linux-kernel - Re:Re: [PATCH] kmalloc_index optimization(code size & runtime stable)

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [day] [month] [year] [list]

Message-ID: <AJkA0AC2CGauZEv3r-l2napW.1.1587114742624.Hmail.bernard@vivo.com>
Date:   Fri, 17 Apr 2020 17:12:22 +0800 (GMT+08:00)
From:   赵军奎 <bernard@...o.com>
To:     Matthew Wilcox <willy@...radead.org>
Cc:     linux-kernel <linux-kernel@...r.kernel.org>,
        "opensource.kernel" <opensource.kernel@...o.com>
Subject: Re:Re: [PATCH] kmalloc_index optimization(code size & runtime stable)




Date: 2020-04-17 11:23:54
To:  Bernard Zhao <bernard@...o.com>
Cc:  Christoph Lameter <cl@...ux.com>,Pekka Enberg <penberg@...nel.org>,David Rientjes <rientjes@...gle.com>,Joonsoo Kim <iamjoonsoo.kim@....com>,Andrew Morton <akpm@...ux-foundation.org>,linux-mm@...ck.org,linux-kernel@...r.kernel.org,kernel@...o.com
Subject: Re: [PATCH] kmalloc_index optimization(code size & runtime stable)>On Thu, Apr 16, 2020 at 07:03:30PM -0700, Bernard Zhao wrote:
>> kmalloc_index inline function code size optimization and runtime
>> performance stability optimization. After optimization, the function
>> kmalloc_index is more stable, the size will never affecte the function`s
>> execution efficiency.
>> And follow test data shows that the performance of new optimization
>> exceeds the original algorithm when applying for more than 512 Bytes
>> (include 512B).And new optimization runtime is more stable than before.
>
>That's all very well and good, but the vast majority of allocations
>are less than 512 bytes in size!  Your numbers show that on average,
>this patch makes the kernel slower!

>


    This is indeed the case, the new algorithm is stable at a time level, but 
there is a certain performance loss for relatively small memory(little than 512).
I will continue to pay attention to this part later.  Thanks.

>>             size        time/Per 100 million times.us
>>                         old fun		new fun with optimise
>> 		8	203777		241934
>> 		16	245611		409278
>> 		32	236384		408419
>> 		64	275499		447732
>> 		128	354909		416439
>> 		256	360472		406598
>> 		512	431072		409168
>> 		1024	463822		407401
>