Re: A89: Line 1111 Emulator

To: assembly-89@lists.ticalc.org
Subject: Re: A89: Line 1111 Emulator
From: Xavier VASSOR <xvassor@mail.dotcom.fr>
Date: Tue, 29 Dec 1998 17:12:23 +0100
Delivered-To: assembly-89-outgoing@towerguard.unix.edu.sollentuna.se
Organization: The Doors Team
References: <010001be333d$a4504c20$1c9bcc98@packard-bell>
Reply-To: assembly-89@lists.ticalc.org


You mean that add.l	#10,a7  is shorter than lea	10(a7),a7  ?
In this case you are wrong; add.l	#10,a7  is longer because of the size
(long word) of the immediate value #10.

 
> I have no idea whether it's faster, but it is, I think, at least two bytes
> shorter since <lea 10(a7),a7> uses relative addressing over absolute.
> Also, <lea 10(a7),a7> _isn't_ reading any memory; a7 is in parenthesis, but
> the lea makes the operand an address register and then ten is added to that
> (a7+10, not (a7)+10).
> 
> >
> >Wouldn't add.l 10, a7 be faster then doing it with lea?
> >Generally, (at least on the Intel processors),  adding an immediate to a
> >register is faster than setting a memory address to a register.
> >
> >Daniel Imfeld
> >

-- 

Xavier VASSOR
---The Doors Team
E-mail:xvassor@mail.dotcom.fr
Doors Homepage: http://start.at/doors
ICQ:10241721

References:

Re: A89: Line 1111 Emulator

From: "Dux Gregis" <assets@eden.rutgers.edu>