Floating point arithmetic model of C...

[Bartc]

Thanks fior the comments guys. I have reverse engineered some 99-C
code into C# and have noticed an unusual bug(it causes the C# to not
work properly). I habve traced it down to a single method and have
compared results with the C version and noticed that small differences
in the results build up (as the calulation i s inside a loop with a +=
operator working on it).

Suggest at this point forgetting the C version, and concentrating on the C#
code that is causing the problem.

And asking about that in a place such as, I don't know, maybe
microsoft.public.dotnet.languages.csharp.

Then it can be seen if the results are reasonable or if there is a bug.

 

[Richard Bos]

Anyone with any ideas?

Yes. Design your programs so that minor variations in floating point
behaviour do not matter in the first place. Anything else will fail
sooner or later.

Richard

 

[lawrence.jones]

n1256.pdf consists of the C99 standard with
the three TCs merged into it; it's not an official ISO document, but
it was edited by a member of the committee, and it's close enough for
most purposes.

For the record, there is one known error in N1256 (other than
"September" being misspelled in the header): The predefined macro
__STDC_MB_MIGHT_NEQ_WC__ should be in 6.10.8p2 (conditional macros)
rather than p1 (required macros).

 

[Jean-Marc Bourguet]

(e-mail address removed) said:


ICBW but ISTR that I spotted that one. Do you do cheques for $2.56?

http://www-cs-faculty.stanford.edu/~knuth/news08.html

Yours,

 

[lawrence.jones]

ICBW but ISTR that I spotted that one. Do you do cheques for $2.56?

As soon as I start collecting royalties, I'll send it along.