Greeting,
FYI, we noticed a 1.2% improvement of pigz.throughput due to commit:
commit: ca7277e7aea43bbfb96a378bac115f6d51c4e1a0 ("mm: do not add anon pages to
LRU")
git://bee.sh.intel.com/git/aaron/linux.git zone_range_lock
in testcase: reaim
on test machine: 256 threads Phi with 96G memory
with following parameters:
runtime: 300s
nr_task: 100%
test: aim9
cpufreq_governor: performance
test-description: REAIM is an updated and improved version of AIM 7 benchmark.
test-url:
https://sourceforge.net/projects/re-aim-7/
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 run job.yaml
=========================================================================================
blocksize/compiler/cpufreq_governor/kconfig/nr_threads/rootfs/tbox_group/testcase/ucode:
512K/gcc-7/performance/x86_64-rhel-7.2/25%/debian-x86_64-2018-04-03.cgz/lkp-ivb-d03/pigz/0x20
commit:
v4.19
ca7277e7ae ("mm: do not add anon pages to LRU")
v4.19 ca7277e7aea43bbfb96a378bac
---------------- --------------------------
fail:runs %reproduction fail:runs
| | |
0:10 1% 0:4
perf-profile.children.cycles-pp.schedule_timeout
%stddev %change %stddev
\ | \
19488588 +1.2% 19718296 pigz.throughput
5706 ± 2% +15.7% 6600 pigz.time.minor_page_faults
15157 +1.2% 15336 pigz.time.voluntary_context_switches
1011862 ± 27% +52.3% 1541212 ± 38% cpuidle.C1E.usage
8.85 ± 28% +4.0 12.88 ± 20% perf-stat.cache-miss-rate%
1011860 ± 27% +52.3% 1541211 ± 38% turbostat.C1E
7.60 ±197% +214.4% 23.91 ± 99% turbostat.CPU%c6
489.11 ± 5% -100.0% 0.00 uptime.boot
1602 ± 6% -100.0% 0.00 uptime.idle
246712 -94.7% 13016 ± 7% meminfo.Active
246708 -94.7% 13012 ± 7% meminfo.Active(anon)
181716 -99.7% 512.00 ±173% meminfo.AnonHugePages
0.01 ±300% +0.1 0.15 ± 28% perf-profile.children.cycles-pp.update_curr
0.08 ±117% +0.2 0.28 ± 71% perf-profile.children.cycles-pp._dl_addr
0.01 ±300% +0.1 0.15 ± 28% perf-profile.self.cycles-pp.update_curr
0.07 ±108% +0.2 0.28 ± 71% perf-profile.self.cycles-pp._dl_addr
61731 -94.7% 3273 ± 7% proc-vmstat.nr_active_anon
60876 -1.5% 59968 proc-vmstat.nr_anon_pages
985.20 -2.4% 961.75 proc-vmstat.nr_page_table_pages
61735 -94.7% 3290 ± 7% proc-vmstat.nr_zone_active_anon
2367750 +1.6% 2405509 proc-vmstat.numa_hit
2367750 +1.6% 2405509 proc-vmstat.numa_local
46131 ± 7% -100.0% 0.00 proc-vmstat.pgalloc_dma32
reaim.time.involuntary_context_switches
270000 +-+----------------------------------------------------------------+
| O |
265000 O-+ O |
| O O O |
260000 +-+ O O |
| O |
255000 +-+ |
| O O O O O O |
250000 +-+ O O O O O |
| O O O O O
245000 +-+ |
| .+.. |
240000 +-+ .+..+..+. |
|..+..+. .+.. .+..+ +.+..+.. .+. .+..+..+.. .+.. |
235000 +-+----------------------------------------------------------------+
reaim.time.user_time
11750 +-+-----------------------------------------------------------------+
| O O O O O O |
11700 +-+O O O O O O O O O O O O O O
O O O O O |
11650 +-+ |
| |
11600 +-+ |
| |
11550 +-+ .+..+.. |
|.. .+. +..+. .+.. .+..+.. .+ |
11500 +-++.. +.+. +..+. +.. +.+. +.. +.+. |
| .. .. .. |
11450 +-+ + + + |
| |
11400 +-+-----------------------------------------------------------------+
reaim.time.system_time
10200 +-+-----------------------------------------------------------------+
| |
10000 +-++..+..+.+..+..+..+..+..+. .+..+..+..+..+.+..+..+..+..+..+.+..+ |
9800 +-+ +. |
| |
9600 +-+ |
| |
9400 +-+ |
| |
9200 +-+ |
9000 +-+ |
| |
8800 +-+ O O O O O O O O |
| O O O O O O O O O O O O O O O
8600 O-+--------O--------------------------------------------------------+
reaim.time.percent_of_cpu_this_job_got
7000 +-+------------------------------------------------------------------+
|..+..+..+..+.+..+..+..+..+..+.. .+.+..+..+..+..+..+..+..+.+..+..+ |
| +. |
6950 +-+ |
| |
| |
6900 +-+ |
| |
6850 +-+ |
| |
| |
6800 +-+ O O O O |
| O O O O O O O O O O O
| O O O O O O O |
6750 O-+-----------------------O--------O---------------------------------+
reaim.time.elapsed_time
308 +-+-------------------------------------------------------------------+
|..+..+..+. +. +. +..+..+..+. +..+.+..+..+. + |
307 +-+ |
306 +-+ |
| |
305 +-+ |
| |
304 +-+ |
| |
303 +-+ |
302 +-+ O O O O O |
| O O O O O O O O O
301 O-+O O O O O O O O O O |
| |
300 +-+-------------------------------------------------------------------+
reaim.time.elapsed_time.max
308 +-+-------------------------------------------------------------------+
|..+..+..+. +. +. +..+..+..+. +..+.+..+..+. + |
307 +-+ |
306 +-+ |
| |
305 +-+ |
| |
304 +-+ |
| |
303 +-+ |
302 +-+ O O O O O |
| O O O O O O O O O
301 O-+O O O O O O O O O O |
| |
300 +-+-------------------------------------------------------------------+
reaim.time.voluntary_context_switches
2.86e+06 +-+--------------------------------------------------------------+
| |
2.85e+06 +-+ .+.+..+..+.+..+..+..+. .+..+. .+ |
2.84e+06 +-++.+..+..+. +..+..+.+. +..+..+..+ |
| |
2.83e+06 +-+ |
| |
2.82e+06 +-+ |
| |
2.81e+06 +-+ |
2.8e+06 +-+ O O O O O O O
| O O O O O O O O O O |
2.79e+06 +-+ O O O O |
O O O O |
2.78e+06 +-+--------------------------------------------------------------+
reaim.parent_time
15.95 +-+-----------------------------------------------------------------+
15.9 +-+ .+.. .+..+..+..+. .+.. .+.. .+.. |
|..+..+..+ +. +. +..+..+..+ +..+..+..+..+ + |
15.85 +-+ |
15.8 +-+ |
| |
15.75 +-+ |
15.7 +-+ |
15.65 +-+ |
| |
15.6 +-+ O |
15.55 +-+ O O O O O O O O O O O O
O O O O O O O O O |
15.5 +-+ O O O |
15.45 +-+-----------------------------------------------------------------+
reaim.child_systime
590 +-+-------------------------------------------------------------------+
|..+. +..+..+..+.+..+..+..+..+..+..+. +..+..+..+.+. +..+..+ |
580 +-+ |
570 +-+ |
| |
560 +-+ |
550 +-+ |
| |
540 +-+ |
530 +-+ |
| |
520 +-+ O O O O O O |
510 +-+O O O O O O O O O O O O
O O O O O O O |
500 +-+-------------------------------------------------------------------+
reaim.child_utime
690 +-+-------------------------------------------------------------------+
| O O O O O O O O O |
688 +-+O O O O O O O O O O O O
686 O-+ O O O |
| |
684 +-+ |
682 +-+ |
| |
680 +-+ |
678 +-+ .+.+.. |
|.. .+. +..+.. .+.. .+..+. .+ |
676 +-++ +..+. +..+. + +..+. + +..+. |
674 +-+ + + + + + + |
| + + + + + + |
672 +-+-------------------------------------------------------------------+
reaim.jobs_per_min
119000 +-+--------O--------O-O--------------------------------------------+
O O O O O O O O O |
118500 +-+ O O O O O O O O O O
| O O O |
118000 +-+ |
| |
117500 +-+ |
| |
117000 +-+ |
| |
116500 +-+ |
| +.. |
116000 +-++..+.+..+..+.. .+. .. .+.+..+..+..+.+..+..+..+..+.+..+ |
| +. +..+ +. |
115500 +-+----------------------------------------------------------------+
reaim.jobs_per_min_child
466 +-+-------------------------------------------------------------------+
| O O |
464 O-+O O O O O O O O |
| O O O O O O O O O O O
462 +-+ O O O |
| |
460 +-+ |
| |
458 +-+ |
| |
456 +-+ |
| |
454 +-+ .+..+.. +.. .+..+.. .+..+.. |
|..+. +..+..+.+..+.. .. .+..+. +..+..+.+. +..+ |
452 +-+-------------------------------------------------------------------+
reaim.std_dev_time
0.7 +-+------------------------------------------------------------------+
| |
0.68 +-++..+.. .+.+..+..+..+..+..+..+..+.+..+.. .+..+..+..+..+. .+..+ |
0.66 +-+ +. +. +. |
| |
0.64 +-+ |
0.62 +-+ |
| |
0.6 +-+ |
0.58 +-+ |
| |
0.56 +-+ |
0.54 +-+ O O O O O O O O O O O O O O
O O O O O O O O O O O |
0.52 +-+------------------------------------------------------------------+
reaim.std_dev_percent
4.6 +-+-------------------------------------------------------------------+
4.5 +-++.. .+..+.. .+..+..+.. .+..+..+.. .+..+..+.+.. .+..+ |
| +..+. + +. +..+. +..+. |
4.4 +-+ |
4.3 +-+ |
| |
4.2 +-+ |
4.1 +-+ |
4 +-+ |
| |
3.9 +-+ |
3.8 +-+ |
| O O O O |
3.7 +-+ O O O O O O O O O O O O O O O
3.6 O-+O--O--O--O----------------------------O----------------------------+
[*] bisect-good sample
[O] bisect-bad sample
Disclaimer:
Results have been estimated based on internal Intel analysis and are provided
for informational purposes only. Any difference in system hardware or software
design or configuration may affect actual performance.
Thanks,
Rong Chen