mersenneforum.org Software for finding prime constellations
 Register FAQ Search Today's Posts Mark Forums Read

2022-08-28, 20:45   #23
ATH
Einyen

Dec 2003
Denmark

3,413 Posts

Quote:
 Originally Posted by Cybertronic I have run now 32 threads...will see, how many conditions in order are true ... up to 32589158477190044730000 Tooks me 4h, so the offset progress (averaged) is 7e16/s and core
Very nice!

I do not understand how you can make it so fast, there must be some modular trick I am missing.

2022-08-28, 21:09   #24
Cybertronic

Jan 2007
Germany

3·11·17 Posts

Quote:
 Originally Posted by ATH Very nice! I do not understand how you can make it so fast, there must be some modular trick I am missing.

It is the 1st time I offer my code, written in Freebasic for a single run.
I'm shure, it is one of the fastest code.
Attached Files
 19TP2_53.txt (77.0 KB, 42 views)

Last fiddled with by Cybertronic on 2022-08-28 at 21:12

 2022-08-29, 02:24 #25 MattcAnderson     "Matthew Anderson" Dec 2010 Oregon, USA 2×19×31 Posts Admissable prime k-tuple checker Hi all, I found a web-page that seems to be germane to this discussion. Chris Caldwell has his 'prime pages' with University of Tennessee at Martin. see this link. https://primes.utm.edu/glossary/includes/ktuple.php Maybee someone (like me) will find some almost prime constellations an put them on a web-page. By almost prime constellation, I mean not as dense as possible, but a similar pattern. Have a nice day. Matt
 2022-08-29, 07:57 #26 Cybertronic     Jan 2007 Germany 56110 Posts status Searching up to 32589158477190044730000 is done, no 19-tuplet found. Best was one hit with 13 conditions in order for 3718239377799223934593. These numbers have 16/19 are prime. 415755805068375308473 14968783430135985396883 28120524069667173601393 Code: 28120524069667173601393+00 is 3-PRP! (0.000000 seconds) 28120524069667173601393+04 is 3-PRP! (0.000000 seconds) 28120524069667173601393+06 is 3-PRP! (0.000000 seconds) 28120524069667173601393+10 is 3-PRP! (0.000000 seconds) 28120524069667173601393+16 is 3-PRP! (0.000000 seconds) 28120524069667173601393+18 is 3-PRP! (0.000000 seconds) 28120524069667173601393+24 is 3-PRP! (0.000000 seconds) 28120524069667173601393+28 is 3-PRP! (0.000000 seconds) 28120524069667173601393+30 is 3-PRP! (0.000000 seconds) 28120524069667173601393+34 is 3-PRP! (0.000000 seconds) 28120524069667173601393+40 is 3-PRP! (0.000000 seconds) 28120524069667173601393+46 is 3-PRP! (0.000000 seconds) 28120524069667173601393+48 is composite: [3C8425ABFBEF829E] (0.000000 seconds) 28120524069667173601393+54 is 3-PRP! (0.000000 seconds) 28120524069667173601393+58 is composite: [1608EA1FFE278867] (0.000000 seconds) 28120524069667173601393+60 is composite: [246C14CD39E07B1F] (0.000000 seconds) 28120524069667173601393+66 is 3-PRP! (0.000000 seconds) 28120524069667173601393+70 is 3-PRP! (0.000000 seconds) 28120524069667173601393+76 is 3-PRP! (0.000000 seconds) Ratio: 1: 2,97 for 23 digits , should be the target have 27 digits, we have ratio ~ 1:3,5 Estimate for 1st 19-tuplet is round: (3,25^13)/1800 = 2400 cycles. 1 cycle = 4h on 32 threads. 2400 cycles is number 7.8e25 and it is a good deal for the smallest nontrivial 19-tuplet ! With finetuning maybe under a year for me.....but not yet. Last fiddled with by Cybertronic on 2022-08-29 at 08:04
 2022-08-29, 11:22 #27 Cybertronic     Jan 2007 Germany 56110 Posts deep checking brought me 17/19 are true Code: 3525391639439773250083+00 is 3-PRP! (0.000000 seconds) 3525391639439773250083+04 is 3-PRP! (0.000000 seconds) 3525391639439773250083+06 is 3-PRP! (0.000000 seconds) 3525391639439773250083+10 is 3-PRP! (0.000000 seconds) 3525391639439773250083+16 is 3-PRP! (0.000000 seconds) 3525391639439773250083+18 is 3-PRP! (0.000000 seconds) 3525391639439773250083+24 is 3-PRP! (0.000000 seconds) 3525391639439773250083+28 is 3-PRP! (0.000000 seconds) 3525391639439773250083+30 is 3-PRP! (0.000000 seconds) 3525391639439773250083+34 is 41813 · 175727 · 479797020767 3525391639439773250083+40 is 25693 · 137212144920397511 3525391639439773250083+46 is 3-PRP! (0.000000 seconds) 3525391639439773250083+48 is 3-PRP! (0.000000 seconds) 3525391639439773250083+54 is 3-PRP! (0.000000 seconds) 3525391639439773250083+58 is 3-PRP! (0.000000 seconds) 3525391639439773250083+60 is 3-PRP! (0.000000 seconds) 3525391639439773250083+66 is 3-PRP! (0.000000 seconds) 3525391639439773250083+70 is 3-PRP! (0.000000 seconds) 3525391639439773250083+76 is 3-PRP! (0.000000 seconds) 11490035915853116358673+00 is 3-PRP! (0.000000 seconds) 11490035915853116358673+04 is 3-PRP! (0.000000 seconds) 11490035915853116358673+06 is 3-PRP! (0.000000 seconds) 11490035915853116358673+10 is 3-PRP! (0.000000 seconds) 11490035915853116358673+16 is 3-PRP! (0.000000 seconds) 11490035915853116358673+18 is 3-PRP! (0.000000 seconds) 11490035915853116358673+24 is 3-PRP! (0.000000 seconds) 11490035915853116358673+28 is 3-PRP! (0.000000 seconds) 11490035915853116358673+30 is 3-PRP! (0.000000 seconds) 11490035915853116358673+34 is 3-PRP! (0.000000 seconds) 11490035915853116358673+40 is 2190756551 · 5244779896063 11490035915853116358673+46 is 3-PRP! (0.000000 seconds) 11490035915853116358673+48 is 391847 · 9577441 · 3061648823 11490035915853116358673+54 is 3-PRP! (0.000000 seconds) 11490035915853116358673+58 is 3-PRP! (0.000000 seconds) 11490035915853116358673+60 is 3-PRP! (0.000000 seconds) 11490035915853116358673+66 is 3-PRP! (0.000000 seconds) 11490035915853116358673+70 is 3-PRP! (0.000000 seconds) 11490035915853116358673+76 is 3-PRP! (0.000000 seconds) Finetuning brought me a tiny speed up. Run the 2nd cycle now, done in 3,9h..up to number 65178316954380089460000 Collect now numbers with 5 conditions...maybe there is to the end 18/19 ??? Last fiddled with by Cybertronic on 2022-08-29 at 11:55
 2022-08-29, 16:43 #28 Cybertronic     Jan 2007 Germany 3×11×17 Posts Searching up to 65'178'316'954'380'089'460'000 is done. No 19-tuplet with pattern d = 0, 4, 6, 10, 16, 18, 24, 28, 30, 34, 40, 46, 48, 54, 58, 60, 66, 70, 76 available. numbers with 16/19 conditions are: 61168575916820634489163 51376068524333215109383 48190289820404657882893 41319342180915627785893 2nd number with first 13 cond. are prime: 57188204801013850971343 Conclution: 1000 modern CPU cores can found this 19-tuplet in 1 or 2 weeks. Last fiddled with by Cybertronic on 2022-08-29 at 17:01
 2022-08-29, 21:46 #29 Cybertronic     Jan 2007 Germany 23116 Posts There is an open question. The unknown smallest non-trivial prime 19-tuplet have a valid 18-tuplet. I found in the archive: 183837276562811649018077773 + d, d = 0, 4, 6, 10, 16, 18, 24, 28, 30, 34, 40, 46, 48, 54, 58, 60, 66, 70 (27 digits, Apr 2010, Jaroslaw Wroblewski) 51342365971531191697537333 + d, d = 0, 4, 6, 10, 16, 18, 24, 28, 30, 34, 40, 46, 48, 54, 58, 60, 66, 70 (26 digits, Mar 2009, Jaroslaw Wroblewski) 44453357465442632103684223 + d, d = 0, 4, 6, 10, 16, 18, 24, 28, 30, 34, 40, 46, 48, 54, 58, 60, 66, 70 (26 digits, Feb 2009, Jaroslaw Wroblewski) 23531820918273007548405133 + d, d = 0, 4, 6, 10, 16, 18, 24, 28, 30, 34, 40, 46, 48, 54, 58, 60, 66, 70 (26 digits, Jan 2009, Jaroslaw Wroblewski) 11298510058634407483251313 + d, d = 0, 4, 6, 10, 16, 18, 24, 28, 30, 34, 40, 46, 48, 54, 58, 60, 66, 70 (26 digits, Dec 2008, Jaroslaw Wroblewski) 1906230835046648293290043 + d, d = 0, 4, 6, 10, 16, 18, 24, 28, 30, 34, 40, 46, 48, 54, 58, 60, 66, 70 (25 digits, 31 Jan 2001, Jörg Waldvogel & Peter Leikauf) / first non-trivial known to given pattern The first starting point is : 1906230835046648293290043 Question: Was found the other 18-tuplets in logical order or not ?! Last fiddled with by Cybertronic on 2022-08-29 at 21:46
 2022-08-30, 09:45 #30 Cybertronic     Jan 2007 Germany 56110 Posts Got the answer. 1906230835046648293290043 is the new starting point and the other known 18-tuplets was byproducts.
 2022-09-07, 05:58 #31 Cybertronic     Jan 2007 Germany 3×11×17 Posts CRGreathouse look for the special prime 19-tuplet X+d,d=0,8,24,48,80,120,168,224,288,360,440,528,624,728,840,960,1088,1224,1368 For number X=653, we get 18/19 are prime. Best runs up to 2'230'928'700'000'000 are 14/19 X=42584685212933 X=1599836864063783 There are round 142000 possible offset for 23# Last fiddled with by Cybertronic on 2022-09-07 at 06:03
2022-09-07, 17:53   #32
CRGreathouse

Aug 2006

5,987 Posts

Quote:
 Originally Posted by Cybertronic It is the 1st time I offer my code, written in Freebasic for a single run. I'm shure, it is one of the fastest code.
How do I get Big-Int overload.bi?

2022-09-07, 18:27   #33
Cybertronic

Jan 2007
Germany

10618 Posts
Big Int

Quote:
 Originally Posted by CRGreathouse How do I get Big-Int overload.bi?

I have optimized now the special sieve incl deep sieving.

Near 64'696'932'300'000'000 (6.4e16) now.
32 threads can scan 6.7e17 per day.

Ratio for 5e16 ~ 1 : 2,4 [condition to condition]

I believe, 2 weeks you must invest .... ~ 1e19 for the 1st hit, deal ?
You can get the singlecore version with individual sieve-ranges.
Attached Files
 Big-Int overload.txt (28.8 KB, 22 views)

Last fiddled with by Cybertronic on 2022-09-07 at 18:34

 Similar Threads Thread Thread Starter Forum Replies Last Post MattcAnderson MattcAnderson 160 2022-07-18 08:34 robert44444uk Prime Gap Searches 45 2022-02-24 18:28 enzocreti enzocreti 3 2020-02-14 12:21 lukerichards Factoring 87 2019-03-28 13:31 CRGreathouse Software 10 2017-07-14 09:45

All times are UTC. The time now is 18:00.

Tue Nov 29 18:00:26 UTC 2022 up 103 days, 15:28, 0 users, load averages: 1.24, 1.16, 1.02