Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1 6/24/83; site alice.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!allegra!alice!ark
From: ark@alice.UucP (Andrew Koenig)
Newsgroups: net.lang.c
Subject: Re: Divisions with shifts
Message-ID: <4789@alice.UUCP>
Date: Fri, 10-Jan-86 17:33:21 EST
Article-I.D.: alice.4789
Posted: Fri Jan 10 17:33:21 1986
Date-Received: Sat, 11-Jan-86 07:20:17 EST
References: <1365@brl-tgr.ARPA>
Organization: Bell Labs, Murray Hill
Lines: 20

> Just shifting using an arithmetic shift may give the wrong answer, but
> you could correct the rounding and still get much faster execution like:
>
>	shiftRightArithmetic	register
>	branchIfPositive	label
>	increment		register
> label:
>
> so -23/2 = -23>>1 + 1 = 11101001>>2 + 1 = 11110100 + 1 = 11110101 = -11
> which is the right answer...
>
>					-Miles

Yet another example of why this problem is harder than it appears.
The algorithm suggested above gives the wrong answer whenever
the remainder from the division is 0.  Thus (-2>>1)+1 is (11111110>>1)+1
is 11111111+1 is 0.

Oh yes, 11101001>>2 + 1 is really 11111101, because addition binds
more tightly than shifting.