50 ±120% interrupts.CPU2.RES:Rescheduling_interrupts 5070 ± 35% +54.3% 7821 ± 14% interrupts.CPU26.NMI:Non-maskable_interrupts 5070 ± 35% +54.3% 7821 ± 14% interrupts.CPU26.PMI:Performance_monitoring_interrupts 158.75 ± 38% +74.3% 276.75 ± 15% interrupts.CPU35.RES:Rescheduling_interrupts 7163 ± 21% -49.6% 3608 ± 35% interrupts.CPU44.NMI:Non-maskable_interrupts 7163 ± 21% -49.6% 3608 ± 35% interrupts.CPU44.PMI:Performance_monitoring_interrupts 57.75 ± 74% +329.4% 248.00 ± 16% interrupts.CPU48.RES:Rescheduling_interrupts 5295 ± 16% +61.1% 8529 ± 3% interrupts.CPU5.NMI:Non-maskable_interrupts 5295 ± 16% +61.1% 8529 ± 3% interrupts.CPU5.PMI:Performance_monitoring_interrupts 7999 ± 16% -53.7% 3703 ± 47% interrupts.CPU60.NMI:Non-maskable_interrupts 7999 ± 16% -53.7% 3703 ± 47% interrupts.CPU60.PMI:Performance_monitoring_interrupts 6556 ± 23% -47.6% 3432 ± 10% interrupts.CPU61.NMI:Non-maskable_interrupts 6556 ± 23% -47.6% 3432 ± 10% interrupts.CPU61.PMI:Performance_monitoring_interrupts 238.00 ± 27% -63.0% 88.00 ± 80% interrupts.CPU61.RES:Rescheduling_interrupts 7585 ± 10% -49.1% 3863 ± 30% interrupts.CPU63.NMI:Non-maskable_interrupts 7585 ± 10% -49.1% 3863 ± 30% interrupts.CPU63.PMI:Performance_monitoring_interrupts 216.25 ± 21% -73.6% 57.00 ±133% interrupts.CPU63.RES:Rescheduling_interrupts 138.25 ± 56% +90.2% 263.00 ± 10% interrupts.CPU65.RES:Rescheduling_interrupts 241.25 ± 31% -56.7% 104.50 ± 96% interrupts.CPU66.RES:Rescheduling_interrupts 7553 ± 15% -47.0% 4004 ± 40% interrupts.CPU76.NMI:Non-maskable_interrupts 7553 ± 15% -47.0% 4004 ± 40% interrupts.CPU76.PMI:Performance_monitoring_interrupts 4934 ± 52% +71.1% 8442 ± 2% interrupts.CPU92.NMI:Non-maskable_interrupts 4934 ± 52% +71.1% 8442 ± 2% interrupts.CPU92.PMI:Performance_monitoring_interrupts 194.00 ± 30% +47.9% 287.00 interrupts.CPU92.RES:Rescheduling_interrupts 23725 ± 8% -27.8% 17131 ± 18% softirqs.CPU1.RCU 23137 ± 12% -31.2% 15917 ± 24% softirqs.CPU101.RCU 21338 ± 16% -28.2% 15329 ± 22% softirqs.CPU104.RCU 18202 ± 7% -17.2% 15064 ± 18% softirqs.CPU112.RCU 21059 ± 18% -26.8% 15417 ± 10% softirqs.CPU113.RCU 21006 ± 17% -30.8% 14529 ± 14% softirqs.CPU115.RCU 27815 ± 36% -33.3% 18546 ± 57% softirqs.CPU121.SCHED 20978 ± 19% -22.2% 16329 ± 27% softirqs.CPU124.RCU 19729 ± 12% -23.7% 15060 ± 14% softirqs.CPU125.RCU 22706 ± 15% -27.8% 16382 ± 12% softirqs.CPU126.RCU 22705 ± 12% -25.4% 16931 ± 27% softirqs.CPU127.RCU 20037 ± 19% -28.2% 14387 ± 12% softirqs.CPU128.RCU 19573 ± 12% -22.8% 15111 ± 4% softirqs.CPU129.RCU 20596 ± 10% -30.9% 14242 ± 14% softirqs.CPU131.RCU 23440 ± 33% +46.3% 34297 ± 9% softirqs.CPU131.SCHED 19106 ± 16% -23.1% 14700 ± 15% softirqs.CPU132.RCU 22821 ± 11% -26.4% 16795 ± 8% softirqs.CPU138.RCU 23505 ± 11% -32.3% 15917 ± 20% softirqs.CPU14.RCU 22836 ± 9% -16.7% 19023 ± 14% softirqs.CPU141.RCU 21767 ± 14% -30.6% 15105 ± 19% softirqs.CPU143.RCU 19267 ± 46% +67.2% 32223 ± 14% softirqs.CPU143.SCHED 24861 ± 4% -41.0% 14660 ± 11% softirqs.CPU144.RCU 11831 ± 47% +177.8% 32871 ± 15% softirqs.CPU144.SCHED 23297 ± 9% -27.0% 17004 ± 6% softirqs.CPU147.RCU 22685 ± 13% -31.4% 15567 ± 18% softirqs.CPU149.RCU 23209 ± 10% -25.8% 17228 ± 8% softirqs.CPU15.RCU 33035 ± 24% -54.6% 15002 ± 56% softirqs.CPU157.SCHED 28385 ± 25% -63.2% 10448 ± 79% softirqs.CPU159.SCHED 20067 ± 7% -12.1% 17638 ± 9% softirqs.CPU160.RCU 22261 ± 14% -35.7% 14306 ± 6% softirqs.CPU161.RCU 20998 ± 43% +67.5% 35179 ± 11% softirqs.CPU161.SCHED 32659 ± 28% -51.0% 16004 ± 66% softirqs.CPU162.SCHED 23692 ± 12% -19.1% 19177 ± 7% softirqs.CPU166.RCU 23350 ± 7% -26.6% 17139 ± 22% softirqs.CPU170.RCU 21357 ± 8% -35.5% 13779 ± 10% softirqs.CPU174.RCU 21886 ± 5% -33.4% 14582 ± 11% softirqs.CPU175.RCU 26609 ± 3% -27.8% 19208 ± 24% softirqs.CPU18.RCU 19898 ± 12% -27.4% 14443 ± 9% softirqs.CPU180.RCU 23109 ± 9% -32.2% 15658 ± 23% softirqs.CPU186.RCU 19719 ± 18% -36.6% 12500 ± 11% softirqs.CPU188.RCU 21919 ± 16% -33.2% 14647 ± 17% softirqs.CPU2.RCU 25187 ± 11% -23.3% 19313 ± 26% softirqs.CPU20.RCU 23983 ± 15% -20.6% 19041 ± 17% softirqs.CPU21.RCU 24148 ± 14% -17.8% 19849 ± 21% softirqs.CPU22.RCU 24732 ± 9% -25.8% 18348 ± 24% softirqs.CPU24.RCU 23282 ± 14% -27.6% 16851 ± 15% softirqs.CPU25.RCU 20694 ± 34% -54.3% 9460 ± 29% softirqs.CPU35.SCHED 25058 ± 13% -27.9% 18059 ± 18% softirqs.CPU4.RCU 20413 ± 18% -28.2% 14659 ± 8% softirqs.CPU43.RCU 23803 ± 17% -32.8% 15994 ± 19% softirqs.CPU44.RCU 24249 ± 11% -22.1% 18886 ± 22% softirqs.CPU46.RCU 32235 ± 14% -65.5% 11106 ± 51% softirqs.CPU48.SCHED 28803 ± 16% -39.1% 17536 ± 30% softirqs.CPU60.RCU 26565 ± 10% -34.4% 17436 ± 22% softirqs.CPU61.RCU 11142 ± 68% +156.3% 28562 ± 24% softirqs.CPU61.SCHED 25175 ± 12% -37.0% 15853 ± 9% softirqs.CPU63.RCU 14961 ± 56% +120.3% 32955 ± 26% softirqs.CPU63.SCHED 23043 ± 39% -62.7% 8602 ± 36% softirqs.CPU65.SCHED 26792 ± 12% -35.8% 17210 ± 14% softirqs.CPU66.RCU 19766 ± 18% -29.6% 13906 ± 9% softirqs.CPU7.RCU 22073 ± 13% -24.4% 16678 ± 23% softirqs.CPU70.RCU 24267 ± 4% -25.3% 18121 ± 19% softirqs.CPU75.RCU 27057 ± 23% -56.9% 11669 ± 75% softirqs.CPU78.SCHED 23392 ± 9% -32.1% 15885 ± 3% softirqs.CPU81.RCU 21645 ± 11% -26.5% 15910 ± 11% softirqs.CPU83.RCU 23556 ± 8% -30.1% 16475 ± 8% softirqs.CPU85.RCU 10105 ± 53% +102.5% 20466 ± 28% softirqs.CPU85.SCHED 24799 ± 15% -27.3% 18022 ± 9% softirqs.CPU88.RCU 19697 ± 58% -71.0% 5715 ± 7% softirqs.CPU92.SCHED 22778 ± 6% -24.3% 17235 ± 15% softirqs.CPU93.RCU 18189 ± 13% -24.4% 13760 ± 10% softirqs.CPU95.RCU 26156 ± 2% -21.7% 20485 ± 13% softirqs.CPU96.RCU 21904 ± 45% -56.8% 9452 ± 57% softirqs.CPU98.SCHED 22162 ± 11% -27.8% 15992 ± 9% softirqs.CPU99.RCU 38457 ± 10% -32.2% 26084 ± 35% softirqs.NET_RX 4109730 -18.0% 3370902 ± 12% softirqs.RCU 47247 ± 4% -20.1% 37765 ± 15% softirqs.TIMER will-it-scale.per_process_ops 1.045e+07 +---------------------------------------------------------------+ | O O O | 1.04e+07 |-+ O O O O O O O O O O O O O O O | | O O O O O O O O O O O O | 1.035e+07 |-+ | | | 1.03e+07 |-+ | | | 1.025e+07 |-+ | | | 1.02e+07 |-+ +.+..+.+ .+. | | + + .+.+..+.+.+.+ +.+.| 1.015e+07 |.+.+.+.+.+ +. .+.+.+.+.+. .+.+.+.+ | | + + + | 1.01e+07 +---------------------------------------------------------------+ will-it-scale.workload 1.005e+09 +---------------------------------------------------------------+ | | 1e+09 |-O O O O O O O O O O | | O O O O O O O O O O O O O O O O | 9.95e+08 |-+ O O O O | | | 9.9e+08 |-+ | | | 9.85e+08 |-+ | | | 9.8e+08 |-+ +. .+.+ | | : +. + +.+..+.+.+.+.+.+.+.| 9.75e+08 |.+.+.+.+. : +. .+.+.+. .+. .+.+.+. + | | + + + + + + | 9.7e+08 +---------------------------------------------------------------+ [*] bisect-good sample [O] bisect-bad sample *************************************************************************************************** lkp-csl-2ap2: 192 threads Intel(R) Xeon(R) Platinum 9242 CPU @ 2.30GHz with 192G memory ========================================================================================= compiler/cpufreq_governor/kconfig/mode/nr_task/rootfs/tbox_group/test/testcase/ucode: gcc-9/performance/x86_64-rhel-8.3/process/50%/debian-10.4-x86_64-20200603.cgz/lkp-csl-2ap2/dup1/will-it-scale/0x5003003 commit: 03b7843215 ("efi: Update implementation of add_links() to create fwnode links") 95f755a4ef ("driver core: Refactor fw_devlink feature") 03b7843215575338 95f755a4ef7b9ccbedf6012b411 ---------------- --------------------------- %stddev %change %stddev \ | \ 4681240 +5.2% 4925857 will-it-scale.per_process_ops 4.494e+08 +5.2% 4.729e+08 will-it-scale.workload 84205 ± 3% -6.6% 78620 ± 2% cpuidle.POLL.time 12.74 +1.4 14.10 mpstat.cpu.all.usr% 38.69 ± 9% +593.5% 268.33 ±101% sched_debug.cfs_rq:/.exec_clock.min 12.00 +8.3% 13.00 vmstat.cpu.us 2342 -1.8% 2299 vmstat.system.cs 1416 ± 4% -9.2% 1287 ± 4% slabinfo.dmaengine-unmap-16.active_objs 1416 ± 4% -9.2% 1287 ± 4% slabinfo.dmaengine-unmap-16.num_objs 6896 ± 4% -10.3% 6185 ± 4% slabinfo.kmalloc-rcl-64.active_objs 6896 ± 4% -10.3% 6185 ± 4% slabinfo.kmalloc-rcl-64.num_objs 2118 ± 2% -13.9% 1824 ± 6% slabinfo.kmalloc-rcl-96.active_objs 2118 ± 2% -13.9% 1824 ± 6% slabinfo.kmalloc-rcl-96.num_objs 13476 ± 4% -12.9% 11741 ± 3% slabinfo.pde_opener.active_objs 13476 ± 4% -12.9% 11741 ± 3% slabinfo.pde_opener.num_objs 2637 ± 63% -81.3% 492.50 ±142% numa-meminfo.node0.Active 2637 ± 63% -81.3% 492.50 ±142% numa-meminfo.node0.Active(anon) 2915 ± 16% -47.8% 1520 ± 19% numa-meminfo.node0.PageTables 87880 ± 10% -20.2% 70124 ± 9% numa-meminfo.node0.SUnreclaim 2153 ± 42% +215.9% 6802 ± 64% numa-meminfo.node2.Shmem 39060 ± 53% +306.7% 158860 ± 45% numa-meminfo.node3.AnonPages 46215 ± 44% +256.0% 164519 ± 42% numa-meminfo.node3.Inactive 46215 ± 44% +256.0% 164519 ± 42% numa-meminfo.node3.Inactive(anon) 750282 ± 10% +20.7% 905405 ± 13% numa-meminfo.node3.MemUsed 658.50 ± 63% -81.4% 122.75 ±142% numa-vmstat.node0.nr_active_anon 728.00 ± 15% -47.9% 379.50 ± 19% numa-vmstat.node0.nr_page_table_pages 21969 ± 10% -20.2% 17530 ± 9% numa-vmstat.node0.nr_slab_unreclaimable 658.50 ± 63% -81.4% 122.75 ±142% numa-vmstat.node0.nr_zone_active_anon 621492 ± 13% -26.7% 455253 ± 16% numa-vmstat.node0.numa_hit 572512 ± 17% -31.5% 391888 ± 15% numa-vmstat.node0.numa_local 538.00 ± 42% +216.0% 1700 ± 64% numa-vmstat.node2.nr_shmem 9760 ± 53% +307.2% 39747 ± 45% numa-vmstat.node3.nr_anon_pages 11513 ± 43% +257.5% 41166 ± 42% numa-vmstat.node3.nr_inactive_anon 11513 ± 43% +257.5% 41166 ± 42% numa-vmstat.node3.nr_zone_inactive_anon 548959 ± 19% +33.0% 730071 ± 31% numa-vmstat.node3.numa_hit 427995 ± 25% +47.7% 632173 ± 36% numa-vmstat.node3.numa_local 2.37 ± 8% -1.5 0.88 ± 8% perf-profile.calltrace.cycles-pp.syscall_enter_from_user_mode.do_syscall_64.entry_SYSCALL_64_after_hwframe.__close 2.33 ± 8% -1.5 0.88 ± 8% perf-profile.calltrace.cycles-pp.syscall_enter_from_user_mode.do_syscall_64.entry_SYSCALL_64_after_hwframe.dup 0.47 ± 57% +0.3 0.74 ± 7% perf-profile.calltrace.cycles-pp.locks_remove_posix.filp_close.__x64_sys_close.do_syscall_64.entry_SYSCALL_64_after_hwframe 1.27 ± 7% +0.7 1.96 ± 8% perf-profile.calltrace.cycles-pp.syscall_return_via_sysret.__close 4.94 ± 8% -3.2 1.77 ± 8% perf-profile.children.cycles-pp.syscall_enter_from_user_mode 0.44 ± 8% -0.2 0.23 ± 8% perf-profile.children.cycles-pp.__x86_indirect_thunk_rax 0.12 ± 18% -0.0 0.08 ± 13% perf-profile.children.cycles-pp.clockevents_program_event 0.11 ± 18% -0.0 0.07 ± 15% perf-profile.children.cycles-pp.ktime_get 0.27 ± 4% +0.1 0.33 ± 18% perf-profile.children.cycles-pp.start_kernel 4.76 ± 8% -3.1 1.62 ± 8% perf-profile.self.cycles-pp.syscall_enter_from_user_mode 1.15 ± 7% -0.5 0.67 ± 8% perf-profile.self.cycles-pp.__x64_sys_close 1.44 ± 9% -0.4 1.05 ± 9% perf-profile.self.cycles-pp.__x64_sys_dup 0.11 ± 21% -0.0 0.07 ± 13% perf-profile.self.cycles-pp.ktime_get 0.09 ± 8% +0.0 0.11 ± 4% perf-profile.self.cycles-pp.rcu_read_unlock_strict 9492 ± 72% -65.5% 3272 ±142% softirqs.CPU10.NET_RX 11789 ± 6% +25.7% 14823 ± 11% softirqs.CPU100.RCU 13884 ± 12% +26.5% 17565 ± 9% softirqs.CPU124.RCU 31918 ± 27% -65.9% 10876 ± 41% softirqs.CPU124.SCHED 15884 ± 11% +22.0% 19382 ± 11% softirqs.CPU13.RCU 30930 ± 28% -48.0% 16070 ± 72% softirqs.CPU132.SCHED 17573 ± 6% +77.1% 31116 ± 23% softirqs.CPU170.SCHED 16544 ± 4% -11.4% 14650 ± 10% softirqs.CPU185.RCU 17192 ± 8% -18.7% 13986 ± 16% softirqs.CPU186.RCU 10342 ± 45% +148.7% 25727 ± 27% softirqs.CPU186.SCHED 18559 ± 11% -22.9% 14300 ± 8% softirqs.CPU2.RCU 11714 ± 45% +150.5% 29341 ± 30% softirqs.CPU2.SCHED 21025 ± 46% -47.9% 10949 ± 77% softirqs.CPU21.SCHED 17639 ± 6% -20.2% 14081 ± 9% softirqs.CPU28.RCU 12427 ± 76% +146.2% 30597 ± 22% softirqs.CPU28.SCHED 5236 ± 6% +242.8% 17948 ± 41% softirqs.CPU4.SCHED 12934 ± 56% +139.0% 30910 ± 17% softirqs.CPU57.SCHED 14047 ± 6% +24.2% 17447 ± 8% softirqs.CPU74.RCU 27112 ± 3% -57.2% 11605 ± 70% softirqs.CPU74.SCHED 14231 ± 13% +15.8% 16478 ± 7% softirqs.CPU79.RCU 12191 ± 4% +28.2% 15634 ± 6% softirqs.CPU90.RCU 33498 ± 11% -47.3% 17657 ± 47% softirqs.CPU90.SCHED 13451 ± 5% +27.0% 17079 ± 13% softirqs.CPU98.RCU 33555 ± 15% -59.5% 13587 ± 56% softirqs.CPU98.SCHED 6.427e+10 +5.1% 6.752e+10 perf-stat.i.branch-instructions 0.90 -0.0 0.88 perf-stat.i.branch-miss-rate% 5.739e+08 +3.4% 5.933e+08 perf-stat.i.branch-misses 2298 -1.8% 2257 perf-stat.i.context-switches 0.88 -5.2% 0.84 perf-stat.i.cpi 1.067e+11 +5.1% 1.121e+11 perf-stat.i.dTLB-loads 110648 +8.9% 120538 perf-stat.i.dTLB-store-misses 7.223e+10 +5.1% 7.591e+10 perf-stat.i.dTLB-stores 6.072e+08 +5.4% 6.401e+08 ± 4% perf-stat.i.iTLB-load-misses 3.356e+11 +5.1% 3.526e+11 perf-stat.i.instructions 1.13 +5.5% 1.19 perf-stat.i.ipc 1.24 ± 2% +7.5% 1.34 ± 3% perf-stat.i.metric.K/sec 1266 +5.1% 1331 perf-stat.i.metric.M/sec 0.03 -3.5% 0.03 perf-stat.overall.MPKI 0.89 -0.0 0.88 perf-stat.overall.branch-miss-rate% 0.88 -5.2% 0.84 perf-stat.overall.cpi 0.00 +0.0 0.00 perf-stat.overall.dTLB-store-miss-rate% 1.13 +5.5% 1.19 perf-stat.overall.ipc 6.405e+10 +5.1% 6.729e+10 perf-stat.ps.branch-instructions 5.719e+08 +3.4% 5.913e+08 perf-stat.ps.branch-misses 2287 -1.8% 2247 perf-stat.ps.context-switches 1.063e+11 +5.1% 1.117e+11 perf-stat.ps.dTLB-loads 110749 +8.7% 120431 perf-stat.ps.dTLB-store-misses 7.199e+10 +5.1% 7.565e+10 perf-stat.ps.dTLB-stores 6.049e+08 +5.4% 6.376e+08 ± 4% perf-stat.ps.iTLB-load-misses 3.345e+11 +5.1% 3.514e+11 perf-stat.ps.instructions 1.013e+14 +5.2% 1.065e+14 perf-stat.total.instructions 18582 ± 74% -66.5% 6229 ±146% interrupts.31:PCI-MSI.524289-edge.eth0-TxRx-0 18582 ± 74% -66.5% 6229 ±146% interrupts.CPU10.31:PCI-MSI.524289-edge.eth0-TxRx-0 2285 ± 40% -27.5% 1655 ± 5% interrupts.CPU103.CAL:Function_call_interrupts 150.00 ± 56% -67.2% 49.25 ±138% interrupts.CPU117.RES:Rescheduling_interrupts 2.00 ± 86% +8587.5% 173.75 ±168% interrupts.CPU123.TLB:TLB_shootdowns 73.00 ±101% +229.5% 240.50 ± 19% interrupts.CPU124.RES:Rescheduling_interrupts 3375 ± 23% +112.2% 7163 ± 20% interrupts.CPU129.NMI:Non-maskable_interrupts 3375 ± 23% +112.2% 7163 ± 20% interrupts.CPU129.PMI:Performance_monitoring_interrupts 2570 ± 19% +131.1% 5940 ± 34% interrupts.CPU132.NMI:Non-maskable_interrupts 2570 ± 19% +131.1% 5940 ± 34% interrupts.CPU132.PMI:Performance_monitoring_interrupts 382.50 ± 68% -78.6% 82.00 ±136% interrupts.CPU15.RES:Rescheduling_interrupts 8363 ± 5% -33.8% 5539 ± 25% interrupts.CPU161.NMI:Non-maskable_interrupts 8363 ± 5% -33.8% 5539 ± 25% interrupts.CPU161.PMI:Performance_monitoring_interrupts 253.00 ± 15% -54.0% 116.50 ± 57% interrupts.CPU186.RES:Rescheduling_interrupts 249.50 ± 16% -64.7% 88.00 ± 64% interrupts.CPU2.RES:Rescheduling_interrupts 4404 ± 25% +89.1% 8327 ± 6% interrupts.CPU21.NMI:Non-maskable_interrupts 4404 ± 25% +89.1% 8327 ± 6% interrupts.CPU21.PMI:Performance_monitoring_interrupts 3516 ± 52% -52.8% 1659 ± 5% interrupts.CPU30.CAL:Function_call_interrupts 8180 ± 9% -42.4% 4710 ± 59% interrupts.CPU33.NMI:Non-maskable_interrupts 8180 ± 9% -42.4% 4710 ± 59% interrupts.CPU33.PMI:Performance_monitoring_interrupts 8601 ± 2% -39.4% 5208 ± 39% interrupts.CPU36.NMI:Non-maskable_interrupts 8601 ± 2% -39.4% 5208 ± 39% interrupts.CPU36.PMI:Performance_monitoring_interrupts 222.75 ± 31% -55.4% 99.25 ±109% interrupts.CPU36.RES:Rescheduling_interrupts 291.75 -40.1% 174.75 ± 38% interrupts.CPU4.RES:Rescheduling_interrupts 7518 ± 17% -30.1% 5251 ± 42% interrupts.CPU66.NMI:Non-maskable_interrupts 7518 ± 17% -30.1% 5251 ± 42% interrupts.CPU66.PMI:Performance_monitoring_interrupts 8034 ± 12% -41.4% 4711 ± 26% interrupts.CPU75.NMI:Non-maskable_interrupts 8034 ± 12% -41.4% 4711 ± 26% interrupts.CPU75.PMI:Performance_monitoring_interrupts 52.50 ± 70% +263.8% 191.00 ± 37% interrupts.CPU90.RES:Rescheduling_interrupts 63.25 ± 65% +243.1% 217.00 ± 34% interrupts.CPU98.RES:Rescheduling_interrupts 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, Oliver Sang