All of lore.kernel.org
 help / color / mirror / Atom feed
From: Huang, Ying <ying.huang@linux.intel.com>
To: lkp@lists.01.org
Subject: Re: [mm/vmstat] 6cdb18ad98: -8.5% will-it-scale.per_thread_ops
Date: Thu, 21 Jan 2016 14:47:28 +0800	[thread overview]
Message-ID: <8737tro2db.fsf@yhuang-dev.intel.com> (raw)
In-Reply-To: <20160107112301.GE4062@osiris>

[-- Attachment #1: Type: text/plain, Size: 7745 bytes --]

Heiko Carstens <heiko.carstens@de.ibm.com> writes:

> On Wed, Jan 06, 2016 at 11:20:55AM +0800, kernel test robot wrote:
>> FYI, we noticed the below changes on
>> 
>> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git master
>> commit 6cdb18ad98a49f7e9b95d538a0614cde827404b8 ("mm/vmstat: fix overflow in mod_zone_page_state()")
>> 
>> 
>> =========================================================================================
>> compiler/cpufreq_governor/kconfig/rootfs/tbox_group/test/testcase:
>>   gcc-4.9/performance/x86_64-rhel/debian-x86_64-2015-02-07.cgz/ivb42/pread1/will-it-scale
>> 
>> commit: 
>>   cc28d6d80f6ab494b10f0e2ec949eacd610f66e3
>>   6cdb18ad98a49f7e9b95d538a0614cde827404b8
>> 
>> cc28d6d80f6ab494 6cdb18ad98a49f7e9b95d538a0 
>> ---------------- -------------------------- 
>>          %stddev     %change         %stddev
>>              \          |                \  
>>    2733943 .  0%      -8.5%    2502129 .  0%  will-it-scale.per_thread_ops
>>       3410 .  0%      -2.0%       3343 .  0%  will-it-scale.time.system_time
>>     340.08 .  0%     +19.7%     406.99 .  0%  will-it-scale.time.user_time
>>   69882822 .  2%     -24.3%   52926191 .  5%  cpuidle.C1-IVT.time
>>     340.08 .  0%     +19.7%     406.99 .  0%  time.user_time
>>     491.25 .  6%     -17.7%     404.25 .  7%  numa-vmstat.node0.nr_alloc_batch
>>       2799 . 20%     -36.6%       1776 .  0%  numa-vmstat.node0.nr_mapped
>>     630.00 .140%    +244.4%       2169 .  1%  numa-vmstat.node1.nr_inactive_anon
>
> Hmm... this is odd. I did review all callers of mod_zone_page_state() and
> couldn't find anything obvious that would go wrong after the int -> long
> change.
>
> I also tried the "pread1_threads" test case from
> https://github.com/antonblanchard/will-it-scale.git
>
> However the results seem to vary a lot after a reboot(!), at least on s390.
>
> So I'm not sure if this is really a regression.

Most part of the regression is restored for v4.4.  But because the changes are
like "V", it is hard to bisect.

=========================================================================================
compiler/cpufreq_governor/kconfig/mode/nr_task/rootfs/tbox_group/test/testcase:
  gcc-4.9/performance/x86_64-rhel/thread/24/debian-x86_64-2015-02-07.cgz/ivb42/pread1/will-it-scale

commit: 
  cc28d6d80f6ab494b10f0e2ec949eacd610f66e3
  6cdb18ad98a49f7e9b95d538a0614cde827404b8
  v4.4

cc28d6d80f6ab494 6cdb18ad98a49f7e9b95d538a0                       v4.4 
---------------- -------------------------- -------------------------- 
         %stddev     %change         %stddev     %change         %stddev
             \          |                \          |                \  
   3083436 ±  0%      -9.6%    2788374 ±  0%      -3.7%    2970130 ±  0%  will-it-scale.per_thread_ops
      6447 ±  0%      -2.2%       6308 ±  0%      -0.3%       6425 ±  0%  will-it-scale.time.system_time
    776.90 ±  0%     +17.9%     915.71 ±  0%      +2.9%     799.12 ±  0%  will-it-scale.time.user_time
    316177 ±  4%      -4.6%     301616 ±  3%     -10.3%     283563 ±  3%  softirqs.RCU
    776.90 ±  0%     +17.9%     915.71 ±  0%      +2.9%     799.12 ±  0%  time.user_time
    777.33 ±  7%     +20.8%     938.67 ±  7%      +7.5%     836.00 ±  8%  slabinfo.blkdev_requests.active_objs
    777.33 ±  7%     +20.8%     938.67 ±  7%      +7.5%     836.00 ±  8%  slabinfo.blkdev_requests.num_objs
  74313962 ± 44%     -16.5%   62053062 ± 41%     -49.9%   37246967 ±  8%  cpuidle.C1-IVT.time
  43381614 ± 79%     +24.4%   53966568 ±111%    +123.9%   97135791 ± 33%  cpuidle.C1E-IVT.time
     97.67 ± 36%     +95.2%     190.67 ± 63%    +122.5%     217.33 ± 41%  cpuidle.C3-IVT.usage
   3679437 ± 69%    -100.0%       0.00 ± -1%    -100.0%       0.00 ± -1%  latency_stats.avg.nfs_wait_on_request.nfs_updatepage.nfs_write_end.generic_perform_write.__generic_file_write_iter.generic_file_write_iter.nfs_file_write.__vfs_write.vfs_write.SyS_write.entry_SYSCALL_64_fastpath
   5177475 ± 82%    -100.0%       0.00 ± -1%    -100.0%       0.00 ± -1%  latency_stats.max.nfs_wait_on_request.nfs_updatepage.nfs_write_end.generic_perform_write.__generic_file_write_iter.generic_file_write_iter.nfs_file_write.__vfs_write.vfs_write.SyS_write.entry_SYSCALL_64_fastpath
  11726393 ±112%    -100.0%       0.00 ± -1%    -100.0%       0.00 ± -1%  latency_stats.sum.nfs_wait_on_request.nfs_updatepage.nfs_write_end.generic_perform_write.__generic_file_write_iter.generic_file_write_iter.nfs_file_write.__vfs_write.vfs_write.SyS_write.entry_SYSCALL_64_fastpath
    178.07 ±  0%      -1.3%     175.79 ±  0%      -0.8%     176.62 ±  0%  turbostat.CorWatt
      0.20 ±  2%     -16.9%       0.16 ± 18%     -11.9%       0.17 ± 17%  turbostat.Pkg%pc6
    207.38 ±  0%      -1.1%     205.13 ±  0%      -0.7%     205.99 ±  0%  turbostat.PkgWatt
      6889 ± 33%     -49.2%       3497 ± 86%     -19.4%       5552 ± 27%  numa-vmstat.node0.nr_active_anon
    483.33 ± 29%     -32.3%     327.00 ± 48%      +0.1%     483.67 ± 29%  numa-vmstat.node0.nr_page_table_pages
     27536 ± 96%     +10.9%      30535 ± 78%    +148.5%      68418 ±  2%  numa-vmstat.node0.numa_other
    214.00 ± 11%     +18.1%     252.67 ±  4%      +2.8%     220.00 ±  9%  numa-vmstat.node1.nr_kernel_stack
    370.67 ± 38%     +42.0%     526.33 ± 30%      -0.2%     370.00 ± 39%  numa-vmstat.node1.nr_page_table_pages
     61177 ± 43%      -5.2%      57976 ± 41%     -66.3%      20644 ± 10%  numa-vmstat.node1.numa_other
     78172 ± 13%     -16.1%      65573 ± 18%      -5.8%      73626 ±  9%  numa-meminfo.node0.Active
     27560 ± 33%     -49.2%      14006 ± 86%     -19.4%      22203 ± 27%  numa-meminfo.node0.Active(anon)
      3891 ± 58%     -38.1%       2407 ±100%     -58.8%       1604 ±110%  numa-meminfo.node0.AnonHugePages
      1934 ± 29%     -32.3%       1309 ± 48%      +0.1%       1936 ± 29%  numa-meminfo.node0.PageTables
     63139 ± 17%     +19.8%      75670 ± 16%      +6.0%      66937 ± 10%  numa-meminfo.node1.Active
      3432 ± 11%     +18.0%       4049 ±  4%      +2.8%       3527 ±  9%  numa-meminfo.node1.KernelStack
      1483 ± 38%     +42.0%       2106 ± 30%      -0.2%       1481 ± 39%  numa-meminfo.node1.PageTables
      1.47 ±  1%     -11.8%       1.30 ±  2%      -7.0%       1.37 ±  3%  perf-profile.cycles-pp.___might_sleep.__might_sleep.find_lock_entry.shmem_getpage_gfp.shmem_file_read_iter
      2.00 ±  2%     -11.3%       1.78 ±  2%      -7.2%       1.86 ±  2%  perf-profile.cycles-pp.__might_sleep.find_lock_entry.shmem_getpage_gfp.shmem_file_read_iter.__vfs_read
      2.30 ±  4%     +33.6%       3.07 ±  0%      -1.9%       2.26 ±  1%  perf-profile.cycles-pp.atime_needs_update.touch_atime.shmem_file_read_iter.__vfs_read.vfs_read
      1.05 ±  1%     -27.7%       0.76 ±  1%      -8.0%       0.96 ±  0%  perf-profile.cycles-pp.current_fs_time.atime_needs_update.touch_atime.shmem_file_read_iter.__vfs_read
      2.21 ±  3%     -11.9%       1.94 ±  2%      -9.4%       2.00 ±  0%  perf-profile.cycles-pp.fput.entry_SYSCALL_64_fastpath
      0.78 ±  2%     +38.5%       1.08 ±  2%     +23.1%       0.96 ±  3%  perf-profile.cycles-pp.fsnotify.vfs_read.sys_pread64.entry_SYSCALL_64_fastpath
      2.87 ±  7%     +42.6%       4.09 ±  1%      -0.3%       2.86 ±  2%  perf-profile.cycles-pp.touch_atime.shmem_file_read_iter.__vfs_read.vfs_read.sys_pread64
      6.68 ±  2%      -7.3%       6.19 ±  1%      -6.7%       6.23 ±  1%  perf-profile.cycles-pp.unlock_page.shmem_file_read_iter.__vfs_read.vfs_read.sys_pread64


Best Regards,
Huang, Ying

WARNING: multiple messages have this Message-ID (diff)
From: "Huang\, Ying" <ying.huang@linux.intel.com>
To: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: lkp@01.org, LKML <linux-kernel@vger.kernel.org>,
	Andrew Morton <akpm@linux-foundation.org>,
	Christoph Lameter <cl@linux.com>,
	Linus Torvalds <torvalds@linux-foundation.org>
Subject: Re: [lkp] [mm/vmstat] 6cdb18ad98: -8.5% will-it-scale.per_thread_ops
Date: Thu, 21 Jan 2016 14:47:28 +0800	[thread overview]
Message-ID: <8737tro2db.fsf@yhuang-dev.intel.com> (raw)
In-Reply-To: <20160107112301.GE4062@osiris> (Heiko Carstens's message of "Thu, 7 Jan 2016 12:23:01 +0100")

Heiko Carstens <heiko.carstens@de.ibm.com> writes:

> On Wed, Jan 06, 2016 at 11:20:55AM +0800, kernel test robot wrote:
>> FYI, we noticed the below changes on
>> 
>> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git master
>> commit 6cdb18ad98a49f7e9b95d538a0614cde827404b8 ("mm/vmstat: fix overflow in mod_zone_page_state()")
>> 
>> 
>> =========================================================================================
>> compiler/cpufreq_governor/kconfig/rootfs/tbox_group/test/testcase:
>>   gcc-4.9/performance/x86_64-rhel/debian-x86_64-2015-02-07.cgz/ivb42/pread1/will-it-scale
>> 
>> commit: 
>>   cc28d6d80f6ab494b10f0e2ec949eacd610f66e3
>>   6cdb18ad98a49f7e9b95d538a0614cde827404b8
>> 
>> cc28d6d80f6ab494 6cdb18ad98a49f7e9b95d538a0 
>> ---------------- -------------------------- 
>>          %stddev     %change         %stddev
>>              \          |                \  
>>    2733943 .  0%      -8.5%    2502129 .  0%  will-it-scale.per_thread_ops
>>       3410 .  0%      -2.0%       3343 .  0%  will-it-scale.time.system_time
>>     340.08 .  0%     +19.7%     406.99 .  0%  will-it-scale.time.user_time
>>   69882822 .  2%     -24.3%   52926191 .  5%  cpuidle.C1-IVT.time
>>     340.08 .  0%     +19.7%     406.99 .  0%  time.user_time
>>     491.25 .  6%     -17.7%     404.25 .  7%  numa-vmstat.node0.nr_alloc_batch
>>       2799 . 20%     -36.6%       1776 .  0%  numa-vmstat.node0.nr_mapped
>>     630.00 .140%    +244.4%       2169 .  1%  numa-vmstat.node1.nr_inactive_anon
>
> Hmm... this is odd. I did review all callers of mod_zone_page_state() and
> couldn't find anything obvious that would go wrong after the int -> long
> change.
>
> I also tried the "pread1_threads" test case from
> https://github.com/antonblanchard/will-it-scale.git
>
> However the results seem to vary a lot after a reboot(!), at least on s390.
>
> So I'm not sure if this is really a regression.

Most part of the regression is restored for v4.4.  But because the changes are
like "V", it is hard to bisect.

=========================================================================================
compiler/cpufreq_governor/kconfig/mode/nr_task/rootfs/tbox_group/test/testcase:
  gcc-4.9/performance/x86_64-rhel/thread/24/debian-x86_64-2015-02-07.cgz/ivb42/pread1/will-it-scale

commit: 
  cc28d6d80f6ab494b10f0e2ec949eacd610f66e3
  6cdb18ad98a49f7e9b95d538a0614cde827404b8
  v4.4

cc28d6d80f6ab494 6cdb18ad98a49f7e9b95d538a0                       v4.4 
---------------- -------------------------- -------------------------- 
         %stddev     %change         %stddev     %change         %stddev
             \          |                \          |                \  
   3083436 ±  0%      -9.6%    2788374 ±  0%      -3.7%    2970130 ±  0%  will-it-scale.per_thread_ops
      6447 ±  0%      -2.2%       6308 ±  0%      -0.3%       6425 ±  0%  will-it-scale.time.system_time
    776.90 ±  0%     +17.9%     915.71 ±  0%      +2.9%     799.12 ±  0%  will-it-scale.time.user_time
    316177 ±  4%      -4.6%     301616 ±  3%     -10.3%     283563 ±  3%  softirqs.RCU
    776.90 ±  0%     +17.9%     915.71 ±  0%      +2.9%     799.12 ±  0%  time.user_time
    777.33 ±  7%     +20.8%     938.67 ±  7%      +7.5%     836.00 ±  8%  slabinfo.blkdev_requests.active_objs
    777.33 ±  7%     +20.8%     938.67 ±  7%      +7.5%     836.00 ±  8%  slabinfo.blkdev_requests.num_objs
  74313962 ± 44%     -16.5%   62053062 ± 41%     -49.9%   37246967 ±  8%  cpuidle.C1-IVT.time
  43381614 ± 79%     +24.4%   53966568 ±111%    +123.9%   97135791 ± 33%  cpuidle.C1E-IVT.time
     97.67 ± 36%     +95.2%     190.67 ± 63%    +122.5%     217.33 ± 41%  cpuidle.C3-IVT.usage
   3679437 ± 69%    -100.0%       0.00 ± -1%    -100.0%       0.00 ± -1%  latency_stats.avg.nfs_wait_on_request.nfs_updatepage.nfs_write_end.generic_perform_write.__generic_file_write_iter.generic_file_write_iter.nfs_file_write.__vfs_write.vfs_write.SyS_write.entry_SYSCALL_64_fastpath
   5177475 ± 82%    -100.0%       0.00 ± -1%    -100.0%       0.00 ± -1%  latency_stats.max.nfs_wait_on_request.nfs_updatepage.nfs_write_end.generic_perform_write.__generic_file_write_iter.generic_file_write_iter.nfs_file_write.__vfs_write.vfs_write.SyS_write.entry_SYSCALL_64_fastpath
  11726393 ±112%    -100.0%       0.00 ± -1%    -100.0%       0.00 ± -1%  latency_stats.sum.nfs_wait_on_request.nfs_updatepage.nfs_write_end.generic_perform_write.__generic_file_write_iter.generic_file_write_iter.nfs_file_write.__vfs_write.vfs_write.SyS_write.entry_SYSCALL_64_fastpath
    178.07 ±  0%      -1.3%     175.79 ±  0%      -0.8%     176.62 ±  0%  turbostat.CorWatt
      0.20 ±  2%     -16.9%       0.16 ± 18%     -11.9%       0.17 ± 17%  turbostat.Pkg%pc6
    207.38 ±  0%      -1.1%     205.13 ±  0%      -0.7%     205.99 ±  0%  turbostat.PkgWatt
      6889 ± 33%     -49.2%       3497 ± 86%     -19.4%       5552 ± 27%  numa-vmstat.node0.nr_active_anon
    483.33 ± 29%     -32.3%     327.00 ± 48%      +0.1%     483.67 ± 29%  numa-vmstat.node0.nr_page_table_pages
     27536 ± 96%     +10.9%      30535 ± 78%    +148.5%      68418 ±  2%  numa-vmstat.node0.numa_other
    214.00 ± 11%     +18.1%     252.67 ±  4%      +2.8%     220.00 ±  9%  numa-vmstat.node1.nr_kernel_stack
    370.67 ± 38%     +42.0%     526.33 ± 30%      -0.2%     370.00 ± 39%  numa-vmstat.node1.nr_page_table_pages
     61177 ± 43%      -5.2%      57976 ± 41%     -66.3%      20644 ± 10%  numa-vmstat.node1.numa_other
     78172 ± 13%     -16.1%      65573 ± 18%      -5.8%      73626 ±  9%  numa-meminfo.node0.Active
     27560 ± 33%     -49.2%      14006 ± 86%     -19.4%      22203 ± 27%  numa-meminfo.node0.Active(anon)
      3891 ± 58%     -38.1%       2407 ±100%     -58.8%       1604 ±110%  numa-meminfo.node0.AnonHugePages
      1934 ± 29%     -32.3%       1309 ± 48%      +0.1%       1936 ± 29%  numa-meminfo.node0.PageTables
     63139 ± 17%     +19.8%      75670 ± 16%      +6.0%      66937 ± 10%  numa-meminfo.node1.Active
      3432 ± 11%     +18.0%       4049 ±  4%      +2.8%       3527 ±  9%  numa-meminfo.node1.KernelStack
      1483 ± 38%     +42.0%       2106 ± 30%      -0.2%       1481 ± 39%  numa-meminfo.node1.PageTables
      1.47 ±  1%     -11.8%       1.30 ±  2%      -7.0%       1.37 ±  3%  perf-profile.cycles-pp.___might_sleep.__might_sleep.find_lock_entry.shmem_getpage_gfp.shmem_file_read_iter
      2.00 ±  2%     -11.3%       1.78 ±  2%      -7.2%       1.86 ±  2%  perf-profile.cycles-pp.__might_sleep.find_lock_entry.shmem_getpage_gfp.shmem_file_read_iter.__vfs_read
      2.30 ±  4%     +33.6%       3.07 ±  0%      -1.9%       2.26 ±  1%  perf-profile.cycles-pp.atime_needs_update.touch_atime.shmem_file_read_iter.__vfs_read.vfs_read
      1.05 ±  1%     -27.7%       0.76 ±  1%      -8.0%       0.96 ±  0%  perf-profile.cycles-pp.current_fs_time.atime_needs_update.touch_atime.shmem_file_read_iter.__vfs_read
      2.21 ±  3%     -11.9%       1.94 ±  2%      -9.4%       2.00 ±  0%  perf-profile.cycles-pp.fput.entry_SYSCALL_64_fastpath
      0.78 ±  2%     +38.5%       1.08 ±  2%     +23.1%       0.96 ±  3%  perf-profile.cycles-pp.fsnotify.vfs_read.sys_pread64.entry_SYSCALL_64_fastpath
      2.87 ±  7%     +42.6%       4.09 ±  1%      -0.3%       2.86 ±  2%  perf-profile.cycles-pp.touch_atime.shmem_file_read_iter.__vfs_read.vfs_read.sys_pread64
      6.68 ±  2%      -7.3%       6.19 ±  1%      -6.7%       6.23 ±  1%  perf-profile.cycles-pp.unlock_page.shmem_file_read_iter.__vfs_read.vfs_read.sys_pread64


Best Regards,
Huang, Ying

  parent reply	other threads:[~2016-01-21  6:47 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-01-06  3:20 [mm/vmstat] 6cdb18ad98: -8.5% will-it-scale.per_thread_ops kernel test robot
2016-01-06  3:20 ` [lkp] " kernel test robot
2016-01-07 11:23 ` Heiko Carstens
2016-01-07 11:23   ` [lkp] " Heiko Carstens
2016-01-08  5:24   ` Huang, Ying
2016-01-08  5:24     ` [LKP] [lkp] " Huang, Ying
2016-01-08 11:13     ` Heiko Carstens
2016-01-08 11:13       ` [LKP] [lkp] " Heiko Carstens
2016-01-21  6:47   ` Huang, Ying [this message]
2016-01-21  6:47     ` Huang, Ying

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=8737tro2db.fsf@yhuang-dev.intel.com \
    --to=ying.huang@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.