Greeting,
FYI, we noticed a 418.8% improvement of stress-ng.ioprio.ops_per_sec due to commit:
commit: d4dd19ec1ea0cf6532d65709325c42b1398614a8 ("f2fs: do not expose unwritten blocks to user by DIO")
https://git.kernel.org/cgit/linux/kernel/git/next/linux-next.git master
in testcase: stress-ng
on test machine: 96 threads 2 sockets Intel(R) Xeon(R) Gold 6252 CPU @ 2.10GHz with 512G memory
with following parameters:
nr_threads: 10%
disk: 1HDD
testtime: 60s
fs: f2fs
class: filesystem
test: ioprio
cpufreq_governor: performance
ucode: 0x5003102
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.
=========================================================================================
class/compiler/cpufreq_governor/disk/fs/kconfig/nr_threads/rootfs/tbox_group/test/testcase/testtime/ucode:
filesystem/gcc-9/performance/1HDD/f2fs/x86_64-rhel-8.3/10%/debian-10.4-x86_64-20200603.cgz/lkp-csl-2sp7/ioprio/stress-ng/60s/0x5003102
commit:
b31bf0f96e ("f2fs: reduce indentation in f2fs_file_write_iter()")
d4dd19ec1e ("f2fs: do not expose unwritten blocks to user by DIO")
b31bf0f96e71a2c8 d4dd19ec1ea0cf6532d65709325
---------------- ---------------------------
%stddev %change %stddev
\ | \
158.83 ? 10% +417.2% 821.50 ? 15% stress-ng.ioprio.ops
2.63 ? 10% +418.8% 13.64 ? 15% stress-ng.ioprio.ops_per_sec
6383 ? 16% +338.5% 27988 ? 6% meminfo.Inactive(file)
27453 ? 5% -37.0% 17302 ? 12% softirqs.TIMER
90.22 +6.2% 95.83 iostat.cpu.idle
7.65 ? 2% -74.2% 1.97 ? 58% iostat.cpu.iowait
7.89 ? 2% -5.9 2.03 ? 58% mpstat.cpu.all.iowait%
0.08 ? 2% +0.0 0.10 ? 4% mpstat.cpu.all.sys%
89.83 +6.5% 95.67 vmstat.cpu.id
2095 +50.9% 3162 ? 6% vmstat.io.bo
7.17 ? 5% -76.7% 1.67 ? 56% vmstat.procs.b
34699 +49.3% 51804 ? 5% proc-vmstat.nr_dirtied
1595 ? 16% +336.3% 6961 ? 5% proc-vmstat.nr_inactive_file
18272 +4.4% 19083 ? 7% proc-vmstat.nr_kernel_stack
47048 +2.5% 48215 ? 2% proc-vmstat.nr_slab_unreclaimable
34641 +49.9% 51938 ? 6% proc-vmstat.nr_written
1595 ? 16% +336.3% 6961 ? 5% proc-vmstat.nr_zone_inactive_file
429745 +2.0% 438139 proc-vmstat.pgalloc_normal
327388 -1.0% 324273 proc-vmstat.pgfault
138652 +50.1% 208086 ? 5% proc-vmstat.pgpgout
0.75 ? 15% -0.4 0.38 ? 72% perf-profile.calltrace.cycles-pp.run_rebalance_domains.__softirqentry_text_start.irq_exit_rcu.sysvec_apic_timer_interrupt.asm_sysvec_apic_timer_interrupt
0.76 ? 15% -0.2 0.55 ? 13% perf-profile.children.cycles-pp.run_rebalance_domains
0.72 ? 12% -0.2 0.54 ? 11% perf-profile.children.cycles-pp.update_blocked_averages
0.07 ? 16% +0.0 0.10 ? 18% perf-profile.children.cycles-pp.schedule_idle
0.01 ?223% +0.1 0.09 ? 17% perf-profile.children.cycles-pp.__submit_merged_write_cond
0.00 +0.1 0.10 ? 52% perf-profile.children.cycles-pp.do_write_page
0.00 +0.1 0.10 ? 56% perf-profile.children.cycles-pp.f2fs_outplace_write_data
0.00 +0.1 0.10 ? 40% perf-profile.children.cycles-pp.f2fs_write_begin
0.01 ?223% +0.1 0.13 ? 59% perf-profile.children.cycles-pp.f2fs_do_write_data_page
0.11 ? 41% +0.1 0.24 ? 34% perf-profile.children.cycles-pp.f2fs_write_cache_pages
0.00 +0.1 0.13 ? 30% perf-profile.children.cycles-pp.generic_perform_write
0.00 +0.1 0.14 ? 31% perf-profile.children.cycles-pp.__generic_file_write_iter
0.00 +0.1 0.14 ? 36% perf-profile.children.cycles-pp.do_iter_write
0.00 +0.1 0.14 ? 36% perf-profile.children.cycles-pp.do_iter_readv_writev
0.00 +0.1 0.14 ? 36% perf-profile.children.cycles-pp.f2fs_file_write_iter
0.13 ? 34% +0.1 0.27 ? 31% perf-profile.children.cycles-pp.do_writepages
0.00 +0.1 0.14 ? 35% perf-profile.children.cycles-pp.vfs_writev
0.12 ? 36% +0.1 0.27 ? 32% perf-profile.children.cycles-pp.f2fs_write_data_pages
0.12 ? 36% +0.1 0.27 ? 31% perf-profile.children.cycles-pp.filemap_fdatawrite_wbc
0.00 +0.2 0.15 ? 38% perf-profile.children.cycles-pp.do_pwritev
0.00 +0.2 0.16 ? 36% perf-profile.children.cycles-pp.pwritev64
0.01 ?223% +0.2 0.17 ? 50% perf-profile.children.cycles-pp.__x64_sys_ioprio_get
0.13 ? 35% +0.2 0.30 ? 26% perf-profile.children.cycles-pp.file_write_and_wait_range
0.01 ?223% +0.2 0.18 ? 45% perf-profile.children.cycles-pp.syscall
0.13 ? 35% +0.2 0.34 ? 24% perf-profile.children.cycles-pp.f2fs_do_sync_file
0.13 ? 35% +0.2 0.34 ? 24% perf-profile.children.cycles-pp.__x64_sys_fsync
0.13 ? 35% +0.2 0.34 ? 24% perf-profile.children.cycles-pp.do_fsync
0.14 ? 35% +0.2 0.36 ? 23% perf-profile.children.cycles-pp.fsync
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