All of lore.kernel.org
 help / color / mirror / Atom feed
From: Xing Zhengjun <zhengjun.xing@linux.intel.com>
To: lkp@lists.01.org
Subject: Re: [xfs] bad77c375e: stress-ng.fallocate.ops_per_sec -10.0% regression
Date: Tue, 30 Nov 2021 14:46:06 +0800	[thread overview]
Message-ID: <96fa41bb-7e9f-ae47-bf34-accef5c36fa8@linux.intel.com> (raw)
In-Reply-To: <36dc5f38-f8ec-591a-8efa-b3bd607ac06e@linux.intel.com>

[-- Attachment #1: Type: text/plain, Size: 7741 bytes --]

Hi Dave,

    Do you have time to look at this? It still existed in v5.16-rc3. Thanks

On 9/3/2021 3:40 PM, Xing Zhengjun wrote:
> Hi Dave,
> 
>     Do you have time to look at this? Thanks.
> 
> On 9/2/2021 3:27 PM, kernel test robot wrote:
>>
>>
>> Greeting,
>>
>> FYI, we noticed a -10.0% regression of stress-ng.fallocate.ops_per_sec 
>> due to commit:
>>
>>
>> commit: bad77c375e8de6c776c848e443f7dc2d0d909be5 ("xfs: CIL checkpoint 
>> flushes caches unconditionally")
>> https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.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: xfs
>>     class: filesystem
>>     test: fallocate
>>     cpufreq_governor: performance
>>     ucode: 0x5003006
>>
>>
>> In addition to that, the commit also has significant impact on the 
>> following tests:
>>
>>
>>
>> If you fix the issue, kindly add following tag
>> Reported-by: kernel test robot <oliver.sang@intel.com>
>>
>>
>> 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 split-job --compatible job.yaml  # generate the yaml 
>> file for lkp run
>>          bin/lkp run                    generated-yaml-file
>>
>> ========================================================================================= 
>>
>> class/compiler/cpufreq_governor/disk/fs/kconfig/nr_threads/rootfs/tbox_group/test/testcase/testtime/ucode: 
>>
>>    
>> filesystem/gcc-9/performance/1HDD/xfs/x86_64-rhel-8.3/10%/debian-10.4-x86_64-20200603.cgz/lkp-csl-2sp7/fallocate/stress-ng/60s/0x5003006 
>>
>>
>> commit:
>>    0431d926b3 ("xfs: async blkdev cache flush")
>>    bad77c375e ("xfs: CIL checkpoint flushes caches unconditionally")
>>
>> 0431d926b399d74f bad77c375e8de6c776c848e443f
>> ---------------- ---------------------------
>>           %stddev     %change         %stddev
>>               \          |                \
>>      618.67           -10.0%     556.67 ±  2%  stress-ng.fallocate.ops
>>       10.31           -10.0%       9.28 ±  2%  
>> stress-ng.fallocate.ops_per_sec
>>      111618            -5.7%     105270        
>> stress-ng.time.voluntary_context_switches
>>      502.67 ±  2%     +52.0%     763.83 ± 47%  
>> interrupts.CPU44.CAL:Function_call_interrupts
>>        2.10 ±  2%     -25.1%       1.57        iostat.cpu.iowait
>>        2.16 ±  2%      -0.5        1.62        mpstat.cpu.all.iowait%
>>        0.09 ± 11%     +45.0%       0.14 ± 34%  
>> perf-sched.sch_delay.avg.ms.worker_thread.kthread.ret_from_fork
>>      151.00 ± 19%     +20.2%     181.50 ±  7%  
>> perf-sched.wait_and_delay.count.smpboot_thread_fn.kthread.ret_from_fork
>>        6749            +6.4%       7178        
>> perf-stat.i.context-switches
>>        6640            +6.4%       7063        
>> perf-stat.ps.context-switches
>>      353877            -1.4%     348924        proc-vmstat.pgfree
>>       39122           -15.1%      33221        proc-vmstat.pgpgout
>>        1225 ± 14%      -9.0%       1115        turbostat.Bzy_MHz
>>       86.82 ± 17%     -10.4%      77.83        turbostat.PkgWatt
>>      602.83           -15.1%     511.83        vmstat.io.bo
>>        6953            +6.1%       7376        vmstat.system.cs
>>        1.33 ±  6%      +0.4        1.70 ± 13%  
>> perf-profile.calltrace.cycles-pp.perf_mux_hrtimer_handler.__hrtimer_run_queues.hrtimer_interrupt.__sysvec_apic_timer_interrupt.sysvec_apic_timer_interrupt 
>>
>>        0.08 ± 22%      +0.1        0.17 ± 23%  
>> perf-profile.children.cycles-pp.wait_for_completion
>>        0.07 ± 17%      +0.1        0.19 ± 64%  
>> perf-profile.children.cycles-pp.submit_bio
>>        0.07 ± 17%      +0.1        0.19 ± 64%  
>> perf-profile.children.cycles-pp.submit_bio_noacct
>>        1.40 ±  8%      +0.3        1.74 ± 12%  
>> perf-profile.children.cycles-pp.perf_mux_hrtimer_handler
>>
>>
>>                           stress-ng.fallocate.ops_per_sec
>>    12 
>> +----------------------------------------------------------------------+
>>       
>> |                                                                      |
>>    10 |.+.++.+.+.+  
>> +.+.+.++.+.+.++.+.+.+.++.+.+.++.+.+.++.+.+.+.++.+.+.++.+.|
>>       | O        :  :                             O O   OO O O O 
>> OO          |
>>       |   OO O   :O :            O             O O    
>> O                      |
>>     8 |-+      O :  : O O OO O O  O O O O OO 
>> O                               |
>>       |          : 
>> :                                                         |
>>     6 |-+        : 
>> :                                                         |
>>       |           
>> ::                                                         |
>>     4 |-+         
>> ::                                                         |
>>       |           
>> ::                                                         |
>>       |           
>> ::                                                         |
>>     2 |-+         
>> :                                                          |
>>       |           
>> :                                                          |
>>     0 
>> +----------------------------------------------------------------------+
>> [*] 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.
>>
>>
>> ---
>> 0DAY/LKP+ Test Infrastructure                   Open Source Technology 
>> Center
>> https://lists.01.org/hyperkitty/list/lkp(a)lists.01.org       Intel 
>> Corporation
>>
>> Thanks,
>> Oliver Sang
>>
> 

-- 
Zhengjun Xing

WARNING: multiple messages have this Message-ID (diff)
From: Xing Zhengjun <zhengjun.xing@linux.intel.com>
To: kernel test robot <oliver.sang@intel.com>,
	Dave Chinner <dchinner@redhat.com>
Cc: "Darrick J. Wong" <djwong@kernel.org>,
	Chandan Babu R <chandanrlinux@gmail.com>,
	Brian Foster <bfoster@redhat.com>,
	Allison Henderson <allison.henderson@oracle.com>,
	LKML <linux-kernel@vger.kernel.org>,
	lkp@lists.01.org, lkp@intel.com
Subject: Re: [LKP] Re: [xfs] bad77c375e: stress-ng.fallocate.ops_per_sec -10.0% regression
Date: Tue, 30 Nov 2021 14:46:06 +0800	[thread overview]
Message-ID: <96fa41bb-7e9f-ae47-bf34-accef5c36fa8@linux.intel.com> (raw)
In-Reply-To: <36dc5f38-f8ec-591a-8efa-b3bd607ac06e@linux.intel.com>

Hi Dave,

    Do you have time to look at this? It still existed in v5.16-rc3. Thanks

On 9/3/2021 3:40 PM, Xing Zhengjun wrote:
> Hi Dave,
> 
>     Do you have time to look at this? Thanks.
> 
> On 9/2/2021 3:27 PM, kernel test robot wrote:
>>
>>
>> Greeting,
>>
>> FYI, we noticed a -10.0% regression of stress-ng.fallocate.ops_per_sec 
>> due to commit:
>>
>>
>> commit: bad77c375e8de6c776c848e443f7dc2d0d909be5 ("xfs: CIL checkpoint 
>> flushes caches unconditionally")
>> https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.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: xfs
>>     class: filesystem
>>     test: fallocate
>>     cpufreq_governor: performance
>>     ucode: 0x5003006
>>
>>
>> In addition to that, the commit also has significant impact on the 
>> following tests:
>>
>>
>>
>> If you fix the issue, kindly add following tag
>> Reported-by: kernel test robot <oliver.sang@intel.com>
>>
>>
>> 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 split-job --compatible job.yaml  # generate the yaml 
>> file for lkp run
>>          bin/lkp run                    generated-yaml-file
>>
>> ========================================================================================= 
>>
>> class/compiler/cpufreq_governor/disk/fs/kconfig/nr_threads/rootfs/tbox_group/test/testcase/testtime/ucode: 
>>
>>    
>> filesystem/gcc-9/performance/1HDD/xfs/x86_64-rhel-8.3/10%/debian-10.4-x86_64-20200603.cgz/lkp-csl-2sp7/fallocate/stress-ng/60s/0x5003006 
>>
>>
>> commit:
>>    0431d926b3 ("xfs: async blkdev cache flush")
>>    bad77c375e ("xfs: CIL checkpoint flushes caches unconditionally")
>>
>> 0431d926b399d74f bad77c375e8de6c776c848e443f
>> ---------------- ---------------------------
>>           %stddev     %change         %stddev
>>               \          |                \
>>      618.67           -10.0%     556.67 ±  2%  stress-ng.fallocate.ops
>>       10.31           -10.0%       9.28 ±  2%  
>> stress-ng.fallocate.ops_per_sec
>>      111618            -5.7%     105270        
>> stress-ng.time.voluntary_context_switches
>>      502.67 ±  2%     +52.0%     763.83 ± 47%  
>> interrupts.CPU44.CAL:Function_call_interrupts
>>        2.10 ±  2%     -25.1%       1.57        iostat.cpu.iowait
>>        2.16 ±  2%      -0.5        1.62        mpstat.cpu.all.iowait%
>>        0.09 ± 11%     +45.0%       0.14 ± 34%  
>> perf-sched.sch_delay.avg.ms.worker_thread.kthread.ret_from_fork
>>      151.00 ± 19%     +20.2%     181.50 ±  7%  
>> perf-sched.wait_and_delay.count.smpboot_thread_fn.kthread.ret_from_fork
>>        6749            +6.4%       7178        
>> perf-stat.i.context-switches
>>        6640            +6.4%       7063        
>> perf-stat.ps.context-switches
>>      353877            -1.4%     348924        proc-vmstat.pgfree
>>       39122           -15.1%      33221        proc-vmstat.pgpgout
>>        1225 ± 14%      -9.0%       1115        turbostat.Bzy_MHz
>>       86.82 ± 17%     -10.4%      77.83        turbostat.PkgWatt
>>      602.83           -15.1%     511.83        vmstat.io.bo
>>        6953            +6.1%       7376        vmstat.system.cs
>>        1.33 ±  6%      +0.4        1.70 ± 13%  
>> perf-profile.calltrace.cycles-pp.perf_mux_hrtimer_handler.__hrtimer_run_queues.hrtimer_interrupt.__sysvec_apic_timer_interrupt.sysvec_apic_timer_interrupt 
>>
>>        0.08 ± 22%      +0.1        0.17 ± 23%  
>> perf-profile.children.cycles-pp.wait_for_completion
>>        0.07 ± 17%      +0.1        0.19 ± 64%  
>> perf-profile.children.cycles-pp.submit_bio
>>        0.07 ± 17%      +0.1        0.19 ± 64%  
>> perf-profile.children.cycles-pp.submit_bio_noacct
>>        1.40 ±  8%      +0.3        1.74 ± 12%  
>> perf-profile.children.cycles-pp.perf_mux_hrtimer_handler
>>
>>
>>                           stress-ng.fallocate.ops_per_sec
>>    12 
>> +----------------------------------------------------------------------+
>>       
>> |                                                                      |
>>    10 |.+.++.+.+.+  
>> +.+.+.++.+.+.++.+.+.+.++.+.+.++.+.+.++.+.+.+.++.+.+.++.+.|
>>       | O        :  :                             O O   OO O O O 
>> OO          |
>>       |   OO O   :O :            O             O O    
>> O                      |
>>     8 |-+      O :  : O O OO O O  O O O O OO 
>> O                               |
>>       |          : 
>> :                                                         |
>>     6 |-+        : 
>> :                                                         |
>>       |           
>> ::                                                         |
>>     4 |-+         
>> ::                                                         |
>>       |           
>> ::                                                         |
>>       |           
>> ::                                                         |
>>     2 |-+         
>> :                                                          |
>>       |           
>> :                                                          |
>>     0 
>> +----------------------------------------------------------------------+
>> [*] 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.
>>
>>
>> ---
>> 0DAY/LKP+ Test Infrastructure                   Open Source Technology 
>> Center
>> https://lists.01.org/hyperkitty/list/lkp@lists.01.org       Intel 
>> Corporation
>>
>> Thanks,
>> Oliver Sang
>>
> 

-- 
Zhengjun Xing

  reply	other threads:[~2021-11-30  6:46 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-09-02  7:27 [xfs] bad77c375e: stress-ng.fallocate.ops_per_sec -10.0% regression kernel test robot
2021-09-02  7:27 ` kernel test robot
2021-09-03  7:40 ` Xing Zhengjun
2021-09-03  7:40   ` Xing Zhengjun
2021-11-30  6:46   ` Xing Zhengjun [this message]
2021-11-30  6:46     ` [LKP] " Xing Zhengjun
2021-12-02 22:04     ` Dave Chinner
2021-12-02 22:04       ` [LKP] " Dave Chinner

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=96fa41bb-7e9f-ae47-bf34-accef5c36fa8@linux.intel.com \
    --to=zhengjun.xing@linux.intel.com \
    --cc=lkp@lists.01.org \
    /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.