Re: Performance benefits of rep/loop?
From: Matt Taylor (para_at_tampabay.rr.com)
Date: 05/26/04
- Previous message: Matt Taylor: "Re: LEA instruction"
- In reply to: flekso: "Re: Performance benefits of rep/loop?"
- Next in thread: Randall Hyde: "Re: Performance benefits of rep/loop?"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Date: Wed, 26 May 2004 09:20:21 +0000 (UTC)
"flekso" <flekso@vip.hr> wrote in message news:c8v5st$vf$1@ls219.htnet.hr...
<snip>
> Y'all talk about loop unrolling, replacing 1 old instruction with 3-5
new..
> But what happens when you write an OS using such techniques, and each of
the
> 21 running threads has code written in such manner, what happens with the
> instruction cache ? And if every process is written like that ?
In that case, it is rather fortunate that code spends 99% of its time in
loops.
-Matt
- Previous message: Matt Taylor: "Re: LEA instruction"
- In reply to: flekso: "Re: Performance benefits of rep/loop?"
- Next in thread: Randall Hyde: "Re: Performance benefits of rep/loop?"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Relevant Pages
|