20101121, 01:14  #12  
"Mark"
Apr 2003
Between here and the
2^{2}×1,607 Posts 
Quote:
p(4347)# % 1006729441 = (41579# % 1006729441) = 31729 41579# % 31729 != 0 because 1006729441 % 31729 = 0. In fact 1006729441 = 31729 ^ 2. I have to add a check to make sure res % p(k) > 0 because I see other cases where res is a multiple of p(k). 

20101121, 02:10  #13 
Jun 2003
5145_{10} Posts 
First off, I'm assuming that where you said "41579# % 31729 != 0", you meant "= 0". I think it might be easier if you don't sieve with semiprimes having factors < p(max).

20101121, 04:06  #14 
"Mark"
Apr 2003
Between here and the
2^{2}×1,607 Posts 

20101121, 13:25  #15 
Account Deleted
"Tim Sorbera"
Aug 2006
San Antonio, TX USA
10254_{8} Posts 
35303549 complete, and all k's had a prime. Here are the primes:
Code:
p(3530)#*p(46)+1 p(3531)#*p(664)1 p(3532)#*p(44)1 p(3533)#*p(207)+1 p(3534)#*p(1708)+1 p(3535)#*p(1)+1 p(3536)#*p(717)1 p(3537)#*p(1077)+1 p(3538)#*p(154)+1 p(3539)#*p(1894)1 p(3540)#*p(67)1 p(3541)#*p(383)+1 p(3542)#*p(3186)+1 p(3543)#*p(816)+1 p(3544)#*p(384)1 p(3545)#*p(1148)1 p(3546)#*p(376)+1 p(3547)#*p(1516)1 p(3548)#*p(860)1 p(3549)#*p(1964)+1 Last fiddled with by MiniGeek on 20101121 at 13:49 
20101121, 14:07  #16 
"Mark"
Apr 2003
Between here and the
2^{2}·1,607 Posts 
After about 10 hours of sieving, the new sieve is already 100x deeper than the old one (1.7e10 vs 1.7e8). The old sieve took about 24 days to get as far as it did. There are 5.75 million candidates left and it is still removing candidates at about 7 per second. There is a long way to go with sieving. The optimal removal rate for the range is about 4 or 5 per minute. I don't know how long it will take to get to that rate.
The only mistake in my sieve is that I could have written to an ABC file and used the number_primes option with PFGW, but didn't. I'm still writing to the k.in and j.in files used by the PFGW script. I'll provide files for the divide side if anyone wants to work on it, but I suggest that everyone holds off until I've sieved deeper. On an ancillary note, although we are trying to disprove the conjecture, should we also try to show that there is a prime of each form for each p(k)? If there is, then that would be a new conjecture. If not (and I suspect not), then that also implies that this conjecture is wrong. 
20101121, 19:48  #17  
"Mark"
Apr 2003
Between here and the
2^{2}·1,607 Posts 
Quote:


20101122, 17:26  #18 
Apr 2004
273_{8} Posts 
p(14)*p(j)+1 is not prime for any j <= 14.

20101122, 17:33  #19 
Apr 2004
11×17 Posts 
p(19)#/p(j)+1 is not prime for any j <= 19

20101122, 17:58  #20 
Apr 2004
11×17 Posts 
p(29)#*p(j)1 is not prime for any j <=29
Last fiddled with by Harvey563 on 20101122 at 18:03 
20101125, 00:48  #21 
"Mark"
Apr 2003
Between here and the
2^{2}×1,607 Posts 
An update
The current sieve is at 1.3e11 with a removal rate of just under 1 per second (about 50 per minute). The removal rate is about 10x faster than what it should be before it's done. There are currently just under 5.3e6 candidates left.
I don't know how much longer sieving will take, but I don't expect a huge percentage of tests to be removed either. I'm going to take 3900 <= k < 4000. 
20101129, 22:02  #22 
"Mark"
Apr 2003
Between here and the
6428_{10} Posts 
Close, but no cigar. I found a k for which there was no j on the divide side that yielded a PRP, but found on on the multiply side after going through 50% of the candidates on that side.

Thread Tools  
Similar Threads  
Thread  Thread Starter  Forum  Replies  Last Post 
What is the problem here?  didgogns  Msieve  1  20161115 03:31 
problem I have  science_man_88  Miscellaneous Math  2  20101010 16:36 
Intel Atom revisited  hj47  Hardware  15  20100708 20:19 
51 problem  Neves  Miscellaneous Math  5  20040210 22:59 
51 problem  Neves  Puzzles  15  20040205 23:11 