linux-mips
[Top] [All Lists]

Re: Endianness...

To: Jun Sun <jsun@mvista.com>
Subject: Re: Endianness...
From: Ralf Baechle <ralf@oss.sgi.com>
Date: Fri, 4 May 2001 09:46:34 -0300
Cc: nick@snowman.net, Matthew Dharm <mdharm@momenco.com>, Linux-MIPS <linux-mips@oss.sgi.com>
In-reply-to: <3AF0724B.D74D9AF9@mvista.com>; from jsun@mvista.com on Wed, May 02, 2001 at 01:47:07PM -0700
References: <Pine.LNX.4.21.0105021603030.22170-100000@ns> <3AF0724B.D74D9AF9@mvista.com>
Sender: owner-linux-mips@oss.sgi.com
User-agent: Mutt/1.2.5i
On Wed, May 02, 2001 at 01:47:07PM -0700, Jun Sun wrote:

> BE is better known perhaps because all SGI workstations are BE.  Generally I
> found networking systems tend to use BE while consumer electronic devices tend
> to use LE (which means there are probably more MIPS CPUs running in LE.)
> 
> So far I have not found much difference in terms of endianess, although
> occassionaly you have to IO swap in drivers for BE machine.

The difference is mostly of religious nature even though I've been told
that various embedded apps can show noticable performance difference due
to byteswapping.  In general I prefer big endian because it tends to
trigger certain bugs in software more than little endian, such as accessing
a memory object with different sizes.

  Ralf

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