Loading ...
Sorry, an error occurred while loading the content.
 

possible software bugs

Expand Messages
  • pbtoau
    Hi, I have started using CPAPSieve to count the number of primes in 10^8th blocks after powers of 10. The version I downloaded is from 9/29/2001. It is
    Message 1 of 2 , Apr 3, 2003
      Hi,

      I have started using CPAPSieve to count the number of primes in
      10^8th blocks after powers of 10. The version I downloaded is from
      9/29/2001. It is blazingly fast, but seems to overstate the count by
      1. e.g. pi(10^23+10^8)-pi(10^23)= 1886441. When I sieve past the
      square root, it still saves 1886442 values. I have looked at other
      ranges with the same plus 1 result. Is there a newer version? How
      can I process the output file to be human readable so I can find the
      extra "prime"? Is there something else available that is better for
      sieveing these contiguous blocks?

      I was using the Factwin program, which I got from the Prime Pages.
      When I entered the prime 8999999999999999983, it gave me the factors
      3272003 and 2750608724992. These produce 8999999999999998976 and are
      obviously not prime themselves. The biggest number it would accept
      was 1 less than 9*10^18 rather than 2^63. Is there a more recent
      working version or something better?

      Finally, I am using fastpix to try to zero in on nth primes, like the
      2 trillionth one (I think it is 61427839512211). It requires big
      steps for big powers of 10. I can get it to do a step that is about
      100 millionth of my x, but unless I get lucky, that can leave me
      quite a ways away form my goal. Is there something comparable with
      finer resolution? In the printout, I understand x and pi(x), but
      what are y, z, A, B, C, D, phi0 and Sigma?

      Best regards,

      David Baugh
    • jim_fougeron
      Type CPAPSieve -? to see a list of options. There is a -L option, which writes the file out as a human readable file (an input file for PFGW). Simply use
      Message 2 of 2 , Apr 4, 2003
        Type CPAPSieve -? to see a list of options.

        There is a -L option, which writes the file out as a "human" readable
        file (an input file for PFGW). Simply use (for file input file 10x23)
        CPAPSieve -I=10x23 -O=10x23.input -L -Mp=3
        Do this after you have trial factored and want to save to readable
        file.

        If starting from "scratch", then simply do
        CPAPSieve -O=10x26 -b=10 -e=26 -k=1 -K=100000001 -Mp=10000000100000 -L
        and the program would save data out in that format to start with.

        Note, as you move forward, trial factoring to sqrt(N) will become
        very time consuming. I would recommend trial factoring a much lower
        level, then running the results through PFGW to pull the primes out
        of the composites.

        Jim.


        --- In primenumbers@yahoogroups.com, "pbtoau" <PbtoAu@A...> wrote:
        > Hi,
        >
        > I have started using CPAPSieve to count the number of primes in
        > 10^8th blocks after powers of 10. The version I downloaded is from
        > 9/29/2001. It is blazingly fast, but seems to overstate the count
        by
        > 1. e.g. pi(10^23+10^8)-pi(10^23)= 1886441. When I sieve past the
        > square root, it still saves 1886442 values. I have looked at other
        > ranges with the same plus 1 result. Is there a newer version? How
        > can I process the output file to be human readable so I can find
        the
        > extra "prime"? Is there something else available that is better
        for
        > sieveing these contiguous blocks?
        >
        > I was using the Factwin program, which I got from the Prime Pages.
        > When I entered the prime 8999999999999999983, it gave me the
        factors
        > 3272003 and 2750608724992. These produce 8999999999999998976 and
        are
        > obviously not prime themselves. The biggest number it would accept
        > was 1 less than 9*10^18 rather than 2^63. Is there a more recent
        > working version or something better?
        >
        > Finally, I am using fastpix to try to zero in on nth primes, like
        the
        > 2 trillionth one (I think it is 61427839512211). It requires big
        > steps for big powers of 10. I can get it to do a step that is
        about
        > 100 millionth of my x, but unless I get lucky, that can leave me
        > quite a ways away form my goal. Is there something comparable with
        > finer resolution? In the printout, I understand x and pi(x), but
        > what are y, z, A, B, C, D, phi0 and Sigma?
        >
        > Best regards,
        >
        > David Baugh
      Your message has been successfully submitted and would be delivered to recipients shortly.