Path: utzoo!attcan!uunet!lll-winken!ncis.llnl.gov!helios.ee.lbl.gov!nosc!ucsd!rutgers!att!ulysses!andante!alice!andrew
From: andrew@alice.UUCP (Andrew Hume)
Newsgroups: comp.lang.c
Subject: Re: BIT Counting: C programming problem
Message-ID: <8837@alice.UUCP>
Date: 26 Jan 89 05:27:39 GMT
References: <225@tityus.UUCP> <35329@think.UUCP> <8398@dasys1.UUCP> <12571@diamond.BBN.COM>
Organization: AT&T Bell Laboratories, Murray Hill NJ
Lines: 12

> >[Barry Margolin suggests using a 64K lookup table for something.]
> >
> >That is a ridiculous idea.  For microscopic savings in speed, you
> >waste 63K of space...what if it needs 400K to run effectively, or 
> >even 640K...Besides, more and more people are going to multiprogramming 
> >environments, even on PCs, like Windows or DesqView.  


the time saving is not microscopic (i measure a 1.8x speedup between
8 bit lookup and 16 bit lookup) unless you don't call the bitcount
routine. i do not encourage mindless devouring of memory but i
figure that i don't much of a time tradeoff to worry about 64KB.