On Sun, Sep 13, 2009 at 10:25 PM, jason <ja...@njkfrudils.plus.com> wrote:
> Hi , did you mean that results should be repeatable across different > arch's ie 64/32bit ? or just repeatable on a set arch? A set architecture would be fine, so people could say, using arch X, with MPIR vY.Z then using this function with this seed will give you the following output... > For a set arch and random seed then everything is repeatable. For a > set random seed and any arch then the urandom functions (uniformly > distributed) are repeatable(or should be!!!) , the rrandom fn's are > not. The mpz_probable_prime_p fn is "not" , although this is only > because on some arch'es it may be faster to do more trial division > than on others and so if you are (un)lucky you will give the function > a number which is composite but passes the miller-rabin test but fails > the trial div test. The mpz_likely_prime_p test is optimized for speed > so has no guarentee , same for mpz_next_likely_prime_p. The likely ones as you said are for speed so that is fine. It would be good to document what you said above about arch/random seed being repeatable or not just so people are aware in case they want to use stuff across different platforms/archs. Jeff. --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "mpir-devel" group. To post to this group, send email to mpir-devel@googlegroups.com To unsubscribe from this group, send email to mpir-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/mpir-devel?hl=en -~----------~----~----~----~------~----~------~--~---