Paul.Zimmermann at inria.fr
Wed Aug 28 16:12:57 CEST 2013
the GMP documentation (5.1.2) says:
`mpf_get_str' doesn't currently use the algorithm described here, it
multiplies or divides by a power of b to move the radix point to the
just above the highest non-zero digit (or at worst one above that
location), then multiplies by b^n to bring out digits. This is O(N^2)
and is certainly not optimal.
First "to move the radix point to *the* just above the highest non-zero digit"
does not seem correct english to me.
Secondly some timings experiments suggest mpf_get_str uses a subquadratic
algorithm, thus it seems the above paragraph is obsolete.
More information about the gmp-bugs