GMP 4.4 remaining tasks

bodrato at bodrato at
Mon Dec 7 12:04:45 CET 2009

Uhm... Paul, are you around there?

> I haven't looked closely neither at your code or at left-to-right
> inversion in general. But maybe you can get some ideas from Alg 3.5
> ApproximateReciprocal (page 121 in version 0.4) in Brent's and
> Zimmermann's MCA book?

The lemma immediately _before_ that algorithm descriptions says:
"Algorithm ApproximateReciprocal satisfies:
                     AX < B^2n < A(X + 2)."
The Proof immediately _after_ declare:
"In all cases we have AX < B^2n < A(X + 1)"

If the condition is "B^2n < A(X+2)", then my implementation _is_ correct
(the approximation is good enough :-). If we need "B^2n < A(X+1)", then I
have to work on it.

Which condition should the result satisfy?



More information about the gmp-devel mailing list