Hello,
kernel test robot noticed a 17.8% improvement of stress-ng.cpu-online.ops_per_sec on:
commit: 2125c0034c5dfd61171b494bd309bb7637bff6eb ("cgroup/cpuset: Make cpuset hotplug processing synchronous")
https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git master
testcase: stress-ng
test machine: 64 threads 2 sockets Intel(R) Xeon(R) Gold 6346 CPU @ 3.10GHz (Ice Lake) with 256G memory
parameters:
nr_threads: 100%
testtime: 60s
test: cpu-online
cpufreq_governor: performance
In addition to that, the commit also has significant impact on the following tests:
+------------------+---------------------------------------------------------------------------------------------+
| testcase: change | stress-ng: stress-ng.cpu-online.ops_per_sec 13.1% improvement |
| test machine | 224 threads 2 sockets Intel(R) Xeon(R) Platinum 8480CTDX (Sapphire Rapids) with 512G memory |
| test parameters | cpufreq_governor=performance |
| | nr_threads=100% |
| | test=cpu-online |
| | testtime=60s |
+------------------+---------------------------------------------------------------------------------------------+
Details are as below:
-------------------------------------------------------------------------------------------------->
The kernel config and materials to reproduce are available at:
https://download.01.org/0day-ci/archive/20240522/202405222215.d537a85f-oliver.sang@xxxxxxxxx
=========================================================================================
compiler/cpufreq_governor/kconfig/nr_threads/rootfs/tbox_group/test/testcase/testtime:
gcc-13/performance/x86_64-rhel-8.3/100%/debian-12-x86_64-20240206.cgz/lkp-icl-2sp9/cpu-online/stress-ng/60s
commit:
4793cb599b ("selftests: cgroup: skip test_cgcore_lesser_ns_open when cgroup2 mounted without nsdelegate")
2125c0034c ("cgroup/cpuset: Make cpuset hotplug processing synchronous")
4793cb599b1bdc3d 2125c0034c5dfd61171b494bd30
---------------- ---------------------------
%stddev %change %stddev
\ | \
2.751e+09 ± 2% -9.7% 2.483e+09 ± 2% cpuidle..time
4.07 ± 12% -20.6% 3.23 ± 11% vmstat.procs.r
2.87 +0.5 3.34 ± 4% mpstat.cpu.all.sys%
16.83 ± 37% -68.3% 5.33 ± 64% mpstat.max_utilization.seconds
1130 ± 23% +72.3% 1948 ± 14% numa-meminfo.node0.Active
1119 ± 24% +69.2% 1894 ± 15% numa-meminfo.node0.Active(anon)
280.28 ± 24% +69.1% 474.04 ± 15% numa-vmstat.node0.nr_active_anon
280.28 ± 24% +69.1% 474.03 ± 15% numa-vmstat.node0.nr_zone_active_anon
0.03 ± 63% +202.3% 0.10 ± 65% perf-stat.i.major-faults
0.03 ± 62% +200.9% 0.09 ± 65% perf-stat.ps.major-faults
171.45 ± 14% -25.8% 127.20 ± 17% sched_debug.cfs_rq:/.load_avg.avg
116.43 ± 10% -31.4% 79.84 ± 23% sched_debug.cfs_rq:/.removed.load_avg.avg
319.20 ± 4% -16.3% 267.32 ± 10% sched_debug.cfs_rq:/.removed.load_avg.stddev
111.20 ± 45% +114.3% 238.25 ± 32% stress-ng.cpu-online.millisecs_per_online_action
516.50 +14.8% 593.00 ± 3% stress-ng.cpu-online.ops
7.72 ± 2% +17.8% 9.09 ± 3% stress-ng.cpu-online.ops_per_sec
***************************************************************************************************
lkp-spr-2sp4: 224 threads 2 sockets Intel(R) Xeon(R) Platinum 8480CTDX (Sapphire Rapids) with 512G memory
=========================================================================================
compiler/cpufreq_governor/kconfig/nr_threads/rootfs/tbox_group/test/testcase/testtime:
gcc-13/performance/x86_64-rhel-8.3/100%/debian-12-x86_64-20240206.cgz/lkp-spr-2sp4/cpu-online/stress-ng/60s
commit:
4793cb599b ("selftests: cgroup: skip test_cgcore_lesser_ns_open when cgroup2 mounted without nsdelegate")
2125c0034c ("cgroup/cpuset: Make cpuset hotplug processing synchronous")
4793cb599b1bdc3d 2125c0034c5dfd61171b494bd30
---------------- ---------------------------
%stddev %change %stddev
\ | \
-13.50 +34.6% -18.17 sched_debug.cpu.nr_uninterruptible.min
59969 +5.2% 63066 vmstat.system.cs
0.54 ± 4% +0.1 0.63 ± 2% mpstat.cpu.all.irq%
0.17 ± 3% +0.0 0.18 ± 2% mpstat.cpu.all.usr%
1678 ± 5% +13.9% 1912 ± 7% perf-stat.i.cpu-migrations
1660 ± 5% +13.9% 1890 ± 7% perf-stat.ps.cpu-migrations
479.33 ± 4% +7.5% 515.50 stress-ng.cpu-online.ops
4.86 ± 4% +13.1% 5.50 ± 3% stress-ng.cpu-online.ops_per_sec
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
