Implementation of weierstrass function

[Phillip Kehler]

Do you know for C implementation of the weierstrass function as
defined from http://en.wikipedia.org/wiki/Weierstrass_function with
guaranteed 18 decimal digits precision (long double) for 0 <= x < 2?

Thanks,
Phillip

 

[Phlip]

Do you know for C implementation of the weierstrass function as
defined from http://en.wikipedia.org/wiki/Weierstrass_function with
guaranteed 18 decimal digits precision (long double) for 0 <= x < 2?

You are asking about a "bigint" or "high precision math library". I think
there's one out there that begins with G. After you find it (by googling for
"high precision library"), typing in the formula will be trivial.

 

[Chris Torek]

Do you know for C implementation of the weierstrass function as
defined from http://en.wikipedia.org/wiki/Weierstrass_function with
guaranteed 18 decimal digits precision (long double) for 0 <= x < 2?

Since the Weierstrass function is not differentiable (despite being
continuous -- it is "infinitely jagged", similar I suppose to Julia
sets for instance) some of the standard ideas, such as using a Taylor
expansion, are right out.

Worse yet, you ask for a "C implementation" that is "guaranteed
18 decimal digits precision", but no C99-conformant compiler is
ever obliged to give you more than 10 decimal digits of precision
for *any* floating-point type. Without going beyond the C standard,
you cannot achieve the desired result at all.

I would suggest none of the newsgroups to which this is posted is
really appropriate, and sci.math.num-analysis is perhaps the right
group.