[VW 7.4.1] Floating point number precision

Hi,

I have a problem with floating point numbers.
When i use a floating point number with more than seven characters (i
use #asFloat),
(for example 1234.567, 3.142856,65265.534 ...etc) my actual number is
rounded off either to have digits less than 8 or returns the exponential
format.
I tried using Double (#asDouble) instead of Float, but it gives wrong
answers.
For example, when i take 3.14 asDouble it gives 3.1400001049042d .So i
thought avoiding the use of #asDouble.

In my application i need to add several floating point numbers. When i
loose the precision by .001 or .0001 for each such number, i have much
difference in the Total amount.

Have anyone experienced such problems?Can anyone suggest a solution or
some hints?

Thanks in advance,

maNi

 
Floating numbers depend on the number of bits (e.g. 32) of your hardware
and the representation of numbers in these bits, e.g. IEEE representation.

It is totally normal that on a 32 bit machine you get max. 6 digits reliable
floating points.

3.14 asDouble leads to a 64 bit representation of 3.14 in e.g. IEEE. The result (3.140000....)
is as exact as the number can be represented in 64 bits, sum of exponents to the power of 2.

You can NEVER rely on that 3.14 or any other floating number is REALLY 3.14.
A deviation is normal.

The most exact you can do is using double and taking care of comparisons
(never ask aNumber = 3.14 but use a delta, e.g. (aNumber - 3.14) abs < 0.0001 or similar).

OR: Convert your numbers to integers with a factor of the accuracy you need.

Karl


__________________________________________________________
Karl Breith
AREVA NP GmbH
Methods & Codes (FDEEC-G)

Freyeslebenstraße 1
91058 Erlangen
Phone: +49 (0) 9131 18-97393
Fax: +49 (0) 9131 18-94045
mail to:  [hidden email]
An AREVA and Siemens company
 
Vorsitzender des Aufsichtsrats: Bertrand Durrande - Geschäftsführer: Dr. Ralf Güldner, Rüdiger Steuerlein
Sitz der Gesellschaft: Erlangen - Registergericht: Fürth, HRB 7817 - www.areva-np.com
 
Wichtiger Hinweis: Diese E-Mail kann Betriebs- oder Geschäftsgeheimnisse, bzw. sonstige vertrauliche Informationen enthalten. Sollten Sie diese E-Mail irrtümlich erhalten haben, ist Ihnen eine Kenntnisnahme des Inhalts, eine Vervielfältigung oder Weitergabe der E-Mail ausdrücklich untersagt. Bitte benachrichtigen Sie uns und vernichten Sie die empfangene E-Mail. Vielen Dank.
 
Important Note: This e-mail may contain trade secrets or privileged, undisclosed or otherwise confidential information. If you have received this e-mail in error, you are hereby notified that any review, copying or distribution of it is strictly prohibited. Please inform us immediately and destroy the original transmittal. Thank you for your cooperation.


-----Ursprüngliche Nachricht-----
Von: Mani S Kartha [mailto:[hidden email]]
Gesendet: Freitag, 24. August 2007 09:43
An: [hidden email]
Betreff: [VW 7.4.1] Floating point number precision

Hi,

I have a problem with floating point numbers.
When i use a floating point number with more than seven characters (i use #asFloat), (for example 1234.567, 3.142856,65265.534 ...etc) my actual number is rounded off either to have digits less than 8 or returns the exponential format.
I tried using Double (#asDouble) instead of Float, but it gives wrong answers.
For example, when i take 3.14 asDouble it gives 3.1400001049042d .So i thought avoiding the use of #asDouble.

In my application i need to add several floating point numbers. When i loose the precision by .001 or .0001 for each such number, i have much difference in the Total amount.

Have anyone experienced such problems?Can anyone suggest a solution or some hints?

Thanks in advance,

maNi

Mani S Kartha wrote: All of these are normal, because most of the time floating point numbers
are by their very nature approximations of actual values.

Also, since they are (essentially) fractions of the form [an integer of
23 bits (floats) or 53 bits (doubles)] * 2^exponent, this means that it
is impossible for them to represent the vast majority of numbers we
write in decimal form without precision loss.

For example, there is no way you can express 0.3 as a floating point
number if the base of the floating point number is 2.

If you need zero precision loss, I'd suggest you use scaled decimal or
fractions.

Floating point arithmetic is extremely well studied.  For further
information, I'd recommend picking up a copy of e.g.: Donald E. Knuth's
The Art of Computer Programming.

Thanks,
Andres.