Message-Id: <3.0.6.32.19990929225059.0079a180@dce03.ipt.br>
X-Sender: csrabak AT dce03 DOT ipt DOT br
X-Mailer: QUALCOMM Windows Eudora Light Version 3.0.6 (32)
Date: Wed, 29 Sep 1999 22:50:59 -0300
To: djgpp AT delorie DOT com
From: "Cesar S. Rabak" <csrabak AT ipt DOT br>
Subject: Re: djgpp and rounding errors
In-Reply-To: <Pine.SUN.3.91.990929103650.11087U-100000@is>
References: <37F173EB DOT 77A2 AT linetap DOT com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Reply-To: djgpp AT delorie DOT com
X-Mailing-List: djgpp AT delorie DOT com
X-Unsubscribes-To: listserv AT delorie DOT com
Precedence: bulk

At 10:37 29/09/99 +0200, Eli Zaretskii wrote:
>
>On Tue, 28 Sep 1999, nebo wrote:
>
>> Does DJGPP fix rounding errors with decimal places?  I know
>> 	some compilers use BCD math to circumvent this problem, but
>> 	I found nothing of the sort in DJGPP.
>
>Please explain more about what you are asking.  What rounding errors
>did you have in mind, and why do they need fixing with BCD?
>
>Floating-point computations in DJGPP run with 80-bit internal
>precision of the numerical co-processor, if that helps.
>
>
Try this small snippet in your machine:

#include <stdlib.h>
#include <stdio.h>

int
main(void)
{
    int i;
    float number = 0.0f;

    for (i=0; i< 10000000; i++)
        number += 0.001;

    printf("Result is %5.2f\n", number);

    return 0;
    
}

In _my_ equipment (a 233MMX Pentium running Win95), gcc 2.8.1:

bash$ gcc -o teste teste.c
bash$ ./teste
Result is 9780.20

This is a problem common to all programming languages, and has to do with
the representation of numbers in binary form. Nor 80 bit helps here. . . 

Hope this clarifies. 

This problem for certain application domains (e.g. when calculations about
money are being done) is more than a nuisance, this is the reason why the
BCD stuff was developed.

HTH

Cesar