(No subject - faster division, multiplication and GCD)

Hans Petter Selasky hps at selasky.org
Mon Aug 29 15:43:46 CEST 2022

On 8/29/22 09:50, Paul Zimmermann wrote:
>         Dear Allan,
> Algorithm 1 page 4 computes one bit at a time. I doubt it can outperform the
> GMP routine, which computes 64 bits at a time (on a 64-bit computer). Please
> make your code available so that people can confirm your claims.
> Best regards,
> Paul Zimmermann

Hi Allan,

After some struggle decoding your PDF attachment, it appears to me that 
your optimization examples depends on lucky numbers for the division and 
multiplication - right?

Did you run your algorithms on random numbers?

If the case is you can optimize multiplication, division and GCD, then 
also you should be able to optimize exponentiation, by simple insert of 
your formulas into this one:



More information about the gmp-discuss mailing list