20050417, 12:32  #56 
"Sander"
Oct 2002
52.345322,5.52471
4A5_{16} Posts 
It seems that the program is not sieveing, but trail factoring to a certain prime, am i right?
I've been playing a bit with SMALL_PRIME and MAX_SMALL_PRIME, which does let the program factor further or less deep, (and spit out more or less possible candidates) but surprisingly, it doesn't really finishes a range faster when i factor less deep. What i want is quickly generate an output file which cancels most candidates by factoring to lets say 1000 or 10000 and continueing with newpgen. This makes it easier to stop/resume, to see how many candidates are removed in respect to prping (important when n gets larger) and it might be faster. 
20050417, 13:12  #57  
Jun 2003
2×2,719 Posts 
Quote:
Quote:
Quote:
Also, if you are searching higher n's, the chance of finding an octo is *really* low! 

20050417, 16:21  #58 
Jun 2003
Suva, Fiji
2^{3}×3×5×17 Posts 
Scarce
So it looks like octos are scarce, but I am wondering if, like twins, they are infinite?
There are a finite number of k to check for each n, and given that each n only multiples the candidates to check by 2, and given the prime occurence rule, 1/ln(x), which has to be scaled by a factor of between 4 and 8 (k.2^n is bigger than 2^nk), then, we should be able to sigma a formula. I am not a mathematician, and I know there are big dangers around playing with inifinity and converging and diverging series, let alone proving anything. If either case you might imagine that there may be dodecaproths (whatever), which are 3 chains, I imagine these might be enormously rare and maybe nonexistent if there are only finite octos. Interesting challenge that to prove that as well! If anyone proves it and writes a paper I want a credit so that I might get an Erdos number!!!! Regards Robert Smith 
20050417, 18:50  #59  
"Sander"
Oct 2002
52.345322,5.52471
29×41 Posts 
Quote:
Quote:
Quote:


20050418, 05:08  #60 
Jun 2003
2×2,719 Posts 
New and improved!
New and improved sieve !
3x to 8x speed up (depending on how deep you sieve). You can sieve specific ranges  start and end values for k. Three executables  fast, medium and deep sieves (p = 10^5, 10^6 and 10^7 resp.) 
20050418, 10:30  #61 
Aug 2004
Melbourne, Australia
98_{16} Posts 
Troubles with the new program.
I've had both success and troubles with the new software. Firstly
22573117995 37 23055820515 37 45547390455 37 66695049135 37 73828108635 37 78745654785 37 108269914095 37 132750210165 37 136640238735 37 were discovered using the new program. However I tested the program on the known n=35 with 231235935 and 422362395 being the only octoproths. Unfortunately it sieved out the latter. (all three versions did it). Furthermore often it seems that terms with small factors (7, 11, 13) are not sieved. For example the following remained. 32803605*2^56+1 = 4884169 * 483961315030365049 32803605*2^561 = 7 * 13 * 25975262110665308033069 32803605*2^(56+1)+1 = 4727497704141086062018561 32803605*2^(56+1)1 = 8220335593 * 575097896023463 2^56+32803605 = 72057594070731541 2^5632803605 = 72057594005124331 2^(56+1)+32803605 = 17147939 * 8404227943 2^(56+1)32803605 = 22878377 * 6299187571 
20050418, 11:27  #62  
Jun 2003
2×2,719 Posts 
Quote:


20050418, 12:09  #63 
Aug 2004
Melbourne, Australia
152_{10} Posts 
More newies...
610118148045 41
802757470515 41 832494696285 41 
20050418, 12:25  #64 
Jun 2003
12476_{8} Posts 
Bug Fix
There was an unnecessary increment of k inside the loop. Fixed it. Can someone test it and confirm ?
@Dougy  You might want to redo the searches done with the last version Last fiddled with by axn on 20050418 at 12:31 
20050418, 14:00  #65 
Jun 2004
2×53 Posts 
Using the new version of axn1's program, I found the following octoproths for k=235 until n=3000000000000 (=3*10^12):
562223326335 235 722744559915 235 926010118305 235 2441346583515 235 I'm looking further on k=235! One question: what's the difference between the 3 sieve executables? I'm currently using octo_deep.exe. Last fiddled with by Templus on 20050418 at 14:04 
20050418, 15:34  #66  
Jun 2003
2×2,719 Posts 
Quote:
PS: The numbers you posted are not octos. They are only prime for 2^n+/k and 2^(n+1)+/k. They are not prime for the other forms, k*2^n+/1 and k*2^(n+1)+/1 Last fiddled with by axn on 20050418 at 15:41 

Thread Tools  
Similar Threads  
Thread  Thread Starter  Forum  Replies  Last Post 
Small Primes for Octoproths <= 155  ValerieVonck  Octoproth Search  100  20070216 23:43 
Found Octoproths  Range Archive  ValerieVonck  Octoproth Search  0  20070214 07:24 
Number of octoproths per n  Greenbank  Octoproth Search  15  20060120 16:29 
Need help with NewPGen(octoproths)  jasong  Software  1  20050510 20:08 