2021-05-25 08:17:51

by kernel test robot

[permalink] [raw]
Subject: 2463a604a8: netperf.Throughput_tps 12.8% improvement



Greeting,

FYI, we noticed a 12.8% improvement of netperf.Throughput_tps due to commit:


commit: 2463a604a86728777ce4284214a52de46a808c9e ("[PATCH v3 2/3] Introduce enums for the SAM, message, host and driver status codes")
url: https://github.com/0day-ci/linux/commits/Bart-Van-Assche/Introduce-enums-for-SCSI-status-codes/20210524-105751
base: https://git.kernel.org/cgit/linux/kernel/git/mkp/scsi.git for-next

in testcase: netperf
on test machine: 192 threads 4 sockets Intel(R) Xeon(R) Platinum 9242 CPU @ 2.30GHz with 192G memory
with following parameters:

ip: ipv4
runtime: 300s
nr_threads: 16
cluster: cs-localhost
test: TCP_CRR
cpufreq_governor: performance
ucode: 0x5003006

test-description: Netperf is a benchmark that can be use to measure various aspect of networking performance.
test-url: http://www.netperf.org/netperf/





Details are as below:
-------------------------------------------------------------------------------------------------->


To reproduce:

git clone https://github.com/intel/lkp-tests.git
cd lkp-tests
bin/lkp install job.yaml # job file is attached in this email
bin/lkp split-job --compatible job.yaml # generate the yaml file for lkp run
bin/lkp run generated-yaml-file

=========================================================================================
cluster/compiler/cpufreq_governor/ip/kconfig/nr_threads/rootfs/runtime/tbox_group/test/testcase/ucode:
cs-localhost/gcc-9/performance/ipv4/x86_64-rhel-8.3/16/debian-10.4-x86_64-20200603.cgz/300s/lkp-csl-2ap3/TCP_CRR/netperf/0x5003006

commit:
c7f8fc6d65 ("libsas: Introduce more SAM status code aliases in enum exec_status")
2463a604a8 ("Introduce enums for the SAM, message, host and driver status codes")

c7f8fc6d650032e7 2463a604a86728777ce4284214a
---------------- ---------------------------
%stddev %change %stddev
\ | \
294711 +12.8% 332486 netperf.Throughput_total_tps
18419 +12.8% 20780 netperf.Throughput_tps
56.03 +16.4% 65.22 netperf.time.user_time
1.723e+08 +12.2% 1.934e+08 netperf.time.voluntary_context_switches
88413544 +12.8% 99745810 netperf.workload
358449 ? 5% +12.1% 401800 ? 4% meminfo.Mapped
0.26 +0.0 0.29 mpstat.cpu.all.usr%
2000541 +4.9% 2098419 vmstat.system.cs
90072013 ? 12% -16.7% 74986487 ? 3% turbostat.C1
4.10 ? 88% -1.8 2.31 ? 3% turbostat.C1%
1.719e+08 ? 11% -15.4% 1.455e+08 cpuidle.C1.usage
3.582e+08 +13.9% 4.081e+08 cpuidle.POLL.time
1.337e+08 +27.1% 1.699e+08 cpuidle.POLL.usage
74012 ? 7% -18.2% 60560 slabinfo.sock_inode_cache.active_objs
1899 ? 7% -18.2% 1555 slabinfo.sock_inode_cache.active_slabs
74115 ? 7% -18.1% 60669 slabinfo.sock_inode_cache.num_objs
1899 ? 7% -18.2% 1555 slabinfo.sock_inode_cache.num_slabs
67730 ?100% +129.6% 155509 ? 18% numa-meminfo.node0.AnonHugePages
108263 ? 85% -85.5% 15705 ?136% numa-meminfo.node1.Inactive
108263 ? 85% -85.5% 15705 ?136% numa-meminfo.node1.Inactive(anon)
1401441 ? 77% -67.7% 452837 ? 7% numa-meminfo.node1.MemUsed
159343 ? 37% +85.2% 295052 ? 20% numa-meminfo.node3.Mapped
2093303 ? 46% +93.5% 4050651 ? 26% numa-meminfo.node3.MemUsed
27050 ? 85% -85.5% 3928 ?136% numa-vmstat.node1.nr_inactive_anon
27050 ? 85% -85.5% 3928 ?136% numa-vmstat.node1.nr_zone_inactive_anon
9751 ? 31% -36.5% 6195 ? 12% numa-vmstat.node2.nr_slab_reclaimable
39894 ? 37% +84.8% 73713 ? 20% numa-vmstat.node3.nr_mapped
864170 ? 33% +71.3% 1480718 ? 22% numa-vmstat.node3.numa_hit
767614 ? 38% +82.2% 1398603 ? 25% numa-vmstat.node3.numa_local
0.01 -15.0% 0.01 ? 5% perf-sched.wait_and_delay.avg.ms.schedule_timeout.wait_woken.sk_wait_data.tcp_recvmsg_locked
2701296 +11.3% 3006865 perf-sched.wait_and_delay.count.schedule_timeout.inet_csk_accept.inet_accept.__sys_accept4_file
7189 ? 9% -15.6% 6069 ? 12% perf-sched.wait_and_delay.max.ms.worker_thread.kthread.ret_from_fork
0.03 ? 3% -16.3% 0.02 perf-sched.wait_time.avg.ms.preempt_schedule_common.__cond_resched.__release_sock.release_sock.inet_accept
0.02 ? 29% -57.1% 0.01 ? 17% perf-sched.wait_time.avg.ms.preempt_schedule_common.__cond_resched.__release_sock.release_sock.sk_wait_data
0.02 ? 43% -40.4% 0.01 ? 9% perf-sched.wait_time.avg.ms.preempt_schedule_common.__cond_resched.lock_sock_nested.tcp_recvmsg.inet_recvmsg
0.01 -25.0% 0.01 perf-sched.wait_time.avg.ms.schedule_timeout.wait_woken.sk_wait_data.tcp_recvmsg_locked
0.03 ? 10% -17.2% 0.03 ? 6% perf-sched.wait_time.max.ms.preempt_schedule_common.__cond_resched.__kmalloc_node.memcg_alloc_page_obj_cgroups.allocate_slab
0.04 ? 5% -18.8% 0.03 ? 4% perf-sched.wait_time.max.ms.preempt_schedule_common.__cond_resched.__release_sock.release_sock.inet_accept
0.19 ?176% -84.1% 0.03 ? 4% perf-sched.wait_time.max.ms.preempt_schedule_common.__cond_resched.__release_sock.release_sock.sk_wait_data
7189 ? 9% -15.6% 6069 ? 12% perf-sched.wait_time.max.ms.worker_thread.kthread.ret_from_fork
927313 +2.9% 954532 proc-vmstat.nr_active_anon
1207711 +2.7% 1240469 proc-vmstat.nr_file_pages
92383 ? 5% +11.8% 103291 ? 4% proc-vmstat.nr_mapped
963105 +3.4% 996180 proc-vmstat.nr_shmem
47165 ? 2% -5.1% 44779 proc-vmstat.nr_slab_reclaimable
927313 +2.9% 954532 proc-vmstat.nr_zone_active_anon
250336 ? 19% +33.9% 335151 ? 11% proc-vmstat.numa_hint_faults
3853978 +7.6% 4148159 proc-vmstat.numa_hit
3593793 +8.2% 3887921 proc-vmstat.numa_local
498965 ? 9% +19.1% 594180 ? 7% proc-vmstat.numa_pte_updates
4924666 +35.0% 6648555 proc-vmstat.pgalloc_normal
1782592 ? 2% +5.4% 1879368 ? 2% proc-vmstat.pgfault
2559244 +62.3% 4153840 ? 3% proc-vmstat.pgfree
4265 ? 11% +28.2% 5467 ? 10% interrupts.CPU1.RES:Rescheduling_interrupts
7221 ? 35% +87.6% 13545 ? 36% interrupts.CPU106.CAL:Function_call_interrupts
7837 ? 20% +57.0% 12306 ? 32% interrupts.CPU107.CAL:Function_call_interrupts
7650 ? 26% +71.5% 13119 ? 25% interrupts.CPU113.CAL:Function_call_interrupts
7336 ? 5% +23.0% 9024 ? 17% interrupts.CPU118.CAL:Function_call_interrupts
6603 ? 19% +83.5% 12119 ? 27% interrupts.CPU139.CAL:Function_call_interrupts
2739 ? 14% -17.6% 2257 ? 15% interrupts.CPU152.RES:Rescheduling_interrupts
5810 ? 31% +69.0% 9818 ? 23% interrupts.CPU17.CAL:Function_call_interrupts
1460 ? 34% +62.2% 2368 ? 29% interrupts.CPU170.NMI:Non-maskable_interrupts
1460 ? 34% +62.2% 2368 ? 29% interrupts.CPU170.PMI:Performance_monitoring_interrupts
1229 ? 13% +20.3% 1478 ? 8% interrupts.CPU18.RES:Rescheduling_interrupts
2319 ? 28% -44.2% 1293 ? 31% interrupts.CPU26.NMI:Non-maskable_interrupts
2319 ? 28% -44.2% 1293 ? 31% interrupts.CPU26.PMI:Performance_monitoring_interrupts
7684 ? 24% +68.3% 12932 ? 31% interrupts.CPU34.CAL:Function_call_interrupts
4986 ? 26% +88.6% 9403 ? 34% interrupts.CPU44.CAL:Function_call_interrupts
1118 ? 49% +133.4% 2609 ? 32% interrupts.CPU5.NMI:Non-maskable_interrupts
1118 ? 49% +133.4% 2609 ? 32% interrupts.CPU5.PMI:Performance_monitoring_interrupts
2106 ? 10% +26.5% 2664 ? 16% interrupts.CPU5.RES:Rescheduling_interrupts
7633 ? 29% +62.5% 12402 ? 16% interrupts.CPU52.CAL:Function_call_interrupts
1076 ? 47% +99.0% 2141 ? 24% interrupts.CPU57.NMI:Non-maskable_interrupts
1076 ? 47% +99.0% 2141 ? 24% interrupts.CPU57.PMI:Performance_monitoring_interrupts
706.67 ? 41% +157.7% 1820 ? 22% interrupts.CPU66.NMI:Non-maskable_interrupts
706.67 ? 41% +157.7% 1820 ? 22% interrupts.CPU66.PMI:Performance_monitoring_interrupts
1193 ? 22% +38.1% 1649 ? 11% interrupts.CPU69.RES:Rescheduling_interrupts
832.67 ? 34% +80.6% 1504 ? 35% interrupts.CPU91.NMI:Non-maskable_interrupts
832.67 ? 34% +80.6% 1504 ? 35% interrupts.CPU91.PMI:Performance_monitoring_interrupts
2303 ? 17% +35.8% 3129 ? 11% interrupts.CPU99.RES:Rescheduling_interrupts
184.17 ? 79% +209.2% 569.50 ? 66% interrupts.TLB:TLB_shootdowns
2407637 ? 17% +19.4% 2874989 ? 6% softirqs.CPU1.NET_RX
170089 ? 12% +18.2% 201090 ? 4% softirqs.CPU1.RCU
109036 ? 11% +21.9% 132886 ? 14% softirqs.CPU103.RCU
1286100 ? 6% +31.4% 1690186 ? 19% softirqs.CPU104.NET_RX
106047 ? 4% +19.7% 126916 ? 15% softirqs.CPU104.RCU
1422375 ? 8% +30.5% 1856460 ? 26% softirqs.CPU105.NET_RX
1342415 ? 20% +24.8% 1675573 ? 8% softirqs.CPU117.NET_RX
112028 ? 12% +15.8% 129770 ? 6% softirqs.CPU117.RCU
871622 ? 13% +28.0% 1115854 ? 17% softirqs.CPU12.NET_RX
83247 ? 10% +16.5% 96992 ? 9% softirqs.CPU12.RCU
118412 ? 8% +21.6% 143979 ? 8% softirqs.CPU122.RCU
1453784 ? 9% +29.8% 1886713 ? 17% softirqs.CPU123.NET_RX
123994 ? 9% +21.4% 150485 ? 10% softirqs.CPU124.RCU
1497798 ? 16% +27.8% 1914309 ? 12% softirqs.CPU130.NET_RX
132936 ? 4% +11.6% 148301 ? 8% softirqs.CPU135.RCU
119189 ? 8% +15.1% 137139 ? 11% softirqs.CPU148.RCU
1514380 ? 9% +23.0% 1862024 ? 14% softirqs.CPU169.NET_RX
115176 ? 9% +14.4% 131723 ? 6% softirqs.CPU170.RCU
753863 ? 15% +20.8% 910510 ? 9% softirqs.CPU18.NET_RX
76149 ? 6% +13.3% 86243 ? 2% softirqs.CPU18.RCU
2338405 ? 9% +21.6% 2843137 ? 7% softirqs.CPU2.NET_RX
164205 ? 6% +18.8% 195118 ? 7% softirqs.CPU2.RCU
212111 ? 10% +16.2% 246464 ? 9% softirqs.CPU24.RCU
2789625 ? 11% +17.3% 3272648 ? 7% softirqs.CPU25.NET_RX
194716 ? 9% +16.0% 225881 ? 7% softirqs.CPU25.RCU
138942 ? 10% +20.2% 167007 ? 10% softirqs.CPU27.RCU
112057 ? 10% +18.0% 132281 ? 8% softirqs.CPU28.RCU
816057 ? 9% +31.8% 1075372 ? 8% softirqs.CPU46.NET_RX
83075 ? 6% +18.9% 98799 ? 5% softirqs.CPU46.RCU
2693920 ? 12% +17.9% 3176816 ? 7% softirqs.CPU49.NET_RX
1179737 ? 13% +25.7% 1482562 ? 14% softirqs.CPU5.NET_RX
102077 ? 8% +16.8% 119190 ? 9% softirqs.CPU5.RCU
1254211 ? 19% +47.1% 1844668 ? 8% softirqs.CPU52.NET_RX
108601 ? 14% +29.0% 140120 ? 7% softirqs.CPU52.RCU
1145938 ? 12% +29.6% 1485285 ? 9% softirqs.CPU55.NET_RX
972858 ? 15% +35.8% 1321088 ? 14% softirqs.CPU57.NET_RX
90294 ? 11% +25.0% 112842 ? 6% softirqs.CPU57.RCU
1064438 ? 17% +30.8% 1392086 ? 10% softirqs.CPU58.NET_RX
93479 ? 10% +19.1% 111345 ? 4% softirqs.CPU59.RCU
914447 ? 14% +44.9% 1324661 ? 13% softirqs.CPU60.NET_RX
85063 ? 9% +30.5% 111008 ? 10% softirqs.CPU60.RCU
959827 ? 19% +45.5% 1396834 ? 12% softirqs.CPU61.NET_RX
89606 ? 14% +27.6% 114320 ? 7% softirqs.CPU61.RCU
90163 ? 9% +14.5% 103199 ? 8% softirqs.CPU64.RCU
833314 ? 15% +30.8% 1090219 ? 9% softirqs.CPU69.NET_RX
81836 ? 9% +22.5% 100235 ? 4% softirqs.CPU69.RCU
786108 ? 24% +43.4% 1127232 ? 14% softirqs.CPU70.NET_RX
1446282 ? 8% +32.5% 1916102 ? 14% softirqs.CPU99.NET_RX
112093 ? 6% +26.9% 142243 ? 9% softirqs.CPU99.RCU
2.7e+08 +13.3% 3.059e+08 softirqs.NET_RX
14.48 ? 3% -15.0% 12.31 ? 2% perf-stat.i.MPKI
8.019e+09 +10.4% 8.856e+09 perf-stat.i.branch-instructions
1.194e+08 +9.7% 1.31e+08 perf-stat.i.branch-misses
16.68 -4.6 12.08 ? 2% perf-stat.i.cache-miss-rate%
93472138 -31.0% 64500000 ? 2% perf-stat.i.cache-misses
5.585e+08 -3.7% 5.38e+08 perf-stat.i.cache-references
2024263 +4.9% 2122993 perf-stat.i.context-switches
1.91 -8.2% 1.75 perf-stat.i.cpi
7.486e+10 +2.2% 7.653e+10 perf-stat.i.cpu-cycles
824.85 +46.9% 1211 ? 2% perf-stat.i.cycles-between-cache-misses
1.203e+10 +10.8% 1.333e+10 perf-stat.i.dTLB-loads
7.006e+09 +10.8% 7.763e+09 perf-stat.i.dTLB-stores
85.68 +1.5 87.17 perf-stat.i.iTLB-load-miss-rate%
83118656 +9.4% 90943725 perf-stat.i.iTLB-load-misses
13547724 -5.5% 12798229 perf-stat.i.iTLB-loads
4.037e+10 +10.6% 4.464e+10 perf-stat.i.instructions
502.31 +1.9% 511.65 perf-stat.i.instructions-per-iTLB-miss
0.54 +8.2% 0.58 perf-stat.i.ipc
0.39 +2.2% 0.40 perf-stat.i.metric.GHz
249.96 -12.4% 218.94 perf-stat.i.metric.K/sec
143.82 +10.4% 158.77 perf-stat.i.metric.M/sec
5691 ? 3% +5.5% 6007 ? 2% perf-stat.i.minor-faults
19727052 -26.1% 14574710 perf-stat.i.node-load-misses
1857523 ? 2% -25.4% 1386452 ? 2% perf-stat.i.node-loads
8801182 +12.1% 9866676 perf-stat.i.node-store-misses
371643 ? 4% +9.1% 405430 ? 5% perf-stat.i.node-stores
5692 ? 3% +5.5% 6008 ? 2% perf-stat.i.page-faults
13.83 -12.9% 12.05 perf-stat.overall.MPKI
16.74 -4.7 11.99 perf-stat.overall.cache-miss-rate%
1.85 -7.5% 1.71 perf-stat.overall.cpi
800.86 +48.2% 1187 ? 2% perf-stat.overall.cycles-between-cache-misses
85.99 +1.7 87.66 perf-stat.overall.iTLB-load-miss-rate%
485.76 +1.1% 490.90 perf-stat.overall.instructions-per-iTLB-miss
0.54 +8.1% 0.58 perf-stat.overall.ipc
139006 -2.0% 136253 perf-stat.overall.path-length
7.993e+09 +10.4% 8.828e+09 perf-stat.ps.branch-instructions
1.19e+08 +9.7% 1.306e+08 perf-stat.ps.branch-misses
93164112 -31.0% 64291099 ? 2% perf-stat.ps.cache-misses
5.567e+08 -3.7% 5.362e+08 perf-stat.ps.cache-references
2017451 +4.9% 2115915 perf-stat.ps.context-switches
7.461e+10 +2.2% 7.628e+10 perf-stat.ps.cpu-cycles
1.199e+10 +10.8% 1.328e+10 perf-stat.ps.dTLB-loads
6.983e+09 +10.8% 7.738e+09 perf-stat.ps.dTLB-stores
82842581 +9.4% 90643962 perf-stat.ps.iTLB-load-misses
13502090 -5.5% 12755686 perf-stat.ps.iTLB-loads
4.024e+10 +10.6% 4.45e+10 perf-stat.ps.instructions
5677 ? 3% +5.6% 5993 ? 2% perf-stat.ps.minor-faults
19662090 -26.1% 14527537 perf-stat.ps.node-load-misses
1851487 ? 2% -25.4% 1382045 ? 2% perf-stat.ps.node-loads
8772419 +12.1% 9834635 perf-stat.ps.node-store-misses
370890 ? 4% +9.1% 404703 ? 5% perf-stat.ps.node-stores
5678 ? 3% +5.6% 5994 ? 2% perf-stat.ps.page-faults
1.229e+13 +10.6% 1.359e+13 perf-stat.total.instructions
51.98 -3.9 48.10 ? 6% perf-profile.calltrace.cycles-pp.do_syscall_64.entry_SYSCALL_64_after_hwframe
52.17 -3.9 48.30 ? 6% perf-profile.calltrace.cycles-pp.entry_SYSCALL_64_after_hwframe
14.68 -2.9 11.82 ? 7% perf-profile.calltrace.cycles-pp.__x64_sys_connect.do_syscall_64.entry_SYSCALL_64_after_hwframe
14.46 -2.9 11.60 ? 7% perf-profile.calltrace.cycles-pp.__inet_stream_connect.inet_stream_connect.__sys_connect.__x64_sys_connect.do_syscall_64
14.67 -2.9 11.81 ? 7% perf-profile.calltrace.cycles-pp.__sys_connect.__x64_sys_connect.do_syscall_64.entry_SYSCALL_64_after_hwframe
14.51 -2.9 11.66 ? 7% perf-profile.calltrace.cycles-pp.inet_stream_connect.__sys_connect.__x64_sys_connect.do_syscall_64.entry_SYSCALL_64_after_hwframe
16.53 -2.0 14.51 ? 6% perf-profile.calltrace.cycles-pp.tcp_v4_rcv.ip_protocol_deliver_rcu.ip_local_deliver_finish.ip_local_deliver.ip_rcv
16.95 -2.0 14.94 ? 6% perf-profile.calltrace.cycles-pp.ip_local_deliver_finish.ip_local_deliver.ip_rcv.__netif_receive_skb_one_core.process_backlog
16.91 -2.0 14.90 ? 6% perf-profile.calltrace.cycles-pp.ip_protocol_deliver_rcu.ip_local_deliver_finish.ip_local_deliver.ip_rcv.__netif_receive_skb_one_core
16.98 -2.0 14.98 ? 6% perf-profile.calltrace.cycles-pp.ip_local_deliver.ip_rcv.__netif_receive_skb_one_core.process_backlog.__napi_poll
7.16 -1.6 5.60 ? 7% perf-profile.calltrace.cycles-pp.tcp_v4_do_rcv.__release_sock.release_sock.__inet_stream_connect.inet_stream_connect
7.19 -1.6 5.64 ? 7% perf-profile.calltrace.cycles-pp.__release_sock.release_sock.__inet_stream_connect.inet_stream_connect.__sys_connect
7.21 -1.5 5.66 ? 7% perf-profile.calltrace.cycles-pp.release_sock.__inet_stream_connect.inet_stream_connect.__sys_connect.__x64_sys_connect
7.03 -1.5 5.50 ? 7% perf-profile.calltrace.cycles-pp.tcp_rcv_synsent_state_process.tcp_rcv_state_process.tcp_v4_do_rcv.__release_sock.release_sock
17.76 -1.5 16.23 ? 6% perf-profile.calltrace.cycles-pp.ip_rcv.__netif_receive_skb_one_core.process_backlog.__napi_poll.net_rx_action
17.98 -1.5 16.50 ? 6% perf-profile.calltrace.cycles-pp.__netif_receive_skb_one_core.process_backlog.__napi_poll.net_rx_action.__softirqentry_text_start
18.38 -1.5 16.90 ? 6% perf-profile.calltrace.cycles-pp.__napi_poll.net_rx_action.__softirqentry_text_start.do_softirq.__local_bh_enable_ip
18.54 -1.5 17.07 ? 6% perf-profile.calltrace.cycles-pp.net_rx_action.__softirqentry_text_start.do_softirq.__local_bh_enable_ip.ip_finish_output2
18.33 -1.5 16.86 ? 6% perf-profile.calltrace.cycles-pp.process_backlog.__napi_poll.net_rx_action.__softirqentry_text_start.do_softirq
7.11 -1.3 5.81 ? 6% perf-profile.calltrace.cycles-pp.tcp_v4_connect.__inet_stream_connect.inet_stream_connect.__sys_connect.__x64_sys_connect
9.28 -1.2 8.08 ? 6% perf-profile.calltrace.cycles-pp.tcp_v4_do_rcv.tcp_v4_rcv.ip_protocol_deliver_rcu.ip_local_deliver_finish.ip_local_deliver
6.24 -1.1 5.10 ? 7% perf-profile.calltrace.cycles-pp.tcp_rcv_state_process.tcp_v4_do_rcv.__release_sock.release_sock.__inet_stream_connect
5.72 -1.0 4.71 ? 6% perf-profile.calltrace.cycles-pp.tcp_connect.tcp_v4_connect.__inet_stream_connect.inet_stream_connect.__sys_connect
9.86 -1.0 8.84 ? 6% perf-profile.calltrace.cycles-pp.__x64_sys_sendto.do_syscall_64.entry_SYSCALL_64_after_hwframe
9.71 -1.0 8.70 ? 6% perf-profile.calltrace.cycles-pp.sock_sendmsg.__sys_sendto.__x64_sys_sendto.do_syscall_64.entry_SYSCALL_64_after_hwframe
9.83 -1.0 8.81 ? 6% perf-profile.calltrace.cycles-pp.__sys_sendto.__x64_sys_sendto.do_syscall_64.entry_SYSCALL_64_after_hwframe
9.44 -1.0 8.44 ? 6% perf-profile.calltrace.cycles-pp.tcp_sendmsg.sock_sendmsg.__sys_sendto.__x64_sys_sendto.do_syscall_64
5.49 -0.9 4.62 ? 7% perf-profile.calltrace.cycles-pp.__tcp_transmit_skb.tcp_rcv_synsent_state_process.tcp_rcv_state_process.tcp_v4_do_rcv.__release_sock
5.25 -0.8 4.47 ? 7% perf-profile.calltrace.cycles-pp.__ip_queue_xmit.__tcp_transmit_skb.tcp_rcv_synsent_state_process.tcp_rcv_state_process.tcp_v4_do_rcv
4.90 -0.7 4.15 ? 7% perf-profile.calltrace.cycles-pp.ip_finish_output2.ip_output.__ip_queue_xmit.__tcp_transmit_skb.tcp_rcv_synsent_state_process
4.79 ? 2% -0.7 4.07 ? 7% perf-profile.calltrace.cycles-pp.tcp_rcv_state_process.tcp_v4_do_rcv.tcp_v4_rcv.ip_protocol_deliver_rcu.ip_local_deliver_finish
4.99 -0.7 4.27 ? 7% perf-profile.calltrace.cycles-pp.ip_output.__ip_queue_xmit.__tcp_transmit_skb.tcp_rcv_synsent_state_process.tcp_rcv_state_process
0.90 -0.6 0.26 ?100% perf-profile.calltrace.cycles-pp.tcp_rcv_state_process.tcp_child_process.tcp_v4_rcv.ip_protocol_deliver_rcu.ip_local_deliver_finish
8.10 -0.6 7.48 ? 6% perf-profile.calltrace.cycles-pp.tcp_sendmsg_locked.tcp_sendmsg.sock_sendmsg.__sys_sendto.__x64_sys_sendto
4.51 -0.6 3.94 ? 6% perf-profile.calltrace.cycles-pp.__tcp_transmit_skb.tcp_connect.tcp_v4_connect.__inet_stream_connect.inet_stream_connect
2.17 -0.5 1.63 ? 7% perf-profile.calltrace.cycles-pp.tcp_conn_request.tcp_rcv_state_process.tcp_v4_do_rcv.tcp_v4_rcv.ip_protocol_deliver_rcu
4.22 -0.5 3.69 ? 6% perf-profile.calltrace.cycles-pp.__ip_queue_xmit.__tcp_transmit_skb.tcp_connect.tcp_v4_connect.__inet_stream_connect
3.77 -0.5 3.27 ? 6% perf-profile.calltrace.cycles-pp.ip_finish_output2.ip_output.__ip_queue_xmit.__tcp_transmit_skb.tcp_connect
3.79 -0.5 3.28 ? 6% perf-profile.calltrace.cycles-pp.ip_output.__ip_queue_xmit.__tcp_transmit_skb.tcp_connect.tcp_v4_connect
2.38 -0.4 1.93 ? 8% perf-profile.calltrace.cycles-pp.tcp_check_req.tcp_v4_rcv.ip_protocol_deliver_rcu.ip_local_deliver_finish.ip_local_deliver
1.28 -0.4 0.91 ? 6% perf-profile.calltrace.cycles-pp.release_sock.tcp_sendmsg.sock_sendmsg.__sys_sendto.__x64_sys_sendto
1.24 ? 2% -0.4 0.87 ? 6% perf-profile.calltrace.cycles-pp.__release_sock.release_sock.tcp_sendmsg.sock_sendmsg.__sys_sendto
1.20 ? 2% -0.4 0.83 ? 6% perf-profile.calltrace.cycles-pp.tcp_v4_do_rcv.__release_sock.release_sock.tcp_sendmsg.sock_sendmsg
1.83 -0.4 1.47 ? 6% perf-profile.calltrace.cycles-pp.tcp_child_process.tcp_v4_rcv.ip_protocol_deliver_rcu.ip_local_deliver_finish.ip_local_deliver
0.88 ? 2% -0.3 0.59 ? 6% perf-profile.calltrace.cycles-pp.tcp_ack.tcp_rcv_established.tcp_v4_do_rcv.__release_sock.release_sock
1.08 ? 2% -0.3 0.81 ? 6% perf-profile.calltrace.cycles-pp.tcp_rcv_established.tcp_v4_do_rcv.__release_sock.release_sock.tcp_sendmsg
1.45 -0.3 1.18 ? 6% perf-profile.calltrace.cycles-pp.inet_create.__sock_create.__sys_socket.__x64_sys_socket.do_syscall_64
2.02 -0.3 1.76 ? 7% perf-profile.calltrace.cycles-pp.tcp_v4_syn_recv_sock.tcp_check_req.tcp_v4_rcv.ip_protocol_deliver_rcu.ip_local_deliver_finish
1.49 -0.2 1.25 ? 8% perf-profile.calltrace.cycles-pp.__release_sock.__tcp_close.tcp_close.inet_release.__sock_release
1.47 -0.2 1.23 ? 8% perf-profile.calltrace.cycles-pp.tcp_v4_do_rcv.__release_sock.__tcp_close.tcp_close.inet_release
1.46 -0.2 1.23 ? 8% perf-profile.calltrace.cycles-pp.tcp_rcv_state_process.tcp_v4_do_rcv.__release_sock.__tcp_close.tcp_close
0.98 ? 2% -0.2 0.81 ? 5% perf-profile.calltrace.cycles-pp.__tcp_transmit_skb.tcp_rcv_established.tcp_v4_do_rcv.tcp_v4_rcv.ip_protocol_deliver_rcu
0.64 ? 2% -0.0 0.59 ? 4% perf-profile.calltrace.cycles-pp.__ip_queue_xmit.__tcp_transmit_skb.tcp_rcv_established.tcp_v4_do_rcv.tcp_v4_rcv
0.75 ? 2% +0.1 0.82 ? 6% perf-profile.calltrace.cycles-pp.sock_alloc_file.__sys_accept4_file.__sys_accept4.__x64_sys_accept.do_syscall_64
0.54 ? 3% +0.1 0.63 ? 8% perf-profile.calltrace.cycles-pp.new_inode_pseudo.sock_alloc.__sock_create.__sys_socket.__x64_sys_socket
0.54 ? 4% +0.1 0.63 ? 7% perf-profile.calltrace.cycles-pp.sock_alloc.__sock_create.__sys_socket.__x64_sys_socket.do_syscall_64
1.27 ? 2% +0.1 1.37 ? 5% perf-profile.calltrace.cycles-pp.inet_accept.__sys_accept4_file.__sys_accept4.__x64_sys_accept.do_syscall_64
0.90 +0.1 1.00 ? 8% perf-profile.calltrace.cycles-pp.tcp_create_openreq_child.tcp_v4_syn_recv_sock.tcp_check_req.tcp_v4_rcv.ip_protocol_deliver_rcu
0.60 ? 5% +0.1 0.71 ? 11% perf-profile.calltrace.cycles-pp.kmem_cache_free.__dentry_kill.__fput.task_work_run.exit_to_user_mode_prepare
0.43 ? 44% +0.1 0.56 ? 6% perf-profile.calltrace.cycles-pp.alloc_file.alloc_file_pseudo.sock_alloc_file.__sys_accept4_file.__sys_accept4
1.10 ? 4% +0.2 1.30 ? 8% perf-profile.calltrace.cycles-pp.__dentry_kill.__fput.task_work_run.exit_to_user_mode_prepare.syscall_exit_to_user_mode
2.68 +0.2 2.92 ? 5% perf-profile.calltrace.cycles-pp.__sys_accept4_file.__sys_accept4.__x64_sys_accept.do_syscall_64.entry_SYSCALL_64_after_hwframe
2.71 +0.2 2.96 ? 5% perf-profile.calltrace.cycles-pp.__sys_accept4.__x64_sys_accept.do_syscall_64.entry_SYSCALL_64_after_hwframe
0.35 ? 70% +0.3 0.60 ? 7% perf-profile.calltrace.cycles-pp.alloc_inode.new_inode_pseudo.sock_alloc.__sock_create.__sys_socket
2.72 +0.3 2.98 ? 5% perf-profile.calltrace.cycles-pp.__x64_sys_accept.do_syscall_64.entry_SYSCALL_64_after_hwframe
0.26 ?100% +0.3 0.55 ? 6% perf-profile.calltrace.cycles-pp.alloc_empty_file.alloc_file.alloc_file_pseudo.sock_alloc_file.__sys_accept4_file
0.72 ? 33% +0.4 1.12 ? 7% perf-profile.calltrace.cycles-pp.__alloc_file.alloc_empty_file.alloc_file.alloc_file_pseudo.sock_alloc_file
52.04 -3.9 48.17 ? 6% perf-profile.children.cycles-pp.do_syscall_64
52.24 -3.9 48.37 ? 6% perf-profile.children.cycles-pp.entry_SYSCALL_64_after_hwframe
19.36 -3.3 16.08 ? 6% perf-profile.children.cycles-pp.tcp_v4_do_rcv
14.67 -2.9 11.81 ? 7% perf-profile.children.cycles-pp.__sys_connect
14.46 -2.9 11.60 ? 7% perf-profile.children.cycles-pp.__inet_stream_connect
14.68 -2.9 11.83 ? 7% perf-profile.children.cycles-pp.__x64_sys_connect
14.51 -2.9 11.66 ? 7% perf-profile.children.cycles-pp.inet_stream_connect
13.41 -2.5 10.90 ? 7% perf-profile.children.cycles-pp.tcp_rcv_state_process
10.21 -2.1 8.10 ? 7% perf-profile.children.cycles-pp.__release_sock
16.66 -2.0 14.65 ? 6% perf-profile.children.cycles-pp.tcp_v4_rcv
17.04 -2.0 15.04 ? 6% perf-profile.children.cycles-pp.ip_protocol_deliver_rcu
17.07 -2.0 15.08 ? 6% perf-profile.children.cycles-pp.ip_local_deliver_finish
17.10 -2.0 15.12 ? 6% perf-profile.children.cycles-pp.ip_local_deliver
1.97 ? 2% -1.9 0.04 ? 45% perf-profile.children.cycles-pp.ipv4_dst_check
9.02 -1.8 7.19 ? 7% perf-profile.children.cycles-pp.release_sock
1.69 ? 2% -1.6 0.08 ? 9% perf-profile.children.cycles-pp.__sk_dst_check
7.04 -1.5 5.50 ? 7% perf-profile.children.cycles-pp.tcp_rcv_synsent_state_process
17.89 -1.5 16.39 ? 6% perf-profile.children.cycles-pp.ip_rcv
18.12 -1.5 16.66 ? 6% perf-profile.children.cycles-pp.__netif_receive_skb_one_core
18.52 -1.5 17.06 ? 6% perf-profile.children.cycles-pp.__napi_poll
18.68 -1.4 17.24 ? 6% perf-profile.children.cycles-pp.net_rx_action
18.47 -1.4 17.04 ? 6% perf-profile.children.cycles-pp.process_backlog
7.11 -1.3 5.82 ? 6% perf-profile.children.cycles-pp.tcp_v4_connect
5.72 -1.0 4.71 ? 6% perf-profile.children.cycles-pp.tcp_connect
9.86 -1.0 8.85 ? 6% perf-profile.children.cycles-pp.__x64_sys_sendto
9.72 -1.0 8.71 ? 6% perf-profile.children.cycles-pp.sock_sendmsg
9.83 -1.0 8.82 ? 6% perf-profile.children.cycles-pp.__sys_sendto
9.44 -1.0 8.45 ? 6% perf-profile.children.cycles-pp.tcp_sendmsg
2.59 ? 2% -0.9 1.73 ? 6% perf-profile.children.cycles-pp.tcp_ack
8.10 -0.6 7.48 ? 6% perf-profile.children.cycles-pp.tcp_sendmsg_locked
1.10 -0.6 0.51 ? 8% perf-profile.children.cycles-pp.tcp_init_transfer
0.72 ? 4% -0.6 0.13 ? 7% perf-profile.children.cycles-pp.tcp_ack_update_rtt
0.60 ? 2% -0.6 0.04 ? 44% perf-profile.children.cycles-pp.inet_sk_rebuild_header
2.18 -0.5 1.64 ? 7% perf-profile.children.cycles-pp.tcp_conn_request
5.60 -0.5 5.12 ? 5% perf-profile.children.cycles-pp.tcp_rcv_established
2.38 -0.4 1.93 ? 7% perf-profile.children.cycles-pp.tcp_check_req
1.46 -0.4 1.03 ? 6% perf-profile.children.cycles-pp.ip_route_output_key_hash_rcu
1.29 ? 2% -0.4 0.87 ? 7% perf-profile.children.cycles-pp.tcp_clean_rtx_queue
1.50 -0.4 1.09 ? 6% perf-profile.children.cycles-pp.ip_route_output_key_hash
0.64 ? 3% -0.4 0.28 ? 6% perf-profile.children.cycles-pp.tcp_v4_init_sock
0.63 ? 3% -0.4 0.27 ? 5% perf-profile.children.cycles-pp.tcp_init_sock
1.83 -0.4 1.47 ? 6% perf-profile.children.cycles-pp.tcp_child_process
0.48 ? 2% -0.4 0.12 ? 7% perf-profile.children.cycles-pp.tcp_connect_init
0.38 ? 3% -0.3 0.06 ? 11% perf-profile.children.cycles-pp.tcp_openreq_init_rwin
0.38 ? 3% -0.3 0.10 ? 9% perf-profile.children.cycles-pp.tcp_update_pacing_rate
1.45 -0.3 1.18 ? 6% perf-profile.children.cycles-pp.inet_create
2.03 -0.3 1.77 ? 7% perf-profile.children.cycles-pp.tcp_v4_syn_recv_sock
1.06 ? 2% -0.3 0.81 ? 6% perf-profile.children.cycles-pp.ip_route_output_flow
0.34 ? 4% -0.3 0.08 ? 8% perf-profile.children.cycles-pp.tcp_parse_options
0.60 ? 2% -0.2 0.35 ? 7% perf-profile.children.cycles-pp.tcp_finish_connect
0.53 ? 4% -0.2 0.37 ? 9% perf-profile.children.cycles-pp.tcp_update_metrics
0.18 ? 5% -0.2 0.03 ? 99% perf-profile.children.cycles-pp.tcp_synack_rtt_meas
0.87 ? 4% -0.1 0.74 ? 9% perf-profile.children.cycles-pp.__hrtimer_run_queues
0.43 ? 2% -0.1 0.30 ? 9% perf-profile.children.cycles-pp.inet_csk_route_req
0.36 ? 2% -0.1 0.26 ? 8% perf-profile.children.cycles-pp.fib_table_lookup
0.42 ? 3% -0.1 0.32 ? 11% perf-profile.children.cycles-pp.tcp_init_metrics
0.24 ? 4% -0.1 0.13 ? 10% perf-profile.children.cycles-pp.fib_lookup_good_nhc
0.48 ? 4% -0.1 0.41 ? 3% perf-profile.children.cycles-pp.tcp_time_wait
0.14 ? 5% -0.1 0.07 ? 11% perf-profile.children.cycles-pp.__inet_lookup_listener
0.42 ? 2% -0.1 0.35 ? 5% perf-profile.children.cycles-pp.inet_csk_route_child_sock
0.29 ? 2% -0.1 0.24 ? 6% perf-profile.children.cycles-pp.inet_ehashfn
0.11 ? 4% -0.0 0.07 ? 11% perf-profile.children.cycles-pp.skb_entail
0.19 ? 5% -0.0 0.17 ? 6% perf-profile.children.cycles-pp.__sys_setsockopt
0.08 ? 5% +0.0 0.10 ? 7% perf-profile.children.cycles-pp.sock_put
0.06 ? 7% +0.0 0.08 ? 9% perf-profile.children.cycles-pp.inet_csk_complete_hashdance
0.07 ? 5% +0.0 0.08 ? 5% perf-profile.children.cycles-pp.___perf_sw_event
0.05 ? 7% +0.0 0.07 ? 7% perf-profile.children.cycles-pp.security_sock_rcv_skb
0.05 +0.0 0.07 ? 7% perf-profile.children.cycles-pp.down_write
0.05 ? 8% +0.0 0.07 ? 11% perf-profile.children.cycles-pp.sk_page_frag_refill
0.18 ? 3% +0.0 0.20 ? 2% perf-profile.children.cycles-pp.inet_ehash_insert
0.07 ? 9% +0.0 0.09 ? 9% perf-profile.children.cycles-pp.tcp_send_mss
0.06 ? 6% +0.0 0.08 ? 6% perf-profile.children.cycles-pp._raw_spin_lock_irq
0.07 ? 6% +0.0 0.09 ? 13% perf-profile.children.cycles-pp.tcp_current_mss
0.07 ? 6% +0.0 0.10 ? 8% perf-profile.children.cycles-pp.ipv4_mtu
0.20 ? 2% +0.0 0.23 ? 5% perf-profile.children.cycles-pp.__skb_clone
0.07 ? 11% +0.0 0.10 ? 8% perf-profile.children.cycles-pp.inet_lookup_ifaddr_rcu
0.13 ? 6% +0.0 0.16 ? 6% perf-profile.children.cycles-pp.kmem_cache_alloc_node
0.10 ? 4% +0.0 0.14 ? 9% perf-profile.children.cycles-pp.inet_twsk_alloc
0.08 ? 4% +0.0 0.11 ? 6% perf-profile.children.cycles-pp.__ip_dev_find
0.17 ? 3% +0.0 0.20 ? 7% perf-profile.children.cycles-pp.__netif_receive_skb_core
0.09 ? 5% +0.0 0.12 ? 10% perf-profile.children.cycles-pp.tcp_schedule_loss_probe
0.02 ?141% +0.0 0.06 ? 8% perf-profile.children.cycles-pp.minmax_running_min
0.22 ? 2% +0.0 0.26 ? 8% perf-profile.children.cycles-pp.kmalloc_reserve
0.01 ?223% +0.0 0.06 ? 8% perf-profile.children.cycles-pp.rb_first
0.10 ? 4% +0.0 0.15 ? 6% perf-profile.children.cycles-pp.__destroy_inode
0.46 ? 4% +0.0 0.51 ? 5% perf-profile.children.cycles-pp.tcp_event_new_data_sent
0.10 ? 5% +0.1 0.15 ? 6% perf-profile.children.cycles-pp.destroy_inode
0.00 +0.1 0.06 ? 6% perf-profile.children.cycles-pp.ipv4_default_advmss
0.53 ? 3% +0.1 0.61 ? 7% perf-profile.children.cycles-pp.loopback_xmit
0.74 +0.1 0.83 ? 8% perf-profile.children.cycles-pp.inet_csk_clone_lock
0.59 ? 3% +0.1 0.69 ? 7% perf-profile.children.cycles-pp.dev_hard_start_xmit
1.27 ? 2% +0.1 1.37 ? 5% perf-profile.children.cycles-pp.inet_accept
0.90 +0.1 1.00 ? 8% perf-profile.children.cycles-pp.tcp_create_openreq_child
0.75 ? 2% +0.1 0.86 ? 7% perf-profile.children.cycles-pp.__alloc_skb
0.04 ? 45% +0.1 0.17 ? 9% perf-profile.children.cycles-pp.ip_build_and_send_pkt
0.69 ? 6% +0.1 0.83 ? 9% perf-profile.children.cycles-pp.get_obj_cgroup_from_current
0.82 ? 3% +0.1 0.96 ? 6% perf-profile.children.cycles-pp.native_queued_spin_lock_slowpath
0.76 ? 3% +0.1 0.90 ? 7% perf-profile.children.cycles-pp.sock_alloc_inode
0.92 ? 2% +0.2 1.08 ? 6% perf-profile.children.cycles-pp.alloc_inode
0.99 ? 2% +0.2 1.15 ? 6% perf-profile.children.cycles-pp.sock_alloc
0.98 ? 2% +0.2 1.14 ? 6% perf-profile.children.cycles-pp.new_inode_pseudo
0.33 ? 4% +0.2 0.50 ? 8% perf-profile.children.cycles-pp.__tcp_get_metrics
0.40 ? 5% +0.2 0.58 ? 8% perf-profile.children.cycles-pp.tcp_get_metrics
1.10 ? 4% +0.2 1.31 ? 9% perf-profile.children.cycles-pp.__dentry_kill
0.18 ? 2% +0.2 0.40 ? 5% perf-profile.children.cycles-pp.__ip_finish_output
2.68 +0.2 2.92 ? 5% perf-profile.children.cycles-pp.__sys_accept4_file
2.71 +0.2 2.96 ? 5% perf-profile.children.cycles-pp.__sys_accept4
2.72 +0.3 2.98 ? 5% perf-profile.children.cycles-pp.__x64_sys_accept
2.12 ? 3% +0.3 2.42 ? 7% perf-profile.children.cycles-pp.kmem_cache_alloc
0.38 ? 2% +0.5 0.84 ? 5% perf-profile.children.cycles-pp.ip_rcv_finish
0.34 ? 2% +0.5 0.81 ? 5% perf-profile.children.cycles-pp.ip_rcv_finish_core
1.93 ? 2% -1.9 0.04 ? 44% perf-profile.self.cycles-pp.ipv4_dst_check
0.66 ? 5% -0.6 0.06 ? 9% perf-profile.self.cycles-pp.tcp_ack_update_rtt
0.58 ? 2% -0.4 0.16 ? 8% perf-profile.self.cycles-pp.tcp_write_xmit
0.98 -0.3 0.64 ? 6% perf-profile.self.cycles-pp.ip_route_output_key_hash_rcu
0.37 ? 3% -0.3 0.10 ? 8% perf-profile.self.cycles-pp.tcp_update_pacing_rate
0.33 ? 4% -0.3 0.07 ? 6% perf-profile.self.cycles-pp.tcp_parse_options
0.27 ? 7% -0.2 0.07 ? 18% perf-profile.self.cycles-pp.tcp_update_metrics
0.24 ? 4% -0.1 0.13 ? 10% perf-profile.self.cycles-pp.fib_lookup_good_nhc
0.12 ? 4% -0.1 0.06 ? 7% perf-profile.self.cycles-pp.tcp_init_transfer
0.28 ? 3% -0.1 0.23 ? 6% perf-profile.self.cycles-pp.inet_ehashfn
0.14 ? 6% -0.0 0.10 ? 15% perf-profile.self.cycles-pp.tcp_conn_request
0.10 ? 4% -0.0 0.06 ? 11% perf-profile.self.cycles-pp.skb_entail
0.18 ? 5% -0.0 0.15 ? 7% perf-profile.self.cycles-pp.__inet_bind
0.15 ? 3% -0.0 0.14 ? 6% perf-profile.self.cycles-pp.tcp_rcv_state_process
0.08 ? 4% +0.0 0.10 ? 7% perf-profile.self.cycles-pp.sock_put
0.08 ? 11% +0.0 0.10 ? 7% perf-profile.self.cycles-pp.finish_task_switch
0.05 ? 7% +0.0 0.07 ? 8% perf-profile.self.cycles-pp.___perf_sw_event
0.06 ? 7% +0.0 0.08 ? 5% perf-profile.self.cycles-pp.ipv4_mtu
0.05 ? 9% +0.0 0.08 ? 6% perf-profile.self.cycles-pp._raw_spin_lock_irq
0.04 ? 44% +0.0 0.07 ? 10% perf-profile.self.cycles-pp.tcp_schedule_loss_probe
0.16 ? 4% +0.0 0.19 ? 6% perf-profile.self.cycles-pp.__skb_clone
0.13 ? 7% +0.0 0.16 ? 6% perf-profile.self.cycles-pp.exit_to_user_mode_prepare
0.06 ? 11% +0.0 0.10 ? 8% perf-profile.self.cycles-pp.inet_lookup_ifaddr_rcu
0.10 ? 5% +0.0 0.13 ? 8% perf-profile.self.cycles-pp.inet_twsk_alloc
0.16 ? 2% +0.0 0.20 ? 8% perf-profile.self.cycles-pp.__netif_receive_skb_core
0.06 ? 8% +0.0 0.10 ? 11% perf-profile.self.cycles-pp.loopback_xmit
0.02 ?141% +0.0 0.06 ? 9% perf-profile.self.cycles-pp.__sys_sendto
0.02 ?141% +0.0 0.06 ? 16% perf-profile.self.cycles-pp.ip_route_output_key_hash
0.01 ?223% +0.0 0.06 ? 9% perf-profile.self.cycles-pp.minmax_running_min
0.12 ? 4% +0.1 0.17 ? 7% perf-profile.self.cycles-pp.sk_clone_lock
0.02 ?141% +0.1 0.09 ? 7% perf-profile.self.cycles-pp.inet_create
0.52 ? 4% +0.1 0.61 ? 8% perf-profile.self.cycles-pp.kmem_cache_alloc
0.14 ? 5% +0.1 0.23 ? 7% perf-profile.self.cycles-pp.tcp_connect
0.40 ? 2% +0.1 0.49 ? 8% perf-profile.self.cycles-pp.tcp_v4_rcv
0.66 ? 7% +0.1 0.77 ? 9% perf-profile.self.cycles-pp.get_obj_cgroup_from_current
0.04 ? 44% +0.1 0.17 ? 9% perf-profile.self.cycles-pp.ip_build_and_send_pkt
0.81 ? 3% +0.1 0.96 ? 6% perf-profile.self.cycles-pp.native_queued_spin_lock_slowpath
0.33 ? 5% +0.2 0.49 ? 8% perf-profile.self.cycles-pp.__tcp_get_metrics
0.16 ? 4% +0.2 0.37 ? 5% perf-profile.self.cycles-pp.__ip_finish_output
0.34 ? 2% +0.5 0.80 ? 6% perf-profile.self.cycles-pp.ip_rcv_finish_core
0.40 +0.6 1.00 ? 5% perf-profile.self.cycles-pp.__ip_queue_xmit



netperf.Throughput_tps

24000 +-------------------------------------------------------------------+
| OO |
23000 |O+OO O O OO OOO OO |
| O O |
22000 |-+ |
| |
21000 |-+ O OOO OO OOO OO OO O O O |
| O O |
20000 |-+ |
| |
19000 |-+ |
|+.++.++.+++ +.++. +.++.+ .+++.++.+++.++.++.+++.++.+|
18000 |-+ + .++ ++ + |
| ++.+++.++.++ |
17000 +-------------------------------------------------------------------+


netperf.Throughput_total_tps

380000 +------------------------------------------------------------------+
370000 |-+ O OO OO OO OO |
|O OO OO O O |
360000 |-+ |
350000 |-+ |
340000 |-+ |
330000 |-+ OO OOO OO OOO OOO OO OOO |
| |
320000 |-+ |
310000 |-+ |
300000 |-+ |
290000 |+.++.+++.++ .+++.+ .+++.+++.++.+++.++.+++.++.+++.++.+|
| + + .+ + + + |
280000 |-+ + + +.+ +.+ |
270000 +------------------------------------------------------------------+


netperf.workload

1.15e+08 +----------------------------------------------------------------+
| OO O |
1.1e+08 |O+OOO O OOO OOO O |
| O |
1.05e+08 |-+ |
| |
1e+08 |-+ OO OOO OO OOO OOO OOO OO |
| |
9.5e+07 |-+ |
| |
9e+07 |-.+ +. |
|+ ++.+++.+ .+++.+ .+++.+++.+++.+++.+++.++.+++.++ +|
8.5e+07 |-+ +.+ +.+++.+++ + |
| + |
8e+07 +----------------------------------------------------------------+


netperf.time.voluntary_context_switches

2.2e+08 +-----------------------------------------------------------------+
|O OO OO O OOO OO |
2.1e+08 |-+ |
| |
2e+08 |-+ |
| O OO OOO OOO OO OOO OOO O |
1.9e+08 |-+ |
| |
1.8e+08 |-+ |
|+.+ ++.+ + +. |
1.7e+08 |-+ +.+ + +.+++. ++.++.+++.+++.++.+++.+ +.++.++ +|
| : .+ + |
1.6e+08 |-+ +.++.+++.+++ |
| |
1.5e+08 +-----------------------------------------------------------------+




3.4e+06 +-----------------------------------------------------------------+
| O |
3.3e+06 |O+ O O O O OOO |
| O O O OO |
3.2e+06 |-+ O O |
3.1e+06 |-+ |
| O OO OOO OOO OO O O |
3e+06 |-+ OO O O |
| O |
2.9e+06 |-+ |
2.8e+06 |-+ |
| |
2.7e+06 |-+ .+ .+++. +.++. + + +. +. ++.++.+ +.+|
|+.+ .+++.+++.++ ++.+ +.++ ++ + +.+ +.+ ++ + + |
2.6e+06 +-----------------------------------------------------------------+


[*] bisect-good sample
[O] bisect-bad sample



Disclaimer:
Results have been estimated based on internal Intel analysis and are provided
for informational purposes only. Any difference in system hardware or software
design or configuration may affect actual performance.


---
0DAY/LKP+ Test Infrastructure Open Source Technology Center
https://lists.01.org/hyperkitty/list/[email protected] Intel Corporation

Thanks,
Oliver Sang


Attachments:
(No filename) (49.48 kB)
config-5.13.0-rc1-00070-g2463a604a867 (176.78 kB)
job-script (8.26 kB)
job.yaml (5.33 kB)
reproduce (1.09 kB)
Download all attachments

2021-05-25 18:03:32

by Bart Van Assche

[permalink] [raw]
Subject: Re: 2463a604a8: netperf.Throughput_tps 12.8% improvement

On 5/24/21 11:44 PM, kernel test robot wrote:
> FYI, we noticed a 12.8% improvement of netperf.Throughput_tps due to commit:
>
> commit: 2463a604a86728777ce4284214a52de46a808c9e ("[PATCH v3 2/3] Introduce enums for the SAM, message, host and driver status codes")
> url: https://github.com/0day-ci/linux/commits/Bart-Van-Assche/Introduce-enums-for-SCSI-status-codes/20210524-105751
> base: https://git.kernel.org/cgit/linux/kernel/git/mkp/scsi.git for-next
>
> in testcase: netperf
> on test machine: 192 threads 4 sockets Intel(R) Xeon(R) Platinum 9242 CPU @ 2.30GHz with 192G memory
> with following parameters:
>
> ip: ipv4
> runtime: 300s
> nr_threads: 16
> cluster: cs-localhost
> test: TCP_CRR
> cpufreq_governor: performance
> ucode: 0x5003006
>
> test-description: Netperf is a benchmark that can be use to measure various aspect of networking performance.
> test-url: http://www.netperf.org/netperf/

The above email reports a performance improvement for the networking
subsystem while my patch only affects the SCSI subsystem and should not
have any performance impact. I'm confused by the above feedback ...

Bart.

2021-05-26 04:09:03

by Xing Zhengjun

[permalink] [raw]
Subject: Re: [LKP] Re: 2463a604a8: netperf.Throughput_tps 12.8% improvement



On 5/26/2021 12:44 AM, Bart Van Assche wrote:
> On 5/24/21 11:44 PM, kernel test robot wrote:
>> FYI, we noticed a 12.8% improvement of netperf.Throughput_tps due to commit:
>>
>> commit: 2463a604a86728777ce4284214a52de46a808c9e ("[PATCH v3 2/3] Introduce enums for the SAM, message, host and driver status codes")
>> url: https://github.com/0day-ci/linux/commits/Bart-Van-Assche/Introduce-enums-for-SCSI-status-codes/20210524-105751
>> base: https://git.kernel.org/cgit/linux/kernel/git/mkp/scsi.git for-next
>>
>> in testcase: netperf
>> on test machine: 192 threads 4 sockets Intel(R) Xeon(R) Platinum 9242 CPU @ 2.30GHz with 192G memory
>> with following parameters:
>>
>> ip: ipv4
>> runtime: 300s
>> nr_threads: 16
>> cluster: cs-localhost
>> test: TCP_CRR
>> cpufreq_governor: performance
>> ucode: 0x5003006
>>
>> test-description: Netperf is a benchmark that can be use to measure various aspect of networking performance.
>> test-url: http://www.netperf.org/netperf/
> The above email reports a performance improvement for the networking
> subsystem while my patch only affects the SCSI subsystem and should not
> have any performance impact. I'm confused by the above feedback ...

I suspect it related with cache alignment, 2463a604a8 changes "u8"
(size:1) to "enum xxx_status" (size: 4),  the cache alignment is better
than before , so cause the improvement.
>
> Bart.
> _______________________________________________
> LKP mailing list -- [email protected]
> To unsubscribe send an email to [email protected]

--
Zhengjun Xing