lkml.org 
[lkml]   [2013]   [Sep]   [11]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [PATCH 0/7] preempt_count rework -v2
On 09/10/2013 09:24 AM, Linus Torvalds wrote:
> On Tue, Sep 10, 2013 at 8:29 AM, Arjan van de Ven <arjan@linux.intel.com> wrote:
>>>
>> also.. yuck on using "dec"
>> "dec" sucks, please use "sub foo ,1" instead
>
> That's a bigger instruction, largely due to the constant.
>
>> (dec sucks because of its broken flags behavior; it creates basically a
>> bubble in the pipeline)
>
> Intel could (and should) just fix that. It's "easy" enough - you just
> need to rename the carry flag as a separate register (and make sure
> that the conditional instructions that don't use it don't think they
> need it).
>
> In fact I thought Intel already did that on their large cores. Are you
> sure they still have trouble with inc/dec?
>

Big cores are fine, but Atom (and presumably Quark) might still suffer a
performance penalty.

-hpa




\
 
 \ /
  Last update: 2013-09-11 18:21    [W:0.090 / U:1.296 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site