From: kernel test robot <oliver.sang@intel.com>
To: Mark Harmstone <mark@harmstone.com>
Cc: <oe-lkp@lists.linux.dev>, <lkp@intel.com>,
David Sterba <dsterba@suse.com>, <linux-btrfs@vger.kernel.org>,
<oliver.sang@intel.com>
Subject: [linux-next:master] [btrfs] 14f9161e87: fio.write_iops 13.1% improvement
Date: Mon, 15 Jun 2026 22:50:33 +0800 [thread overview]
Message-ID: <202606152223.d5cf620b-lkp@intel.com> (raw)
Hello,
kernel test robot noticed a 13.1% improvement of fio.write_iops on:
commit: 14f9161e872072075284b8afa4c3f929e199a0f6 ("btrfs: don't force DIO writes to be serialized")
https://git.kernel.org/cgit/linux/kernel/git/next/linux-next.git master
testcase: fio-basic
config: x86_64-rhel-9.4
compiler: gcc-14
test machine: 64 threads 2 sockets Intel(R) Xeon(R) Gold 6346 CPU @ 3.10GHz (Ice Lake) with 256G memory
parameters:
runtime: 300s
disk: 1HDD
fs: btrfs
nr_task: 1
test_size: 128G
rw: write
bs: 4k
ioengine: falloc
cpufreq_governor: performance
Details are as below:
-------------------------------------------------------------------------------------------------->
The kernel config and materials to reproduce are available at:
https://download.01.org/0day-ci/archive/20260615/202606152223.d5cf620b-lkp@intel.com
=========================================================================================
bs/compiler/cpufreq_governor/disk/fs/ioengine/kconfig/nr_task/rootfs/runtime/rw/tbox_group/test_size/testcase:
4k/gcc-14/performance/1HDD/btrfs/falloc/x86_64-rhel-9.4/1/debian-13-x86_64-20250902.cgz/300s/write/lkp-icl-2sp9/128G/fio-basic
commit:
9bce95edb1 ("btrfs: move large data folios out of experimental features")
14f9161e87 ("btrfs: don't force DIO writes to be serialized")
9bce95edb1b4d280 14f9161e872072075284b8afa4c
---------------- ---------------------------
%stddev %change %stddev
\ | \
6154 +13.1% 6960 fio.write_bw_MBps
464.67 -16.1% 390.00 fio.write_clat_90%_ns
468.00 -16.1% 392.67 fio.write_clat_95%_ns
474.00 -15.8% 399.33 fio.write_clat_99%_ns
458.34 -15.9% 385.59 fio.write_clat_mean_ns
1575465 +13.1% 1781998 fio.write_iops
134710 ± 4% -7.6% 124437 ± 2% numa-numastat.node1.numa_hit
134733 ± 4% -7.6% 124458 ± 2% numa-vmstat.node1.numa_hit
156867 ± 2% -10.6% 140262 turbostat.IRQ
1.473e+09 -10.6% 1.317e+09 cpuidle..time
103066 ± 3% -11.2% 91502 cpuidle..usage
84.50 +5.2% 88.91 ± 2% perf-stat.i.cpu-migrations
80.85 +4.7% 84.65 ± 2% perf-stat.ps.cpu-migrations
323.49 ± 2% -7.4% 299.52 ± 4% sched_debug.cfs_rq:/.runnable_avg.stddev
323.01 ± 2% -7.3% 299.39 ± 4% sched_debug.cfs_rq:/.util_avg.stddev
253020 ± 5% -5.1% 240055 ± 3% proc-vmstat.numa_hit
186828 ± 7% -6.9% 173889 ± 5% proc-vmstat.numa_local
273378 ± 5% -5.8% 257473 ± 4% proc-vmstat.pgalloc_normal
156448 ± 9% -10.3% 140365 ± 6% proc-vmstat.pgfault
147056 ± 10% -10.6% 131405 ± 9% proc-vmstat.pgfree
8024 ± 22% -20.8% 6358 ± 11% proc-vmstat.pgreuse
43.66 ±101% -37.0 6.70 ±199% perf-profile.calltrace.cycles-pp.zap_pmd_range.__zap_vma_range.unmap_vmas.exit_mmap.__mmput
43.66 ±101% -37.0 6.70 ±199% perf-profile.calltrace.cycles-pp.zap_pte_range.zap_pmd_range.__zap_vma_range.unmap_vmas.exit_mmap
43.48 ±101% -36.9 6.61 ±202% perf-profile.calltrace.cycles-pp.__zap_vma_range.unmap_vmas.exit_mmap.__mmput.exit_mm
43.48 ±101% -36.7 6.79 ±196% perf-profile.calltrace.cycles-pp.unmap_vmas.exit_mmap.__mmput.exit_mm.do_exit
43.48 ±101% -35.4 8.12 ±206% perf-profile.calltrace.cycles-pp.__mmput.exit_mm.do_exit.do_group_exit.get_signal
43.48 ±101% -35.4 8.12 ±206% perf-profile.calltrace.cycles-pp.exit_mm.do_exit.do_group_exit.get_signal.arch_do_signal_or_restart
44.01 ± 99% -33.8 10.18 ±196% perf-profile.calltrace.cycles-pp.exit_mmap.__mmput.exit_mm.do_exit.do_group_exit
43.48 ±101% -33.2 10.31 ±163% perf-profile.calltrace.cycles-pp.arch_do_signal_or_restart.exit_to_user_mode_loop.do_syscall_64.entry_SYSCALL_64_after_hwframe
43.48 ±101% -33.2 10.31 ±163% perf-profile.calltrace.cycles-pp.do_exit.do_group_exit.get_signal.arch_do_signal_or_restart.exit_to_user_mode_loop
43.48 ±101% -33.2 10.31 ±163% perf-profile.calltrace.cycles-pp.do_group_exit.get_signal.arch_do_signal_or_restart.exit_to_user_mode_loop.do_syscall_64
43.48 ±101% -33.2 10.31 ±163% perf-profile.calltrace.cycles-pp.exit_to_user_mode_loop.do_syscall_64.entry_SYSCALL_64_after_hwframe
43.48 ±101% -33.2 10.31 ±163% perf-profile.calltrace.cycles-pp.get_signal.arch_do_signal_or_restart.exit_to_user_mode_loop.do_syscall_64.entry_SYSCALL_64_after_hwframe
26.09 ±111% -21.1 5.00 ±200% perf-profile.calltrace.cycles-pp.zap_present_ptes.zap_pte_range.zap_pmd_range.__zap_vma_range.unmap_vmas
12.03 ±120% -10.4 1.61 ±208% perf-profile.calltrace.cycles-pp.__tlb_batch_free_encoded_pages.tlb_flush_mmu.zap_pte_range.zap_pmd_range.__zap_vma_range
12.03 ±120% -10.4 1.61 ±208% perf-profile.calltrace.cycles-pp.free_pages_and_swap_cache.__tlb_batch_free_encoded_pages.tlb_flush_mmu.zap_pte_range.zap_pmd_range
12.03 ±120% -10.4 1.61 ±208% perf-profile.calltrace.cycles-pp.tlb_flush_mmu.zap_pte_range.zap_pmd_range.__zap_vma_range.unmap_vmas
9.05 ± 94% -9.0 0.00 perf-profile.calltrace.cycles-pp.folio_remove_rmap_ptes.zap_present_ptes.zap_pte_range.zap_pmd_range.__zap_vma_range
4.48 ±102% -2.1 2.34 ±150% perf-profile.calltrace.cycles-pp.asm_exc_page_fault
43.66 ±101% -37.0 6.70 ±199% perf-profile.children.cycles-pp.__zap_vma_range
43.66 ±101% -37.0 6.70 ±199% perf-profile.children.cycles-pp.zap_pmd_range
43.66 ±101% -37.0 6.70 ±199% perf-profile.children.cycles-pp.zap_pte_range
43.66 ±101% -36.8 6.88 ±193% perf-profile.children.cycles-pp.unmap_vmas
44.01 ± 99% -33.8 10.18 ±196% perf-profile.children.cycles-pp.exit_mm
44.19 ± 98% -33.7 10.46 ±190% perf-profile.children.cycles-pp.__mmput
44.19 ± 98% -33.7 10.46 ±190% perf-profile.children.cycles-pp.exit_mmap
43.48 ±101% -33.2 10.31 ±163% perf-profile.children.cycles-pp.arch_do_signal_or_restart
43.48 ±101% -33.2 10.31 ±163% perf-profile.children.cycles-pp.exit_to_user_mode_loop
43.48 ±101% -33.2 10.31 ±163% perf-profile.children.cycles-pp.get_signal
28.12 ±105% -23.1 5.00 ±200% perf-profile.children.cycles-pp.zap_present_ptes
12.03 ±120% -10.4 1.61 ±208% perf-profile.children.cycles-pp.tlb_flush_mmu
12.21 ±117% -10.3 1.88 ±176% perf-profile.children.cycles-pp.__tlb_batch_free_encoded_pages
12.21 ±117% -10.3 1.88 ±176% perf-profile.children.cycles-pp.free_pages_and_swap_cache
8.87 ± 98% -8.9 0.00 perf-profile.children.cycles-pp.folio_remove_rmap_ptes
4.84 ±101% -2.1 2.70 ±143% perf-profile.children.cycles-pp.asm_exc_page_fault
4.13 ±107% -2.0 2.15 ±158% perf-profile.children.cycles-pp.handle_mm_fault
4.13 ±107% -1.9 2.24 ±153% perf-profile.children.cycles-pp.do_user_addr_fault
4.13 ±107% -1.8 2.34 ±150% perf-profile.children.cycles-pp.exc_page_fault
19.24 ±114% -17.5 1.79 ±185% perf-profile.self.cycles-pp.zap_present_ptes
8.70 ±101% -8.7 0.00 perf-profile.self.cycles-pp.folio_remove_rmap_ptes
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
reply other threads:[~2026-06-15 14:50 UTC|newest]
Thread overview: [no followups] expand[flat|nested] mbox.gz Atom feed
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=202606152223.d5cf620b-lkp@intel.com \
--to=oliver.sang@intel.com \
--cc=dsterba@suse.com \
--cc=linux-btrfs@vger.kernel.org \
--cc=lkp@intel.com \
--cc=mark@harmstone.com \
--cc=oe-lkp@lists.linux.dev \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.