hi, Linus, we reported "[linus:master] [x86] adfcf4231b: blogbench.read_score -10.9% regression" on https://lore.kernel.org/lkml/202305041446.71d46724-yujie.liu@intel.com/ however, as you pointed out, the blogbench is a "*horrifically* bad benchmark for this case", Feng Tang also made a debug patch to confirm this. now we noticed 47ee3f1dd9 is fix patch for adfcf4231b, just as we found for adfcf4231b, this commit could also cause performance regression or improvement for different cases. actually only this case is a regression, all others are improvement (we normally pick regression as title to report). below are detail data. hope they could be for your information about the possible performance impact of this change. Hello, kernel test robot noticed a -21.7% regression of phoronix-test-suite.ior.2MB.DefaultTestDirectory.mb_s on: commit: 47ee3f1dd93bcbe031539b1ecdaafb44b661c772 ("x86: re-introduce support for ERMS copies for user space accesses") https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git master testcase: phoronix-test-suite test machine: 96 threads 2 sockets Intel(R) Xeon(R) Gold 6252 CPU @ 2.10GHz (Cascade Lake) with 512G memory parameters: test: ior-1.1.1 cpufreq_governor: performance 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 11.8% improvement | | test machine | 104 threads 2 sockets (Skylake) with 192G memory | | test parameters | cpufreq_governor=performance | | | mode=process | | | nr_task=16 | | | test=pread3 | +------------------+-----------------------------------------------------------------------------------------------------------+ | testcase: change | nepim: nepim.tcp.avg.kbps_out 9.5% improvement | | test machine | 8 threads 1 sockets Intel(R) Core(TM) i7-4770 CPU @ 3.40GHz (Haswell) with 16G memory | | test parameters | cluster=cs-localhost | | | cpufreq_governor=performance | | | nr_threads=40% | | | protocol=tcp6 | | | runtime=300s | +------------------+-----------------------------------------------------------------------------------------------------------+ | testcase: change | phoronix-test-suite: phoronix-test-suite.intel-mpi.IMB-MPI1Exchange.average_mbytes_sec 21.6% improvement | | test machine | 16 threads 1 sockets Intel(R) Xeon(R) E-2278G CPU @ 3.40GHz (Coffee Lake) with 32G memory | | test parameters | cpufreq_governor=performance | | | option_a=IMB-MPI1 Exchange | | | test=intel-mpi-1.0.1 | +------------------+-----------------------------------------------------------------------------------------------------------+ | testcase: change | will-it-scale: will-it-scale.per_process_ops 4.6% improvement | | test machine | 104 threads 2 sockets (Skylake) with 192G memory | | test parameters | cpufreq_governor=performance | | | mode=process | | | nr_task=50% | | | test=readseek1 | +------------------+-----------------------------------------------------------------------------------------------------------+ | testcase: change | phoronix-test-suite: phoronix-test-suite.tiobench.Write.64MB.8.mb_s 43.5% improvement | | test machine | 16 threads 1 sockets Intel(R) Xeon(R) E-2278G CPU @ 3.40GHz (Coffee Lake) with 32G memory | | test parameters | cpufreq_governor=performance | | | option_a=Write | | | option_b=64MB | | | option_c=8 | | | test=tiobench-1.3.1 | +------------------+-----------------------------------------------------------------------------------------------------------+ | testcase: change | phoronix-test-suite: phoronix-test-suite.tiobench.RandomWrite.32MB.32.mb_s 49.5% improvement | | test machine | 16 threads 1 sockets Intel(R) Xeon(R) E-2278G CPU @ 3.40GHz (Coffee Lake) with 32G memory | | test parameters | cpufreq_governor=performance | | | option_a=Random Write | | | option_b=32MB | | | option_c=4 | | | test=tiobench-1.3.1 | +------------------+-----------------------------------------------------------------------------------------------------------+ | testcase: change | lmbench3: lmbench3.AF_UNIX.sock.stream.bandwidth.MB/sec 9.2% improvement | | test machine | 48 threads 2 sockets Intel(R) Xeon(R) CPU E5-2697 v2 @ 2.70GHz (Ivy Bridge-EP) with 112G memory | | test parameters | cpufreq_governor=performance | | | mode=development | | | nr_threads=1 | | | test=UNIX | | | test_memory_size=50% | +------------------+-----------------------------------------------------------------------------------------------------------+ | testcase: change | phoronix-test-suite: phoronix-test-suite.tiobench.RandomWrite.64MB.32.mb_s 51.1% improvement | | test machine | 16 threads 1 sockets Intel(R) Xeon(R) E-2278G CPU @ 3.40GHz (Coffee Lake) with 32G memory | | test parameters | cpufreq_governor=performance | | | option_a=Random Write | | | option_b=64MB | | | option_c=4 | | | test=tiobench-1.3.1 | +------------------+-----------------------------------------------------------------------------------------------------------+ | testcase: change | phoronix-test-suite: phoronix-test-suite.tiobench.Write.64MB.32.mb_s 58.5% improvement | | test machine | 16 threads 1 sockets Intel(R) Xeon(R) E-2278G CPU @ 3.40GHz (Coffee Lake) with 32G memory | | test parameters | cpufreq_governor=performance | | | option_a=Write | | | option_b=64MB | | | option_c=4 | | | test=tiobench-1.3.1 | +------------------+-----------------------------------------------------------------------------------------------------------+ | testcase: change | nepim: nepim.tcp.avg.kbps_out 10.3% improvement | | test machine | 8 threads 1 sockets Intel(R) Core(TM) i7-4770K CPU @ 3.50GHz (Haswell) with 8G memory | | test parameters | cluster=cs-localhost | | | cpufreq_governor=performance | | | nr_threads=25% | | | protocol=tcp | | | runtime=300s | +------------------+-----------------------------------------------------------------------------------------------------------+ | testcase: change | phoronix-test-suite: phoronix-test-suite.tiobench.Write.32MB.32.mb_s 57.3% improvement | | test machine | 16 threads 1 sockets Intel(R) Xeon(R) E-2278G CPU @ 3.40GHz (Coffee Lake) with 32G memory | | test parameters | cpufreq_governor=performance | | | option_a=Write | | | option_b=32MB | | | option_c=4 | | | test=tiobench-1.3.1 | +------------------+-----------------------------------------------------------------------------------------------------------+ | testcase: change | phoronix-test-suite: phoronix-test-suite.tiobench.RandomWrite.64MB.8.mb_s 45.1% improvement | | test machine | 16 threads 1 sockets Intel(R) Xeon(R) E-2278G CPU @ 3.40GHz (Coffee Lake) with 32G memory | | test parameters | cpufreq_governor=performance | | | option_a=Random Write | | | option_b=64MB | | | option_c=8 | | | test=tiobench-1.3.1 | +------------------+-----------------------------------------------------------------------------------------------------------+ | testcase: change | phoronix-test-suite: phoronix-test-suite.x11perf.500pxPutImageSquare.operations___second 8.4% improvement | | test machine | 12 threads 1 sockets Intel(R) Core(TM) i7-8700 CPU @ 3.20GHz (Coffee Lake) with 32G memory | | test parameters | cpufreq_governor=performance | | | need_x=true | | | option_a=500px PutImage Square | | | test=x11perf-1.1.1 | +------------------+-----------------------------------------------------------------------------------------------------------+ | testcase: change | will-it-scale: will-it-scale.per_thread_ops 14.1% improvement | | test machine | 104 threads 2 sockets (Skylake) with 192G memory | | test parameters | cpufreq_governor=performance | | | mode=thread | | | nr_task=50% | | | test=pread1 | +------------------+-----------------------------------------------------------------------------------------------------------+ | testcase: change | will-it-scale: will-it-scale.per_thread_ops 7.8% improvement | | test machine | 104 threads 2 sockets (Skylake) with 192G memory | | test parameters | cpufreq_governor=performance | | | mode=thread | | | nr_task=100% | | | test=readseek1 | +------------------+-----------------------------------------------------------------------------------------------------------+ If you fix the issue, kindly add following tag | Reported-by: kernel test robot | Closes: https://lore.kernel.org/oe-lkp/202306041623.cdc379d-oliver.sang@intel.com 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/rootfs/tbox_group/test/testcase: gcc-11/performance/x86_64-rhel-8.3/debian-x86_64-phoronix/lkp-csl-2sp7/ior-1.1.1/phoronix-test-suite commit: 0d85b27b0c ("Merge tag '6.4-rc3-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6") 47ee3f1dd9 ("x86: re-introduce support for ERMS copies for user space accesses") 0d85b27b0cc6b5cf 47ee3f1dd93bcbe031539b1ecda ---------------- --------------------------- %stddev %change %stddev \ | \ 3626 -22.0% 2828 phoronix-test-suite.ior.2MB./opt/rootfs.mb_s 3586 -21.7% 2808 phoronix-test-suite.ior.2MB.DefaultTestDirectory.mb_s 4487378 -3.5% 4330977 perf-stat.i.cache-misses 76327545 -1.2% 75402170 perf-stat.i.cache-references 292483 ± 3% -16.0% 245543 ± 5% perf-stat.i.node-stores 5.88 -0.1 5.74 perf-stat.overall.cache-miss-rate% 1462 +3.5% 1513 perf-stat.overall.cycles-between-cache-misses 35.88 +4.9 40.77 ± 5% perf-stat.overall.node-store-miss-rate% 4387788 -3.6% 4229799 perf-stat.ps.cache-misses 285917 ± 4% -16.2% 239711 ± 5% perf-stat.ps.node-stores *************************************************************************************************** lkp-skl-fpga01: 104 threads 2 sockets (Skylake) with 192G memory ========================================================================================= compiler/cpufreq_governor/kconfig/mode/nr_task/rootfs/tbox_group/test/testcase: gcc-11/performance/x86_64-rhel-8.3/process/16/debian-11.1-x86_64-20220510.cgz/lkp-skl-fpga01/pread3/will-it-scale commit: 0d85b27b0c ("Merge tag '6.4-rc3-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6") 47ee3f1dd9 ("x86: re-introduce support for ERMS copies for user space accesses") 0d85b27b0cc6b5cf 47ee3f1dd93bcbe031539b1ecda ---------------- --------------------------- %stddev %change %stddev \ | \ 6.01 +0.7 6.71 ± 2% mpstat.cpu.all.usr% 0.22 -42.4% 0.13 ± 3% turbostat.IPC 13556369 +11.8% 15162679 ± 2% will-it-scale.16.processes 847272 +11.8% 947666 ± 2% will-it-scale.per_process_ops 13556369 +11.8% 15162679 ± 2% will-it-scale.workload 1.30 ± 2% +75.8% 2.28 ± 3% perf-stat.i.MPKI 4.457e+09 -10.9% 3.969e+09 ± 2% perf-stat.i.branch-instructions 1.15 +0.3 1.41 perf-stat.i.branch-miss-rate% 51357767 +9.3% 56148787 perf-stat.i.branch-misses 1718 -0.8% 1704 perf-stat.i.context-switches 1.27 +76.6% 2.25 ± 2% perf-stat.i.cpi 0.12 +0.1 0.27 perf-stat.i.dTLB-load-miss-rate% 14889205 +10.7% 16475643 ± 2% perf-stat.i.dTLB-load-misses 1.248e+10 -50.3% 6.202e+09 ± 2% perf-stat.i.dTLB-loads 0.00 +0.0 0.00 ± 2% perf-stat.i.dTLB-store-miss-rate% 1.009e+10 -65.0% 3.534e+09 ± 2% perf-stat.i.dTLB-stores 14680606 +14.5% 16815386 ± 5% perf-stat.i.iTLB-load-misses 33983782 ± 5% +14.8% 39004142 ± 4% perf-stat.i.iTLB-loads 3.708e+10 -43.3% 2.101e+10 ± 2% perf-stat.i.instructions 2528 -50.4% 1254 ± 3% perf-stat.i.instructions-per-iTLB-miss 0.78 -43.3% 0.44 ± 2% perf-stat.i.ipc 790.06 ± 3% +5.7% 835.02 ± 2% perf-stat.i.metric.K/sec 259.91 -49.3% 131.77 ± 2% perf-stat.i.metric.M/sec 1.30 ± 2% +75.3% 2.27 ± 3% perf-stat.overall.MPKI 1.15 +0.3 1.41 perf-stat.overall.branch-miss-rate% 1.27 +76.4% 2.25 ± 2% perf-stat.overall.cpi 0.12 +0.1 0.26 perf-stat.overall.dTLB-load-miss-rate% 0.00 +0.0 0.00 ± 2% perf-stat.overall.dTLB-store-miss-rate% 2525 -50.4% 1251 ± 3% perf-stat.overall.instructions-per-iTLB-miss 0.78 -43.3% 0.44 ± 2% perf-stat.overall.ipc 822348 -49.3% 416945 perf-stat.overall.path-length 4.442e+09 -10.9% 3.956e+09 ± 2% perf-stat.ps.branch-instructions 51175390 +9.3% 55945707 perf-stat.ps.branch-misses 14839395 +10.7% 16420381 ± 2% perf-stat.ps.dTLB-load-misses 1.244e+10 -50.3% 6.181e+09 ± 2% perf-stat.ps.dTLB-loads 1.006e+10 -65.0% 3.522e+09 ± 2% perf-stat.ps.dTLB-stores 14630964 +14.5% 16757456 ± 5% perf-stat.ps.iTLB-load-misses 33871661 ± 5% +14.8% 38879673 ± 4% perf-stat.ps.iTLB-loads 3.695e+10 -43.3% 2.094e+10 ± 2% perf-stat.ps.instructions 1.115e+13 -43.3% 6.322e+12 ± 2% perf-stat.total.instructions 11.76 ± 4% -7.2 4.59 ± 36% perf-profile.calltrace.cycles-pp.rep_movs_alternative.copyout._copy_to_iter.copy_page_to_iter.shmem_file_read_iter 12.22 ± 4% -7.1 5.13 ± 31% perf-profile.calltrace.cycles-pp.copyout._copy_to_iter.copy_page_to_iter.shmem_file_read_iter.vfs_read 12.65 ± 4% -7.0 5.61 ± 28% perf-profile.calltrace.cycles-pp._copy_to_iter.copy_page_to_iter.shmem_file_read_iter.vfs_read.__x64_sys_pread64 12.94 ± 4% -7.0 5.94 ± 26% perf-profile.calltrace.cycles-pp.copy_page_to_iter.shmem_file_read_iter.vfs_read.__x64_sys_pread64.do_syscall_64 20.53 ± 3% -6.3 14.28 ± 9% perf-profile.calltrace.cycles-pp.shmem_file_read_iter.vfs_read.__x64_sys_pread64.do_syscall_64.entry_SYSCALL_64_after_hwframe 25.48 ± 4% -5.7 19.80 ± 6% perf-profile.calltrace.cycles-pp.vfs_read.__x64_sys_pread64.do_syscall_64.entry_SYSCALL_64_after_hwframe.__libc_pread 27.64 ± 3% -5.5 22.16 ± 5% perf-profile.calltrace.cycles-pp.__x64_sys_pread64.do_syscall_64.entry_SYSCALL_64_after_hwframe.__libc_pread 41.36 ± 4% -4.0 37.37 ± 2% perf-profile.calltrace.cycles-pp.do_syscall_64.entry_SYSCALL_64_after_hwframe.__libc_pread 1.09 ± 3% +0.1 1.22 ± 4% perf-profile.calltrace.cycles-pp.__fget_light.__x64_sys_pread64.do_syscall_64.entry_SYSCALL_64_after_hwframe.__libc_pread 1.60 ± 2% +0.1 1.74 ± 3% perf-profile.calltrace.cycles-pp.filemap_get_entry.shmem_get_folio_gfp.shmem_file_read_iter.vfs_read.__x64_sys_pread64 0.95 ± 3% +0.2 1.11 ± 10% perf-profile.calltrace.cycles-pp.__fsnotify_parent.vfs_read.__x64_sys_pread64.do_syscall_64.entry_SYSCALL_64_after_hwframe 1.62 ± 4% +0.2 1.78 ± 3% perf-profile.calltrace.cycles-pp.touch_atime.shmem_file_read_iter.vfs_read.__x64_sys_pread64.do_syscall_64 1.97 ± 4% +0.2 2.18 ± 4% perf-profile.calltrace.cycles-pp.entry_SYSCALL_64_safe_stack.__libc_pread 0.35 ± 70% +0.2 0.58 ± 4% perf-profile.calltrace.cycles-pp.syscall_enter_from_user_mode.do_syscall_64.entry_SYSCALL_64_after_hwframe.__libc_pread 2.85 ± 2% +0.3 3.18 ± 3% perf-profile.calltrace.cycles-pp.shmem_get_folio_gfp.shmem_file_read_iter.vfs_read.__x64_sys_pread64.do_syscall_64 6.65 ± 4% +0.7 7.35 ± 4% perf-profile.calltrace.cycles-pp.__entry_text_start.__libc_pread 12.30 ± 5% +1.3 13.62 ± 4% perf-profile.calltrace.cycles-pp.syscall_exit_to_user_mode.do_syscall_64.entry_SYSCALL_64_after_hwframe.__libc_pread 13.67 ± 5% +1.5 15.13 ± 4% perf-profile.calltrace.cycles-pp.syscall_return_via_sysret.__libc_pread 11.77 ± 4% -7.1 4.63 ± 35% perf-profile.children.cycles-pp.rep_movs_alternative 12.47 ± 4% -7.0 5.42 ± 29% perf-profile.children.cycles-pp.copyout 12.67 ± 4% -7.0 5.64 ± 28% perf-profile.children.cycles-pp._copy_to_iter 12.96 ± 4% -7.0 5.96 ± 26% perf-profile.children.cycles-pp.copy_page_to_iter 20.61 ± 3% -6.2 14.38 ± 9% perf-profile.children.cycles-pp.shmem_file_read_iter 25.58 ± 4% -5.7 19.90 ± 5% perf-profile.children.cycles-pp.vfs_read 27.65 ± 3% -5.5 22.18 ± 5% perf-profile.children.cycles-pp.__x64_sys_pread64 41.52 ± 4% -4.0 37.53 ± 2% perf-profile.children.cycles-pp.do_syscall_64 0.54 ± 5% +0.1 0.61 ± 4% perf-profile.children.cycles-pp.syscall_enter_from_user_mode 1.01 ± 4% +0.1 1.12 ± 3% perf-profile.children.cycles-pp.entry_SYSRETQ_unsafe_stack 1.09 ± 4% +0.1 1.22 ± 4% perf-profile.children.cycles-pp.__fget_light 1.62 ± 2% +0.1 1.76 ± 3% perf-profile.children.cycles-pp.filemap_get_entry 0.95 ± 3% +0.2 1.12 ± 10% perf-profile.children.cycles-pp.__fsnotify_parent 1.64 ± 4% +0.2 1.82 ± 3% perf-profile.children.cycles-pp.touch_atime 2.89 ± 2% +0.3 3.23 ± 3% perf-profile.children.cycles-pp.shmem_get_folio_gfp 6.55 ± 4% +0.7 7.24 ± 4% perf-profile.children.cycles-pp.__entry_text_start 12.44 ± 5% +1.3 13.76 ± 4% perf-profile.children.cycles-pp.syscall_exit_to_user_mode 13.81 ± 5% +1.5 15.28 ± 4% perf-profile.children.cycles-pp.syscall_return_via_sysret 11.60 ± 4% -7.1 4.46 ± 37% perf-profile.self.cycles-pp.rep_movs_alternative 0.48 ± 5% +0.1 0.55 ± 4% perf-profile.self.cycles-pp.syscall_enter_from_user_mode 0.45 ± 7% +0.1 0.52 ± 2% perf-profile.self.cycles-pp.current_time 0.26 ± 5% +0.1 0.34 ± 11% perf-profile.self.cycles-pp.xas_load 0.87 ± 4% +0.1 0.97 ± 3% perf-profile.self.cycles-pp.entry_SYSRETQ_unsafe_stack 1.11 ± 4% +0.1 1.21 ± 3% perf-profile.self.cycles-pp.__libc_pread 0.82 ± 4% +0.1 0.92 ± 5% perf-profile.self.cycles-pp.copyout 1.09 ± 3% +0.1 1.21 ± 4% perf-profile.self.cycles-pp.__fget_light 1.13 ± 3% +0.2 1.29 ± 4% perf-profile.self.cycles-pp.shmem_get_folio_gfp 0.91 ± 4% +0.2 1.07 ± 10% perf-profile.self.cycles-pp.__fsnotify_parent 5.71 ± 4% +0.6 6.31 ± 4% perf-profile.self.cycles-pp.__entry_text_start 7.81 ± 5% +0.9 8.68 ± 5% perf-profile.self.cycles-pp.entry_SYSCALL_64_after_hwframe 11.95 ± 5% +1.3 13.21 ± 4% perf-profile.self.cycles-pp.syscall_exit_to_user_mode 13.79 ± 5% +1.5 15.26 ± 4% perf-profile.self.cycles-pp.syscall_return_via_sysret *************************************************************************************************** lkp-hsw-d03: 8 threads 1 sockets Intel(R) Core(TM) i7-4770 CPU @ 3.40GHz (Haswell) with 16G memory ========================================================================================= cluster/compiler/cpufreq_governor/kconfig/nr_threads/protocol/rootfs/runtime/tbox_group/testcase: cs-localhost/gcc-11/performance/x86_64-rhel-8.3/40%/tcp6/debian-11.1-x86_64-20220510.cgz/300s/lkp-hsw-d03/nepim commit: 0d85b27b0c ("Merge tag '6.4-rc3-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6") 47ee3f1dd9 ("x86: re-introduce support for ERMS copies for user space accesses") 0d85b27b0cc6b5cf 47ee3f1dd93bcbe031539b1ecda ---------------- --------------------------- %stddev %change %stddev \ | \ 40382 +5.8% 42719 vmstat.system.cs 29.19 +1.1% 29.51 turbostat.CorWatt 0.28 -45.7% 0.15 turbostat.IPC 51340235 +9.5% 56202298 proc-vmstat.numa_hit 51317600 +9.5% 56198357 proc-vmstat.numa_local 4.088e+08 +9.5% 4.478e+08 proc-vmstat.pgalloc_normal 4.088e+08 +9.5% 4.478e+08 proc-vmstat.pgfree 7501256 +9.5% 8214645 nepim.tcp.avg.kbps_in 7501474 +9.5% 8214867 nepim.tcp.avg.kbps_out 28758 +9.4% 31464 nepim.tcp.avg.rcv_s 28615 +9.5% 31337 nepim.tcp.avg.snd_s 1895 ± 21% +21.9% 2309 ± 13% nepim.time.involuntary_context_switches 14.45 ± 5% -2.3 12.20 ± 5% perf-profile.calltrace.cycles-pp.skb_do_copy_data_nocache.tcp_sendmsg_locked.tcp_sendmsg.sock_write_iter.vfs_write 14.05 ± 4% -2.2 11.86 ± 5% perf-profile.calltrace.cycles-pp._copy_from_iter.skb_do_copy_data_nocache.tcp_sendmsg_locked.tcp_sendmsg.sock_write_iter 13.88 ± 5% -2.1 11.75 ± 5% perf-profile.calltrace.cycles-pp.copyin._copy_from_iter.skb_do_copy_data_nocache.tcp_sendmsg_locked.tcp_sendmsg 13.84 ± 4% -2.1 11.71 ± 5% perf-profile.calltrace.cycles-pp.rep_movs_alternative.copyin._copy_from_iter.skb_do_copy_data_nocache.tcp_sendmsg_locked 34.14 ± 3% -3.4 30.78 ± 3% perf-profile.children.cycles-pp.rep_movs_alternative 14.46 ± 5% -2.3 12.20 ± 5% perf-profile.children.cycles-pp.skb_do_copy_data_nocache 14.06 ± 4% -2.2 11.86 ± 5% perf-profile.children.cycles-pp._copy_from_iter 13.92 ± 4% -2.2 11.76 ± 5% perf-profile.children.cycles-pp.copyin 0.07 ± 21% +0.0 0.11 ± 17% perf-profile.children.cycles-pp._raw_spin_lock_irq 0.03 ±127% +0.1 0.09 ± 36% perf-profile.children.cycles-pp.rcu_all_qs 0.31 ± 6% +0.1 0.38 ± 8% perf-profile.children.cycles-pp.aa_sk_perm 33.97 ± 3% -3.5 30.51 ± 4% perf-profile.self.cycles-pp.rep_movs_alternative 0.15 ± 12% -0.0 0.10 ± 22% perf-profile.self.cycles-pp._copy_from_iter 0.10 ± 13% +0.0 0.14 ± 18% perf-profile.self.cycles-pp.process_backlog 0.07 ± 18% +0.0 0.11 ± 16% perf-profile.self.cycles-pp._raw_spin_lock_irq 0.03 ±124% +0.1 0.09 ± 20% perf-profile.self.cycles-pp.__release_sock 0.32 ± 11% +0.1 0.38 ± 10% perf-profile.self.cycles-pp._raw_spin_lock_bh 0.28 ± 6% +0.1 0.35 ± 9% perf-profile.self.cycles-pp.aa_sk_perm 0.77 ± 12% +0.4 1.19 ± 8% perf-profile.self.cycles-pp.tcp_sendmsg_locked 25.41 +57.6% 40.05 perf-stat.i.MPKI 8.761e+08 -13.6% 7.568e+08 perf-stat.i.branch-instructions 1.71 +0.3 2.03 perf-stat.i.branch-miss-rate% 13.05 ± 2% +3.7 16.74 perf-stat.i.cache-miss-rate% 23909839 +5.5% 25223669 ± 2% perf-stat.i.cache-misses 1.834e+08 -17.8% 1.507e+08 perf-stat.i.cache-references 40658 +5.9% 43055 perf-stat.i.context-switches 1.02 +91.3% 1.95 perf-stat.i.cpi 315.87 -5.2% 299.30 perf-stat.i.cycles-between-cache-misses 0.10 ± 3% +0.1 0.17 perf-stat.i.dTLB-load-miss-rate% 2.492e+09 ± 4% -39.8% 1.5e+09 perf-stat.i.dTLB-loads 0.04 +0.0 0.07 perf-stat.i.dTLB-store-miss-rate% 717270 +9.2% 783320 perf-stat.i.dTLB-store-misses 2.022e+09 -47.7% 1.057e+09 perf-stat.i.dTLB-stores 897381 +7.0% 959782 perf-stat.i.iTLB-loads 7.305e+09 -46.8% 3.888e+09 perf-stat.i.instructions 0.98 -47.3% 0.52 perf-stat.i.ipc 806.16 ± 2% -58.0% 338.57 ± 2% perf-stat.i.metric.K/sec 699.04 ± 2% -37.6% 435.98 perf-stat.i.metric.M/sec 18383434 +28.2% 23575415 perf-stat.i.node-loads 4791695 ± 3% -81.1% 904896 ± 6% perf-stat.i.node-stores 25.11 +54.4% 38.77 perf-stat.overall.MPKI 1.83 +0.3 2.17 perf-stat.overall.branch-miss-rate% 13.04 ± 2% +3.7 16.74 perf-stat.overall.cache-miss-rate% 1.02 +88.1% 1.91 perf-stat.overall.cpi 310.68 -5.1% 294.84 perf-stat.overall.cycles-between-cache-misses 0.10 ± 3% +0.1 0.17 perf-stat.overall.dTLB-load-miss-rate% 0.04 +0.0 0.07 perf-stat.overall.dTLB-store-miss-rate% 0.98 -46.8% 0.52 perf-stat.overall.ipc 8.732e+08 -13.6% 7.543e+08 perf-stat.ps.branch-instructions 23830434 +5.5% 25139918 ± 2% perf-stat.ps.cache-misses 1.828e+08 -17.8% 1.502e+08 perf-stat.ps.cache-references 40523 +5.9% 42913 perf-stat.ps.context-switches 2.484e+09 ± 4% -39.8% 1.495e+09 perf-stat.ps.dTLB-loads 714887 +9.2% 780718 perf-stat.ps.dTLB-store-misses 2.016e+09 -47.7% 1.053e+09 perf-stat.ps.dTLB-stores 894402 +7.0% 956596 perf-stat.ps.iTLB-loads 7.281e+09 -46.8% 3.875e+09 perf-stat.ps.instructions 18322375 +28.2% 23497100 perf-stat.ps.node-loads 4775792 ± 3% -81.1% 901896 ± 6% perf-stat.ps.node-stores 2.194e+12 -46.8% 1.167e+12 perf-stat.total.instructions *************************************************************************************************** lkp-cfl-e1: 16 threads 1 sockets Intel(R) Xeon(R) E-2278G CPU @ 3.40GHz (Coffee Lake) with 32G memory ========================================================================================= compiler/cpufreq_governor/kconfig/option_a/rootfs/tbox_group/test/testcase: gcc-11/performance/x86_64-rhel-8.3/IMB-MPI1 Exchange/debian-x86_64-phoronix/lkp-cfl-e1/intel-mpi-1.0.1/phoronix-test-suite commit: 0d85b27b0c ("Merge tag '6.4-rc3-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6") 47ee3f1dd9 ("x86: re-introduce support for ERMS copies for user space accesses") 0d85b27b0cc6b5cf 47ee3f1dd93bcbe031539b1ecda ---------------- --------------------------- %stddev %change %stddev \ | \ 678.50 ± 17% -45.7% 368.50 ± 62% turbostat.C10 9064 +21.6% 11018 phoronix-test-suite.intel-mpi.IMB-MPI1Exchange.average_mbytes_sec 177.33 -7.6% 163.84 phoronix-test-suite.intel-mpi.IMB-MPI1Exchange.average_usec 8.86 ± 97% -7.8 1.00 ±142% perf-profile.calltrace.cycles-pp.begin_new_exec.load_elf_binary.search_binary_handler.exec_binprm.bprm_execve 6.77 ± 79% -6.2 0.56 ±223% perf-profile.calltrace.cycles-pp.__mmput.exec_mmap.begin_new_exec.load_elf_binary.search_binary_handler 6.77 ± 79% -6.2 0.56 ±223% perf-profile.calltrace.cycles-pp.exec_mmap.begin_new_exec.load_elf_binary.search_binary_handler.exec_binprm 7.08 ± 81% -4.1 2.96 ±158% perf-profile.calltrace.cycles-pp.zap_pmd_range.unmap_page_range.unmap_vmas.exit_mmap.__mmput 7.08 ± 81% -4.1 2.96 ±158% perf-profile.calltrace.cycles-pp.zap_pte_range.zap_pmd_range.unmap_page_range.unmap_vmas.exit_mmap 5.80 ± 77% -2.8 2.96 ±158% perf-profile.calltrace.cycles-pp.unmap_page_range.unmap_vmas.exit_mmap.__mmput.exit_mm 6.14 ± 66% -2.3 3.84 ±142% perf-profile.calltrace.cycles-pp.unmap_vmas.exit_mmap.__mmput.exit_mm.do_exit 8.86 ± 97% -7.8 1.00 ±142% perf-profile.children.cycles-pp.begin_new_exec 6.77 ± 79% -6.2 0.56 ±223% perf-profile.children.cycles-pp.exec_mmap 9.06 ± 80% -5.2 3.84 ±142% perf-profile.children.cycles-pp.unmap_vmas 7.08 ± 81% -4.1 2.96 ±158% perf-profile.children.cycles-pp.unmap_page_range 7.08 ± 81% -4.1 2.96 ±158% perf-profile.children.cycles-pp.zap_pmd_range 7.08 ± 81% -4.1 2.96 ±158% perf-profile.children.cycles-pp.zap_pte_range 4.10 ± 62% -2.5 1.59 ±157% perf-profile.children.cycles-pp.release_pages 3.63 ± 85% -1.0 2.62 ±149% perf-profile.children.cycles-pp.tlb_finish_mmu 3.63 ± 85% -1.0 2.62 ±149% perf-profile.children.cycles-pp.tlb_batch_pages_flush 8776475 -7.0% 8158830 perf-stat.i.cache-misses 2.749e+08 -3.7% 2.647e+08 perf-stat.i.cache-references 1.55 +5.2% 1.63 ± 2% perf-stat.i.cpi 9.08e+09 -1.5% 8.944e+09 perf-stat.i.dTLB-loads 0.03 ± 3% +0.0 0.04 ± 4% perf-stat.i.dTLB-store-miss-rate% 4.758e+09 -2.5% 4.64e+09 perf-stat.i.dTLB-stores 216085 -2.3% 211119 perf-stat.i.iTLB-loads 2.562e+10 -1.5% 2.524e+10 perf-stat.i.instructions 1.13 -2.8% 1.10 perf-stat.i.ipc 983.28 -2.5% 958.80 perf-stat.i.metric.M/sec 750185 +203.4% 2276128 perf-stat.i.node-loads 1661980 +28.8% 2141116 perf-stat.i.node-stores 10.73 -2.3% 10.49 perf-stat.overall.MPKI 3.20 -0.1 3.08 perf-stat.overall.cache-miss-rate% 2088 +6.7% 2227 perf-stat.overall.cycles-between-cache-misses 0.00 ± 9% -0.0 0.00 ± 10% perf-stat.overall.node-load-miss-rate% 8684309 -7.1% 8069768 perf-stat.ps.cache-misses 2.718e+08 -3.7% 2.617e+08 perf-stat.ps.cache-references 8.975e+09 -1.5% 8.84e+09 perf-stat.ps.dTLB-loads 4.703e+09 -2.5% 4.587e+09 perf-stat.ps.dTLB-stores 213618 -2.3% 208713 perf-stat.ps.iTLB-loads 2.532e+10 -1.5% 2.495e+10 perf-stat.ps.instructions 742356 +203.3% 2251351 perf-stat.ps.node-loads 1645341 +28.7% 2117834 perf-stat.ps.node-stores 2.261e+12 -1.3% 2.232e+12 perf-stat.total.instructions *************************************************************************************************** lkp-skl-fpga01: 104 threads 2 sockets (Skylake) with 192G memory ========================================================================================= compiler/cpufreq_governor/kconfig/mode/nr_task/rootfs/tbox_group/test/testcase: gcc-11/performance/x86_64-rhel-8.3/process/50%/debian-11.1-x86_64-20220510.cgz/lkp-skl-fpga01/readseek1/will-it-scale commit: 0d85b27b0c ("Merge tag '6.4-rc3-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6") 47ee3f1dd9 ("x86: re-introduce support for ERMS copies for user space accesses") 0d85b27b0cc6b5cf 47ee3f1dd93bcbe031539b1ecda ---------------- --------------------------- %stddev %change %stddev \ | \ 5492754 ± 14% -17.6% 4524182 ± 14% sched_debug.cfs_rq:/.spread0.max 0.16 -43.8% 0.09 turbostat.IPC 14210 ± 2% -9.0% 12931 turbostat.POLL 26574376 +4.6% 27809208 will-it-scale.52.processes 511045 +4.6% 534792 will-it-scale.per_process_ops 26574376 +4.6% 27809208 will-it-scale.workload 0.06 +79.4% 0.11 ± 12% perf-stat.i.MPKI 1.024e+10 -13.3% 8.877e+09 perf-stat.i.branch-instructions 1.34 +0.3 1.62 perf-stat.i.branch-miss-rate% 1.374e+08 +4.5% 1.436e+08 perf-stat.i.branch-misses 1.80 +71.0% 3.07 perf-stat.i.cpi 0.20 +0.2 0.40 perf-stat.i.dTLB-load-miss-rate% 53209330 +4.6% 55668630 perf-stat.i.dTLB-load-misses 2.683e+10 -48.3% 1.388e+10 perf-stat.i.dTLB-loads 0.00 +0.0 0.00 perf-stat.i.dTLB-store-miss-rate% 37320 +2.7% 38319 perf-stat.i.dTLB-store-misses 2.138e+10 -61.8% 8.175e+09 perf-stat.i.dTLB-stores 58111654 +3.6% 60212884 perf-stat.i.iTLB-load-misses 8.1e+10 -41.5% 4.738e+10 perf-stat.i.instructions 1395 -43.6% 787.58 perf-stat.i.instructions-per-iTLB-miss 0.56 -41.5% 0.33 perf-stat.i.ipc 563.05 -47.0% 298.51 perf-stat.i.metric.M/sec 178317 +3.2% 183996 ± 2% perf-stat.i.node-load-misses 0.06 +79.8% 0.11 ± 11% perf-stat.overall.MPKI 1.34 +0.3 1.62 perf-stat.overall.branch-miss-rate% 1.80 +71.0% 3.07 perf-stat.overall.cpi 0.20 +0.2 0.40 perf-stat.overall.dTLB-load-miss-rate% 0.00 +0.0 0.00 perf-stat.overall.dTLB-store-miss-rate% 1394 -43.6% 786.84 perf-stat.overall.instructions-per-iTLB-miss 0.56 -41.5% 0.33 perf-stat.overall.ipc 916671 -44.1% 512426 perf-stat.overall.path-length 1.02e+10 -13.3% 8.847e+09 perf-stat.ps.branch-instructions 1.37e+08 +4.5% 1.431e+08 perf-stat.ps.branch-misses 53029639 +4.6% 55478925 perf-stat.ps.dTLB-load-misses 2.674e+10 -48.3% 1.383e+10 perf-stat.ps.dTLB-loads 37232 +2.6% 38204 perf-stat.ps.dTLB-store-misses 2.131e+10 -61.8% 8.147e+09 perf-stat.ps.dTLB-stores 57916551 +3.6% 60006718 perf-stat.ps.iTLB-load-misses 8.073e+10 -41.5% 4.721e+10 perf-stat.ps.instructions 177742 +3.2% 183382 ± 2% perf-stat.ps.node-load-misses 2.436e+13 -41.5% 1.425e+13 perf-stat.total.instructions 7.47 -3.1 4.40 perf-profile.calltrace.cycles-pp.rep_movs_alternative.copyout._copy_to_iter.copy_page_to_iter.shmem_file_read_iter 7.73 -3.0 4.68 perf-profile.calltrace.cycles-pp.copyout._copy_to_iter.copy_page_to_iter.shmem_file_read_iter.vfs_read 8.07 -3.0 5.03 perf-profile.calltrace.cycles-pp._copy_to_iter.copy_page_to_iter.shmem_file_read_iter.vfs_read.ksys_read 8.25 -3.0 5.22 perf-profile.calltrace.cycles-pp.copy_page_to_iter.shmem_file_read_iter.vfs_read.ksys_read.do_syscall_64 12.67 -2.8 9.90 perf-profile.calltrace.cycles-pp.shmem_file_read_iter.vfs_read.ksys_read.do_syscall_64.entry_SYSCALL_64_after_hwframe 15.82 -2.6 13.24 perf-profile.calltrace.cycles-pp.vfs_read.ksys_read.do_syscall_64.entry_SYSCALL_64_after_hwframe.read 16.94 -2.5 14.43 perf-profile.calltrace.cycles-pp.ksys_read.do_syscall_64.entry_SYSCALL_64_after_hwframe.read 26.01 -2.1 23.93 perf-profile.calltrace.cycles-pp.do_syscall_64.entry_SYSCALL_64_after_hwframe.read 30.63 -1.8 28.83 perf-profile.calltrace.cycles-pp.entry_SYSCALL_64_after_hwframe.read 45.06 -1.1 43.92 perf-profile.calltrace.cycles-pp.read 0.55 ± 2% +0.0 0.59 ± 2% perf-profile.calltrace.cycles-pp.shmem_file_llseek.ksys_lseek.do_syscall_64.entry_SYSCALL_64_after_hwframe.llseek 0.67 ± 2% +0.0 0.71 perf-profile.calltrace.cycles-pp.filemap_get_entry.shmem_get_folio_gfp.shmem_file_read_iter.vfs_read.ksys_read 0.59 ± 2% +0.0 0.63 ± 2% perf-profile.calltrace.cycles-pp.__fsnotify_parent.vfs_read.ksys_read.do_syscall_64.entry_SYSCALL_64_after_hwframe 0.67 ± 3% +0.0 0.72 perf-profile.calltrace.cycles-pp.__fget_light.__fdget_pos.ksys_read.do_syscall_64.entry_SYSCALL_64_after_hwframe 0.76 +0.0 0.81 perf-profile.calltrace.cycles-pp.atime_needs_update.touch_atime.shmem_file_read_iter.vfs_read.ksys_read 0.82 ± 3% +0.1 0.88 perf-profile.calltrace.cycles-pp.__fdget_pos.ksys_read.do_syscall_64.entry_SYSCALL_64_after_hwframe.read 0.97 ± 2% +0.1 1.03 perf-profile.calltrace.cycles-pp.touch_atime.shmem_file_read_iter.vfs_read.ksys_read.do_syscall_64 1.60 +0.1 1.70 perf-profile.calltrace.cycles-pp.shmem_get_folio_gfp.shmem_file_read_iter.vfs_read.ksys_read.do_syscall_64 1.62 +0.1 1.72 perf-profile.calltrace.cycles-pp.ksys_lseek.do_syscall_64.entry_SYSCALL_64_after_hwframe.llseek 0.78 +0.1 0.90 ± 22% perf-profile.calltrace.cycles-pp.entry_SYSCALL_64_safe_stack.read 0.76 ± 2% +0.1 0.90 ± 23% perf-profile.calltrace.cycles-pp.entry_SYSCALL_64_safe_stack.llseek 4.14 +0.2 4.33 perf-profile.calltrace.cycles-pp.__entry_text_start.read 4.08 +0.3 4.37 perf-profile.calltrace.cycles-pp.__entry_text_start.llseek 7.80 +0.4 8.15 perf-profile.calltrace.cycles-pp.syscall_exit_to_user_mode.do_syscall_64.entry_SYSCALL_64_after_hwframe.read 8.57 +0.4 8.97 perf-profile.calltrace.cycles-pp.syscall_return_via_sysret.read 7.60 +0.5 8.05 perf-profile.calltrace.cycles-pp.syscall_exit_to_user_mode.do_syscall_64.entry_SYSCALL_64_after_hwframe.llseek 10.43 +0.6 11.04 perf-profile.calltrace.cycles-pp.do_syscall_64.entry_SYSCALL_64_after_hwframe.llseek 8.37 +0.6 9.00 perf-profile.calltrace.cycles-pp.syscall_return_via_sysret.llseek 15.05 +0.8 15.85 perf-profile.calltrace.cycles-pp.entry_SYSCALL_64_after_hwframe.llseek 28.78 +2.2 30.95 perf-profile.calltrace.cycles-pp.llseek 7.47 -3.1 4.42 perf-profile.children.cycles-pp.rep_movs_alternative 8.08 -3.0 5.05 perf-profile.children.cycles-pp._copy_to_iter 7.90 -3.0 4.88 perf-profile.children.cycles-pp.copyout 8.26 -3.0 5.24 perf-profile.children.cycles-pp.copy_page_to_iter 12.73 -2.8 9.96 perf-profile.children.cycles-pp.shmem_file_read_iter 15.89 -2.6 13.31 perf-profile.children.cycles-pp.vfs_read 16.99 -2.5 14.49 perf-profile.children.cycles-pp.ksys_read 36.59 -1.5 35.12 perf-profile.children.cycles-pp.do_syscall_64 45.17 -1.1 44.06 perf-profile.children.cycles-pp.read 46.04 -1.0 45.07 perf-profile.children.cycles-pp.entry_SYSCALL_64_after_hwframe 0.10 ± 4% +0.0 0.12 ± 4% perf-profile.children.cycles-pp.__cond_resched 0.20 ± 3% +0.0 0.22 perf-profile.children.cycles-pp.folio_test_hugetlb 0.32 +0.0 0.34 perf-profile.children.cycles-pp.__x64_sys_read 0.31 +0.0 0.33 ± 2% perf-profile.children.cycles-pp.__x64_sys_lseek 0.26 ± 3% +0.0 0.29 ± 3% perf-profile.children.cycles-pp.syscall_exit_to_user_mode_prepare 0.35 ± 4% +0.0 0.39 perf-profile.children.cycles-pp.current_time 0.68 +0.0 0.72 ± 2% perf-profile.children.cycles-pp.syscall_enter_from_user_mode 0.55 ± 2% +0.0 0.59 ± 2% perf-profile.children.cycles-pp.shmem_file_llseek 0.59 +0.0 0.63 ± 2% perf-profile.children.cycles-pp.__fsnotify_parent 0.79 +0.0 0.84 perf-profile.children.cycles-pp.atime_needs_update 0.68 ± 2% +0.0 0.73 perf-profile.children.cycles-pp.filemap_get_entry 1.26 +0.1 1.32 perf-profile.children.cycles-pp.entry_SYSRETQ_unsafe_stack 0.99 +0.1 1.05 perf-profile.children.cycles-pp.touch_atime 1.37 ± 3% +0.1 1.44 perf-profile.children.cycles-pp.__fget_light 1.66 ± 3% +0.1 1.76 perf-profile.children.cycles-pp.__fdget_pos 1.64 +0.1 1.73 perf-profile.children.cycles-pp.ksys_lseek 1.54 +0.1 1.63 perf-profile.children.cycles-pp.entry_SYSCALL_64_safe_stack 1.62 +0.1 1.72 perf-profile.children.cycles-pp.shmem_get_folio_gfp 8.09 +0.5 8.58 perf-profile.children.cycles-pp.__entry_text_start 15.54 +0.8 16.34 perf-profile.children.cycles-pp.syscall_exit_to_user_mode 17.10 +1.0 18.13 perf-profile.children.cycles-pp.syscall_return_via_sysret 29.24 +1.8 31.07 perf-profile.children.cycles-pp.llseek 7.34 -3.0 4.29 perf-profile.self.cycles-pp.rep_movs_alternative 0.21 +0.0 0.23 ± 3% perf-profile.self.cycles-pp.touch_atime 0.30 +0.0 0.32 perf-profile.self.cycles-pp.__x64_sys_lseek 0.66 +0.0 0.68 perf-profile.self.cycles-pp.llseek 0.20 ± 3% +0.0 0.22 perf-profile.self.cycles-pp.folio_test_hugetlb 0.30 ± 3% +0.0 0.33 ± 3% perf-profile.self.cycles-pp.__fdget_pos 0.17 ± 4% +0.0 0.20 ± 5% perf-profile.self.cycles-pp.syscall_exit_to_user_mode_prepare 0.57 ± 2% +0.0 0.60 perf-profile.self.cycles-pp.filemap_get_entry 0.60 +0.0 0.63 ± 2% perf-profile.self.cycles-pp.syscall_enter_from_user_mode 0.84 +0.0 0.88 perf-profile.self.cycles-pp.do_syscall_64 0.54 ± 4% +0.0 0.57 perf-profile.self.cycles-pp.copyout 0.56 +0.0 0.60 ± 2% perf-profile.self.cycles-pp.__fsnotify_parent 0.83 ± 2% +0.0 0.86 perf-profile.self.cycles-pp.shmem_get_folio_gfp 0.53 ± 2% +0.0 0.57 ± 2% perf-profile.self.cycles-pp.shmem_file_llseek 0.62 +0.0 0.66 perf-profile.self.cycles-pp.entry_SYSCALL_64_safe_stack 0.71 +0.1 0.76 ± 2% perf-profile.self.cycles-pp.read 1.10 +0.1 1.15 perf-profile.self.cycles-pp.entry_SYSRETQ_unsafe_stack 1.29 +0.1 1.35 ± 2% perf-profile.self.cycles-pp.shmem_file_read_iter 1.32 ± 3% +0.1 1.39 perf-profile.self.cycles-pp.__fget_light 1.05 +0.1 1.14 perf-profile.self.cycles-pp.vfs_read 7.04 +0.4 7.47 perf-profile.self.cycles-pp.__entry_text_start 9.74 +0.5 10.26 perf-profile.self.cycles-pp.entry_SYSCALL_64_after_hwframe 14.94 +0.8 15.70 perf-profile.self.cycles-pp.syscall_exit_to_user_mode 17.08 +1.0 18.10 perf-profile.self.cycles-pp.syscall_return_via_sysret *************************************************************************************************** lkp-cfl-e1: 16 threads 1 sockets Intel(R) Xeon(R) E-2278G CPU @ 3.40GHz (Coffee Lake) with 32G memory ========================================================================================= compiler/cpufreq_governor/kconfig/option_a/option_b/option_c/rootfs/tbox_group/test/testcase: gcc-11/performance/x86_64-rhel-8.3/Write/64MB/8/debian-x86_64-phoronix/lkp-cfl-e1/tiobench-1.3.1/phoronix-test-suite commit: 0d85b27b0c ("Merge tag '6.4-rc3-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6") 47ee3f1dd9 ("x86: re-introduce support for ERMS copies for user space accesses") 0d85b27b0cc6b5cf 47ee3f1dd93bcbe031539b1ecda ---------------- --------------------------- %stddev %change %stddev \ | \ 12535 ± 2% +43.5% 17982 ± 2% phoronix-test-suite.tiobench.Write.64MB.8.mb_s 2132097 ± 11% -26.0% 1578355 ± 9% perf-stat.i.node-stores 313.06 ± 7% +19.9% 375.47 ± 2% perf-stat.overall.cycles-between-cache-misses 2086550 ± 12% -26.6% 1531685 ± 11% perf-stat.ps.node-stores 3.75 ±102% -3.7 0.00 perf-profile.calltrace.cycles-pp.do_cow_fault.do_fault.__handle_mm_fault.handle_mm_fault.do_user_addr_fault 5.83 ± 78% -3.2 2.68 ±171% perf-profile.calltrace.cycles-pp.unmap_vmas.exit_mmap.__mmput.exit_mm.do_exit 5.83 ± 78% -3.2 2.68 ±171% perf-profile.calltrace.cycles-pp.unmap_page_range.unmap_vmas.exit_mmap.__mmput.exit_mm 8.58 ±107% -8.0 0.60 ±223% perf-profile.children.cycles-pp.tlb_finish_mmu 8.58 ±107% -8.0 0.60 ±223% perf-profile.children.cycles-pp.tlb_batch_pages_flush 8.58 ±107% -6.5 2.11 ±160% perf-profile.children.cycles-pp.release_pages 3.75 ±102% -3.7 0.00 perf-profile.children.cycles-pp.do_cow_fault *************************************************************************************************** lkp-cfl-e1: 16 threads 1 sockets Intel(R) Xeon(R) E-2278G CPU @ 3.40GHz (Coffee Lake) with 32G memory ========================================================================================= compiler/cpufreq_governor/kconfig/option_a/option_b/option_c/rootfs/tbox_group/test/testcase: gcc-11/performance/x86_64-rhel-8.3/Random Write/32MB/4/debian-x86_64-phoronix/lkp-cfl-e1/tiobench-1.3.1/phoronix-test-suite commit: 0d85b27b0c ("Merge tag '6.4-rc3-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6") 47ee3f1dd9 ("x86: re-introduce support for ERMS copies for user space accesses") 0d85b27b0cc6b5cf 47ee3f1dd93bcbe031539b1ecda ---------------- --------------------------- %stddev %change %stddev \ | \ 1.70 ± 2% -0.3 1.39 ± 3% mpstat.cpu.all.sys% 32.33 ± 4% -18.6% 26.33 ± 4% phoronix-test-suite.time.percent_of_cpu_this_job_got 107378 +49.5% 160566 ± 2% phoronix-test-suite.tiobench.RandomWrite.32MB.32.mb_s 10052508 ± 3% -16.7% 8372597 ± 4% perf-stat.i.cache-misses 4.595e+09 ± 2% -6.3% 4.304e+09 ± 2% perf-stat.i.cpu-cycles 3.647e+08 ± 2% -5.6% 3.444e+08 ± 2% perf-stat.i.dTLB-stores 0.29 ± 2% -6.3% 0.27 ± 2% perf-stat.i.metric.GHz 0.00 ± 22% +0.0 0.01 ± 11% perf-stat.i.node-load-miss-rate% 5.67 ± 19% +282.5% 21.70 ± 7% perf-stat.i.node-load-misses 6.24 ± 31% +232.9% 20.76 ± 9% perf-stat.i.node-store-misses 3545056 ± 3% -20.5% 2818290 ± 7% perf-stat.i.node-stores 457.85 ± 5% +12.5% 514.94 ± 3% perf-stat.overall.cycles-between-cache-misses 0.00 ± 18% +0.0 0.01 ± 11% perf-stat.overall.node-load-miss-rate% 0.00 ± 29% +0.0 0.00 ± 12% perf-stat.overall.node-store-miss-rate% 9684141 ± 3% -16.7% 8068153 ± 4% perf-stat.ps.cache-misses 4.427e+09 ± 2% -6.3% 4.15e+09 perf-stat.ps.cpu-cycles 3.514e+08 ± 3% -5.5% 3.32e+08 ± 2% perf-stat.ps.dTLB-stores 5.47 ± 19% +282.5% 20.91 ± 7% perf-stat.ps.node-load-misses 6.00 ± 31% +233.3% 20.01 ± 9% perf-stat.ps.node-store-misses 3414551 ± 3% -20.4% 2718058 ± 8% perf-stat.ps.node-stores 8.14 ± 97% -8.1 0.00 perf-profile.calltrace.cycles-pp.do_fault.__handle_mm_fault.handle_mm_fault.do_user_addr_fault.exc_page_fault 7.55 ± 83% -7.5 0.00 perf-profile.calltrace.cycles-pp.free_swap_cache.free_pages_and_swap_cache.tlb_batch_pages_flush.tlb_finish_mmu.exit_mmap 6.48 ±137% -6.5 0.00 perf-profile.calltrace.cycles-pp.do_read_fault.do_fault.__handle_mm_fault.handle_mm_fault.do_user_addr_fault 9.33 ± 62% -6.3 3.02 ±173% perf-profile.calltrace.cycles-pp.tlb_finish_mmu.exit_mmap.__mmput.exit_mm.do_exit 9.33 ± 62% -6.3 3.02 ±173% perf-profile.calltrace.cycles-pp.tlb_batch_pages_flush.tlb_finish_mmu.exit_mmap.__mmput.exit_mm 6.82 ± 73% -4.3 2.49 ±164% perf-profile.calltrace.cycles-pp.load_elf_binary.search_binary_handler.exec_binprm.bprm_execve.do_execveat_common 8.14 ± 97% -2.7 5.45 ±168% perf-profile.calltrace.cycles-pp.__handle_mm_fault.handle_mm_fault.do_user_addr_fault.exc_page_fault.asm_exc_page_fault 5.48 ±114% -2.5 3.02 ±173% perf-profile.calltrace.cycles-pp.release_pages.tlb_batch_pages_flush.tlb_finish_mmu.exit_mmap.__mmput 9.34 ± 91% -2.3 7.04 ±141% perf-profile.calltrace.cycles-pp.poll_idle.cpuidle_enter_state.cpuidle_enter.cpuidle_idle_call.do_idle 5.44 ± 81% +0.0 5.45 ±168% perf-profile.calltrace.cycles-pp.asm_exc_page_fault 5.44 ± 81% +0.0 5.45 ±168% perf-profile.calltrace.cycles-pp.exc_page_fault.asm_exc_page_fault 5.44 ± 81% +0.0 5.45 ±168% perf-profile.calltrace.cycles-pp.do_user_addr_fault.exc_page_fault.asm_exc_page_fault 7.65 ± 65% +0.1 7.78 ±202% perf-profile.calltrace.cycles-pp.do_group_exit.__x64_sys_exit_group.do_syscall_64.entry_SYSCALL_64_after_hwframe 7.65 ± 65% +0.1 7.78 ±202% perf-profile.calltrace.cycles-pp.do_exit.do_group_exit.__x64_sys_exit_group.do_syscall_64.entry_SYSCALL_64_after_hwframe 7.65 ± 65% +0.1 7.78 ±202% perf-profile.calltrace.cycles-pp.exit_mm.do_exit.do_group_exit.__x64_sys_exit_group.do_syscall_64 7.65 ± 65% +0.1 7.78 ±202% perf-profile.calltrace.cycles-pp.__mmput.exit_mm.do_exit.do_group_exit.__x64_sys_exit_group 7.65 ± 65% +0.1 7.78 ±202% perf-profile.calltrace.cycles-pp.__x64_sys_exit_group.do_syscall_64.entry_SYSCALL_64_after_hwframe 4.67 ± 91% +0.9 5.56 ±223% perf-profile.calltrace.cycles-pp.rep_movs_alternative.copyin.copy_page_from_iter_atomic.generic_perform_write.__generic_file_write_iter 4.67 ± 91% +0.9 5.56 ±223% perf-profile.calltrace.cycles-pp.copy_page_from_iter_atomic.generic_perform_write.__generic_file_write_iter.generic_file_write_iter.vfs_write 4.67 ± 91% +0.9 5.56 ±223% perf-profile.calltrace.cycles-pp.copyin.copy_page_from_iter_atomic.generic_perform_write.__generic_file_write_iter.generic_file_write_iter 4.40 ±110% +1.0 5.45 ±168% perf-profile.calltrace.cycles-pp.handle_mm_fault.do_user_addr_fault.exc_page_fault.asm_exc_page_fault 8.14 ± 97% -8.1 0.00 perf-profile.children.cycles-pp.do_fault 10.72 ± 74% -7.7 3.02 ±173% perf-profile.children.cycles-pp.tlb_finish_mmu 10.72 ± 74% -7.7 3.02 ±173% perf-profile.children.cycles-pp.tlb_batch_pages_flush 6.48 ±137% -6.5 0.00 perf-profile.children.cycles-pp.do_read_fault 5.23 ± 84% -5.2 0.00 perf-profile.children.cycles-pp.free_pages_and_swap_cache 5.23 ± 84% -5.2 0.00 perf-profile.children.cycles-pp.free_swap_cache 6.82 ± 73% -4.3 2.49 ±164% perf-profile.children.cycles-pp.load_elf_binary 3.97 ± 78% -4.0 0.00 perf-profile.children.cycles-pp.page_cache_ra_unbounded 5.49 ±114% -2.5 3.02 ±173% perf-profile.children.cycles-pp.release_pages 9.34 ± 91% -2.3 7.04 ±141% perf-profile.children.cycles-pp.poll_idle 8.14 ± 97% -2.0 6.09 ±155% perf-profile.children.cycles-pp.__handle_mm_fault 7.65 ± 65% +0.1 7.78 ±202% perf-profile.children.cycles-pp.__x64_sys_exit_group 4.67 ± 91% +0.9 5.56 ±223% perf-profile.children.cycles-pp.rep_movs_alternative 4.67 ± 91% +0.9 5.56 ±223% perf-profile.children.cycles-pp.copy_page_from_iter_atomic 4.67 ± 91% +0.9 5.56 ±223% perf-profile.children.cycles-pp.copyin 9.34 ± 91% -2.3 7.04 ±141% perf-profile.self.cycles-pp.poll_idle 4.67 ± 91% +0.9 5.56 ±223% perf-profile.self.cycles-pp.rep_movs_alternative *************************************************************************************************** lkp-ivb-2ep1: 48 threads 2 sockets Intel(R) Xeon(R) CPU E5-2697 v2 @ 2.70GHz (Ivy Bridge-EP) with 112G memory ========================================================================================= compiler/cpufreq_governor/kconfig/mode/nr_threads/rootfs/tbox_group/test/test_memory_size/testcase: gcc-11/performance/x86_64-rhel-8.3/development/1/debian-11.1-x86_64-20220510.cgz/lkp-ivb-2ep1/UNIX/50%/lmbench3 commit: 0d85b27b0c ("Merge tag '6.4-rc3-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6") 47ee3f1dd9 ("x86: re-introduce support for ERMS copies for user space accesses") 0d85b27b0cc6b5cf 47ee3f1dd93bcbe031539b1ecda ---------------- --------------------------- %stddev %change %stddev \ | \ 6375 +9.2% 6959 lmbench3.AF_UNIX.sock.stream.bandwidth.MB/sec 597101 ± 9% +65.9% 990600 ± 37% sched_debug.cpu.max_idle_balance_cost.max 16099 ± 65% +411.0% 82275 ± 79% sched_debug.cpu.max_idle_balance_cost.stddev 178220 ± 10% -65.7% 61065 ± 6% turbostat.C1 0.14 ± 4% -0.1 0.06 ± 7% turbostat.C1% 10200875 +9.2% 11144066 ± 2% proc-vmstat.numa_hit 10145413 +8.8% 11033665 ± 2% proc-vmstat.numa_local 46796092 ± 2% +6.2% 49719663 proc-vmstat.pgalloc_normal 46721572 ± 2% +6.3% 49645401 proc-vmstat.pgfree 0.24 ±223% +2.4 2.62 ± 73% perf-profile.calltrace.cycles-pp.do_sys_openat2.__x64_sys_openat.do_syscall_64.entry_SYSCALL_64_after_hwframe 2.74 ± 65% -1.9 0.82 ± 86% perf-profile.children.cycles-pp.__schedule 1.44 ± 58% -1.0 0.42 ±109% perf-profile.children.cycles-pp.newidle_balance 0.47 ±115% +1.0 1.44 ± 62% perf-profile.children.cycles-pp._raw_spin_lock 0.84 ± 75% +1.3 2.14 ± 19% perf-profile.children.cycles-pp.vsnprintf 0.72 ± 75% +1.4 2.08 ± 18% perf-profile.children.cycles-pp.seq_printf 1.32 ± 96% +1.6 2.96 ± 23% perf-profile.children.cycles-pp.link_path_walk 0.47 ±115% +1.0 1.44 ± 62% perf-profile.self.cycles-pp._raw_spin_lock 25.20 ± 16% +123.4% 56.28 ± 7% perf-stat.i.MPKI 1.081e+08 ± 4% +35.4% 1.463e+08 ± 6% perf-stat.i.cache-references 0.63 ± 5% +0.1 0.68 ± 2% perf-stat.i.dTLB-load-miss-rate% 1.648e+09 ± 4% -14.7% 1.406e+09 ± 6% perf-stat.i.dTLB-loads 1571133 ± 4% -8.6% 1435719 ± 5% perf-stat.i.dTLB-store-misses 1.187e+09 ± 4% -20.9% 9.395e+08 ± 6% perf-stat.i.dTLB-stores 6.106e+09 ± 4% -22.7% 4.719e+09 ± 5% perf-stat.i.instructions 8412 ± 4% -26.8% 6159 ± 3% perf-stat.i.instructions-per-iTLB-miss 0.66 ± 3% -23.5% 0.51 ± 4% perf-stat.i.ipc 17.70 ± 2% +75.1% 30.98 ± 2% perf-stat.overall.MPKI 4.73 +0.2 4.93 perf-stat.overall.branch-miss-rate% 20.30 ± 3% -5.6 14.73 ± 5% perf-stat.overall.cache-miss-rate% 1.46 ± 2% +30.4% 1.90 ± 2% perf-stat.overall.cpi 0.51 ± 8% +0.1 0.59 ± 5% perf-stat.overall.dTLB-load-miss-rate% 0.13 ± 2% +0.0 0.15 ± 3% perf-stat.overall.dTLB-store-miss-rate% 7715 ± 3% -23.8% 5877 ± 3% perf-stat.overall.instructions-per-iTLB-miss 0.69 ± 2% -23.3% 0.53 ± 2% perf-stat.overall.ipc 1.064e+08 ± 4% +35.4% 1.441e+08 ± 6% perf-stat.ps.cache-references 1.623e+09 ± 4% -14.6% 1.386e+09 ± 6% perf-stat.ps.dTLB-loads 1547538 ± 4% -8.6% 1414466 ± 5% perf-stat.ps.dTLB-store-misses 1.169e+09 ± 4% -20.8% 9.255e+08 ± 6% perf-stat.ps.dTLB-stores 6.015e+09 ± 4% -22.7% 4.651e+09 ± 5% perf-stat.ps.instructions 4.099e+11 ± 2% -21.3% 3.225e+11 ± 4% perf-stat.total.instructions *************************************************************************************************** lkp-cfl-e1: 16 threads 1 sockets Intel(R) Xeon(R) E-2278G CPU @ 3.40GHz (Coffee Lake) with 32G memory ========================================================================================= compiler/cpufreq_governor/kconfig/option_a/option_b/option_c/rootfs/tbox_group/test/testcase: gcc-11/performance/x86_64-rhel-8.3/Random Write/64MB/4/debian-x86_64-phoronix/lkp-cfl-e1/tiobench-1.3.1/phoronix-test-suite commit: 0d85b27b0c ("Merge tag '6.4-rc3-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6") 47ee3f1dd9 ("x86: re-introduce support for ERMS copies for user space accesses") 0d85b27b0cc6b5cf 47ee3f1dd93bcbe031539b1ecda ---------------- --------------------------- %stddev %change %stddev \ | \ 2.40 -0.5 1.90 ± 5% mpstat.cpu.all.sys% 207080 +51.1% 312861 phoronix-test-suite.tiobench.RandomWrite.64MB.32.mb_s 15.34 ±126% -14.1 1.28 ±223% perf-profile.calltrace.cycles-pp.wp_page_copy.__handle_mm_fault.handle_mm_fault.do_user_addr_fault.exc_page_fault 15.34 ±126% -14.1 1.28 ±223% perf-profile.children.cycles-pp.wp_page_copy 79033 ± 4% +15.2% 91012 ± 10% turbostat.C8 12.50 ± 29% -45.3% 6.83 ± 44% turbostat.C9 194443 ± 87% -79.8% 39236 ± 8% sched_debug.cfs_rq:/.load.stddev 12884 ± 8% +34.7% 17354 ± 28% sched_debug.cfs_rq:/.min_vruntime.max 2307 ± 10% +32.4% 3055 ± 11% sched_debug.cfs_rq:/.min_vruntime.stddev 518.00 ± 15% +21.4% 628.86 ± 7% sched_debug.cfs_rq:/.util_avg.avg 279473 ± 12% -31.4% 191767 ± 11% sched_debug.cpu.avg_idle.stddev 9.64 ± 4% -1.4 8.21 ± 11% perf-stat.i.cache-miss-rate% 14015078 ± 2% -21.6% 10993421 ± 4% perf-stat.i.cache-misses 5.125e+09 ± 2% -8.3% 4.699e+09 ± 2% perf-stat.i.cpu-cycles 4.042e+08 ± 2% -6.3% 3.788e+08 ± 2% perf-stat.i.dTLB-stores 1294820 ± 2% +12.8% 1460653 ± 12% perf-stat.i.iTLB-load-misses 1172786 ± 2% +16.2% 1362610 ± 15% perf-stat.i.iTLB-loads 0.32 ± 2% -8.3% 0.29 ± 2% perf-stat.i.metric.GHz 0.00 ± 31% +0.0 0.01 ± 15% perf-stat.i.node-load-miss-rate% 6.62 ± 16% +241.5% 22.62 ± 15% perf-stat.i.node-load-misses 5.72 ± 18% +313.5% 23.65 ± 16% perf-stat.i.node-store-misses 12.94 -3.2 9.77 ± 8% perf-stat.overall.cache-miss-rate% 365.74 +17.0% 427.97 ± 2% perf-stat.overall.cycles-between-cache-misses 2561 -15.3% 2169 ± 22% perf-stat.overall.instructions-per-iTLB-miss 0.00 ± 16% +0.0 0.01 ± 25% perf-stat.overall.node-load-miss-rate% 0.00 ± 17% +0.0 0.00 ± 19% perf-stat.overall.node-store-miss-rate% 13486705 ± 2% -20.9% 10664323 ± 3% perf-stat.ps.cache-misses 4.932e+09 ± 2% -7.5% 4.56e+09 perf-stat.ps.cpu-cycles 3.89e+08 ± 2% -5.5% 3.675e+08 ± 2% perf-stat.ps.dTLB-stores 1245754 ± 2% +14.0% 1419784 ± 14% perf-stat.ps.iTLB-load-misses 1128129 ± 2% +17.4% 1324815 ± 17% perf-stat.ps.iTLB-loads 6.37 ± 16% +244.6% 21.95 ± 15% perf-stat.ps.node-load-misses 5.50 ± 18% +317.3% 22.96 ± 16% perf-stat.ps.node-store-misses *************************************************************************************************** lkp-cfl-e1: 16 threads 1 sockets Intel(R) Xeon(R) E-2278G CPU @ 3.40GHz (Coffee Lake) with 32G memory ========================================================================================= compiler/cpufreq_governor/kconfig/option_a/option_b/option_c/rootfs/tbox_group/test/testcase: gcc-11/performance/x86_64-rhel-8.3/Write/64MB/4/debian-x86_64-phoronix/lkp-cfl-e1/tiobench-1.3.1/phoronix-test-suite commit: 0d85b27b0c ("Merge tag '6.4-rc3-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6") 47ee3f1dd9 ("x86: re-introduce support for ERMS copies for user space accesses") 0d85b27b0cc6b5cf 47ee3f1dd93bcbe031539b1ecda ---------------- --------------------------- %stddev %change %stddev \ | \ 2.29 ± 2% -0.6 1.73 ± 2% mpstat.cpu.all.sys% 36.00 ± 2% -30.6% 25.00 ± 2% phoronix-test-suite.time.percent_of_cpu_this_job_got 13930 +58.5% 22083 phoronix-test-suite.tiobench.Write.64MB.32.mb_s 15352488 ± 3% -19.6% 12345834 perf-stat.i.cache-misses 67514100 ± 2% -6.2% 63302150 perf-stat.i.cache-references 3.915e+09 ± 3% -11.7% 3.455e+09 ± 3% perf-stat.i.cpu-cycles 2.867e+08 ± 4% -9.7% 2.588e+08 ± 3% perf-stat.i.dTLB-stores 0.46 ± 2% +7.4% 0.49 ± 3% perf-stat.i.ipc 0.24 ± 3% -11.8% 0.22 ± 3% perf-stat.i.metric.GHz 0.00 ± 15% +0.0 0.01 ± 16% perf-stat.i.node-load-miss-rate% 7.27 ± 16% +232.0% 24.13 ± 17% perf-stat.i.node-load-misses 282123 ± 5% +16.2% 327849 ± 4% perf-stat.i.node-loads 8.60 ± 23% +195.5% 25.42 ± 17% perf-stat.i.node-store-misses 6748445 ± 3% -20.7% 5351270 ± 4% perf-stat.i.node-stores 25.14 ± 3% -6.4% 23.54 ± 3% perf-stat.overall.MPKI 22.72 -3.2 19.49 perf-stat.overall.cache-miss-rate% 1.46 ± 2% -11.9% 1.28 perf-stat.overall.cpi 255.14 +9.7% 279.81 ± 2% perf-stat.overall.cycles-between-cache-misses 0.02 ± 3% +0.0 0.03 ± 2% perf-stat.overall.dTLB-store-miss-rate% 0.69 ± 2% +13.5% 0.78 perf-stat.overall.ipc 0.00 ± 15% +0.0 0.01 ± 18% perf-stat.overall.node-load-miss-rate% 0.00 ± 24% +0.0 0.00 ± 20% perf-stat.overall.node-store-miss-rate% 14688148 ± 3% -19.5% 11824921 perf-stat.ps.cache-misses 64639446 ± 2% -6.2% 60656902 perf-stat.ps.cache-references 3.747e+09 ± 3% -11.7% 3.309e+09 ± 2% perf-stat.ps.cpu-cycles 2.743e+08 ± 4% -9.6% 2.479e+08 ± 2% perf-stat.ps.dTLB-stores 6.95 ± 16% +232.7% 23.11 ± 17% perf-stat.ps.node-load-misses 269991 ± 5% +16.4% 314138 ± 4% perf-stat.ps.node-loads 8.22 ± 23% +196.0% 24.34 ± 16% perf-stat.ps.node-store-misses 6457429 ± 3% -20.6% 5127035 ± 4% perf-stat.ps.node-stores *************************************************************************************************** lkp-hsw-d04: 8 threads 1 sockets Intel(R) Core(TM) i7-4770K CPU @ 3.50GHz (Haswell) with 8G memory ========================================================================================= cluster/compiler/cpufreq_governor/kconfig/nr_threads/protocol/rootfs/runtime/tbox_group/testcase: cs-localhost/gcc-11/performance/x86_64-rhel-8.3/25%/tcp/debian-11.1-x86_64-20220510.cgz/300s/lkp-hsw-d04/nepim commit: 0d85b27b0c ("Merge tag '6.4-rc3-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6") 47ee3f1dd9 ("x86: re-introduce support for ERMS copies for user space accesses") 0d85b27b0cc6b5cf 47ee3f1dd93bcbe031539b1ecda ---------------- --------------------------- %stddev %change %stddev \ | \ 45715 +10.7% 50623 vmstat.system.cs 955743 +9.6% 1047078 sched_debug.cpu.nr_switches.avg 1531197 ± 6% +15.1% 1762126 ± 5% sched_debug.cpu.nr_switches.max 5782803 +12.6% 6513418 ± 2% turbostat.C1 2.44 ± 2% +0.2 2.62 ± 4% turbostat.C1% 0.28 -45.7% 0.15 turbostat.IPC 50588669 +10.4% 55825973 proc-vmstat.numa_hit 50591236 +10.4% 55832674 proc-vmstat.numa_local 25784 ± 3% +13.8% 29331 proc-vmstat.pgactivate 4.029e+08 +10.3% 4.445e+08 proc-vmstat.pgalloc_normal 4.028e+08 +10.3% 4.445e+08 proc-vmstat.pgfree 11101771 +10.3% 12243890 nepim.tcp.avg.kbps_in 11101986 +10.3% 12244118 nepim.tcp.avg.kbps_out 42677 +10.4% 47099 nepim.tcp.avg.rcv_s 42350 +10.3% 46707 nepim.tcp.avg.snd_s 1924 ± 27% -44.9% 1061 ± 40% nepim.time.involuntary_context_switches 76.00 ± 3% -6.8% 70.80 ± 3% nepim.time.percent_of_cpu_this_job_got 196.47 ± 5% -11.4% 174.00 ± 5% nepim.time.system_time 32.86 ± 7% +21.7% 40.00 ± 5% nepim.time.user_time 3702890 ± 27% +57.8% 5842771 ± 15% nepim.time.voluntary_context_switches 25.11 +59.0% 39.91 perf-stat.i.MPKI 8.617e+08 -12.8% 7.514e+08 perf-stat.i.branch-instructions 1.71 +0.2 1.95 perf-stat.i.branch-miss-rate% 12.34 +3.9 16.28 ± 2% perf-stat.i.cache-miss-rate% 21804011 ± 2% +10.2% 24021171 perf-stat.i.cache-misses 1.768e+08 -16.5% 1.477e+08 perf-stat.i.cache-references 46033 +10.7% 50946 perf-stat.i.context-switches 1.03 +90.8% 1.96 perf-stat.i.cpi 340.91 -9.0% 310.31 perf-stat.i.cycles-between-cache-misses 0.10 ± 7% +0.1 0.19 ± 7% perf-stat.i.dTLB-load-miss-rate% 2.379e+09 ± 2% -37.5% 1.486e+09 ± 3% perf-stat.i.dTLB-loads 0.04 +0.0 0.08 perf-stat.i.dTLB-store-miss-rate% 772508 +9.2% 843638 perf-stat.i.dTLB-store-misses 2.009e+09 -47.8% 1.049e+09 ± 2% perf-stat.i.dTLB-stores 1029493 +7.2% 1104044 perf-stat.i.iTLB-loads 7.126e+09 -46.3% 3.828e+09 perf-stat.i.instructions 5394 ± 12% -50.5% 2671 ± 17% perf-stat.i.instructions-per-iTLB-miss 0.97 -47.1% 0.51 perf-stat.i.ipc 654.57 ± 3% -47.7% 342.06 perf-stat.i.metric.K/sec 680.50 -36.5% 432.13 perf-stat.i.metric.M/sec 18110220 +24.9% 22615284 perf-stat.i.node-loads 3374408 ± 5% -78.6% 723088 ± 5% perf-stat.i.node-stores 24.81 +55.5% 38.59 perf-stat.overall.MPKI 1.83 +0.3 2.09 perf-stat.overall.branch-miss-rate% 12.33 +3.9 16.26 ± 2% perf-stat.overall.cache-miss-rate% 1.02 +87.4% 1.92 perf-stat.overall.cpi 335.05 -8.6% 306.10 perf-stat.overall.cycles-between-cache-misses 0.10 ± 7% +0.1 0.18 ± 7% perf-stat.overall.dTLB-load-miss-rate% 0.04 +0.0 0.08 perf-stat.overall.dTLB-store-miss-rate% 5293 ± 12% -51.6% 2561 ± 18% perf-stat.overall.instructions-per-iTLB-miss 0.98 -46.6% 0.52 perf-stat.overall.ipc 8.588e+08 -12.8% 7.489e+08 perf-stat.ps.branch-instructions 21731682 ± 2% +10.2% 23941402 perf-stat.ps.cache-misses 1.762e+08 -16.5% 1.472e+08 perf-stat.ps.cache-references 45881 +10.7% 50777 perf-stat.ps.context-switches 2.371e+09 ± 2% -37.5% 1.481e+09 ± 3% perf-stat.ps.dTLB-loads 769946 +9.2% 840836 perf-stat.ps.dTLB-store-misses 2.002e+09 -47.8% 1.046e+09 ± 2% perf-stat.ps.dTLB-stores 1026078 +7.2% 1100384 perf-stat.ps.iTLB-loads 7.102e+09 -46.3% 3.815e+09 perf-stat.ps.instructions 18050146 +24.9% 22540180 perf-stat.ps.node-loads 3363213 ± 5% -78.6% 720692 ± 5% perf-stat.ps.node-stores 2.143e+12 -46.3% 1.15e+12 perf-stat.total.instructions 20.63 ± 4% -2.7 17.94 ± 4% perf-profile.calltrace.cycles-pp.rep_movs_alternative.copyout._copy_to_iter.__skb_datagram_iter.skb_copy_datagram_iter 20.76 ± 4% -2.7 18.08 ± 4% perf-profile.calltrace.cycles-pp._copy_to_iter.__skb_datagram_iter.skb_copy_datagram_iter.tcp_recvmsg_locked.tcp_recvmsg 20.67 ± 4% -2.7 17.99 ± 4% perf-profile.calltrace.cycles-pp.copyout._copy_to_iter.__skb_datagram_iter.skb_copy_datagram_iter.tcp_recvmsg_locked 29.55 ± 3% -2.3 27.28 ± 4% perf-profile.calltrace.cycles-pp.sock_recvmsg.sock_read_iter.vfs_read.ksys_read.do_syscall_64 29.67 ± 3% -2.3 27.40 ± 4% perf-profile.calltrace.cycles-pp.sock_read_iter.vfs_read.ksys_read.do_syscall_64.entry_SYSCALL_64_after_hwframe 30.11 ± 3% -2.2 27.87 ± 4% perf-profile.calltrace.cycles-pp.vfs_read.ksys_read.do_syscall_64.entry_SYSCALL_64_after_hwframe.read 30.29 ± 3% -2.2 28.10 ± 4% perf-profile.calltrace.cycles-pp.ksys_read.do_syscall_64.entry_SYSCALL_64_after_hwframe.read.oop_sys_run_once 1.20 ± 3% +0.2 1.36 ± 6% perf-profile.calltrace.cycles-pp.syscall_exit_to_user_mode.do_syscall_64.entry_SYSCALL_64_after_hwframe.write.oop_sys_run_once 1.09 ± 10% +0.2 1.29 ± 5% perf-profile.calltrace.cycles-pp.syscall_exit_to_user_mode.do_syscall_64.entry_SYSCALL_64_after_hwframe.__select 1.08 ± 10% +0.2 1.29 ± 9% perf-profile.calltrace.cycles-pp.syscall_exit_to_user_mode.do_syscall_64.entry_SYSCALL_64_after_hwframe.sigprocmask.main 2.32 ± 10% +0.4 2.71 ± 5% perf-profile.calltrace.cycles-pp.sigprocmask.main.__libc_start_main 2.72 ± 10% +0.4 3.16 ± 4% perf-profile.calltrace.cycles-pp.__libc_start_main 2.69 ± 10% +0.4 3.13 ± 4% perf-profile.calltrace.cycles-pp.main.__libc_start_main 1.70 ± 12% +0.6 2.26 ± 4% perf-profile.calltrace.cycles-pp.tcp_rcv_established.tcp_v4_do_rcv.tcp_v4_rcv.ip_protocol_deliver_rcu.ip_local_deliver_finish 6.89 ± 8% +1.2 8.12 ± 9% perf-profile.calltrace.cycles-pp.__select 34.20 ± 2% -3.7 30.50 ± 4% perf-profile.children.cycles-pp.rep_movs_alternative 20.77 ± 4% -2.7 18.08 ± 4% perf-profile.children.cycles-pp._copy_to_iter 20.68 ± 4% -2.7 17.99 ± 4% perf-profile.children.cycles-pp.copyout 22.03 ± 4% -2.7 19.37 ± 4% perf-profile.children.cycles-pp.skb_copy_datagram_iter 22.00 ± 4% -2.6 19.35 ± 4% perf-profile.children.cycles-pp.__skb_datagram_iter 29.17 ± 3% -2.3 26.83 ± 4% perf-profile.children.cycles-pp.tcp_recvmsg 26.95 ± 3% -2.3 24.62 ± 4% perf-profile.children.cycles-pp.tcp_recvmsg_locked 29.55 ± 3% -2.3 27.29 ± 4% perf-profile.children.cycles-pp.sock_recvmsg 29.68 ± 3% -2.3 27.41 ± 4% perf-profile.children.cycles-pp.sock_read_iter 30.14 ± 3% -2.2 27.90 ± 4% perf-profile.children.cycles-pp.vfs_read 30.33 ± 3% -2.2 28.14 ± 4% perf-profile.children.cycles-pp.ksys_read 0.37 ± 9% -0.1 0.30 ± 8% perf-profile.children.cycles-pp.tcp_queue_rcv 0.30 ± 10% -0.1 0.24 ± 6% perf-profile.children.cycles-pp.tcp_try_coalesce 0.09 ± 23% -0.1 0.04 ± 87% perf-profile.children.cycles-pp.alloc_pages 0.21 ± 7% +0.0 0.25 ± 8% perf-profile.children.cycles-pp.tcp_current_mss 0.01 ±200% +0.1 0.07 ± 15% perf-profile.children.cycles-pp.perf_rotate_context 0.11 ± 15% +0.1 0.17 ± 10% perf-profile.children.cycles-pp.__fdelt_warn 0.29 ± 4% +0.1 0.35 ± 9% perf-profile.children.cycles-pp.security_socket_recvmsg 0.64 ± 5% +0.1 0.73 ± 3% perf-profile.children.cycles-pp.mem_cgroup_charge_skmem 0.48 ± 7% +0.1 0.58 ± 9% perf-profile.children.cycles-pp.entry_SYSCALL_64_safe_stack 0.34 ± 16% +0.1 0.45 ± 20% perf-profile.children.cycles-pp.poll_freewait 1.70 ± 2% +0.2 1.90 ± 4% perf-profile.children.cycles-pp.syscall_return_via_sysret 2.39 ± 10% +0.4 2.79 ± 5% perf-profile.children.cycles-pp.sigprocmask 2.72 ± 10% +0.4 3.16 ± 4% perf-profile.children.cycles-pp.__libc_start_main 2.72 ± 10% +0.4 3.16 ± 4% perf-profile.children.cycles-pp.main 4.71 ± 6% +0.7 5.38 ± 5% perf-profile.children.cycles-pp.syscall_exit_to_user_mode 6.98 ± 8% +1.2 8.20 ± 9% perf-profile.children.cycles-pp.__select 33.94 ± 2% -3.7 30.23 ± 4% perf-profile.self.cycles-pp.rep_movs_alternative 0.07 ± 14% +0.0 0.08 ± 9% perf-profile.self.cycles-pp.tcp_data_queue 0.06 ± 7% +0.0 0.08 ± 11% perf-profile.self.cycles-pp.apparmor_socket_sendmsg 0.21 ± 15% +0.0 0.25 ± 6% perf-profile.self.cycles-pp.vfs_write 0.29 ± 9% +0.0 0.33 ± 3% perf-profile.self.cycles-pp.entry_SYSRETQ_unsafe_stack 0.04 ± 50% +0.0 0.09 ± 42% perf-profile.self.cycles-pp.datagram_poll 0.04 ± 83% +0.0 0.09 ± 11% perf-profile.self.cycles-pp.__fdelt_chk@plt 0.26 ± 7% +0.1 0.31 ± 11% perf-profile.self.cycles-pp.aa_sk_perm 0.16 ± 20% +0.1 0.22 ± 12% perf-profile.self.cycles-pp.tcp_rcv_established 0.14 ± 27% +0.1 0.19 ± 13% perf-profile.self.cycles-pp.skb_page_frag_refill 0.05 ± 85% +0.1 0.12 ± 14% perf-profile.self.cycles-pp.enqueue_entity 1.70 ± 3% +0.2 1.89 ± 4% perf-profile.self.cycles-pp.syscall_return_via_sysret 0.72 ± 6% +0.5 1.23 ± 7% perf-profile.self.cycles-pp.tcp_sendmsg_locked 4.59 ± 6% +0.7 5.26 ± 5% perf-profile.self.cycles-pp.syscall_exit_to_user_mode *************************************************************************************************** lkp-cfl-e1: 16 threads 1 sockets Intel(R) Xeon(R) E-2278G CPU @ 3.40GHz (Coffee Lake) with 32G memory ========================================================================================= compiler/cpufreq_governor/kconfig/option_a/option_b/option_c/rootfs/tbox_group/test/testcase: gcc-11/performance/x86_64-rhel-8.3/Write/32MB/4/debian-x86_64-phoronix/lkp-cfl-e1/tiobench-1.3.1/phoronix-test-suite commit: 0d85b27b0c ("Merge tag '6.4-rc3-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6") 47ee3f1dd9 ("x86: re-introduce support for ERMS copies for user space accesses") 0d85b27b0cc6b5cf 47ee3f1dd93bcbe031539b1ecda ---------------- --------------------------- %stddev %change %stddev \ | \ 1.57 ± 2% -0.3 1.25 ± 2% mpstat.cpu.all.sys% 273236 +1.5% 277363 proc-vmstat.pgfault 20.83 -26.4% 15.33 ± 3% phoronix-test-suite.time.percent_of_cpu_this_job_got 13836 +57.3% 21761 phoronix-test-suite.tiobench.Write.32MB.32.mb_s 4.99 ± 79% -3.9 1.04 ±223% perf-profile.calltrace.cycles-pp.next_uptodate_page.filemap_map_pages.do_read_fault.do_fault.__handle_mm_fault 4.99 ± 79% -3.9 1.04 ±223% perf-profile.children.cycles-pp.next_uptodate_page 4.71 ± 77% -3.1 1.62 ±145% perf-profile.children.cycles-pp.__fput 4.71 ± 77% -3.1 1.62 ±145% perf-profile.children.cycles-pp.task_work_run 8.88 ± 48% +12.2 21.06 ± 44% perf-profile.children.cycles-pp.__mmput 8.88 ± 48% +12.2 21.06 ± 44% perf-profile.children.cycles-pp.exit_mmap 4.99 ± 79% -3.9 1.04 ±223% perf-profile.self.cycles-pp.next_uptodate_page 12.51 ± 2% -1.2 11.32 perf-stat.i.cache-miss-rate% 10967720 -14.1% 9423790 ± 2% perf-stat.i.cache-misses 62094269 -4.0% 59580926 perf-stat.i.cache-references 3.296e+09 ± 2% -7.4% 3.052e+09 ± 2% perf-stat.i.cpu-cycles 0.47 ± 2% +5.1% 0.49 ± 2% perf-stat.i.ipc 0.21 ± 2% -7.5% 0.19 ± 2% perf-stat.i.metric.GHz 0.00 ± 12% +0.0 0.01 ± 15% perf-stat.i.node-load-miss-rate% 6.58 ± 13% +230.0% 21.71 ± 15% perf-stat.i.node-load-misses 268579 ± 3% +7.6% 288964 ± 3% perf-stat.i.node-loads 7.18 ± 35% +215.4% 22.65 ± 12% perf-stat.i.node-store-misses 3675951 -23.7% 2805440 ± 2% perf-stat.i.node-stores 24.86 -3.2% 24.06 perf-stat.overall.MPKI 17.66 -1.9 15.81 perf-stat.overall.cache-miss-rate% 1.32 ± 2% -6.7% 1.23 perf-stat.overall.cpi 300.56 ± 2% +7.7% 323.83 perf-stat.overall.cycles-between-cache-misses 0.03 ± 2% +0.0 0.03 ± 2% perf-stat.overall.dTLB-store-miss-rate% 0.76 ± 2% +7.1% 0.81 perf-stat.overall.ipc 0.00 ± 13% +0.0 0.01 ± 13% perf-stat.overall.node-load-miss-rate% 0.00 ± 35% +0.0 0.00 ± 13% perf-stat.overall.node-store-miss-rate% 10494435 -14.0% 9020387 ± 2% perf-stat.ps.cache-misses 59412983 -4.0% 57045833 perf-stat.ps.cache-references 3.154e+09 ± 2% -7.4% 2.921e+09 ± 2% perf-stat.ps.cpu-cycles 6.29 ± 13% +230.1% 20.77 ± 15% perf-stat.ps.node-load-misses 256921 ± 3% +7.7% 276598 ± 3% perf-stat.ps.node-loads 6.87 ± 35% +215.5% 21.68 ± 12% perf-stat.ps.node-store-misses 3518377 -23.7% 2685938 ± 2% perf-stat.ps.node-stores *************************************************************************************************** lkp-cfl-e1: 16 threads 1 sockets Intel(R) Xeon(R) E-2278G CPU @ 3.40GHz (Coffee Lake) with 32G memory ========================================================================================= compiler/cpufreq_governor/kconfig/option_a/option_b/option_c/rootfs/tbox_group/test/testcase: gcc-11/performance/x86_64-rhel-8.3/Random Write/64MB/8/debian-x86_64-phoronix/lkp-cfl-e1/tiobench-1.3.1/phoronix-test-suite commit: 0d85b27b0c ("Merge tag '6.4-rc3-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6") 47ee3f1dd9 ("x86: re-introduce support for ERMS copies for user space accesses") 0d85b27b0cc6b5cf 47ee3f1dd93bcbe031539b1ecda ---------------- --------------------------- %stddev %change %stddev \ | \ 203084 +45.1% 294676 ± 2% phoronix-test-suite.tiobench.RandomWrite.64MB.8.mb_s 1819666 ± 8% -19.4% 1466212 ± 9% perf-stat.i.node-stores 361.37 ± 3% +13.6% 410.37 ± 2% perf-stat.overall.cycles-between-cache-misses 1775294 ± 9% -19.6% 1427506 ± 10% perf-stat.ps.node-stores 36.61 ± 4% -23.8 12.85 ±143% perf-profile.calltrace.cycles-pp.secondary_startup_64_no_verify 34.62 ± 9% -22.8 11.80 ±141% perf-profile.calltrace.cycles-pp.cpuidle_enter_state.cpuidle_enter.cpuidle_idle_call.do_idle.cpu_startup_entry 34.08 ± 5% -22.3 11.80 ±141% perf-profile.calltrace.cycles-pp.cpuidle_idle_call.do_idle.cpu_startup_entry.start_secondary.secondary_startup_64_no_verify 34.08 ± 5% -22.3 11.80 ±141% perf-profile.calltrace.cycles-pp.cpuidle_enter.cpuidle_idle_call.do_idle.cpu_startup_entry.start_secondary 34.87 ± 7% -22.0 12.85 ±143% perf-profile.calltrace.cycles-pp.start_secondary.secondary_startup_64_no_verify 34.87 ± 7% -22.0 12.85 ±143% perf-profile.calltrace.cycles-pp.cpu_startup_entry.start_secondary.secondary_startup_64_no_verify 34.87 ± 7% -22.0 12.85 ±143% perf-profile.calltrace.cycles-pp.do_idle.cpu_startup_entry.start_secondary.secondary_startup_64_no_verify 31.16 ± 14% -19.4 11.80 ±141% perf-profile.calltrace.cycles-pp.intel_idle.cpuidle_enter_state.cpuidle_enter.cpuidle_idle_call.do_idle 5.67 ± 72% -5.7 0.00 perf-profile.calltrace.cycles-pp.do_fault.__handle_mm_fault.handle_mm_fault.do_user_addr_fault.exc_page_fault 4.14 ± 72% -4.1 0.00 perf-profile.calltrace.cycles-pp.rep_movs_alternative.copyin.copy_page_from_iter_atomic.generic_perform_write.__generic_file_write_iter 12.06 ± 49% -3.2 8.89 ±147% perf-profile.calltrace.cycles-pp.__handle_mm_fault.handle_mm_fault.do_user_addr_fault.exc_page_fault.asm_exc_page_fault 4.14 ± 72% -3.1 1.04 ±223% perf-profile.calltrace.cycles-pp.copyin.copy_page_from_iter_atomic.generic_perform_write.__generic_file_write_iter.generic_file_write_iter 4.14 ± 72% -3.1 1.04 ±223% perf-profile.calltrace.cycles-pp.copy_page_from_iter_atomic.generic_perform_write.__generic_file_write_iter.generic_file_write_iter.vfs_write 7.45 ± 64% -3.1 4.38 ±168% perf-profile.calltrace.cycles-pp.__mmput.exit_mm.do_exit.do_group_exit.__x64_sys_exit_group 7.45 ± 64% -3.1 4.38 ±168% perf-profile.calltrace.cycles-pp.exit_mm.do_exit.do_group_exit.__x64_sys_exit_group.do_syscall_64 7.80 ± 61% -2.6 5.21 ±175% perf-profile.calltrace.cycles-pp.load_elf_binary.search_binary_handler.exec_binprm.bprm_execve.do_execveat_common 11.40 ± 23% -2.0 9.38 ±195% perf-profile.calltrace.cycles-pp.__x64_sys_execve.do_syscall_64.entry_SYSCALL_64_after_hwframe 11.40 ± 23% -2.0 9.38 ±195% perf-profile.calltrace.cycles-pp.do_execveat_common.__x64_sys_execve.do_syscall_64.entry_SYSCALL_64_after_hwframe 9.98 ± 54% -1.1 8.89 ±147% perf-profile.calltrace.cycles-pp.asm_exc_page_fault 9.98 ± 54% -1.1 8.89 ±147% perf-profile.calltrace.cycles-pp.exc_page_fault.asm_exc_page_fault 9.98 ± 54% -1.1 8.89 ±147% perf-profile.calltrace.cycles-pp.do_user_addr_fault.exc_page_fault.asm_exc_page_fault 10.21 ± 27% -0.8 9.38 ±195% perf-profile.calltrace.cycles-pp.bprm_execve.do_execveat_common.__x64_sys_execve.do_syscall_64.entry_SYSCALL_64_after_hwframe 10.21 ± 27% -0.8 9.38 ±195% perf-profile.calltrace.cycles-pp.exec_binprm.bprm_execve.do_execveat_common.__x64_sys_execve.do_syscall_64 10.21 ± 27% -0.8 9.38 ±195% perf-profile.calltrace.cycles-pp.search_binary_handler.exec_binprm.bprm_execve.do_execveat_common.__x64_sys_execve 4.14 ± 72% -0.8 3.33 ±223% perf-profile.calltrace.cycles-pp.wp_page_copy.__handle_mm_fault.handle_mm_fault.do_user_addr_fault.exc_page_fault 9.46 ± 62% -0.6 8.89 ±147% perf-profile.calltrace.cycles-pp.handle_mm_fault.do_user_addr_fault.exc_page_fault.asm_exc_page_fault 5.72 ± 87% -0.5 5.21 ±175% perf-profile.calltrace.cycles-pp.begin_new_exec.load_elf_binary.search_binary_handler.exec_binprm.bprm_execve 7.60 ± 63% +0.1 7.71 ±189% perf-profile.calltrace.cycles-pp.zap_pte_range.zap_pmd_range.unmap_page_range.unmap_vmas.exit_mmap 7.60 ± 63% +0.1 7.71 ±189% perf-profile.calltrace.cycles-pp.zap_pmd_range.unmap_page_range.unmap_vmas.exit_mmap.__mmput 7.34 ± 50% +0.2 7.50 ±142% perf-profile.calltrace.cycles-pp.exit_mm.do_exit.do_group_exit.get_signal.arch_do_signal_or_restart 7.34 ± 50% +0.2 7.50 ±142% perf-profile.calltrace.cycles-pp.__mmput.exit_mm.do_exit.do_group_exit.get_signal 6.21 ± 70% +1.5 7.71 ±189% perf-profile.calltrace.cycles-pp.unmap_vmas.exit_mmap.__mmput.exit_mm.do_exit 6.21 ± 70% +1.5 7.71 ±189% perf-profile.calltrace.cycles-pp.unmap_page_range.unmap_vmas.exit_mmap.__mmput.exit_mm 35.81 ± 5% -24.0 11.80 ±141% perf-profile.children.cycles-pp.cpuidle_idle_call 35.81 ± 5% -24.0 11.80 ±141% perf-profile.children.cycles-pp.cpuidle_enter 35.81 ± 5% -24.0 11.80 ±141% perf-profile.children.cycles-pp.cpuidle_enter_state 36.61 ± 4% -23.8 12.85 ±143% perf-profile.children.cycles-pp.secondary_startup_64_no_verify 36.61 ± 4% -23.8 12.85 ±143% perf-profile.children.cycles-pp.cpu_startup_entry 36.61 ± 4% -23.8 12.85 ±143% perf-profile.children.cycles-pp.do_idle 34.87 ± 7% -22.0 12.85 ±143% perf-profile.children.cycles-pp.start_secondary 31.16 ± 14% -19.4 11.80 ±141% perf-profile.children.cycles-pp.intel_idle 5.67 ± 72% -5.7 0.00 perf-profile.children.cycles-pp.do_fault 5.18 ± 83% -5.2 0.00 perf-profile.children.cycles-pp.rep_movs_alternative 12.58 ± 44% -3.7 8.89 ±147% perf-profile.children.cycles-pp.exc_page_fault 12.58 ± 44% -3.7 8.89 ±147% perf-profile.children.cycles-pp.do_user_addr_fault 12.06 ± 49% -3.2 8.89 ±147% perf-profile.children.cycles-pp.__handle_mm_fault 12.06 ± 49% -3.2 8.89 ±147% perf-profile.children.cycles-pp.handle_mm_fault 4.14 ± 72% -3.1 1.04 ±223% perf-profile.children.cycles-pp.copyin 4.14 ± 72% -3.1 1.04 ±223% perf-profile.children.cycles-pp.copy_page_from_iter_atomic 7.80 ± 61% -2.6 5.21 ±175% perf-profile.children.cycles-pp.load_elf_binary 11.40 ± 23% -2.0 9.38 ±195% perf-profile.children.cycles-pp.__x64_sys_execve 11.40 ± 23% -2.0 9.38 ±195% perf-profile.children.cycles-pp.do_execveat_common 10.21 ± 27% -0.8 9.38 ±195% perf-profile.children.cycles-pp.bprm_execve 10.21 ± 27% -0.8 9.38 ±195% perf-profile.children.cycles-pp.exec_binprm 10.21 ± 27% -0.8 9.38 ±195% perf-profile.children.cycles-pp.search_binary_handler 4.14 ± 72% -0.8 3.33 ±223% perf-profile.children.cycles-pp.wp_page_copy 5.72 ± 87% -0.5 5.21 ±175% perf-profile.children.cycles-pp.begin_new_exec 8.12 ± 54% -0.4 7.71 ±189% perf-profile.children.cycles-pp.zap_pte_range 8.12 ± 54% -0.4 7.71 ±189% perf-profile.children.cycles-pp.unmap_vmas 8.12 ± 54% -0.4 7.71 ±189% perf-profile.children.cycles-pp.unmap_page_range 8.12 ± 54% -0.4 7.71 ±189% perf-profile.children.cycles-pp.zap_pmd_range 4.26 ± 74% +1.3 5.56 ±223% perf-profile.children.cycles-pp.task_work_run 31.16 ± 14% -19.4 11.80 ±141% perf-profile.self.cycles-pp.intel_idle 4.14 ± 72% -4.1 0.00 perf-profile.self.cycles-pp.rep_movs_alternative *************************************************************************************************** lkp-cfl-d2: 12 threads 1 sockets Intel(R) Core(TM) i7-8700 CPU @ 3.20GHz (Coffee Lake) with 32G memory ========================================================================================= compiler/cpufreq_governor/kconfig/need_x/option_a/rootfs/tbox_group/test/testcase: gcc-11/performance/x86_64-rhel-8.3/true/500px PutImage Square/debian-x86_64-phoronix/lkp-cfl-d2/x11perf-1.1.1/phoronix-test-suite commit: 0d85b27b0c ("Merge tag '6.4-rc3-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6") 47ee3f1dd9 ("x86: re-introduce support for ERMS copies for user space accesses") 0d85b27b0cc6b5cf 47ee3f1dd93bcbe031539b1ecda ---------------- --------------------------- %stddev %change %stddev \ | \ 59773 -3.3% 57826 proc-vmstat.pgreuse 9554 ± 11% -26.3% 7038 ± 20% sched_debug.cfs_rq:/.min_vruntime.min 237.81 ± 14% -15.9% 199.90 ± 2% uptime.boot 2378 ± 12% -14.6% 2031 ± 2% uptime.idle 1992 +6.1% 2114 vmstat.io.bi 18164 +2.3% 18590 vmstat.system.in 150.06 -5.7% 141.56 phoronix-test-suite.time.elapsed_time 150.06 -5.7% 141.56 phoronix-test-suite.time.elapsed_time.max 108.51 -9.2% 98.54 phoronix-test-suite.time.system_time 5626 +8.4% 6101 phoronix-test-suite.x11perf.500pxPutImageSquare.operations___second 1.90 ± 2% -0.2 1.66 ± 3% turbostat.C1% 2.85 +0.2 3.00 turbostat.C1E% 0.07 +0.0 0.08 turbostat.CPUGFX% 4.27 ± 4% -22.1% 3.33 turbostat.Pkg%pc2 1.72 +1.6% 1.75 turbostat.RAMWatt 40.01 ± 3% -3.4 36.60 ± 4% perf-profile.calltrace.cycles-pp.cpuidle_enter_state.cpuidle_enter.cpuidle_idle_call.do_idle.cpu_startup_entry 1.76 ± 4% -0.7 1.04 ± 21% perf-profile.calltrace.cycles-pp.__sysvec_apic_timer_interrupt.sysvec_apic_timer_interrupt.asm_sysvec_apic_timer_interrupt.cpuidle_enter_state.cpuidle_enter 1.71 ± 3% -0.7 1.00 ± 21% perf-profile.calltrace.cycles-pp.hrtimer_interrupt.__sysvec_apic_timer_interrupt.sysvec_apic_timer_interrupt.asm_sysvec_apic_timer_interrupt.cpuidle_enter_state 1.40 ± 2% -0.6 0.80 ± 19% perf-profile.calltrace.cycles-pp.__hrtimer_run_queues.hrtimer_interrupt.__sysvec_apic_timer_interrupt.sysvec_apic_timer_interrupt.asm_sysvec_apic_timer_interrupt 42.14 ± 3% -3.6 38.57 ± 4% perf-profile.children.cycles-pp.cpuidle_idle_call 40.22 ± 3% -3.5 36.72 ± 4% perf-profile.children.cycles-pp.cpuidle_enter_state 40.25 ± 3% -3.5 36.76 ± 4% perf-profile.children.cycles-pp.cpuidle_enter 2.08 ± 3% -0.8 1.30 ± 18% perf-profile.children.cycles-pp.__sysvec_apic_timer_interrupt 2.01 ± 4% -0.8 1.24 ± 18% perf-profile.children.cycles-pp.hrtimer_interrupt 1.66 ± 4% -0.6 1.01 ± 17% perf-profile.children.cycles-pp.__hrtimer_run_queues 1.05 ± 7% -0.4 0.60 ± 20% perf-profile.children.cycles-pp.tick_sched_timer 0.90 ± 9% -0.4 0.54 ± 19% perf-profile.children.cycles-pp.tick_sched_handle 0.78 ± 11% -0.3 0.50 ± 18% perf-profile.children.cycles-pp.update_process_times 0.44 ± 11% -0.2 0.27 ± 21% perf-profile.children.cycles-pp.scheduler_tick 0.26 ± 18% -0.1 0.14 ± 30% perf-profile.children.cycles-pp.tick_irq_enter 0.26 ± 18% -0.1 0.14 ± 28% perf-profile.children.cycles-pp.irq_enter_rcu 0.14 ± 13% -0.1 0.06 ± 54% perf-profile.children.cycles-pp.rcu_sched_clock_irq 0.11 ± 18% -0.1 0.04 ± 71% perf-profile.children.cycles-pp.update_irq_load_avg 0.23 ± 18% -0.1 0.16 ± 13% perf-profile.children.cycles-pp.perf_mux_hrtimer_handler 0.11 ± 21% -0.1 0.04 ± 72% perf-profile.children.cycles-pp.rcu_pending 0.18 ± 16% -0.1 0.11 ± 31% perf-profile.children.cycles-pp.update_rq_clock_task 0.09 ± 17% -0.1 0.03 ±101% perf-profile.children.cycles-pp.tick_check_oneshot_broadcast_this_cpu 0.16 ± 17% -0.0 0.12 ± 17% perf-profile.children.cycles-pp.perf_rotate_context 0.10 ± 10% -0.0 0.06 ± 53% perf-profile.children.cycles-pp.tick_nohz_stop_idle 0.64 ± 17% +0.2 0.81 ± 8% perf-profile.children.cycles-pp._raw_spin_lock 0.34 ± 18% -0.1 0.24 ± 32% perf-profile.self.cycles-pp.cpuidle_enter_state 0.11 ± 18% -0.1 0.04 ± 71% perf-profile.self.cycles-pp.update_irq_load_avg 0.12 ± 33% -0.1 0.06 ± 52% perf-profile.self.cycles-pp.__hrtimer_run_queues 0.08 ± 14% -0.1 0.03 ±101% perf-profile.self.cycles-pp.tick_check_oneshot_broadcast_this_cpu 0.62 ± 18% +0.2 0.80 ± 7% perf-profile.self.cycles-pp._raw_spin_lock 184.41 +40.9% 259.89 ± 2% perf-stat.i.MPKI 8.998e+08 -9.4% 8.148e+08 perf-stat.i.branch-instructions 1.83 ± 2% +0.1 1.94 ± 2% perf-stat.i.branch-miss-rate% 15153824 ± 5% +6.2% 16088855 ± 3% perf-stat.i.branch-misses 1.497e+09 -7.3% 1.388e+09 perf-stat.i.cache-references 1.29 ± 3% +33.7% 1.73 perf-stat.i.cpi 7.394e+09 +1.3% 7.488e+09 perf-stat.i.cpu-cycles 8076 ± 3% +6.3% 8588 perf-stat.i.cycles-between-cache-misses 2.321e+09 -42.5% 1.334e+09 perf-stat.i.dTLB-loads 1.992e+09 -50.7% 9.825e+08 perf-stat.i.dTLB-stores 7.709e+09 -32.2% 5.227e+09 perf-stat.i.instructions 15364 ± 14% -28.2% 11033 ± 11% perf-stat.i.instructions-per-iTLB-miss 0.99 -32.0% 0.68 perf-stat.i.ipc 2.37 ± 2% +6.3% 2.53 ± 2% perf-stat.i.major-faults 0.62 +1.3% 0.62 perf-stat.i.metric.GHz 559.06 -32.6% 376.59 perf-stat.i.metric.M/sec 2727 +3.0% 2810 perf-stat.i.minor-faults 176938 ± 4% +8.6% 192161 perf-stat.i.node-stores 2730 +3.0% 2813 perf-stat.i.page-faults 194.17 +36.8% 265.59 ± 2% perf-stat.overall.MPKI 1.68 ± 4% +0.3 1.97 ± 2% perf-stat.overall.branch-miss-rate% 0.17 ± 4% +0.0 0.19 ± 4% perf-stat.overall.cache-miss-rate% 0.96 +49.4% 1.43 perf-stat.overall.cpi 0.03 ± 5% +0.0 0.05 ± 5% perf-stat.overall.dTLB-load-miss-rate% 0.00 ± 17% +0.0 0.01 ± 28% perf-stat.overall.dTLB-store-miss-rate% 11594 ± 19% -31.3% 7967 ± 11% perf-stat.overall.instructions-per-iTLB-miss 1.04 -33.0% 0.70 perf-stat.overall.ipc 8.938e+08 -9.5% 8.09e+08 perf-stat.ps.branch-instructions 15056160 ± 5% +6.1% 15976861 ± 3% perf-stat.ps.branch-misses 1.487e+09 -7.3% 1.378e+09 perf-stat.ps.cache-references 7.345e+09 +1.2% 7.435e+09 perf-stat.ps.cpu-cycles 2.305e+09 -42.5% 1.325e+09 perf-stat.ps.dTLB-loads 1.978e+09 -50.7% 9.756e+08 perf-stat.ps.dTLB-stores 7.657e+09 -32.2% 5.19e+09 perf-stat.ps.instructions 2.36 ± 2% +6.2% 2.51 ± 2% perf-stat.ps.major-faults 2710 +3.0% 2791 perf-stat.ps.minor-faults 175787 ± 4% +8.6% 190826 perf-stat.ps.node-stores 2712 +3.0% 2793 perf-stat.ps.page-faults 1.158e+12 -36.1% 7.395e+11 perf-stat.total.instructions *************************************************************************************************** lkp-skl-fpga01: 104 threads 2 sockets (Skylake) with 192G memory ========================================================================================= compiler/cpufreq_governor/kconfig/mode/nr_task/rootfs/tbox_group/test/testcase: gcc-11/performance/x86_64-rhel-8.3/thread/50%/debian-11.1-x86_64-20220510.cgz/lkp-skl-fpga01/pread1/will-it-scale commit: 0d85b27b0c ("Merge tag '6.4-rc3-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6") 47ee3f1dd9 ("x86: re-introduce support for ERMS copies for user space accesses") 0d85b27b0cc6b5cf 47ee3f1dd93bcbe031539b1ecda ---------------- --------------------------- %stddev %change %stddev \ | \ 19.54 +2.8 22.30 mpstat.cpu.all.usr% 117438 ± 69% -66.1% 39831 ± 4% numa-meminfo.node1.FilePages 29350 ± 69% -66.1% 9943 ± 4% numa-vmstat.node1.nr_file_pages 0.23 -43.5% 0.13 turbostat.IPC 43809356 +14.1% 49998501 will-it-scale.52.threads 842487 +14.1% 961509 will-it-scale.per_thread_ops 43809356 +14.1% 49998501 will-it-scale.workload 264681 ± 21% -23.5% 202416 ± 2% sched_debug.cpu.clock.avg 264689 ± 21% -23.5% 202424 ± 2% sched_debug.cpu.clock.max 264673 ± 21% -23.5% 202407 ± 2% sched_debug.cpu.clock.min 260519 ± 21% -23.0% 200690 ± 3% sched_debug.cpu.clock_task.avg 262320 ± 21% -23.3% 201218 ± 2% sched_debug.cpu.clock_task.max 566960 ± 12% -11.5% 501642 sched_debug.cpu.max_idle_balance_cost.max 8195 ±120% -98.0% 160.33 ±186% sched_debug.cpu.max_idle_balance_cost.stddev 264673 ± 21% -23.5% 202407 ± 2% sched_debug.cpu_clk 264068 ± 21% -23.6% 201802 ± 2% sched_debug.ktime 0.04 +68.4% 0.07 ± 2% perf-stat.i.MPKI 1.447e+10 -8.8% 1.321e+10 perf-stat.i.branch-instructions 0.94 +0.2 1.17 perf-stat.i.branch-miss-rate% 1.36e+08 +13.6% 1.544e+08 perf-stat.i.branch-misses 1.22 +73.6% 2.12 perf-stat.i.cpi 0.11 +0.1 0.24 perf-stat.i.dTLB-load-miss-rate% 43870070 +14.1% 50047738 perf-stat.i.dTLB-load-misses 4.026e+10 -49.3% 2.04e+10 perf-stat.i.dTLB-loads 0.00 +0.0 0.00 perf-stat.i.dTLB-store-miss-rate% 32603 +8.4% 35341 perf-stat.i.dTLB-store-misses 3.274e+10 -63.9% 1.183e+10 perf-stat.i.dTLB-stores 63088748 ± 2% +13.6% 71686582 perf-stat.i.iTLB-load-misses 1.19e+11 -42.4% 6.854e+10 perf-stat.i.instructions 1900 ± 2% -49.3% 963.97 perf-stat.i.instructions-per-iTLB-miss 0.82 -42.4% 0.47 perf-stat.i.ipc 841.01 -48.0% 436.92 perf-stat.i.metric.M/sec 0.04 +68.7% 0.07 perf-stat.overall.MPKI 0.94 +0.2 1.17 perf-stat.overall.branch-miss-rate% 1.22 +73.7% 2.12 perf-stat.overall.cpi 0.11 +0.1 0.24 perf-stat.overall.dTLB-load-miss-rate% 0.00 +0.0 0.00 perf-stat.overall.dTLB-store-miss-rate% 1888 ± 2% -49.3% 957.20 perf-stat.overall.instructions-per-iTLB-miss 0.82 -42.4% 0.47 perf-stat.overall.ipc 817740 -49.6% 412119 perf-stat.overall.path-length 1.442e+10 -8.8% 1.316e+10 perf-stat.ps.branch-instructions 1.355e+08 +13.6% 1.539e+08 perf-stat.ps.branch-misses 43722670 +14.1% 49880640 perf-stat.ps.dTLB-load-misses 4.013e+10 -49.3% 2.034e+10 perf-stat.ps.dTLB-loads 32528 +8.4% 35259 perf-stat.ps.dTLB-store-misses 3.263e+10 -63.9% 1.179e+10 perf-stat.ps.dTLB-stores 62833990 ± 2% +13.6% 71389669 perf-stat.ps.iTLB-load-misses 1.186e+11 -42.4% 6.832e+10 perf-stat.ps.instructions 3.582e+13 -42.5% 2.061e+13 perf-stat.total.instructions 12.31 -8.9 3.41 perf-profile.calltrace.cycles-pp.rep_movs_alternative.copyout._copy_to_iter.copy_page_to_iter.shmem_file_read_iter 12.74 -8.8 3.94 perf-profile.calltrace.cycles-pp.copyout._copy_to_iter.copy_page_to_iter.shmem_file_read_iter.vfs_read 13.30 -8.7 4.58 perf-profile.calltrace.cycles-pp._copy_to_iter.copy_page_to_iter.shmem_file_read_iter.vfs_read.__x64_sys_pread64 13.59 -8.7 4.90 perf-profile.calltrace.cycles-pp.copy_page_to_iter.shmem_file_read_iter.vfs_read.__x64_sys_pread64.do_syscall_64 20.90 -7.8 13.12 perf-profile.calltrace.cycles-pp.shmem_file_read_iter.vfs_read.__x64_sys_pread64.do_syscall_64.entry_SYSCALL_64_after_hwframe 25.72 -7.2 18.51 perf-profile.calltrace.cycles-pp.vfs_read.__x64_sys_pread64.do_syscall_64.entry_SYSCALL_64_after_hwframe.__libc_pread64 28.82 -6.8 22.05 perf-profile.calltrace.cycles-pp.__x64_sys_pread64.do_syscall_64.entry_SYSCALL_64_after_hwframe.__libc_pread64 43.04 -5.0 38.08 perf-profile.calltrace.cycles-pp.do_syscall_64.entry_SYSCALL_64_after_hwframe.__libc_pread64 50.67 -3.9 46.78 perf-profile.calltrace.cycles-pp.entry_SYSCALL_64_after_hwframe.__libc_pread64 0.54 ± 2% +0.1 0.61 ± 2% perf-profile.calltrace.cycles-pp.folio_unlock.shmem_file_read_iter.vfs_read.__x64_sys_pread64.do_syscall_64 0.54 +0.1 0.62 ± 2% perf-profile.calltrace.cycles-pp.syscall_enter_from_user_mode.do_syscall_64.entry_SYSCALL_64_after_hwframe.__libc_pread64 1.13 ± 2% +0.1 1.26 perf-profile.calltrace.cycles-pp.filemap_get_entry.shmem_get_folio_gfp.shmem_file_read_iter.vfs_read.__x64_sys_pread64 0.98 +0.1 1.12 perf-profile.calltrace.cycles-pp.__fsnotify_parent.vfs_read.__x64_sys_pread64.do_syscall_64.entry_SYSCALL_64_after_hwframe 1.24 +0.1 1.38 ± 2% perf-profile.calltrace.cycles-pp.atime_needs_update.touch_atime.shmem_file_read_iter.vfs_read.__x64_sys_pread64 1.59 +0.2 1.77 ± 2% perf-profile.calltrace.cycles-pp.touch_atime.shmem_file_read_iter.vfs_read.__x64_sys_pread64.do_syscall_64 0.35 ± 70% +0.2 0.58 ± 3% perf-profile.calltrace.cycles-pp.current_time.atime_needs_update.touch_atime.shmem_file_read_iter.vfs_read 1.60 +0.3 1.87 ± 5% perf-profile.calltrace.cycles-pp.__fget_light.__x64_sys_pread64.do_syscall_64.entry_SYSCALL_64_after_hwframe.__libc_pread64 1.99 ± 2% +0.3 2.27 perf-profile.calltrace.cycles-pp.entry_SYSCALL_64_safe_stack.__libc_pread64 2.67 +0.3 3.00 perf-profile.calltrace.cycles-pp.shmem_get_folio_gfp.shmem_file_read_iter.vfs_read.__x64_sys_pread64.do_syscall_64 0.00 +0.5 0.53 ± 2% perf-profile.calltrace.cycles-pp.fput.__x64_sys_pread64.do_syscall_64.entry_SYSCALL_64_after_hwframe.__libc_pread64 6.78 +0.9 7.70 perf-profile.calltrace.cycles-pp.__entry_text_start.__libc_pread64 12.88 +1.7 14.56 perf-profile.calltrace.cycles-pp.syscall_exit_to_user_mode.do_syscall_64.entry_SYSCALL_64_after_hwframe.__libc_pread64 14.10 +1.9 15.96 perf-profile.calltrace.cycles-pp.syscall_return_via_sysret.__libc_pread64 12.31 -8.9 3.44 perf-profile.children.cycles-pp.rep_movs_alternative 13.04 -8.8 4.28 perf-profile.children.cycles-pp.copyout 13.32 -8.7 4.61 perf-profile.children.cycles-pp._copy_to_iter 13.61 -8.7 4.92 perf-profile.children.cycles-pp.copy_page_to_iter 21.00 -7.8 13.23 perf-profile.children.cycles-pp.shmem_file_read_iter 25.82 -7.2 18.62 perf-profile.children.cycles-pp.vfs_read 28.83 -6.8 22.07 perf-profile.children.cycles-pp.__x64_sys_pread64 43.17 -4.9 38.24 perf-profile.children.cycles-pp.do_syscall_64 50.98 -3.9 47.12 perf-profile.children.cycles-pp.entry_SYSCALL_64_after_hwframe 0.11 ± 3% +0.0 0.13 ± 3% perf-profile.children.cycles-pp.folio_mark_accessed 0.11 +0.0 0.12 ± 4% perf-profile.children.cycles-pp.__pthread_enable_asynccancel 0.09 ± 4% +0.0 0.10 ± 4% perf-profile.children.cycles-pp.rw_verify_area 0.15 ± 3% +0.0 0.18 ± 2% perf-profile.children.cycles-pp.__cond_resched 0.34 ± 2% +0.0 0.38 ± 2% perf-profile.children.cycles-pp.folio_test_hugetlb 0.24 ± 12% +0.0 0.29 ± 2% perf-profile.children.cycles-pp.aa_file_perm 0.17 ± 10% +0.1 0.23 ± 21% perf-profile.children.cycles-pp.syscall_exit_to_user_mode_prepare 0.47 +0.1 0.53 ± 2% perf-profile.children.cycles-pp.fput 0.54 ± 2% +0.1 0.61 ± 2% perf-profile.children.cycles-pp.folio_unlock 0.57 +0.1 0.65 ± 2% perf-profile.children.cycles-pp.syscall_enter_from_user_mode 0.58 ± 2% +0.1 0.66 ± 4% perf-profile.children.cycles-pp.current_time 1.16 ± 2% +0.1 1.29 perf-profile.children.cycles-pp.filemap_get_entry 0.99 +0.1 1.13 perf-profile.children.cycles-pp.__fsnotify_parent 1.03 +0.1 1.18 perf-profile.children.cycles-pp.entry_SYSRETQ_unsafe_stack 1.28 +0.2 1.44 ± 2% perf-profile.children.cycles-pp.atime_needs_update 1.24 ± 2% +0.2 1.41 ± 2% perf-profile.children.cycles-pp.entry_SYSCALL_64_safe_stack 1.61 +0.2 1.80 ± 2% perf-profile.children.cycles-pp.touch_atime 1.60 +0.3 1.87 ± 5% perf-profile.children.cycles-pp.__fget_light 2.70 +0.3 3.04 perf-profile.children.cycles-pp.shmem_get_folio_gfp 6.68 +0.9 7.58 perf-profile.children.cycles-pp.__entry_text_start 12.97 +1.7 14.63 perf-profile.children.cycles-pp.syscall_exit_to_user_mode 14.24 +1.9 16.11 perf-profile.children.cycles-pp.syscall_return_via_sysret 12.09 -8.9 3.23 perf-profile.self.cycles-pp.rep_movs_alternative 0.11 +0.0 0.12 ± 3% perf-profile.self.cycles-pp.__pthread_enable_asynccancel 0.09 ± 4% +0.0 0.10 ± 3% perf-profile.self.cycles-pp.rw_verify_area 0.14 ± 2% +0.0 0.16 ± 4% perf-profile.self.cycles-pp.testcase 0.36 ± 4% +0.0 0.39 ± 2% perf-profile.self.cycles-pp.touch_atime 0.34 ± 2% +0.0 0.38 ± 2% perf-profile.self.cycles-pp.folio_test_hugetlb 0.22 ± 14% +0.0 0.26 ± 3% perf-profile.self.cycles-pp.aa_file_perm 0.30 ± 4% +0.0 0.35 ± 2% perf-profile.self.cycles-pp._copy_to_iter 0.46 ± 2% +0.0 0.52 perf-profile.self.cycles-pp.current_time 0.47 +0.1 0.52 perf-profile.self.cycles-pp.fput 0.57 ± 3% +0.1 0.63 ± 2% perf-profile.self.cycles-pp.do_syscall_64 0.48 +0.1 0.55 ± 3% perf-profile.self.cycles-pp.entry_SYSCALL_64_safe_stack 0.50 ± 2% +0.1 0.57 ± 2% perf-profile.self.cycles-pp.syscall_enter_from_user_mode 0.58 ± 2% +0.1 0.65 ± 2% perf-profile.self.cycles-pp.atime_needs_update 0.54 ± 2% +0.1 0.61 ± 2% perf-profile.self.cycles-pp.folio_unlock 0.13 ± 12% +0.1 0.20 ± 24% perf-profile.self.cycles-pp.syscall_exit_to_user_mode_prepare 0.91 +0.1 1.01 perf-profile.self.cycles-pp.__x64_sys_pread64 0.95 ± 2% +0.1 1.06 perf-profile.self.cycles-pp.filemap_get_entry 0.90 +0.1 1.03 perf-profile.self.cycles-pp.entry_SYSRETQ_unsafe_stack 0.89 +0.1 1.02 ± 2% perf-profile.self.cycles-pp.copyout 0.94 +0.1 1.07 perf-profile.self.cycles-pp.__fsnotify_parent 1.24 +0.2 1.42 perf-profile.self.cycles-pp.__libc_pread64 1.37 +0.2 1.56 perf-profile.self.cycles-pp.shmem_get_folio_gfp 1.76 +0.2 2.00 ± 2% perf-profile.self.cycles-pp.vfs_read 2.13 ± 2% +0.3 2.39 perf-profile.self.cycles-pp.shmem_file_read_iter 1.59 +0.3 1.86 ± 5% perf-profile.self.cycles-pp.__fget_light 5.81 +0.8 6.60 perf-profile.self.cycles-pp.__entry_text_start 8.05 +1.1 9.16 perf-profile.self.cycles-pp.entry_SYSCALL_64_after_hwframe 12.47 +1.6 14.06 perf-profile.self.cycles-pp.syscall_exit_to_user_mode 14.21 +1.9 16.08 perf-profile.self.cycles-pp.syscall_return_via_sysret *************************************************************************************************** lkp-skl-fpga01: 104 threads 2 sockets (Skylake) with 192G memory ========================================================================================= compiler/cpufreq_governor/kconfig/mode/nr_task/rootfs/tbox_group/test/testcase: gcc-11/performance/x86_64-rhel-8.3/thread/100%/debian-11.1-x86_64-20220510.cgz/lkp-skl-fpga01/readseek1/will-it-scale commit: 0d85b27b0c ("Merge tag '6.4-rc3-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6") 47ee3f1dd9 ("x86: re-introduce support for ERMS copies for user space accesses") 0d85b27b0cc6b5cf 47ee3f1dd93bcbe031539b1ecda ---------------- --------------------------- %stddev %change %stddev \ | \ 0.01 ± 3% +0.0 0.01 ± 4% mpstat.cpu.all.soft% 53591 -2.8% 52083 proc-vmstat.pgactivate 0.10 -40.0% 0.06 turbostat.IPC 5.12 ± 38% +1410.7% 77.33 ±138% sched_debug.cfs_rq:/.removed.load_avg.avg 28.32 ± 18% +2025.5% 601.98 ±136% sched_debug.cfs_rq:/.removed.load_avg.stddev 30834962 +7.8% 33231698 will-it-scale.104.threads 296489 +7.8% 319535 will-it-scale.per_thread_ops 30834962 +7.8% 33231698 will-it-scale.workload 0.04 ± 5% +295.1% 0.17 ±129% perf-stat.i.MPKI 1.331e+10 -8.9% 1.212e+10 perf-stat.i.branch-instructions 1.38 +0.3 1.71 perf-stat.i.branch-miss-rate% 1.834e+08 +12.5% 2.063e+08 perf-stat.i.branch-misses 2.90 +59.9% 4.64 perf-stat.i.cpi 165.59 +2.2% 169.19 perf-stat.i.cpu-migrations 0.19 +0.2 0.37 perf-stat.i.dTLB-load-miss-rate% 61714012 +7.5% 66344685 perf-stat.i.dTLB-load-misses 3.26e+10 -44.4% 1.812e+10 perf-stat.i.dTLB-loads 0.00 +0.0 0.00 ± 50% perf-stat.i.dTLB-store-miss-rate% 60456 +2.3% 61857 perf-stat.i.dTLB-store-misses 2.576e+10 -58.1% 1.079e+10 perf-stat.i.dTLB-stores 92796911 ± 4% +14.6% 1.063e+08 perf-stat.i.iTLB-load-misses 79032012 ± 3% +16.7% 92191583 ± 2% perf-stat.i.iTLB-loads 9.913e+10 -37.5% 6.199e+10 perf-stat.i.instructions 1073 ± 4% -45.4% 585.88 perf-stat.i.instructions-per-iTLB-miss 0.34 -37.4% 0.22 perf-stat.i.ipc 784.59 ± 4% +9.7% 860.49 ± 4% perf-stat.i.metric.K/sec 689.14 -42.7% 394.59 perf-stat.i.metric.M/sec 10907 ± 6% -8.6% 9974 ± 6% perf-stat.i.node-stores 0.04 ± 5% +68.8% 0.07 ± 7% perf-stat.overall.MPKI 1.38 +0.3 1.70 perf-stat.overall.branch-miss-rate% 2.90 +59.8% 4.64 perf-stat.overall.cpi 0.19 +0.2 0.36 perf-stat.overall.dTLB-load-miss-rate% 0.00 +0.0 0.00 perf-stat.overall.dTLB-store-miss-rate% 1070 ± 4% -45.5% 583.23 perf-stat.overall.instructions-per-iTLB-miss 0.34 -37.4% 0.22 perf-stat.overall.ipc 967334 -41.9% 562134 perf-stat.overall.path-length 1.326e+10 -8.9% 1.208e+10 perf-stat.ps.branch-instructions 1.828e+08 +12.5% 2.056e+08 perf-stat.ps.branch-misses 165.01 +2.1% 168.48 perf-stat.ps.cpu-migrations 61509195 +7.5% 66124018 perf-stat.ps.dTLB-load-misses 3.249e+10 -44.4% 1.806e+10 perf-stat.ps.dTLB-loads 60319 +2.3% 61694 perf-stat.ps.dTLB-store-misses 2.568e+10 -58.1% 1.075e+10 perf-stat.ps.dTLB-stores 92483996 ± 4% +14.6% 1.059e+08 perf-stat.ps.iTLB-load-misses 78775497 ± 3% +16.6% 91888236 ± 2% perf-stat.ps.iTLB-loads 9.88e+10 -37.5% 6.179e+10 perf-stat.ps.instructions 2.983e+13 -37.4% 1.868e+13 perf-stat.total.instructions 32.98 -7.8 25.20 perf-profile.calltrace.cycles-pp.ksys_read.do_syscall_64.entry_SYSCALL_64_after_hwframe.__libc_read 43.20 -7.4 35.79 perf-profile.calltrace.cycles-pp.do_syscall_64.entry_SYSCALL_64_after_hwframe.__libc_read 26.83 -6.8 20.03 perf-profile.calltrace.cycles-pp.vfs_read.ksys_read.do_syscall_64.entry_SYSCALL_64_after_hwframe.__libc_read 48.41 -6.1 42.34 perf-profile.calltrace.cycles-pp.entry_SYSCALL_64_after_hwframe.__libc_read 9.66 -5.8 3.90 ± 2% perf-profile.calltrace.cycles-pp.rep_movs_alternative.copyout._copy_to_iter.copy_page_to_iter.shmem_file_read_iter 20.63 -5.5 15.14 perf-profile.calltrace.cycles-pp.shmem_file_read_iter.vfs_read.ksys_read.do_syscall_64.entry_SYSCALL_64_after_hwframe 10.11 -5.1 4.96 ± 5% perf-profile.calltrace.cycles-pp.copyout._copy_to_iter.copy_page_to_iter.shmem_file_read_iter.vfs_read 10.70 -5.0 5.69 ± 2% perf-profile.calltrace.cycles-pp._copy_to_iter.copy_page_to_iter.shmem_file_read_iter.vfs_read.ksys_read 11.03 -5.0 6.07 ± 2% perf-profile.calltrace.cycles-pp.copy_page_to_iter.shmem_file_read_iter.vfs_read.ksys_read.do_syscall_64 64.33 -4.1 60.23 perf-profile.calltrace.cycles-pp.__libc_read 3.50 -0.6 2.92 ± 2% perf-profile.calltrace.cycles-pp.shmem_get_folio_gfp.shmem_file_read_iter.vfs_read.ksys_read.do_syscall_64 1.24 ± 4% -0.6 0.67 ± 3% perf-profile.calltrace.cycles-pp.mutex_unlock.ksys_read.do_syscall_64.entry_SYSCALL_64_after_hwframe.__libc_read 1.34 ± 2% -0.5 0.79 ± 3% perf-profile.calltrace.cycles-pp.fput.ksys_read.do_syscall_64.entry_SYSCALL_64_after_hwframe.__libc_read 1.30 ± 2% -0.5 0.84 perf-profile.calltrace.cycles-pp.__fsnotify_parent.vfs_read.ksys_read.do_syscall_64.entry_SYSCALL_64_after_hwframe 1.19 ± 2% -0.3 0.86 perf-profile.calltrace.cycles-pp.filemap_get_entry.shmem_get_folio_gfp.shmem_file_read_iter.vfs_read.ksys_read 12.58 -0.3 12.25 perf-profile.calltrace.cycles-pp.syscall_return_via_sysret.__libc_lseek64 1.60 ± 4% -0.2 1.44 perf-profile.calltrace.cycles-pp.security_file_permission.vfs_read.ksys_read.do_syscall_64.entry_SYSCALL_64_after_hwframe 1.41 ± 4% -0.1 1.26 perf-profile.calltrace.cycles-pp.apparmor_file_permission.security_file_permission.vfs_read.ksys_read.do_syscall_64 2.35 +0.1 2.42 perf-profile.calltrace.cycles-pp.__fget_light.__fdget_pos.ksys_read.do_syscall_64.entry_SYSCALL_64_after_hwframe 2.88 +0.1 3.01 perf-profile.calltrace.cycles-pp.__fdget_pos.ksys_read.do_syscall_64.entry_SYSCALL_64_after_hwframe.__libc_read 4.08 +0.3 4.34 perf-profile.calltrace.cycles-pp.ksys_lseek.do_syscall_64.entry_SYSCALL_64_after_hwframe.__libc_lseek64 9.05 +0.3 9.32 perf-profile.calltrace.cycles-pp.syscall_exit_to_user_mode.do_syscall_64.entry_SYSCALL_64_after_hwframe.__libc_read 0.26 ±100% +0.3 0.56 ± 2% perf-profile.calltrace.cycles-pp.current_time.atime_needs_update.touch_atime.shmem_file_read_iter.vfs_read 3.28 +0.3 3.62 perf-profile.calltrace.cycles-pp.entry_SYSCALL_64_safe_stack.__libc_read 7.39 +0.7 8.06 perf-profile.calltrace.cycles-pp.syscall_exit_to_user_mode.do_syscall_64.entry_SYSCALL_64_after_hwframe.__libc_lseek64 3.86 +0.8 4.68 perf-profile.calltrace.cycles-pp.__entry_text_start.__libc_lseek64 3.09 +0.9 4.03 perf-profile.calltrace.cycles-pp.__entry_text_start.__libc_read 12.47 +1.0 13.48 perf-profile.calltrace.cycles-pp.do_syscall_64.entry_SYSCALL_64_after_hwframe.__libc_lseek64 9.73 +1.3 11.00 perf-profile.calltrace.cycles-pp.syscall_return_via_sysret.__libc_read 17.06 +1.4 18.50 perf-profile.calltrace.cycles-pp.entry_SYSCALL_64_after_hwframe.__libc_lseek64 2.23 +4.1 6.30 perf-profile.calltrace.cycles-pp.entry_SYSCALL_64_safe_stack.__libc_lseek64 35.19 +4.1 39.27 perf-profile.calltrace.cycles-pp.__libc_lseek64 33.06 -7.8 25.28 perf-profile.children.cycles-pp.ksys_read 26.95 -6.8 20.14 perf-profile.children.cycles-pp.vfs_read 55.83 -6.4 49.44 perf-profile.children.cycles-pp.do_syscall_64 20.76 -5.5 15.26 perf-profile.children.cycles-pp.shmem_file_read_iter 9.68 -5.3 4.36 ± 2% perf-profile.children.cycles-pp.rep_movs_alternative 10.30 -5.2 5.09 ± 2% perf-profile.children.cycles-pp.copyout 10.71 -5.0 5.72 ± 2% perf-profile.children.cycles-pp._copy_to_iter 11.05 -5.0 6.10 ± 2% perf-profile.children.cycles-pp.copy_page_to_iter 65.89 -4.6 61.25 perf-profile.children.cycles-pp.entry_SYSCALL_64_after_hwframe 64.93 -4.1 60.84 perf-profile.children.cycles-pp.__libc_read 1.02 ± 5% -0.6 0.42 ± 3% perf-profile.children.cycles-pp.fsnotify_perm 3.54 -0.6 2.95 ± 2% perf-profile.children.cycles-pp.shmem_get_folio_gfp 1.77 ± 3% -0.5 1.25 ± 3% perf-profile.children.cycles-pp.fput 1.82 ± 4% -0.5 1.32 ± 4% perf-profile.children.cycles-pp.mutex_unlock 1.34 -0.5 0.87 perf-profile.children.cycles-pp.__fsnotify_parent 1.22 ± 2% -0.3 0.88 perf-profile.children.cycles-pp.filemap_get_entry 1.62 ± 3% -0.2 1.45 perf-profile.children.cycles-pp.security_file_permission 1.42 ± 4% -0.1 1.28 perf-profile.children.cycles-pp.apparmor_file_permission 0.24 ± 2% -0.1 0.16 ± 4% perf-profile.children.cycles-pp.aa_file_perm 0.22 ± 15% -0.1 0.16 ± 13% perf-profile.children.cycles-pp.make_vfsuid 0.26 -0.1 0.21 ± 3% perf-profile.children.cycles-pp.xas_load 0.16 ± 3% -0.0 0.12 ± 4% perf-profile.children.cycles-pp.xas_start 0.32 -0.0 0.30 perf-profile.children.cycles-pp.__cond_resched 0.19 +0.0 0.20 perf-profile.children.cycles-pp.folio_test_hugetlb 0.28 ± 2% +0.0 0.30 perf-profile.children.cycles-pp.folio_unlock 0.09 ± 6% +0.0 0.11 ± 5% perf-profile.children.cycles-pp.make_vfsgid 0.29 +0.0 0.31 ± 2% perf-profile.children.cycles-pp.testcase 0.19 ± 3% +0.0 0.21 ± 6% perf-profile.children.cycles-pp.generic_file_llseek_size 0.08 ± 4% +0.0 0.11 ± 4% perf-profile.children.cycles-pp.read@plt 0.41 +0.0 0.45 ± 2% perf-profile.children.cycles-pp.exit_to_user_mode_prepare 0.54 ± 3% +0.1 0.59 ± 2% perf-profile.children.cycles-pp.current_time 1.08 +0.1 1.17 perf-profile.children.cycles-pp.entry_SYSRETQ_unsafe_stack 0.64 ± 2% +0.2 0.80 perf-profile.children.cycles-pp.syscall_enter_from_user_mode 4.14 +0.3 4.40 perf-profile.children.cycles-pp.ksys_lseek 16.52 +1.0 17.47 perf-profile.children.cycles-pp.syscall_exit_to_user_mode 22.43 +1.0 23.39 perf-profile.children.cycles-pp.syscall_return_via_sysret 3.03 +2.3 5.31 perf-profile.children.cycles-pp.entry_SYSCALL_64_safe_stack 8.91 +3.3 12.23 perf-profile.children.cycles-pp.__entry_text_start 35.78 +4.2 39.95 perf-profile.children.cycles-pp.__libc_lseek64 9.53 -5.3 4.24 ± 2% perf-profile.self.cycles-pp.rep_movs_alternative 1.01 ± 5% -0.6 0.42 ± 3% perf-profile.self.cycles-pp.fsnotify_perm 1.73 ± 3% -0.5 1.22 ± 2% perf-profile.self.cycles-pp.fput 1.75 ± 4% -0.5 1.25 ± 4% perf-profile.self.cycles-pp.mutex_unlock 1.30 -0.5 0.83 ± 2% perf-profile.self.cycles-pp.__fsnotify_parent 0.96 ± 2% -0.3 0.67 perf-profile.self.cycles-pp.filemap_get_entry 2.13 ± 2% -0.2 1.90 ± 3% perf-profile.self.cycles-pp.shmem_get_folio_gfp 0.22 ± 4% -0.1 0.14 ± 3% perf-profile.self.cycles-pp.aa_file_perm 0.20 ± 14% -0.1 0.14 ± 11% perf-profile.self.cycles-pp.make_vfsuid 0.14 ± 2% -0.0 0.12 ± 4% perf-profile.self.cycles-pp.xas_start 0.20 ± 4% -0.0 0.18 ± 3% perf-profile.self.cycles-pp.security_file_permission 0.10 ± 3% -0.0 0.08 perf-profile.self.cycles-pp.xas_load 0.22 ± 2% -0.0 0.20 ± 2% perf-profile.self.cycles-pp.__cond_resched 0.08 ± 5% +0.0 0.10 ± 3% perf-profile.self.cycles-pp.make_vfsgid 0.26 ± 3% +0.0 0.29 perf-profile.self.cycles-pp.folio_unlock 0.19 ± 3% +0.0 0.21 ± 4% perf-profile.self.cycles-pp.generic_file_llseek_size 0.33 ± 3% +0.0 0.36 perf-profile.self.cycles-pp.ksys_lseek 0.38 +0.0 0.41 perf-profile.self.cycles-pp.exit_to_user_mode_prepare 0.47 +0.0 0.51 ± 2% perf-profile.self.cycles-pp.__fdget_pos 0.42 ± 5% +0.1 0.48 ± 2% perf-profile.self.cycles-pp.current_time 0.29 ± 3% +0.1 0.35 ± 2% perf-profile.self.cycles-pp.copy_page_to_iter 1.03 +0.1 1.08 perf-profile.self.cycles-pp.__libc_read 0.00 +0.1 0.06 ± 7% perf-profile.self.cycles-pp.read@plt 0.96 +0.1 1.03 perf-profile.self.cycles-pp.entry_SYSRETQ_unsafe_stack 0.72 +0.1 0.82 perf-profile.self.cycles-pp.copyout 0.70 +0.1 0.80 perf-profile.self.cycles-pp.__libc_lseek64 0.52 ± 2% +0.1 0.64 perf-profile.self.cycles-pp.syscall_enter_from_user_mode 0.54 ± 2% +0.2 0.70 ± 3% perf-profile.self.cycles-pp.entry_SYSCALL_64_safe_stack 0.42 +0.2 0.64 ± 7% perf-profile.self.cycles-pp._copy_to_iter 15.71 +0.9 16.60 perf-profile.self.cycles-pp.syscall_exit_to_user_mode 22.41 +1.0 23.37 perf-profile.self.cycles-pp.syscall_return_via_sysret 10.40 +1.8 12.15 perf-profile.self.cycles-pp.entry_SYSCALL_64_after_hwframe 7.90 +3.2 11.13 perf-profile.self.cycles-pp.__entry_text_start 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