mersenneforum.org Cullen-Williams primes and Woodall-Williams primes
 Register FAQ Search Today's Posts Mark Forums Read

 2020-10-24, 16:28 #1 sweety439     "99(4^34019)99 palind" Nov 2016 (P^81993)SZ base 36 C2F16 Posts Cullen-Williams primes and Woodall-Williams primes The Cullen-Williams number base b is (b-1)*b^(b-1)+1, which is both Cullen number base b (n*b^n+1, some author requires n>=b-1, and for this number n is exactly b-1) and 2nd Williams number base b ((b-1)*b^n+1) The Woodall-Williams number base b is (b-1)*b^(b-1)-1, which is both Woodall number base b (n*b^n-1, some author requires n>=b-1, and for this number n is exactly b-1) and 1st Williams number base b ((b-1)*b^n-1) The Cullen-Williams number base b, (b-1)*b^(b-1)+1 is prime for b = 2, 3, 4, 10, 11, 15, 34, 37, ... (they are exactly the smallest Cullen prime base b for b = 2, 3, 11, 37, and they are exactly the smallest 2nd Williams prime base b for b = 2 and 11) The Woodall-Williams number base b, (b-1)*b^(b-1)-1 is prime for 3, 4, 8, 15, 44, 82, ... (they are exactly the smallest Woodall prime base b for b = 82, and they are exactly the smallest 2nd Williams prime base b for b = 15 and 82) What are the next Cullen-Williams prime and the next Woodall-Williams prime?
 2020-10-24, 19:24 #2 Dylan14     "Dylan" Mar 2017 2×293 Posts Do you have search limits for these forms?
 2020-10-25, 20:23 #3 rogue     "Mark" Apr 2003 Between here and the 23·809 Posts Must not be too deeply searched. A pfgw script to b = 1000 yields the PRPs (944-1)*944^(944-1)-1 and (1622-1)*1622^(1622-1)-1 Here is the script. Use -f to trial factor before PRP testing. ABC2 ($a-1)*$a^($a-1)+1 | ($a-1)*$a^($a-1)-1 a: from 1 to Running to a higher value to see if anything else shows up. Last fiddled with by rogue on 2020-10-25 at 20:24
 2020-10-25, 23:35 #4 Batalov     "Serge" Mar 2008 Phi(4,2^7658614+1)/2 963310 Posts Faster still is to use plain ABC. ABC2 ($a-1)*$a^($a-1)+1 uses generic FFT. Instead, run something like this: cat > a1.abc ABC$a*$b^$c$d ^D seq 1 20000 | awk '{print$1-1,$1,$1-1,"+1"}' >> a1.abc pfgw64 -N -f -l a1.abc OEIS: 271718 , 191568
 2020-10-26, 12:03 #5 rogue     "Mark" Apr 2003 Between here and the 23·809 Posts I stopped searching at b=12000 and am stopping. Someone else can take it further. There *might* be value in someone using sr1sieve with a script to find factors rather than using pfgw to find factors. Last fiddled with by rogue on 2020-10-26 at 12:04
 2020-10-28, 00:36 #6 Trilo     "W. Byerly" Aug 2013 81*2^2797443-1 5×23 Posts Continuing Woodall-Williams from b=12000.
 2021-12-01, 14:19 #7 Trilo     "W. Byerly" Aug 2013 81*2^2797443-1 1638 Posts Woodall-Williams Series is now to b=100000, no new primes. I wrote a (slow) python sieve for "Generalized Woodall-Williams/Cullen-Williams" numbers of the form (b+x)*b^(b+y) +/- 1 for constant x, y (woodall- williams series has x, y = -1.) using fbncsieve. If there is any interest I'll release the source code. Seached (b-1)*b^(b+1) +/- 1 both to 20000: (b-1)*b^(b+1) - 1 is prime for b= 1, 5, 18, 6073 (b-1)*b^(b+1) + 1 is prime for b= 2, 4. Who will be the first to find a number of this form large enough for the top 5000 list?
 2021-12-01, 17:39 #8 kar_bon     Mar 2006 Germany 1011011100002 Posts 17*18^19-1 is not prime, so your b is 19 not 18. Same for b=5 and b=2 is also a prime for the first form. So: (b-1)*b^(b+1) - 1 is prime for b= 1, 2, 6, 19, 6073. and (b-1)*b^(b+1) + 1 is prime for b= 3, 5. Last fiddled with by kar_bon on 2021-12-01 at 17:47 Reason: others
2021-12-02, 18:27   #9
rogue

"Mark"
Apr 2003
Between here and the

11001010010002 Posts

Quote:
 Originally Posted by Trilo Woodall-Williams Series is now to b=100000, no new primes. I wrote a (slow) python sieve for "Generalized Woodall-Williams/Cullen-Williams" numbers of the form (b+x)*b^(b+y) +/- 1 for constant x, y (woodall- williams series has x, y = -1.) using fbncsieve. If there is any interest I'll release the source code. Seached (b-1)*b^(b+1) +/- 1 both to 20000: (b-1)*b^(b+1) - 1 is prime for b= 1, 5, 18, 6073 (b-1)*b^(b+1) + 1 is prime for b= 2, 4. Who will be the first to find a number of this form large enough for the top 5000 list?
Are you continuing to work on this? Can you share your script?

 Similar Threads Thread Thread Starter Forum Replies Last Post rogue And now for something completely different 40 2020-12-08 13:21 Batalov And now for something completely different 15 2019-11-27 15:11 geoff Open Projects 55 2019-05-11 23:30 Citrix And now for something completely different 1 2017-10-26 09:12 jasong Marin's Mersenne-aries 1 2007-11-18 23:17

All times are UTC. The time now is 14:20.

Thu Dec 9 14:20:46 UTC 2021 up 139 days, 8:49, 0 users, load averages: 1.75, 1.57, 1.43