Re: [PATCH] ipc/msg.c: mitigate the lock contention with percpu counter

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

To: "Sun, Jiebin" <jiebin.sun@xxxxxxxxx>
Subject: Re: [PATCH] ipc/msg.c: mitigate the lock contention with percpu counter
From: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
Date: Wed, 7 Sep 2022 13:43:20 -0700
Cc: Tim Chen <tim.c.chen@xxxxxxxxxxxxxxx>, Shakeel Butt <shakeelb@xxxxxxxxxx>, vasily.averin@xxxxxxxxx, Dennis Zhou <dennis@xxxxxxxxxx>, Tejun Heo <tj@xxxxxxxxxx>, Christoph Lameter <cl@xxxxxxxxx>, "Eric W. Biederman" <ebiederm@xxxxxxxxxxxx>, Alexey Gladkov <legion@xxxxxxxxxx>, Manfred Spraul <manfred@xxxxxxxxxxxxxxxx>, alexander.mikhalitsyn@xxxxxxxxxxxxx, Linux MM <linux-mm@xxxxxxxxx>, LKML <linux-kernel@xxxxxxxxxxxxxxx>, "Chen, Tim C" <tim.c.chen@xxxxxxxxx>, Feng Tang <feng.tang@xxxxxxxxx>, Huang Ying <ying.huang@xxxxxxxxx>, tianyou.li@xxxxxxxxx, wangyang.guo@xxxxxxxxx
In-reply-to: <a4d10f9e-ab8f-ffad-5eea-48316c436f41@intel.com>
References: <20220902152243.479592-1-jiebin.sun@intel.com> <CALvZod44uUFnwfF4StC24t+d1s_XE10hkmSCgb04FjtTATo6xQ@mail.gmail.com> <048517e7f95aa8460cd47a169f3dfbd8e9b70d5c.camel@linux.intel.com> <a4d10f9e-ab8f-ffad-5eea-48316c436f41@intel.com>

On Wed, 7 Sep 2022 17:39:47 +0800 "Sun, Jiebin" <jiebin.sun@xxxxxxxxx> wrote:

> I have sent out the patch v4 which use percpu_counter_add_batch. If we use
> a tuned large batch size (1024),

Oh.  Why not simply use a batch size of INT_MAX?

> the performance gain is 3.17x (patch v4)
> vs 3.38x (patch v3) previously in stress-ng -- message. It still has
> significant performance improvement and also good balance between
> performance gain and overflow issue.

References:
- [PATCH] ipc/msg.c: mitigate the lock contention with percpu counter
  - From: Jiebin Sun
- Re: [PATCH] ipc/msg.c: mitigate the lock contention with percpu counter
  - From: Shakeel Butt
- Re: [PATCH] ipc/msg.c: mitigate the lock contention with percpu counter
  - From: Tim Chen
- Re: [PATCH] ipc/msg.c: mitigate the lock contention with percpu counter
  - From: Sun, Jiebin

Prev by Date: Re: [v2 PATCH 2/2] powerpc/64s/radix: don't need to broadcast IPI for radix pmd collapse flush
Next by Date: Re: [PATCH 00/13] Fix the DAX-gup mistake
Previous by thread: Re: [PATCH] ipc/msg.c: mitigate the lock contention with percpu counter
Next by thread: [PATCH v4] ipc/msg: mitigate the lock contention with percpu counter
Index(es):
- Date
- Thread

[Index of Archives] [Linux ARM Kernel] [Linux ARM] [Linux Omap] [Fedora ARM] [IETF Annouce] [Bugtraq] [Linux OMAP] [Linux MIPS] [eCos] [Asterisk Internet PBX] [Linux API]