Any interest (Wolfram or general) in fast multiply?
Does anyone know if Wolfram (or anyone else for that matter) has any
interest in replacing the _Integer multiplication N^1.58 code with N
logN code (i.e. Karatsuba algorithm with some form of discrete
transform such as Fourier)? I am interested in large primes, and when
you get to 10 or 20 thousand digits, a factor of Sqrt[N] would be
greatly appreciated... For a look at the discrete transform stuff see
the WEB pages dealing with GIMPS (Great Internet Mersenne Prime Search)
or Proth.exe (finds Proth's theorem primes of size tens of thousands of
digits...) ... thanks for any replies (a beginner who remembers using
Macsyma in 1973 .. vive la difference!!!)
Prev by Date:
Next by Date:
Prev by thread:
Next by thread:
Re: Any interest (Wolfram or general) in fast multiply?