realbasic-betas.mbox
[Top] [All Lists]

Re: runtime failure

To: REALbasic Betas <realbasic-betas at lists dot realsoftware dot com>
Subject: Re: runtime failure
From: Thomas Tempelmann <listuser at tempel dot org>
Date: Sun, 29 Dec 2002 17:54:26 +0100
Theodore H. Smith wrote:

>I don't think reading 1 
>char from the RAM is any slower at non aligned addresses than aligned 
>addresses.

But are you aware that writing may be tricky? If one byte should be
set, but the memory is organized in 32 bit size "words", an additional
read cycle is necessary before the data can be modified in RAM.
For that reason, it may in special cases be more effective to only
use 32 bit variables even if only one bit is used. But those are
very rare cases anyways, with the modern complex architecture of
multi-level caches (meaning you hardly have any control over this).

Thomas


-- 
Thomas Tempelmann,
http://www.tempel.org/rb/



---
A searchable archive of this list is available at:
<http://dbserver.realsoftware.com/KBDB/search.php>

Unsubscribe:
<mailto:realbasic-betas-off at lists dot realsoftware dot com>

Subscribe to the digest:
<mailto:realbasic-betas-digest at lists dot realsoftware dot com>
.


<Prev in Thread] Current Thread [Next in Thread>