|
2008-09-25, 11:48
|
#12 | |
May 2007
Kansas; USA
101×103 Posts |
Quote:
http://dc.rieselsieve.com/viewtopic....3&t=9&p=44#p44 H, if people are serious about getting RS rolling again, I want no part of "poaching" the project. As I state, the post was intended to ruffle feathers because nothing else was doing the trick. I've had 3 different people contact me about potentially getting it moving again here. After reading the threads there, I can now see why. The admins will not respond in a timely manner to their posts about a status of the project. Regardless of how state-of-the art that RS was when it started, people deserve better than to be left in the dark like they have after a site crashes suddenly. Like I say there, can we please focus on the situation and not the style or manner of the people involved? Thanks, Gary |
|
|
|
|
2008-09-25, 15:33
|
#13 |
|
Apprentice Crank
Mar 2006
2×227 Posts |
A long while ago, I participated in an effort to solve the Riesel Problem. We got to reserve individual k's at that time, and the k I got was k=234343. Then Riesel Sieve came along. Nearly all individual searchers, including myself, had to give up our k's because we couldn't keep up with their effort.
When Riesel Sieve came down a few weeks ago, my first thought was "did they find a prime for k=234343 yet? When I saw that they hadn't, and that CRUS might take over, I wanted to reserve it again. So, could I make a reservation for k=234343 if this forum continues where Riesel Sieve left off? Wait. Nevermind. The n's are searched way too high now. I'm already stuck with my two other reserved k's, k=173 and k=313. Even if I release them, each candidate will take at least a few hours to test, and k=234343 will progress verrrrrrry sloooowwwly. Oh well, it seemed like a good idea for a few minutes. Yes, I know that I'm not really adding anything helpful here, but I just wanted to share my thoughts. |
|
|
|
2008-09-25, 16:42
|
#14 |
Sep 2004
2·5·283 Posts |
What we can do at least is to host a llrnet server for RieselSieve, I don't mind to host one. I just want to see the project alive and kicking.
|
|
|
|
|
2008-09-25, 17:05
|
#15 |
"Phil"
Sep 2002
Tracktown, U.S.A.
3·373 Posts |
It does seem a pity that the project has gone down so soon after finding its first megadigit prime. The sieving work done under BOINC was incredible, and it certainly would be a shame to waste that work by not continuing the primality testing. It sounds like the owners of the project have had offers of help to get the server going again so it seems to me that they owe the project contributors a simple yes/no answer as to whether they intend to get the project going again and when. Granted, it is a hobby pursued for fun, but it is also something that many people have invested in, so there is also the aspect of responsibility toward the participants. To me, there would be absolutely no shame in saying "no, I am not able to continue coordinating this project at this time" and turning it over to someone else. On the other hand, not communicating about the present status of the project will create bad feelings about this project in particular, and perhaps distributed computing projects in general.
|
|
|
|
|
2008-09-25, 17:29
|
#16 | |
|
Sep 2004
283010 Posts |
Lee's post from Sep 02 2008:
Quote:
|
|
|
|
|
|
2008-09-25, 19:08
|
#17 | |
|
May 2007
Kansas; USA
242438 Posts |
Quote:
Thank you Phil. VERY VERY well put! May I quote your post here in the dc-rieselsieve forum? That is all that I and I believe others are asking for. For the owners/admins of the project to step forth and say: "Folks, I'm sorry that we don't have the time nor the resources to continue the project.". If they do that, I believe that many people would have much greater respect for them bowing out of the prime search world in at least a semi-graceful manner. Thanks, Gary |
|
|
|
|
|
2008-09-25, 19:26
|
#18 | |
|
May 2007
Kansas; USA
101×103 Posts |
Quote:
I read this before posting my 'ultimatum style' post and is exactly why I posted it in the style that I did. Saying that it is 'laziness' that is causing one to not restart a monsterous project and then not following up with people in over 3 weeks is simply inexcusable. At least in my opinion, it is. Perhaps others feel differently. (lol) As Phil so eloquently alluded to: It may be a hobby but it is one that many people have invested much time, effort, and money into and people deserve better than this kind of response. BTW, yesterday KEP forwarded me their fairly recent sieved file. Rest assured that all of their sieving effort will not go to waste! Let me say this also: I have no intention of doing anything on this only at CRUS. RS is too big for the relatively new effort here and we don't have the infrastructure set up to handle it quickly. Sure we could set up and run the servers but without proper automated stats, we wouldn't get near the volume of searching that we would need to make any measureable progress. I would want to get PrimeGrid involved. Rytis is likely more 'politically correct' than yours truly  and so may give RS a couple of more months. My main thinking here is that we push the former RS admins to get this thing going or to turn it over to others much more quickly than they appear to be currently doing. Than we can help coordinate the transition efforts and determine an action plan for the future. If that involves running some servers here and doing some manual stats updates, that is what we will do.My main thinking on prime search projects: Set goals and time lines for completing things and you will have people going after them. Without them, projects can go off in all different directions. Gary |
|
|
|
|
|
2008-09-29, 05:18
|
#19 | |
|
Apprentice Crank
Mar 2006
2×227 Posts |
Quote:
Anyway, why do you and hhh believe that the Riesel conjecture won't be proven in our lifetimes? IMO, the only way we won't live to see the Riesel conjecture proven is if: 1.) One of those k's < 509203 really has no prime, no matter how high of an n you test. 2.) Civilization is destroyed through an asteroid impact, nuclear war, global pandemic, etc before 2100. 3.) Very few or no people are interested in trying to prove the conjecture anymore. 4.) Moore's law grinds to a complete halt tomorrow, and computer speed never increases again. 5.) The average life expectancy drops to 45. I don't have a PHD in math or anything, but I'm pretty sure the chance of #1 being true is under 10%. The odds of #2-5 happening are also quite low. |
|
|
|
|
|
2008-09-30, 06:09
|
#20 | |
|
May 2007
Kansas; USA
242438 Posts |
Quote:
To determine how high the highest prime is likely to be on this conjecture, we need to go back to the history of the conjecture and see what percentage of k's were eliminated with each doubling of the n-value. It's very low because base 2 is not a very prime base like base 3. For historical reference and a personal estimate of how high we will need to test to prove this conjecture, here are the various numbers of k's remaining at each power-of-2 n-range: 2^15; 266 (i.e. at n=2^15, there were 266 k's remaining) 2^16; 204 2^17; 171 2^18; 136 2^19; 111 2^20; 90 2^21; 72 2^22; 64 (not complete yet) Therefore for n=2^15 thru 2^21 (last completed power-of-2 n-range), we can see that on average, there is a 19.57% reduction in k's remaining each time the n-value doubles. Formula: 1-(266/72)^(1/6). Extrapolating we can conclude future estimates of k's remaining at the following n-values: 2^22; 57.9 2^23; 46.6 2^24; 37.5 2^25; 30.1 2^26; 24.2 2^28; 15.7 2^30; 10.1 2^32; 6.6 2^34; 4.2 2^36; 2.7 2^38; 1.78 2^40; 1.15 2^42; 0.74 2^44; 0.48 So, it looks like we will likely have to search to somewhere between n=2^40 and 2^44. This would be somewhere between n=1.1T and 17.6T!! I'm sorry folks, but unless a new method of computing primes is determined, IMHO we just are not going to be finding trillion-bit primes in any of our lifetimes, even if my lifetime is shorter (I'm 47) than many of yours and computing capacity doubles every 18 months and the interest remains high for the effort and an asteriod does not strike us. I'll take it one step further: I think it's barely better than a 50-50 chance that the Sierp conjecture with only 6 k's remaining will be proven in MY remaining lifetime; speculated to be ~35 more years. Does that sound crazy? Using the above calulations with 6 k's remaining at n=~13M (on most k's), SOB will have to search to n=13M*2^12, i.e. n=53G! Now, I know there is some effort underway to look at 1-billion bit candidates but for some reason, something tells me that will be a very long effort! Of course SOB could get lucky and find several primes real quick but that's all it would be; luck. But those are just my opinions. I could be wrong. Gary P.S. If you think this is bad, you should see estimates for base 19 and other very composite bases with high conjectures; try n>10^15 or higher! P.S.S. Technically the above calculations assume that all k-values have the same weight. If one or more of the k's has an extraordinarly low weight vs. the others, the search limit estimates should be higher! Imagine 1 k remaining at n=1 trillion. Now imagine that the 1 k remaining has a weight of only 10! It's likely going to be many doublings of the n-value before a prime is found. Last fiddled with by gd_barnes on 2008-09-30 at 06:15 |
|
|
|
|
|
2008-09-30, 18:38
|
#21 |
Mar 2004
Belgium
292 Posts |
I have found some info about the k status:
http://dc.rieselsieve.com/viewtopic....3&t=9&p=56#p56 I will repost it here: Code:
From http://web.archive.org/web/20061021144805/stats.rieselsieve.com/
Sieve:
Total Ranges Outstanding: 42
Total G Outstanding: 128804
Min G Oustanding: 144325
Max G Outstanding: 417775
Next Range to Begin With: 417775
Most G Outstanding: boinc with 125000
Oldest Range: MCarp:2006-06-11,167046-167071
K Status: not max outstanding k
http://web.archive.org/web/20061021145152/stats.rieselsieve.com/kstats.php
K's :D
K Weight Min N Max N Completed Total Tests in Queue
1 342673 310 2288919 3321615 873 => prime
2 93839 288 2365248 3321528 769
3 40597 225 2289105 3321909 765
4 362609 268 2360688 3321808 763
5 415267 296 2361169 3321809 761
6 319511 280 2365030 3321454 742
7 368411 268 2360522 3321866 702
8 474491 256 2291442 2401458 642
9 365159 256 2365128 3321488 638
10 364903 272 2360615 3321687 617
11 402539 235 2360652 3321764 596
12 26773 218 2360751 3321887 568 => prime
13 315929 209 2360788 3321876 560
14 215443 200 2360811 3321459 560
15 107347 201 2361093 2401749 531
16 65531 203 2365214 2401406 526
17 67117 190 2289925 3321457 526
18 191249 177 2371536 3321840 516
19 384539 183 2360696 3321824 514
20 146561 191 2365138 3321874 508
21 273809 189 2360532 3321876 500
22 371893 166 2360715 3321819 471
23 409753 177 2360675 3321899 470
24 428639 146 2364988 3321652 449
25 363343 158 2290307 3321707 437
26 470173 161 2360823 3321807 431
27 226153 147 2356955 3321803 427
28 344759 158 2360520 3321888 420
29 234343 159 2360591 3321743 394
30 97139 148 2365180 3321868 378
31 121889 151 2365016 3321536 378
32 23669 154 2360052 3321864 373
33 252191 145 2365950 3321870 373
34 46663 118 2360519 2401631 361
35 250027 120 2365273 3321793 351
36 353159 146 2365324 3321484 350
37 206231 130 2361218 3321626 348
38 81041 118 2360710 3321190 343
39 206039 129 2289496 3321160 337
40 325123 129 2365199 3321671 336
41 129007 110 2365073 3321569 334
42 386801 126 2365126 2401438 315
43 38473 105 2361227 3321083 301
44 502573 107 2360651 3321419 297
45 123547 112 2365145 3321737 277
46 327671 101 2245934 3321758 261
47 485557 89 2365125 3321537 249
48 494743 92 2360511 3321711 249
49 74699 91 2360596 3321580 227
50 143047 90 2365021 3321397 224
51 469949 89 2291708 3321788 215
52 397027 83 2354381 2398949 211
53 304207 66 2360645 3321701 180
54 192971 57 2361202 3321658 166
55 477583 60 2361091 3321499 159
56 161669 57 2371736 3321776 153
57 245561 45 2372174 2398526 126
58 31859 172 3321828 121
59 324011 153 3321634 117
60 398023 130 3321531 111
61 162941 123 2291950 3321766 103
K's 1, 12 found prime after 2006-12
external k's
1 2293 202 max0526 3097679 3400751
2 113983 398 maefly 2953575 3321855 => prime
3 336839 390 maefly 2953556 3321836
4 9221 379 maefly 2953518 3321834
5 444637 39 maefly 2954829 2952669
6 141941 439 maefly 2953506 3321894
7 485767 346 maefly 2953717 3321685 => prime
8 342847 38 (inactive) 4077289 4077289
|
|
|
|
|
2008-10-14, 02:51
|
#22 | |
|
May 2007
Kansas; USA
101·103 Posts |
Quote:
Thanks for the info. This is helpful but I need to get a little clarification: Can someone clarify what "Min n" and "Max n" mean here? What we need is the n-value where ALL exponents below that n-value have been tested. That's what I thought "Min n" was. If so, the search limit is much lower on many k's than what we have discussed here so far (n=2.36M vs. n=2.9M-3M). To clarify what I thought further: I thought that "Max n" was the highest n-value handed out for testing but not necessarily completed yet. In other words, there could be many exponents below that value still untested. We need to start at the n-value where there are no untested exponents below it. Gary Last fiddled with by gd_barnes on 2008-10-14 at 02:53 |
|
|
|
|
 |
| Thread Tools | |
Similar Threads
|
||||
| Thread | Thread Starter | Forum | Replies | Last Post |
| Riesel Base 5 LLR | em99010pepe | Sierpinski/Riesel Base 5 | 8 | 2010-06-08 21:21 |
| Discussion about new Riesel base 3 subforum | KEP | Conjectures 'R Us | 20 | 2008-11-15 10:15 |
| Sierpinski/Riesel Base 10 | rogue | Conjectures 'R Us | 11 | 2007-12-17 05:08 |
| Sierpinski / Riesel - Base 23 | michaf | Conjectures 'R Us | 2 | 2007-12-17 05:04 |
| Riesel Base 5 discussion | rogue | Sierpinski/Riesel Base 5 | 2 | 2005-12-21 08:00 |
