tables of p(n) and/or pi(n)

[Hans Havermann]

Joshua Zucker wrote:

> The easy compression fits 30 numbers per byte, by using one bit for
> prime/not prime for each of the number equal to 1, 7, 11, 13, 17, 19,
> 23, 29 mod 30.
>
> So it's not 30 primes per byte, but rather 30 numbers.
>
> With that scheme, since pi(3*10^13) is roughly 10^12, it would take
> only (3*10^13)/30 bytes to store the primality of all of them, so
> indeed, 1TB is about right (not 10TB as in your corrected post).

To be fair, I'm not just storing primality, but also relatively fast  
recovery of pi(prime) values.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://list.seqfan.eu/pipermail/seqfan/attachments/20050919/ab87afaa/attachment-0001.htm>