Greeting, FYI, we noticed a -5.7% regression of will-it-scale.per_thread_ops due to commit: commit: 054aa8d439b9185d4f5eb9a90282d1ce74772969 ("fget: check that the fd still exists after getting a ref to it") https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git master in testcase: will-it-scale on test machine: 48 threads 2 sockets Intel(R) Xeon(R) CPU E5-2697 v2 @ 2.70GHz with 112G memory with following parameters: nr_task: 50% mode: thread test: poll2 cpufreq_governor: performance ucode: 0x42e 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_thread_ops -6.0% regression | | test machine | 48 threads 2 sockets Intel(R) Xeon(R) CPU E5-2697 v2 @ 2.70GHz with 112G memory | | test parameters | cpufreq_governor=performance | | | mode=thread | | | nr_task=16 | | | test=poll2 | | | ucode=0x42e | +------------------+---------------------------------------------------------------------------------+ If you fix the issue, kindly add following tag Reported-by: kernel test robot Details are as below: --------------------------------------------------------------------------------------------------> To reproduce: git clone https://github.com/intel/lkp-tests.git cd lkp-tests sudo bin/lkp install job.yaml # job file is attached in this email bin/lkp split-job --compatible job.yaml # generate the yaml file for lkp run sudo bin/lkp run generated-yaml-file # if come across any failure that blocks the test, # please remove ~/.lkp and /lkp dir to run from a clean state. ========================================================================================= compiler/cpufreq_governor/kconfig/mode/nr_task/rootfs/tbox_group/test/testcase/ucode: gcc-9/performance/x86_64-rhel-8.3/thread/50%/debian-10.4-x86_64-20200603.cgz/lkp-ivb-2ep1/poll2/will-it-scale/0x42e commit: 5f58da2bef ("Merge tag 'drm-fixes-2021-12-03-1' of git://anongit.freedesktop.org/drm/drm") 054aa8d439 ("fget: check that the fd still exists after getting a ref to it") 5f58da2befa58edf 054aa8d439b9185d4f5eb9a9028 ---------------- --------------------------- %stddev %change %stddev \ | \ 6666720 -5.7% 6288280 will-it-scale.24.threads 277779 -5.7% 262011 will-it-scale.per_thread_ops 6666720 -5.7% 6288280 will-it-scale.workload 173.02 +1.0% 174.71 turbostat.CorWatt 27.16 ± 10% +4.3 31.51 ± 2% perf-profile.calltrace.cycles-pp.__fget_light.do_sys_poll.__x64_sys_poll.do_syscall_64.entry_SYSCALL_64_after_hwframe 22.91 ± 10% +4.4 27.34 ± 2% perf-profile.calltrace.cycles-pp.__fget_files.__fget_light.do_sys_poll.__x64_sys_poll.do_syscall_64 26.33 ± 10% +4.4 30.70 ± 2% perf-profile.children.cycles-pp.__fget_light 22.92 ± 10% +4.4 27.35 ± 2% perf-profile.children.cycles-pp.__fget_files 22.70 ± 10% +4.4 27.11 ± 2% perf-profile.self.cycles-pp.__fget_files 17234 ± 13% -33.0% 11544 ± 26% softirqs.CPU18.RCU 6001 ± 83% +418.0% 31085 ± 39% softirqs.CPU18.SCHED 14695 ± 4% -20.3% 11714 ± 12% softirqs.CPU27.RCU 15579 ± 16% -23.0% 12000 ± 10% softirqs.CPU35.RCU 10842 ± 9% +49.4% 16201 ± 14% softirqs.CPU42.RCU 37992 ± 13% -66.8% 12604 ±101% softirqs.CPU42.SCHED 8899 ± 35% +87.0% 16640 ± 41% softirqs.CPU47.SCHED 5611 ± 47% -53.2% 2624 ± 35% interrupts.CPU12.NMI:Non-maskable_interrupts 5611 ± 47% -53.2% 2624 ± 35% interrupts.CPU12.PMI:Performance_monitoring_interrupts 5372 ± 42% -57.6% 2277 ± 49% interrupts.CPU13.NMI:Non-maskable_interrupts 5372 ± 42% -57.6% 2277 ± 49% interrupts.CPU13.PMI:Performance_monitoring_interrupts 6058 ± 38% -52.6% 2871 ± 87% interrupts.CPU18.NMI:Non-maskable_interrupts 6058 ± 38% -52.6% 2871 ± 87% interrupts.CPU18.PMI:Performance_monitoring_interrupts 284.33 ± 14% -65.3% 98.67 ± 97% interrupts.CPU18.RES:Rescheduling_interrupts 1323 ± 6% -20.4% 1053 ± 26% interrupts.CPU22.CAL:Function_call_interrupts 2175 ± 47% +176.5% 6015 ± 38% interrupts.CPU42.NMI:Non-maskable_interrupts 2175 ± 47% +176.5% 6015 ± 38% interrupts.CPU42.PMI:Performance_monitoring_interrupts 102.33 ±136% +233.9% 341.67 ± 48% interrupts.CPU42.TLB:TLB_shootdowns 312.83 ± 34% -69.3% 96.00 ±110% interrupts.CPU43.TLB:TLB_shootdowns 5386 ± 36% -55.6% 2393 ± 27% interrupts.CPU45.NMI:Non-maskable_interrupts 5386 ± 36% -55.6% 2393 ± 27% interrupts.CPU45.PMI:Performance_monitoring_interrupts 0.20 -0.0 0.19 perf-stat.i.branch-miss-rate% 44316621 -5.5% 41893733 perf-stat.i.branch-misses 0.74 -12.9% 0.65 perf-stat.i.cpi 0.19 ± 3% -0.0 0.14 ± 8% perf-stat.i.dTLB-load-miss-rate% 2.296e+10 +25.4% 2.88e+10 perf-stat.i.dTLB-loads 0.10 -0.0 0.08 perf-stat.i.dTLB-store-miss-rate% 14444127 -5.8% 13613178 ± 2% perf-stat.i.dTLB-store-misses 1.458e+10 +21.9% 1.777e+10 perf-stat.i.dTLB-stores 7406987 -4.6% 7067379 perf-stat.i.iTLB-load-misses 9.764e+10 +14.6% 1.118e+11 perf-stat.i.instructions 13170 +20.0% 15808 perf-stat.i.instructions-per-iTLB-miss 1.35 +14.9% 1.55 perf-stat.i.ipc 1277 +15.0% 1469 perf-stat.i.metric.M/sec 0.19 -0.0 0.17 perf-stat.overall.branch-miss-rate% 0.74 -13.0% 0.64 perf-stat.overall.cpi 0.19 ± 3% -0.0 0.14 ± 8% perf-stat.overall.dTLB-load-miss-rate% 0.10 -0.0 0.08 perf-stat.overall.dTLB-store-miss-rate% 13182 +20.1% 15826 perf-stat.overall.instructions-per-iTLB-miss 1.35 +14.9% 1.55 perf-stat.overall.ipc 4411089 +21.5% 5360082 perf-stat.overall.path-length 44188606 -5.4% 41787951 perf-stat.ps.branch-misses 2.289e+10 +25.4% 2.871e+10 perf-stat.ps.dTLB-loads 14395868 -5.8% 13567459 ± 2% perf-stat.ps.dTLB-store-misses 1.453e+10 +21.9% 1.771e+10 perf-stat.ps.dTLB-stores 7382256 -4.6% 7043603 perf-stat.ps.iTLB-load-misses 9.731e+10 +14.6% 1.115e+11 perf-stat.ps.instructions 2.941e+13 +14.6% 3.371e+13 perf-stat.total.instructions *************************************************************************************************** lkp-ivb-2ep1: 48 threads 2 sockets Intel(R) Xeon(R) CPU E5-2697 v2 @ 2.70GHz with 112G memory ========================================================================================= compiler/cpufreq_governor/kconfig/mode/nr_task/rootfs/tbox_group/test/testcase/ucode: gcc-9/performance/x86_64-rhel-8.3/thread/16/debian-10.4-x86_64-20200603.cgz/lkp-ivb-2ep1/poll2/will-it-scale/0x42e commit: 5f58da2bef ("Merge tag 'drm-fixes-2021-12-03-1' of git://anongit.freedesktop.org/drm/drm") 054aa8d439 ("fget: check that the fd still exists after getting a ref to it") 5f58da2befa58edf 054aa8d439b9185d4f5eb9a9028 ---------------- --------------------------- %stddev %change %stddev \ | \ 4461267 -6.0% 4194062 will-it-scale.16.threads 278828 -6.0% 262128 will-it-scale.per_thread_ops 4461267 -6.0% 4194062 will-it-scale.workload 105.14 ± 56% +129.9% 241.71 ± 18% interrupts.CPU15.RES:Rescheduling_interrupts 162.86 ± 46% -75.6% 39.71 ± 69% interrupts.CPU39.RES:Rescheduling_interrupts 26439 ± 38% -69.0% 8190 ± 43% softirqs.CPU15.SCHED 18393 ± 51% +96.6% 36159 ± 10% softirqs.CPU39.SCHED 15.89 ± 45% -9.7 6.18 ± 95% turbostat.C1E% 49.22 ± 13% +8.3 57.51 ± 7% turbostat.C6% 0.23 -0.0 0.21 ± 2% perf-stat.i.branch-miss-rate% 34323793 -4.1% 32902170 perf-stat.i.branch-misses 0.76 -12.5% 0.66 perf-stat.i.cpi 1.545e+10 +25.2% 1.934e+10 perf-stat.i.dTLB-loads 9.781e+09 +21.6% 1.19e+10 perf-stat.i.dTLB-stores 5312601 -5.1% 5039573 perf-stat.i.iTLB-load-misses 6.538e+10 +14.5% 7.489e+10 perf-stat.i.instructions 12305 +20.6% 14845 perf-stat.i.instructions-per-iTLB-miss 1.33 +14.4% 1.52 perf-stat.i.ipc 857.08 +14.9% 984.90 perf-stat.i.metric.M/sec 0.22 -0.0 0.21 perf-stat.overall.branch-miss-rate% 0.75 -12.6% 0.66 perf-stat.overall.cpi 12310 +20.7% 14859 perf-stat.overall.instructions-per-iTLB-miss 1.33 +14.4% 1.52 perf-stat.overall.ipc 4411591 +21.9% 5379108 perf-stat.overall.path-length 34213688 -4.1% 32797737 perf-stat.ps.branch-misses 1.539e+10 +25.2% 1.927e+10 perf-stat.ps.dTLB-loads 9.749e+09 +21.6% 1.186e+10 perf-stat.ps.dTLB-stores 5294998 -5.1% 5022765 perf-stat.ps.iTLB-load-misses 6.516e+10 +14.5% 7.464e+10 perf-stat.ps.instructions 1.968e+13 +14.6% 2.256e+13 perf-stat.total.instructions 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. --- 0DAY/LKP+ Test Infrastructure Open Source Technology Center https://lists.01.org/hyperkitty/list/lkp@lists.01.org Intel Corporation Thanks, Oliver Sang