From mboxrd@z Thu Jan 1 00:00:00 1970 Content-Type: multipart/mixed; boundary="===============9138920132208475712==" MIME-Version: 1.0 From: Carel Si To: lkp@lists.01.org Subject: Re: [x86/mm/tlb] 2f4305b19f: will-it-scale.per_thread_ops 23.3% improvement Date: Thu, 25 Nov 2021 13:02:22 +0800 Message-ID: <20211125050221.GA26741@linux.intel.com> In-Reply-To: <7030ECD9-B2A8-4483-AEC1-549C14D5558B@vmware.com> List-Id: --===============9138920132208475712== Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Hi Amit, On Sun, Nov 07, 2021 at 09:47:46PM +0000, Nadav Amit wrote: > = > = > > On Nov 7, 2021, at 6:28 AM, kernel test robot = wrote: > > = > > = > > = > > Greeting, > > = > > FYI, we noticed a 23.3% improvement of will-it-scale.per_thread_ops due= to commit: > > = > > = > > commit: 2f4305b19fe6a2a261d76c21856c5598f7d878fe ("x86/mm/tlb: Privatiz= e cpu_tlbstate") > > https://nam04.safelinks.protection.outlook.com/?url=3Dhttps%3A%2F%2Fgit= .kernel.org%2Fcgit%2Flinux%2Fkernel%2Fgit%2Ftorvalds%2Flinux.git&data= =3D04%7C01%7Cnamit%40vmware.com%7Cfb2bd9a47e794ec3898508d9a1faf794%7Cb39138= ca3cee4b4aa4d6cd83d9dd62f0%7C0%7C0%7C637718921590793957%7CUnknown%7CTWFpbGZ= sb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1= 000&sdata=3Dcw6sFeT6qAQSSaG7nCIWm19Yi9RPZqyr2bQPwSBuzzs%3D&reserved= =3D0 master > > = > > will-it-scale.per_thread_ops = = > > = > > 4000 +----------------------------------------------------------------= ----+ = > > | = | = > > 3500 |-O O O O O O O O O O O O O O OO O O O O O O O O O O O O O O = O O | = > > 3000 |-+ .+. .+. = | = > > |.+.+.+.+.+.+ +.+.+.+.+.+.+.+ ++ +.+.+.+.+.+.+.+.+.+.+.+ = | = > > 2500 |-+ : : = | = > > | : : = | = > > 2000 |-+ : : = | = > > | : : = | = > > 1500 |-+ : : = | = > > 1000 |-+ : : = | = > > | : : = | = > > 500 |-+ : = | = > > | : = | = > > 0 +----------------------------------------------------------------= ----+ = > = > Am I to understand that the following commit somehow reverted the perform= ance > improvement of this patch? The graph shows it as a =E2=80=9Cspike=E2=80= =9D, no? The performance improvement is exists in 635de956a7 ("Merge tag = 'x86-mm-2021-04-29' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/ti= p"), in which the fbc has been merged, and v5.13-rc1, in which the fbc has been = merged in mainline, and v5.14. We found the improvement has reverted between v5.15-rc5 and v5.14, has trig= gered bisection for it. Thanks. =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D compiler/cpufreq_governor/kconfig/mode/nr_task/rootfs/tbox_group/test/testc= ase/ucode: gcc-9/performance/x86_64-rhel-8.3/thread/100%/debian-10.4-x86_64-20200603= .cgz/lkp-hsw-4ex1/tlb_flush3/will-it-scale/0x16 commit: = 4ce94eabac ("x86/mm/tlb: Flush remote and local TLBs concurrently") >> pa= rent 2f4305b19f ("x86/mm/tlb: Privatize cpu_tlbstate") >> fbc 635de956a7 >> fbc merged v5.13-rc1 >> fbc merged to main= line v5.14 v5.15-rc5 4ce94eabac16b1d2 2f4305b19fe6a2a261d76c21856 635de956a7f5a6ffcb04f29d706 = v5.13-rc1 v5.14 v5.= 15-rc5 = ---------------- --------------------------- --------------------------- --= ------------------------- --------------------------- ---------------------= ------ = %stddev %change %stddev %change %stddev = %change %stddev %change %stddev %change %= stddev \ | \ | \ = | \ | \ | = \ = 2796 +23.3% 3448 +21.0% 3382 = +21.4% 3393 +20.4% 3366 +10.0% 3075= =C2=B1 9% will-it-scale.per_thread_ops * 635de956a7f5a6 Merge tag 'x86-mm-2021-04-29' of git://git.kernel.org/pu= b/scm/linux/kernel/git/tip/tip |\ = | * a500fc918f7b8d Merge branch 'locking/core' into x86/mm, to resolve co= nflict | |\ = | * | d43f17a1da2537 smp: Micro-optimize smp_call_function_many_cond() | * | a5aa5ce3005972 smp: Inline on_each_cpu_cond() and on_each_cpu() | * | 1608e4cf31b88c x86/mm/tlb: Remove unnecessary uses of the inline keyw= ord | * | 291c4011dd7ac0 cpumask: Mark functions as pure | * | 09c5272e48614a x86/mm/tlb: Do not make is_lazy dirty for no reason | * | 2f4305b19fe6a2 x86/mm/tlb: Privatize cpu_tlbstate >>= > fbc | * | 4ce94eabac16b1 x86/mm/tlb: Flush remote and local TLBs concurrently > = > = > _______________________________________________ > LKP mailing list -- lkp(a)lists.01.org > To unsubscribe send an email to lkp-leave(a)lists.01.org --===============9138920132208475712==--