Path: utzoo!attcan!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!sdd.hp.com!usc!jarthur!nntp-server.caltech.edu!juliet.caltech.edu!gceych
From: gceych@juliet.caltech.edu (Eychaner, Glenn C.)
Newsgroups: comp.lang.c
Subject: Re: which bits are set
Message-ID: <1990Dec18.171839.15135@nntp-server.caltech.edu>
Date: 18 Dec 90 17:04:30 GMT
References: <3047:Dec1618:51:1590@kramden.acf.nyu.edu> <1990Dec17.071404.6544@kithrup.COM> <15598:Dec1804:57:0390@kramden.acf.nyu.edu> <1990Dec18.093905.17196@kithrup.COM>
Sender: news@nntp-server.caltech.edu
Reply-To: gceych@juliet.caltech.edu
Organization: California Institute of Technology
Lines: 42
News-Software: VAX/VMS VNEWS 1.3-4x
Nntp-Posting-Host: juliet.caltech.edu

In article <1990Dec18.093905.17196@kithrup.COM>, sef@kithrup.COM (Sean Eric Fagan) writes...
>In article <15598:Dec1804:57:0390@kramden.acf.nyu.edu> brnstnd@kramden.acf.nyu.edu (Dan Bernstein) writes:
>>Eek. Why do you want to make this so slow?
> 
Ok, to throw my two cents in, this is the algorithm I use...

int bit_set_array [SIZE];

void test_for_bits (int test_int)
{
int j = 0;

while (test_int) {
    bit_set_array [j] = test_int & 1;
    test_int >>= 1;
    j++;
    }
for (; j < SIZE; j++) bit_set_array [j] = 0;
}

Note that this has the advantage of dropping out if test_int is small.
Alternatively:

while (j < SIZE) {

has the advantage of not filling the array at the end.
(I hope this is right.  It's a conversion from my parity calculation program,
 which is:)

parity = 0;
while (test_int) {
    parity ^= test_int & 1;
    test_int >>= 1;
    }
return (parity);

I love bitwise operators!

Glenn Eychaner   |Eychaner@SunCub.Caltech.edu |Remember: It is easier to ride a
40386 N Shore Ln |gceych@iago.caltech.edu     |camel through the eye of a needle
Big Bear City, CA| Big Bear Solar Observatory |than to drive a Buick through the
            92314| !*** G O  N I N E R S ***! |hole in a doughnut.