Greeting, FYI, we noticed a -16% regression of vm-scalability.throughput due to commit: commit: 2aa6d036b716c9242222e054d4ef34905ad45fd3 ("mm: numa: avoid waiting on freed migrated pages") https://git.kernel.org/cgit/linux/kernel/git/stable/linux-stable-rc.git linux-4.9.y in testcase: vm-scalability on test machine: 88 threads Intel(R) Xeon(R) CPU E5-2699 v4 @ 2.20GHz with 128G memory with following parameters: runtime: 300s size: 512G test: anon-wx-rand-mt cpufreq_governor: performance 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/01org/lkp-tests.git cd lkp-tests bin/lkp install job.yaml # job file is attached in this email bin/lkp run job.yaml testcase/path_params/tbox_group/run: vm-scalability/300s-512G-anon-wx-rand-mt-performance/lkp-bdw-ep2 08cb8e5f83fd2d4f 2aa6d036b716c9242222e054d4 ---------------- -------------------------- %stddev change %stddev \ | \ 0.04 ± 12% 218% 0.12 ± 21% vm-scalability.median_stddev 0.04 ± 8% 206% 0.11 ± 20% vm-scalability.stddev 4561044 -16% 3832033 vm-scalability.throughput 51928 -17% 43276 vm-scalability.median 26584 54% 41052 vm-scalability.time.voluntary_context_switches 152761 28% 194891 vm-scalability.time.minor_page_faults 274 14% 313 vm-scalability.time.elapsed_time 274 14% 313 vm-scalability.time.elapsed_time.max 22749 9% 24858 vm-scalability.time.user_time 8295 -4% 7926 vm-scalability.time.percent_of_cpu_this_job_got 24.84 -34% 16.36 ± 3% vm-scalability.time.system_time 10072 ± 78% -8e+03 1637 ±113% latency_stats.sum.futex_wait_queue_me.futex_wait.do_futex.SyS_futex.entry_SYSCALL_64_fastpath 97849 96905 vmstat.system.in 22591267 ± 5% 76% 39830428 ± 5% perf-stat.iTLB-loads 0.00 60% 0.00 ± 6% perf-stat.dTLB-load-miss-rate% 82808556 45% 1.2e+08 ± 4% perf-stat.dTLB-load-misses 0.02 38% 0.02 ± 3% perf-stat.branch-miss-rate% 39251 35% 52915 perf-stat.cpu-migrations 4.301e+08 26% 5.405e+08 ± 3% perf-stat.branch-misses 738430 17% 866092 perf-stat.page-faults 738428 17% 866089 perf-stat.minor-faults 6.329e+13 9% 6.93e+13 perf-stat.cpu-cycles 88.72 87.58 perf-stat.cache-miss-rate% 53.02 -8% 49.04 perf-stat.node-store-miss-rate% 1.415e+11 -8% 1.3e+11 ± 3% perf-stat.cache-references 1.377e+12 -9% 1.256e+12 ± 3% perf-stat.dTLB-stores 3.416e+12 -9% 3.108e+12 ± 3% perf-stat.dTLB-loads 1.243e+13 -9% 1.131e+13 ± 3% perf-stat.instructions 2.863e+12 -9% 2.604e+12 ± 3% perf-stat.branch-instructions 1.255e+11 -9% 1.138e+11 ± 3% perf-stat.cache-misses 62.61 ± 4% -15% 53.06 ± 3% perf-stat.iTLB-load-miss-rate% 6.642e+10 -16% 5.573e+10 ± 4% perf-stat.node-store-misses 0.20 -17% 0.16 perf-stat.ipc 329433 ± 8% -23% 252456 ± 9% perf-stat.instructions-per-iTLB-miss vm-scalability.throughput 4.7e+06 ++----------------------------------------------------------------+ 4.6e+06 *+. *..*...*.. .*.. .*... | | *... .. *.. ..*.. *... .*..*...*. *. *..* 4.5e+06 ++ * *..*. .. *. | 4.4e+06 ++ * | 4.3e+06 ++ | 4.2e+06 ++ | | | 4.1e+06 ++ | 4e+06 ++ O | 3.9e+06 ++ O O O O O O | 3.8e+06 O+ O O | | O O O | 3.7e+06 ++ O O | 3.6e+06 ++----------------------------------------------------------------+ vm-scalability.median 54000 ++------------------------------------------------------------------+ *.. .*.. .*.. | 52000 ++ *... .*...*. *... *... ..*.. .*...*. . .* | *. *..*...*.. .. *..*. *. *. | 50000 ++ * | | | 48000 ++ | | | 46000 ++ | | O O O | 44000 ++ O O O O O | O O | 42000 ++ O O O | | O O | 40000 ++------------------------------------------------------------------+ [*] 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, Xiaolong