At 12:58 AM 10/13/2006, Allen Poapst wrote:

>My formula uses minimally all primes smaller or equal to the sqrt(x) to give 
>an accurate solution to pi(x). I am not sure if there is an accurate formula 
>for pi(x) which ALWAYS works, if there is than even the math profs at my 
>school have not heard of it.

Hi Allen,

Although you sent your first message some time ago, it appears that you
haven't received much information about the long history of methods for
computing pi(x).  I hope you will find some of the following to be of use.

Your description of your formula sounds very much like Legendre's formula,
which has been known for about two hundred years; it too requires knowledge
of the primes up to sqrt(x).  More efficient formulas of a similar type were
developed by Meissel and Lehmer, requiring knowledge of the primes only up
to the cube root of x.  These methods are described in detail in the first
chapter of Hans Riesel's book, _Prime Numbers and Computer Methods for
Factorization_, along with a few others.

Still better methods have been developed in the past two decades; using
these, the exact number of primes less than 4*10^22 has been determined.
(See http://primes.utm.edu/howmany.shtml .)

There are indeed "formulas" for pi(n) (see 
http://primes.utm.edu/notes/faq/p_n.html 
and http://en.wikipedia.org/wiki/Prime_formula ), but they generally work
by encoding the computations for some standard method of detecting
primes (trial division, or, even worse, Wilson's theorem) into a
summation.  They're not a practical means of computing anything, but
they do exist.

Hope that helps.

-- 
Fred W. Helenius
[EMAIL PROTECTED]


_______________________________________________
Prime mailing list
[email protected]
http://hogranch.com/mailman/listinfo/prime

Reply via email to