Sender: nate AT cartsys DOT com
Message-ID: <36BA705B.3F0D8C68@cartsys.com>
Date: Thu, 04 Feb 1999 20:15:23 -0800
From: Nate Eldredge <nate AT cartsys DOT com>
X-Mailer: Mozilla 4.08 [en] (X11; I; Linux 2.0.36 i586)
MIME-Version: 1.0
To: djgpp AT delorie DOT com
Subject: Re: Question about long long math on intel archs
References: <3 DOT 0 DOT 6 DOT 32 DOT 19990202153421 DOT 0091a590 AT pop DOT netaddress DOT com> <3 DOT 0 DOT 6 DOT 32 DOT 19990204025117 DOT 008ee990 AT pop DOT netaddress DOT com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Reply-To: djgpp AT delorie DOT com

Paul Derbyshire wrote:
> 
> At 05:18 PM 2/2/99 -0800, you wrote:
> >movl y, %eax
> >addl %eax, x
> >movl y+4, %eax
> >adcl %eax, x+4
> 
> I assume adcl uses the carry from the first add?

Correct.

> This leaves one question. If the result of the adds is left in eax it never
> gets saved and the second add clobbers the result of the first one. Or does
> addl/adcl put the result in the address at right, x and then x+4?

Yes.  Move a long from y to eax, then add it to x.  eax is just a
scratch register since we have no mem-to-mem adds.
-- 

Nate Eldredge
nate AT cartsys DOT com