## Re: Sophie heuristics

Expand Messages
• ... cool - I see how they got the predicted # less than N (after figuring out to use ln instead of log on my graph calc). Now I m just wondering what program
Message 1 of 3 , Jun 7, 2002
• 0 Attachment
> see section 3.5 and Table 6 of Chris Caldwell's note:
> http://www.utm.edu/~caldwell/preprints/Heuristics.pdf

cool - I see how they got the predicted # less than N (after figuring
out to use ln instead of log on my graph calc). Now I'm just
wondering what program you guys use to calculate these equations - on
my graphing clac I get infinity or undef for a lot of answers bc. the
#'s are so large.
Right now I'm working on Sophie Germain primes with 10544 digits, so
I was going to calculate the # of Sophie Germain primes w/ 10544
digits = 2C * \$(ln(y)*ln(2y))^-1 dy (where \$ = integral) w/ limits
10^10543 (lower) and 10^10544 (upper) to estimate the # of s.g.
primes between 10^10543 and 10^10544 (ie out of a possible 9 *
10^10543), then divide 9 * 10^10543 by the answer the see how many
#'s I'd have to search through on average before I found a s.g.
prime. Unfortunately my calc gives me garbage answers - any
reccommended prog's?
• ... With such a narrow range (on a logarithmic scale) you make safely estimate the integral as the integrand times the increment. David Broadhurst
Message 2 of 3 , Jun 7, 2002
• 0 Attachment
> Unfortunately my calc gives me garbage answers
With such a narrow range (on a logarithmic scale)
you make safely estimate the integral
as the integrand times the increment.