Sieving discussion thread

[cipher]

n=500,000
range= 0-50G
sieving depth= 138.0 T
candidates= 19,638,289
rate= 1 k every 0.8 seconds

[Denahar]

Just some playing with numbers:

During the last 5 days while sieving through the range 2300T-2450T I have removed 36.468 k's. That equals a total processing time on the same host testing with LLR of approximately 70 days.
That's impressive and now I do better understand the reason behind sieving, too.

-D

[cipher]

But if we find a Twin prime tom for n=333333 where k is around 2.1G (highly unlikely) all that deep sieving will be pointless.
cipher

[Skligmund]

It just means that we will get there quicker. The higher up the twin, the more efficient sieving is over LLR. If it is up at 40G or so, then all this seiving will have saved us months.

[cipher]

n=500,000
range= 0-50G
sieving depth= 228.8 T
candidates= 19,042,103
Avg K per 1M = 381

[cipher]

n=500,000
range= 0-50G
sieving depth= 300 T
candidates= 18,732,832
Avg K per 1M = 374

[cipher]

Gribozavr what utility do you use to remove n's from the main file when you just have .del file? how do you do it can you please explain?
thanks
cipher

[gribozavr]

I use a bit modified scripts from 321 project:
get_latest_sieve_file.sh

#!/bin/sh

cat *.del | cut -f3 -d= | sort -n | uniq > delete_file
perl remove_deleted.pl > latest_sieve_file

cat *.del | sort -n | uniq | perl check_divisors.pl > check_file

cat check_file | gp

remove_deleted.pl

#!/usr/bin/perl
open(DELETE_FILE, "delete_file");
open(SIEVE_FILE, "twin333333-16.txt");

$sieve_head = <SIEVE_FILE>;
print($sieve_head);

$delete_value = '';
while($delete_value == '')
{
  $delete_value = <DELETE_FILE>;
}

while($output = <SIEVE_FILE>)
{
  @sieve_line = split(' ', $output);
  while(($sieve_line[0] > $delete_value) && ($delete_value != EOF))
  {
    $delete_value=<DELETE_FILE>;
  }
  if($sieve_line[0] != $delete_value)
  {
    print $output;
  }
}

close(SIEVE_FILE);
close(DELETE_FILE);

check_divisors.pl

#!/usr/bin/perl -w
while(<>)
{
  if(m/^p=(\d+) divides k=(\d+)\r?$/)
  {
    print("if((lift($2*Mod(2,$1)^333333-1)!=0)&&(lift($2*Mod(2,$1)^333333+1)!=0),print(\"$1 $2 bad\"))\n");
  }
  else
  {
    print STDERR "What's $_";
  }
}

I also use some scripts to extract a file for LLR and to check a del file's sieving range.

get_llr_file.pl

#! /usr/bin/perl

use warnings;
use strict;

our $min = $ARGV[0];
our $max = $ARGV[1];

$_ = <STDIN>;
print $_;

while($_ = <STDIN>)
{
  if(/^(\d+) \d+\r?$/)
  {
    if(($1 >= $min) && ($1 <= $max))
    {
      print $_;
    }
    if($1 > $max)
    {
      exit;
    }
  }
}

get_p_range.pl

#!/usr/bin/perl -w
our $max_delta = 14000000000;
our $prev_p = 0;
our $in_gap = 1;
while(<>)
{
  if(m/^p=(\d+) divides k=(\d+)\r?$/)
  {
    if($prev_p)
    {
      if($in_gap)
      {
        print("range start: $prev_p\n");
        $in_gap = 0;
      }
      if(($1 - $prev_p) >= $max_delta)
      {
        print("range end:   $prev_p\n\n");
        $in_gap = 1;
      }
    }
    $prev_p = $1;
  }
  else
  {
    print STDERR "What's $_";
  }
}

print("range end:   $prev_p\n\n");

MooooMoo, you might want to move these posts to a separate topic, the scripts in question could be useful for someone else.

[cipher]

very helpful
thanks
cipher

[MooooMoo]

Cipher,

Could you post the 1-50G file when it gets to 500T? I'm going to make it available for distributed sieving then, since the first tests for n=500,000 will be made available once the n=333,333 primes start dropping off the top-5000 list.

This is because TPS and PG currently have over 200 primes in the top-5000 list, and we're currently in the #3 spot of all projects (in terms of number of primes found). Of course, we don't want to slip back and have no top 5000 primes, so testing candidates with an exponent of n=500,000 will prevent this from happening.

Also, the n=500,000 tests will be reserved manually on this forum (because n=333333 remains the main effort), until a twin is found for n=333,333. After that, n=500,000 becomes the main effort, and all of the tests will be made available at PrimeGrid.

[cipher]

Allright MooooMOO you got it i will post it when it reaches 500T.
cipher

P.S: To anyone who can answer this i checked the primes.utm.edu page for any prime with n=500000 i found

  883f 4468523*2^500000-1               150522 L81  2006 
  884f 3868043*2^500000-1               150522 L81  2006 
  885f 3809145*2^500000-1               150522 L81  2006 
  886  3404927*2^500000-1               150522 L81  2006 
  887  2500725*2^500000-1               150522 L81  2006 
  888  1972625*2^500000-1               150522 L81  2006 
  889  1627497*2^500000-1               150522 L81  2006 
  890  1035779*2^500000-1               150522 L81  2006 
  891  995547*2^500000-1                150521 L81  2006 
  892  367899*2^500000-1                150521 L81  2006

Sadly When i check the first 1-5M Range for n=500000 for the k=prime from the sieve file its missing all the Prime k=4468523 and so on... are missing from the sieve file. any one has any idea why, is it missing from the sieve file? I uploaded the 5M.txt the range, so if anyone wants to look and can help answer that.

Thanks
cipher
FYI: i am testing the 5M.txt Range for n=500000 just as a sample.