mersenneforum.org  

Go Back   mersenneforum.org > Great Internet Mersenne Prime Search > Data

Reply
 
Thread Tools
Old 2020-09-03, 14:10   #3312
S485122
 
S485122's Avatar
 
Sep 2006
Brussels, Belgium

2·7·113 Posts
Default

Quote:
Originally Posted by Prime95 View Post
...
I've been working on clearing out the 93M TF problem. I should have it all cleared by month's end.
The same problem exists for double checking : 18 exponents that where in CAT 1 DC have been handed out for TF "Manual Testing" to relatively slow (judging by the rate at which they complete) workers. CAT 0 and CAT 1 assignments should not be available for manual testing whatever the work type requested, not only unavailable for manual PRP or LL testing but also unavailable for TF, P-1, ECM or whatever.

I am also not sure the recycling rules as stated on the PrimeNet site are enforced. For instance there are a bit less than 60 assignments in CAT 1 or O that have not been reported about for more than 30 days, the rules state that "Assignments are recycled if assignment ... does not report in for 30 days".

More than 400 assignments are older than 120 days while CAT 2 recycling rule states that "Assignments are recycled if ... or when the exponent moves midway into the first category and the assignment is more than 120 days old."

And so on.

Jacob
S485122 is offline   Reply With Quote
Old 2020-09-03, 19:30   #3313
Uncwilly
6809 > 6502
 
Uncwilly's Avatar
 
"""""""""""""""""""
Aug 2003
101×103 Posts

8,731 Posts
Default

This list shows dozens in the DC Cat 1 range with no expiration dates and some with dates beyond the Cat 1 "Must be completed in 60 days" limit.
https://www.mersenne.org/assignments...tf=1&exfirst=1

Last fiddled with by Uncwilly on 2020-09-03 at 19:30
Uncwilly is online now   Reply With Quote
Old 2020-09-11, 02:58   #3314
Uncwilly
6809 > 6502
 
Uncwilly's Avatar
 
"""""""""""""""""""
Aug 2003
101×103 Posts

8,731 Posts
Default

Based upon the numbers highlighted in red, it appears that we have had 10,000 exponents that have been PRP'ed and had their Certs run.

Code:
----------=-----=-- | -----=-----=-----=-----=----- | -----=-----=-----=----- | -----=-----=-----=----- |
  Exponent Range    | Composite  | Status Unproven  |        Assigned         |        Available        |
   Start   Count  P |   F    DC  |LL/PRP  ERR NO-LL |  TF    P-1  LL/PRP  DC  |  TF    P-1  LL/PRP  DC  |
----------=-----=-- | -----=-----=-----=-----=----- | -----=-----=-----=----- | -----=-----=-----=----- |
 
  90000000 54705    | 35760   286 18659             |                 3    34 |                   18622 |
  91000000 54578    | 35525   288 18741     1    23 |                24    40 |                   18701 |
  92000000 54444    | 35501    64 18831          48 |                29    19 |                19 18812 |
  93000000 54423    | 35334   246 18599         244 |               244     6 |                   18582 |
  94000000 54645    | 35664    69 18768     2   142 |               143     2 |                 1 18766 |
  95000000 54452    | 35397    31 18826     4   194 |     1         176     6 |                21 18820 |
  96000000 54364    | 35356    44 18930          34 |                 7     2 |                27 18928 |
  97000000 54431    | 35481   882 17956         112 |                72     7 |                40 17949 |
  98000000 54127    | 35443  3343 13847     4  1490 |   887         442     6 |               165 13837 |
  99000000 54332    | 35257  5093  7529     1  6452 |         156  1107    38 |              5190  7457 |

 100000000 54208    | 34581    21  1898     1 17707 |   714   459    19    21 |       12735  3781  1877 |
 101000000 54316    | 34068    23  3158     5 17062 | 13963         108     3 |        2855   142  3154 |
 102000000 54304    | 34206    43  3186     1 16868 | 15478         118     5 |        1065   209  3180 |
 103000000 54206    | 34035    83  2721     1 17366 | 15712         268    11 |        1170   221  2706 |
 104000000 54207    | 33880   197  1828     2 18300 | 17593         690     4 |          18     3  1822 |
 105000000 54071    | 33891   171  2031     7 17971 | 16423         632     7 |               926  2021 |
 106000000 54061    | 33721    75  2712    10 17543 | 16482         655     9 |               420  2698 |
 107000000 54135    | 34065   739  2438    13 16880 | 14925        1797     7 |          27   151  2423 |
 108000000 54215    | 33554    44   531     2 20084 | 18794   462   475     1 |         239   117   529 |
 109000000 54131    | 33760     4   656     3 19708 | 18233         703     2 |         613   165   651 |
Uncwilly is online now   Reply With Quote
Old 2020-09-13, 23:43   #3315
Uncwilly
6809 > 6502
 
Uncwilly's Avatar
 
"""""""""""""""""""
Aug 2003
101×103 Posts

8,731 Posts
Default

A thread for Curtis Cooper exists: https://www.mersenneforum.org/showthread.php?t=10870
If a user wants to start a Ben Delo thread...

Oops, just did it: https://www.mersenneforum.org/showthread.php?t=25959

Last fiddled with by Uncwilly on 2020-09-13 at 23:46
Uncwilly is online now   Reply With Quote
Old 2020-09-15, 19:21   #3316
ATH
Einyen
 
ATH's Avatar
 
Dec 2003
Denmark

1011100101112 Posts
Default

Now there are TF assignments in the DC Cat 0 and Cat 1 again:

https://www.mersenne.org/assignments...first=1&exp1=1

This should not happen, it only inspires / force us to poach these exponents.
ATH is offline   Reply With Quote
Old 2020-09-15, 19:31   #3317
Uncwilly
6809 > 6502
 
Uncwilly's Avatar
 
"""""""""""""""""""
Aug 2003
101×103 Posts

8,731 Posts
Default

Quote:
Originally Posted by ATH View Post
Now there are TF assignments in the DC Cat 0 and Cat 1 again:

https://www.mersenne.org/assignments...first=1&exp1=1

This should not happen, it only inspires / force us to poach these exponents.
The bounds on this page: https://www.mersenne.org/manual_gpu_assignment/
need to be set to Cat4 to Cat4+10M
Uncwilly is online now   Reply With Quote
Old 2020-09-18, 17:33   #3318
Uncwilly
6809 > 6502
 
Uncwilly's Avatar
 
"""""""""""""""""""
Aug 2003
101×103 Posts

210338 Posts
Default

A nice little milestone happened yesterday. The DC vs LL/PRP ratio in the 99M range passed 1. That is more exponents have had the DC (Cert) done than only a FTC. 98M won't have this happen until the DC wave hits it. 100M and up can have this happen.

Also, early October we should clear the 92M milestone. And 93M looks like it will fall in October. 94M should be done by the end of November. And 95M by the end of the year looks like a strong likelihood.
Yeah, go team!!

Code:
----------=-----=-- | -----=-----=-----=-----=----- | -----=-----=-----=----- | -----=-----=-----=----- |
  Exponent Range    | Composite  | Status Unproven  |        Assigned         |        Available        |
   Start   Count  P |   F    DC  |LL/PRP  ERR NO-LL |  TF    P-1  LL/PRP  DC  |  TF    P-1  LL/PRP  DC  |
----------=-----=-- | -----=-----=-----=-----=----- | -----=-----=-----=----- | -----=-----=-----=----- |
  90000000 54705    | 35760   287 18658             |                 3    34 |                   18621 |
  91000000 54578    | 35525   300 18737          16 |                16    39 |                   18698 |
  92000000 54444    | 35501    69 18840          34 |                34    19 |                   18821 |
  93000000 54423    | 35334   348 18638         103 |               103     6 |                   18613 |
  94000000 54645    | 35664    72 18826          83 |                83     2 |                   18824 |
  95000000 54452    | 35397    40 18878     3   134 |     1         136     6 |                   18872 |
  96000000 54364    | 35356    49 18930          29 |                29     2 |                   18928 |
  97000000 54431    | 35481   891 17968          91 |                91     7 |                   17961 |
  98000000 54127    | 35449  4169 13930     4   575 |    78         393     2 |               108 13924 |
  99000000 54332    | 35260  8051  7660     1  3360 |          86  1295    56 |              1980  7561 |
 100000000 54208    | 34651    21  1900     1 17635 |   100  1047    19    21 |       11318  5152  1879 |
 101000000 54316    | 34082    23  3166     5 17040 | 13521    60    96     3 |        3223   146  3162 |
 102000000 54304    | 34214   127  3193     1 16769 | 15387    40   107     5 |        1025   212  3187 |
 103000000 54206    | 34036    92  2770       17308 | 15681         201     5 |        1201   227  2763 |
 104000000 54207    | 33885   240  1918     2 18162 | 17507         615     2 |          40     4  1913 |
 105000000 54071    | 33919   187  2133     7 17825 | 16128         711     5 |               996  2124 |
 106000000 54061    | 33739   109  2749    10 17454 | 16306         570     9 |               591  2736 |
 107000000 54135    | 34075   779  2544    11 16726 | 14880        1716     9 |               146  2524 |
 108000000 54215    | 33567    53   555     3 20037 | 17910   433   792     1 |         897    10   552 |
 109000000 54131    | 33791     4   712     3 19621 | 18186         617     2 |         647   179   705 |
Uncwilly is online now   Reply With Quote
Old 2020-09-18, 20:52   #3319
kriesel
 
kriesel's Avatar
 
"TF79LL86GIMPS96gpu17"
Mar 2017
US midwest

2×3×13×59 Posts
Default

Quote:
Originally Posted by Uncwilly View Post
A nice little milestone happened yesterday. The DC vs LL/PRP ratio in the 99M range passed 1. That is more exponents have had the DC (Cert) done than only a FTC. 98M won't have this happen until the DC wave hits it. 100M and up can have this happen.

Also, early October we should clear the 92M milestone. And 93M looks like it will fall in October. 94M should be done by the end of November. And 95M by the end of the year looks like a strong likelihood.
98M LLDC was forecast for ~2033 back in late 2018, by interpolation of the simple extrapolation at https://www.mersenneforum.org/showpo...5&postcount=11
Also forecast 94M first test completion at end of 2020. If things go faster than the forecast, so much the better.
kriesel is offline   Reply With Quote
Old 2020-09-19, 20:28   #3320
petrw1
1976 Toyota Corona years forever!
 
petrw1's Avatar
 
"Wayne"
Nov 2006
Saskatchewan, Canada

23·7·79 Posts
Default

Quote:
Originally Posted by kriesel View Post
98M LLDC was forecast for ~2033 back in late 2018, by interpolation of the simple extrapolation at https://www.mersenneforum.org/showpo...5&postcount=11
Also forecast 94M first test completion at end of 2020. If things go faster than the forecast, so much the better.
I can see us finishing all of 92-99 this year ... if those who grab the assignments behave.
petrw1 is offline   Reply With Quote
Old 2020-09-28, 14:30   #3321
kriesel
 
kriesel's Avatar
 
"TF79LL86GIMPS96gpu17"
Mar 2017
US midwest

10001111110102 Posts
Default 5 left to 92M first test

https://www.mersenne.org/assignments...xtf=1&exdchk=1 shows exponents by Friend and Zwen respectively, between 98 & 99% and stalled. An attempt to reach them by PM indicates no forum account for either. They don't appear in top 500 producers in the various categories, so no chance of picking up a usable email address or web link for them there, as has sometimes worked for reaching others in the past.
edit:
Expired LL FT attempts may be better followed up by PRP/GEC/proof.
If the expired first LL finally reports in, it would take another full LL (more in case of a mismatch, due to ~2% error rate per LL test; 1.04 more LL tests on average) to confirm.
If the expired first LL does not finally report in, it would take another 2 or more full LL runs to confirm; ~2.04 on average.
But in either case, a PRP/GEC/proof completed takes only a cert to conclude testing the exponent, ~1.01 tests total. And confirmation occurs in hours typically, not months or years. At high reliability.

Last fiddled with by kriesel on 2020-09-28 at 15:01
kriesel is offline   Reply With Quote
Old 2020-09-28, 14:45   #3322
Uncwilly
6809 > 6502
 
Uncwilly's Avatar
 
"""""""""""""""""""
Aug 2003
101×103 Posts

8,731 Posts
Default

And only 168 (this includes 4 errors that need to be rerun) will get us up to 98M. There are some that obviously won't finish before the expire. They might need a very early (even premature) doublecheck (since these are FT LL's an early LL DC won't waste cycles.).


On a different note: Over the weekend I noticed that Curtis Cooper is now running v30 on at least 1 machine. This bodes well. As does Ryan P hitting the 54M range hard with a bunch of DCs.
Uncwilly is online now   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
Newer X64 build needed Googulator Msieve 73 2020-08-30 07:47
Performance of cuda-ecm on newer hardware? fivemack GMP-ECM 14 2015-02-12 20:10
Cause this don't belong in the milestone thread bcp19 Data 30 2012-09-08 15:09
Newer msieves are slow on Core i7 mklasson Msieve 9 2009-02-18 12:58
Use of large memory pages possible with newer linux kernels Dresdenboy Software 3 2003-12-08 14:47

All times are UTC. The time now is 11:07.

Sun Oct 25 11:07:57 UTC 2020 up 45 days, 8:18, 0 users, load averages: 1.47, 1.46, 1.50

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2020, Jelsoft Enterprises Ltd.

This forum has received and complied with 0 (zero) government requests for information.

Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.
A copy of the license is included in the FAQ.