Floating Point Calculation Problem

Thread view:

Enable EMail Alerts

Start New Thread

Thread rating:

DJ - 29 Sep 2006 21:16 GMT

Hello,

In an application that worked fine until recently I have started to have
problems with calculations involving two doubles. For example: -24.708 +
26.0 = 1.2920000000000016, however, -24.708 + 8.0 = -16.708. What would
cause this? I don't think the numbers are so large that it would be the
usual floating point calculation error because of rounding. How can I fix
it? I'd rather not change every calcuation in my program to include a
math.round...

.Net Framework 1.1 version 1.1.4322 SP1
Visual Basic - Visual Studio .Net 2003
Windows XP

Thanks,

Reply to this Message

Chris Saunders - 29 Sep 2006 22:42 GMT

The size of your numbers are not relevant to this issue.
Here is an URL with an article you should probably read if you wish to
better
understand this issue: http://docs.sun.com/source/806-3568/ncg_goldberg.html

I know nothing about your application but you may indeed need something
along the lines you suggested.

Regards
Chris Saunders

> Hello,
>
[quoted text clipped - 11 lines]
>
> Thanks,

Reply to this Message

Jon Skeet [C# MVP] - 30 Sep 2006 19:13 GMT

> In an application that worked fine until recently I have started to have
> problems with calculations involving two doubles. For example: -24.708 +
[quoted text clipped - 3 lines]
> it? I'd rather not change every calcuation in my program to include a
> math.round...

See http://www.pobox.com/~skeet/csharp/floatingpoint.html

Signature

Jon Skeet - <skeet@pobox.com>
http://www.pobox.com/~skeet Blog: http://www.msmvps.com/jon.skeet
If replying to the group, please do not mail me too

Reply to this Message

Nick Malik [Microsoft] - 01 Oct 2006 05:11 GMT

> Hello,
>
> In an application that worked fine until recently I have started to have
> problems with calculations involving two doubles. For example: -24.708 +
> 26.0 = 1.2920000000000016, however, -24.708 + 8.0 = -16.708. What would
> cause this?

I don't see a problem here. Why do you?

> I don't think the numbers are so large that it would be the
> usual floating point calculation error because of rounding.

Who said size has anything at all to do with this calculation? Rounding
doesn't just happen for very large numbers. It happens for all of them.

> How can I fix
> it?

Decimal datatype

> I'd rather not change every calcuation in my program to include a
> math.round...
[quoted text clipped - 4 lines]
>
> Thanks,

Reply to this Message

Rate this thread:

Floating Point Calculation Problem

Free Magazines