Greeting,
FYI, we noticed a 91.4% improvement of aim7.jobs-per-min due to commit:
commit: 84b89e5d943d8d6b997fcebb4609373e1b3e4b03 ("f2fs: add auto tuning for small devices")
https://git.kernel.org/cgit/linux/kernel/git/jaegeuk/f2fs.git dev-test
in testcase: aim7
on test machine: 40 threads Intel(R) Xeon(R) CPU E5-2690 v2 @ 3.00GHz with 384G memory
with following parameters:
disk: 4BRD_12G
md: RAID1
fs: f2fs
test: sync_disk_rw
load: 600
cpufreq_governor: performance
test-description: AIM7 is a traditional UNIX system level benchmark suite which is used to test and measure the performance of multiuser system.
test-url: https://sourceforge.net/projects/aimbench/files/aim-suite7/
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
=========================================================================================
compiler/cpufreq_governor/disk/fs/kconfig/load/md/rootfs/tbox_group/test/testcase:
gcc-7/performance/4BRD_12G/f2fs/x86_64-rhel-7.2/600/RAID1/debian-x86_64-2016-08-31.cgz/lkp-ivb-ep01/sync_disk_rw/aim7
commit:
079396270b ("f2fs: add mount option for segment allocation policy")
84b89e5d94 ("f2fs: add auto tuning for small devices")
079396270b0f1e98 84b89e5d943d8d6b997fcebb46
---------------- --------------------------
%stddev %change %stddev
\ | \
2085 +91.4% 3992 aim7.jobs-per-min
1726 -47.8% 901.87 aim7.time.elapsed_time
1726 -47.8% 901.87 aim7.time.elapsed_time.max
4.453e+08 -28.1% 3.201e+08 aim7.time.file_system_outputs
6353142 +2.8% 6530834 aim7.time.involuntary_context_switches
55733 -12.0% 49024 ? 2% aim7.time.minor_page_faults
10088 -31.9% 6873 aim7.time.system_time
137.87 -14.1% 118.39 aim7.time.user_time
2.318e+08 -34.8% 1.512e+08 aim7.time.voluntary_context_switches
573596 -46.8% 304950 interrupts.CAL:Function_call_interrupts
13.99 +4.7 18.70 mpstat.cpu.sys%
0.24 +0.1 0.37 mpstat.cpu.usr%
3963108 ? 4% -42.1% 2292711 ? 3% softirqs.RCU
3796151 -44.2% 2117326 softirqs.SCHED
7874455 -39.6% 4758552 softirqs.TIMER
11719191 -55.6% 5199095 ? 2% numa-numastat.node0.local_node
11723711 -55.6% 5202967 ? 2% numa-numastat.node0.numa_hit
12092691 -57.0% 5204988 ? 2% numa-numastat.node1.local_node
12095947 -56.9% 5208887 ? 2% numa-numastat.node1.numa_hit
126642 +37.9% 174684 vmstat.io.bo
3.48e+08 +12.5% 3.914e+08 vmstat.memory.free
5.00 +45.0% 7.25 ? 5% vmstat.procs.r
297230 +33.3% 396251 vmstat.system.cs
151400 -13.5% 130952 meminfo.Active
21942 -80.1% 4362 meminfo.Active(file)
103327 +10.9% 114591 meminfo.Dirty
3.466e+08 +12.5% 3.899e+08 meminfo.MemAvailable
3.48e+08 +12.5% 3.914e+08 meminfo.MemFree
164848 -58.0% 69270 meminfo.SReclaimable
27687 ? 3% -17.2% 22925 ? 4% meminfo.Shmem
255899 -37.9% 158857 meminfo.Slab
1.067e+10 ? 2% -33.3% 7.116e+09 cpuidle.C1.time
1.07e+08 -32.8% 71977775 cpuidle.C1.usage
7.845e+09 -28.5% 5.613e+09 cpuidle.C1E.time
49422496 -23.5% 37827986 cpuidle.C1E.usage
1.45e+10 -42.4% 8.344e+09 cpuidle.C3.time
69214148 -29.4% 48831739 cpuidle.C3.usage
2.166e+10 -73.6% 5.717e+09 cpuidle.C6.time
66088990 -61.5% 25470289 cpuidle.C6.usage
46247917 ? 16% -64.2% 16560051 ? 6% cpuidle.POLL.time
288470 ? 13% -69.2% 88832 cpuidle.POLL.usage
22682 ? 3% +147.1% 56051 slabinfo.ext4_io_end.active_objs
354.50 ? 3% +150.8% 889.25 slabinfo.ext4_io_end.active_slabs
22718 ? 3% +150.6% 56935 slabinfo.ext4_io_end.num_objs
354.50 ? 3% +150.8% 889.25 slabinfo.ext4_io_end.num_slabs
1343 ? 4% +11.7% 1500 ? 2% slabinfo.f2fs_inode_cache.active_objs
1343 ? 4% +11.7% 1500 ? 2% slabinfo.f2fs_inode_cache.num_objs
750.75 +15.3% 865.50 slabinfo.names_cache.num_objs
6774 ? 4% -48.4% 3498 ? 8% slabinfo.pool_workqueue.active_objs
6776 ? 3% -48.3% 3500 ? 8% slabinfo.pool_workqueue.num_objs
197904 -86.4% 26826 slabinfo.radix_tree_node.active_objs
7071 -86.4% 962.50 slabinfo.radix_tree_node.active_slabs
198023 -86.4% 26967 slabinfo.radix_tree_node.num_objs
7071 -86.4% 962.50 slabinfo.radix_tree_node.num_slabs
258.75 +23.1% 318.50 turbostat.Avg_MHz
21.58 +4.9 26.52 turbostat.Busy%
1.07e+08 -32.8% 71973266 turbostat.C1
15.44 ? 2% +4.3 19.70 turbostat.C1%
49422345 -23.5% 37827909 turbostat.C1E
11.35 +4.2 15.54 turbostat.C1E%
69213989 -29.4% 48831644 turbostat.C3
66087305 -61.5% 25468764 turbostat.C6
31.34 -15.5 15.82 turbostat.C6%
3.65 -65.1% 1.27 turbostat.CPU%c3
6.04 ? 2% -69.0% 1.88 ? 2% turbostat.CPU%c6
75467541 -47.5% 39623922 turbostat.IRQ
0.31 ? 18% +120.2% 0.68 ? 18% turbostat.Pkg%pc2
138214 -47.7% 72250 turbostat.SMI
64816 ? 42% -48.3% 33504 ? 9% numa-meminfo.node0.Active
11005 -81.1% 2082 ? 10% numa-meminfo.node0.Active(file)
51333 +12.0% 57487 numa-meminfo.node0.Dirty
170715 -7.5% 157887 ? 2% numa-meminfo.node0.Inactive
9647 ? 2% -69.6% 2929 ?114% numa-meminfo.node0.Inactive(anon)
13828 -25.4% 10309 ? 19% numa-meminfo.node0.Mapped
1.745e+08 +12.2% 1.958e+08 numa-meminfo.node0.MemFree
23538613 ? 4% -90.3% 2286601 ? 17% numa-meminfo.node0.MemUsed
80101 ? 5% -57.5% 34053 ? 7% numa-meminfo.node0.SReclaimable
51286 ? 10% -21.7% 40165 ? 16% numa-meminfo.node0.SUnreclaim
131388 ? 7% -43.5% 74219 ? 9% numa-meminfo.node0.Slab
10935 -79.1% 2281 ? 9% numa-meminfo.node1.Active(file)
51850 +10.6% 57350 numa-meminfo.node1.Dirty
1259 ? 20% +459.8% 7051 ? 47% numa-meminfo.node1.Inactive(anon)
6317 ? 65% +77.5% 11215 ? 36% numa-meminfo.node1.KernelStack
1.735e+08 +12.8% 1.956e+08 numa-meminfo.node1.MemFree
24679786 ? 4% -89.8% 2511754 ? 15% numa-meminfo.node1.MemUsed
84745 ? 4% -58.4% 35215 ? 7% numa-meminfo.node1.SReclaimable
124508 ? 7% -32.0% 84640 ? 7% numa-meminfo.node1.Slab
5485 -80.1% 1090 proc-vmstat.nr_active_file
54892885 -27.7% 39667532 proc-vmstat.nr_dirtied
25863 +11.0% 28717 proc-vmstat.nr_dirty
8654369 +12.5% 9737487 proc-vmstat.nr_dirty_background_threshold
17329899 +12.5% 19498783 proc-vmstat.nr_dirty_threshold
86998864 +12.5% 97854032 proc-vmstat.nr_free_pages
6922 ? 3% -17.2% 5732 ? 4% proc-vmstat.nr_shmem
41211 -58.0% 17315 proc-vmstat.nr_slab_reclaimable
54691555 -27.8% 39466623 proc-vmstat.nr_written
5485 -80.1% 1090 proc-vmstat.nr_zone_active_file
25609 +11.9% 28645 proc-vmstat.nr_zone_write_pending
24152 ? 3% -28.1% 17368 ? 6% proc-vmstat.numa_hint_faults
13032 ? 5% -27.0% 9510 ? 8% proc-vmstat.numa_hint_faults_local
23847081 -56.2% 10441460 proc-vmstat.numa_hit
23839301 -56.2% 10433680 proc-vmstat.numa_local
31136 ? 5% -31.9% 21201 ? 3% proc-vmstat.numa_pte_updates
63833 ? 2% -79.2% 13291 ? 2% proc-vmstat.pgactivate
23961726 -56.2% 10492221 proc-vmstat.pgalloc_normal
52104 ? 5% -83.4% 8648 ? 2% proc-vmstat.pgdeactivate
4368571 -47.3% 2302031 proc-vmstat.pgfault
11372855 -13.9% 9789645 proc-vmstat.pgfree
2.188e+08 -27.8% 1.579e+08 proc-vmstat.pgpgout
104071 ? 5% -83.5% 17174 ? 3% proc-vmstat.pgrotated
2751 -81.1% 520.00 ? 10% numa-vmstat.node0.nr_active_file
13654269 -28.4% 9776555 numa-vmstat.node0.nr_dirtied
12855 +11.9% 14380 numa-vmstat.node0.nr_dirty
43631346 +12.2% 48944373 numa-vmstat.node0.nr_free_pages
2411 ? 2% -69.6% 732.00 ?114% numa-vmstat.node0.nr_inactive_anon
3551 -24.7% 2672 ? 19% numa-vmstat.node0.nr_mapped
20025 ? 5% -57.5% 8512 ? 7% numa-vmstat.node0.nr_slab_reclaimable
12821 ? 10% -21.7% 10040 ? 16% numa-vmstat.node0.nr_slab_unreclaimable
13592188 -28.5% 9713197 numa-vmstat.node0.nr_written
2751 -81.1% 520.00 ? 10% numa-vmstat.node0.nr_zone_active_file
2411 ? 2% -69.6% 732.00 ?114% numa-vmstat.node0.nr_zone_inactive_anon
12729 +12.7% 14348 numa-vmstat.node0.nr_zone_write_pending
8773276 ? 2% -64.9% 3080220 ? 3% numa-vmstat.node0.numa_hit
8768509 ? 2% -64.9% 3076129 ? 3% numa-vmstat.node0.numa_local
2733 -79.1% 570.00 ? 9% numa-vmstat.node1.nr_active_file
13633751 -28.0% 9817880 numa-vmstat.node1.nr_dirtied
12985 +10.6% 14362 numa-vmstat.node1.nr_dirty
43367610 +12.8% 48909638 numa-vmstat.node1.nr_free_pages
314.75 ? 20% +460.3% 1763 ? 47% numa-vmstat.node1.nr_inactive_anon
6317 ? 65% +77.5% 11214 ? 36% numa-vmstat.node1.nr_kernel_stack
21185 ? 4% -58.4% 8803 ? 7% numa-vmstat.node1.nr_slab_reclaimable
13571355 -28.1% 9754541 numa-vmstat.node1.nr_written
2733 -79.1% 570.00 ? 9% numa-vmstat.node1.nr_zone_active_file
314.75 ? 20% +460.3% 1763 ? 47% numa-vmstat.node1.nr_zone_inactive_anon
12856 +11.5% 14330 numa-vmstat.node1.nr_zone_write_pending
9085874 ? 2% -65.3% 3154882 ? 3% numa-vmstat.node1.numa_hit
8908436 ? 2% -66.6% 2976162 ? 3% numa-vmstat.node1.numa_local
1.647e+12 -34.5% 1.078e+12 perf-stat.branch-instructions
2.86 -0.8 2.05 perf-stat.branch-miss-rate%
4.703e+10 -53.1% 2.207e+10 perf-stat.branch-misses
3.988e+10 -38.3% 2.461e+10 perf-stat.cache-misses
2.15e+11 -38.4% 1.324e+11 perf-stat.cache-references
5.135e+08 -30.2% 3.582e+08 perf-stat.context-switches
2.36 -2.8% 2.30 perf-stat.cpi
1.732e+13 -35.8% 1.112e+13 perf-stat.cpu-cycles
38022672 -48.1% 19727798 perf-stat.cpu-migrations
2.76 -0.9 1.85 ? 9% perf-stat.dTLB-load-miss-rate%
5.165e+10 ? 2% -55.2% 2.314e+10 ? 9% perf-stat.dTLB-load-misses
1.821e+12 ? 2% -32.4% 1.231e+12 perf-stat.dTLB-loads
0.17 ? 4% -0.0 0.15 ? 6% perf-stat.dTLB-store-miss-rate%
1.815e+09 ? 5% -43.7% 1.021e+09 ? 7% perf-stat.dTLB-store-misses
1.05e+12 -36.2% 6.702e+11 perf-stat.dTLB-stores
8.035e+09 -31.0% 5.546e+09 perf-stat.iTLB-load-misses
1.691e+09 ? 2% -27.8% 1.22e+09 ? 6% perf-stat.iTLB-loads
7.334e+12 -34.0% 4.84e+12 perf-stat.instructions
912.90 -4.4% 872.78 perf-stat.instructions-per-iTLB-miss
0.42 +2.8% 0.44 perf-stat.ipc
4278365 -47.3% 2254777 perf-stat.minor-faults
47.56 -1.1 46.50 perf-stat.node-load-miss-rate%
1.607e+10 -42.8% 9.183e+09 perf-stat.node-load-misses
1.771e+10 -40.3% 1.057e+10 perf-stat.node-loads
40.19 +1.5 41.73 perf-stat.node-store-miss-rate%
1.267e+10 -25.3% 9.46e+09 perf-stat.node-store-misses
1.885e+10 -29.9% 1.321e+10 perf-stat.node-stores
4278418 -47.3% 2254809 perf-stat.page-faults
aim7.jobs-per-min
5000 +-+-----------------------------O--O-O-O-----------------------------+
4500 O-O O O O O O O |
| O O O O O |
4000 +-+ O O O O |
3500 +-+ |
| |
3000 +-+ |
2500 +-+ |
2000 +-+.+.+..+ .+.+.+.+.+.+..+.+.+.+.+.+..+.+.+.|
| : +.+.+..+.+ +.+.+..+ |
1500 +-+ : : : : |
1000 +-+ : : : : |
| : : : : |
500 +-+ : : : |
0 +-+--------O-O-------------------------------------------------------+
interrupts.CAL:Function_call_interrupts
600000 +-+----------------------------------------------------------------+
| + + + : : +.+.+.+.+.+ + +..+.+.+.+.+.|
500000 +-+ : : : : |
| : : : : |
| : : : : |
400000 +-+ : : : : |
| : : : : |
300000 +-+ : : : : O O O O |
O O O O O: :O O O O O:O:O O O O O O |
200000 +-+ : : : : |
| : : : : |
| : : : |
100000 +-+ : : : |
| : : : |
0 +-+-------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,
Xiaolong