#6 Arbitrary-precision real numbers have multiplication limit.

[Jason McGuiness]

For arbitrary-precision real numbers (double.h & cpp),
and this affects the integers (integer.h & cpp) too, they
have a limit of around 10^9 digits. This is because the
multiplication is currently implemented using an FFT.
For larger numbers of digits, a Toombe-Coombs
algorithm is required, see Knuth Vol 2. Unfortunately his
implementation is tricky to understand.

[Jason McGuiness]

Logged In: YES
user_id=513437

Pardon me - I mean Toom-Cook algorithm.