From: Martin Ambuhl <mambuhl AT earthlink DOT net>
Newsgroups: comp.os.msdos.djgpp
Subject: Re: cos(90 * (PI / 180) = 6.12......?
Date: Fri, 14 May 1999 22:51:53 -0400
Content-Transfer-Encoding: 7bit
References: <926562032 DOT 737 DOT 64 AT news DOT remarQ DOT com>
X-Posted-Path-Was: not-for-mail
X-Accept-Language: en
Content-Type: text/plain; charset=us-ascii
X-ELN-Date: 15 May 1999 02:50:39 GMT
X-ELN-Insert-Date: Fri May 14 19:55:15 1999
Organization: Nocturnal Aviation
Lines: 42
Mime-Version: 1.0
NNTP-Posting-Host: dialup-209.246.66.218.newyork2.level3.net
Message-ID: <373CE149.8C7EB282@earthlink.net>
X-Mailer: Mozilla 4.51 [en] (Win95; I)
To: djgpp AT delorie DOT com
DJ-Gateway: from newsgroup comp.os.msdos.djgpp
Reply-To: djgpp AT delorie DOT com

Shaggs wrote:
> 
> wtf is up with this, all my other cosine values seem to be normal, but this
> comes up (although i havent checked 270).
> 
> any ideas? please help ASAP

There is nothing strange here...

#undef __STRICT_ANSI__
#include <math.h>
#include <stdio.h>
#include <float.h>
int main(void)
{
    double angle = M_PI / 2, result = cos(M_PI / 2);
    printf("\"Shaggs\" <jud AT tminet DOT com> reports "
           "cos(90 degrees == %g)==6.12...\n"
           "but we get cos(%g)==%g\n"
           "Note that \"Shaggs\" did not report the exponent.\n"
           "It is important to remember that even though DBL_MIN==%g,\n"
           "in computations it is best to treat numbers smaller than\n"
           "a small multiple (usually 4-6) of DBL_EPSILON==%g as zero\n"
           "and (cos(%g)==%g)/(DBL_EPSILON==%g) == %g\n",
           angle, angle, result, DBL_MIN, DBL_EPSILON,
           angle, result, DBL_EPSILON, result / DBL_EPSILON);
    return 0;
}


"Shaggs" <jud AT tminet DOT com> reports cos(90 degrees == 1.5708)==6.12...
but we get cos(1.5708)==6.1232e-17
Note that "Shaggs" did not report the exponent.
It is important to remember that even though DBL_MIN==2.22507e-308,
in computations it is best to treat numbers smaller than
a small multiple (usually 4-6) of DBL_EPSILON==2.22045e-16 as zero
and (cos(1.5708)==6.1232e-17)/(DBL_EPSILON==2.22045e-16) == 0.275764


-- 
Martin Ambuhl (mambuhl AT earthlink DOT net)
Note: mambuhl AT tiac DOT net will soon be inactive