reproducibility of GMP random functions vs limb size and GMP version

Paul Zimmermann Paul.Zimmermann at
Fri Sep 17 02:10:43 CEST 2010

       Dear Pedro,

thank you for the details about GMP random functions. I would suggest to
add those very interesting details in the Algorithms chapter of the GMP

I have one question:

> mpz_urandomm extracts the exact number of random bits necessary to
> accomodate the given limit, exactly as mpz_urandomb does. If the result
> is greater than or equal to the limit, all bits are discarded and a new
> extraction is performed. There is a limit to the iteration count to
> prevent ill generators from locking GMP.

does it mean that if N=2^n+1 for large n, and I call mpz_urandomm with
modulus N, we will most likely hit the limit?


More information about the gmp-devel mailing list