n=1.7M and n=3.322M Twin-Sophie record attempts

[Happy5214]

Originally Posted by MooMoo2

Based on the 9 hour, 35 minute age of that test, the page was likely archived on Tue December 27 11:27:27 (server time)

The Wayback Machine embeds the archive timestamp in the URL, so https://web.archive.org/web/20221227112659/http://noprimeleftbehind.net:12000/all.html was archived at exactly 2022-12-27T11:26:59Z (ISO 8601 format), or December 27 at 11:26:59 UTC.

[whengryphonsfly]

Originally Posted by MooMoo2

It appears that when the page was archived, the only outstanding candidate was 906063*2^490562-1:

Candidate		User		Client	Team	Date Assigned			Age (hh:mm)
906063*2^490562-1	whengryphonsfly	alpha05		Tue Dec 27 01:52:27 2022	9:35

Based on the 9 hour, 35 minute age of that test, the page was likely archived on Tue December 27 11:27:27 (server time)
The 906063*2^490562-1 test was returned on December 27, 20:18:02:

user=whengryphonsfly
[2022-12-27 20:18:02]
906063*2^490562-1 is not prime.  Res64: FB745BD8E481399A  Time : 0.0 sec.

I looked up the files at: http://www.noprimeleftbehind.net/tps/results/llrnet/ and saw that you submitted 17,870 tests and found 3 primes between 2022-12-27 20:18:02 and when the stats page went down early on 2023-01-03. There were no tests completed by any other users. If we were to use 2022-12-27 11:27:28 instead of 2022-12-27 20:18:02, the number goes up by 1,954 to 19,824 (all completed by you).

The 19,824 additional tests with 3 additional primes is a reasonable assumption. Your totals would then be 89,134 tests with 21 primes found, and all other users would stay the same.

I'm not sure I would trust those timestamps. I would trust the Wayback Machine URL timestamp and the Date Assigned timestamp, but I would not trust the timestamps listed in http://www.noprimeleftbehind.net/tps/results/llrnet/. My stop option at the time was to complete and return in-progress work units, so 9:35 elapsed doesn't make sense - except for the fact that I semi-regularly received malformed work units (as in https://www.mersenneforum.org/showpo...&postcount=242). The problem is if the work unit was assigned at 1:52 UTC and returned at 20:18 UTC, it couldn't have expired and been reassigned in that timeframe.

In fact, I'm pretty sure the results timestamps are local time (UTC-6). Looking at the 1/2 result page of 1/1, one can see that it implies I was up until a little bit past midnight, turned off my computer, and resumed after I woke up - all of which is true. All of the results pages I looked at seem to corroborate this; they only make sense based on my schedule if they are local time and not UTC.

(This does leave the question of the Last Modified timestamp, which reports as the same both in normal and private windows. I therefore assume it's UTC, but then I wonder how it puts my completed work units from before 18:00 on the same page as after 18:00? Some of the early results pages clearly show the timestamp can be something other than 00:01, so maybe it's grouping by the date of the candidate and just not recording the last modified time anymore, only the last modified date? But it also puts the completed work units of the first minute of the next day in the same results page; see the 1/1 report of 12/31 for instance. Ultimately, .)

Assuming the results timestamps are local time, that means all tests from the 12/28 report of 12/27 on were not on the archived page; the page was archived at 5:26 local time on the 27th. This would be 19933 additional tests completed for 3 additional primes, for a total of 89243 tests completed and 21 primes. This does leave the total score a mystery, or it would if it weren't for the fact that it was the one statistic I was tracking very closely; my final score was 19416818.

Ultimately, I think my original point (rebuilding the tables) doesn't matter that much anymore. The point was making sure that everyone who contributed got credit, but I think people will know to check the Wayback Machine and/or the forum if they actually want that information.

[gd_barnes]

I saw that there might have been a teeny bit of interest in having the stats back out there on port 12000.

Behold: The stats are back. :-)

I took the WayBack page that you guys provided and added it to the current stats. Fortunately Whengry was not processing any work at that moment so I was able to add his Wayback and current totals together without losing any tests in the meantime. Kruoli was the only other one processing work at that exact moment but he had no stats prior to when I erroneously wiped them out so I didn't have to touch his totals.

The only issue now is the work that Whengry did between the time that the stats were wiped out and the time of the WayBack page. MooMoo, I see that you did some calculations on what Whengry did during this interval and you were maybe able to determine the exact results that may have been lost. Would it be possible to provide me with a file with those results?

With the missed results in hand, I can easily update the total tests done and primes found. For the score, I should be able to easily enough back into the formula used by the server to score each test to ultimately determine the score of the entire file. I'll then add that score to his current score.

[The Carnivore]

Originally Posted by gd_barnes

Behold: The stats are back. :-)

Yay!

Do you think you can add the n=480K-500K LLRNet stats as well? Those stats are at: https://www.mersenneforum.org/showpo...7&postcount=13 , and if you were to combine them, we'd have a complete stats page for n=480K-500K.

[MooMoo2]

Originally Posted by gd_barnes

I saw that there might have been a teeny bit of interest in having the stats back out there on port 12000.

Behold: The stats are back. :-)

I took the WayBack page that you guys provided and added it to the current stats. Fortunately Whengry was not processing any work at that moment so I was able to add his Wayback and current totals together without losing any tests in the meantime. Kruoli was the only other one processing work at that exact moment but he had no stats prior to when I erroneously wiped them out so I didn't have to touch his totals.

The only issue now is the work that Whengry did between the time that the stats were wiped out and the time of the WayBack page. MooMoo, I see that you did some calculations on what Whengry did during this interval and you were maybe able to determine the exact results that may have been lost. Would it be possible to provide me with a file with those results?

With the missed results in hand, I can easily update the total tests done and primes found. For the score, I should be able to easily enough back into the formula used by the server to score each test to ultimately determine the score of the entire file. I'll then add that score to his current score.

Thanks!

The WayBack page was archived on 12/27 at 11:27 GMT time (5:27 server time), and the server was updated on January 3. Since the first test on 12/27 was after 5:27 server time, the missing results files are at:
http://noprimeleftbehind.net/tps/res..._tps_12000.txt (2129 tests by Whengry)
http://noprimeleftbehind.net/tps/res..._tps_12000.txt (2574 tests by Whengry)
http://noprimeleftbehind.net/tps/res..._tps_12000.txt (3158 tests by Whengry)
http://noprimeleftbehind.net/tps/res..._tps_12000.txt (2945 tests by Whengry)
http://noprimeleftbehind.net/tps/res..._tps_12000.txt (3114 tests by Whengry)
http://noprimeleftbehind.net/tps/res..._tps_12000.txt (2623 tests by Whengry)
http://noprimeleftbehind.net/tps/res..._tps_12000.txt (3390 tests by Whengry)

The total is 19,933 tests and 3 primes by whengryphonsfly. I've attached a combined file for convenience.

Originally Posted by gd_barnes

With a better PRPnet server allowing for the execution of more modern LLR, perhaps it will increase interest here.

It sure has

14,755 tests were done today, which is the highest daily total since late September 2011. Nice job, everyone!

[gd_barnes]

Originally Posted by The Carnivore

Yay!

Do you think you can add the n=480K-500K LLRNet stats as well? Those stats are at: https://www.mersenneforum.org/showpo...7&postcount=13 , and if you were to combine them, we'd have a complete stats page for n=480K-500K.

I'd rather not. The stats on that page are compiled automatically only for the PRPnet server.

Originally Posted by MooMoo2

Thanks!

The WayBack page was archived on 12/27 at 11:27 GMT time (5:27 server time), and the server was updated on January 3. Since the first test on 12/27 was after 5:27 server time, the missing results files are at:
<snip>
The total is 19,933 tests and 3 primes by whengryphonsfly. I've attached a combined file for convenience.

Thanks for all of the detail. It helped tremendously! I have updated the stats to include all of Whengry's results in that time.

The formula used by the PRPnet server for score is:

Length of test ^ 2 / 1e8

Note:
Length of test = int [exponent * log (2) + log (k) + 1]

A quick parsing of the k and n-value for each of the results in your attached file from Whengry came up with a total score of 4336437. That was added to the score that he already had from the WayBack page and current totals.

[gd_barnes]

With the increased interest here, how about we finish n=480K-500K up to k=1M ? That would finish the entire file that was sent to me. If everyone agrees and since we are going through it fairly quickly, I'll load k=920K-960K into the server later today.

[odicin]

Originally Posted by gd_barnes

Good news: Due to the corruption, I dropped the database and created a new one with newer PRPnet version 5.3.2.

Good to hear, so it is much easier for me to spend some cores on it again. Especially the newer AMD processors were disadvantaged because of the missing FMA3 in gnum of the old LLR.

Regards Odi

[MooMoo2]

Originally Posted by gd_barnes

With the increased interest here, how about we finish n=480K-500K up to k=1M ?

That's fine with me. Maybe we'll find a twin between k=920K and k=1M

[The Carnivore]

Originally Posted by gd_barnes

With the increased interest here, how about we finish n=480K-500K up to k=1M ? That would finish the entire file that was sent to me. If everyone agrees and since we are going through it fairly quickly, I'll load k=920K-960K into the server later today.

Even though I voted for moving on to the quad-sieved candidates in the "New Exponents for TPS" poll, I'm not opposed to testing n=480K-500K until k=1M. Best-case scenario, we find a variable n-range twin, worst-case scenario, we get a higher sieve depth for the quad-sieved candidates.

[MooMoo2]

Originally Posted by gd_barnes

I am looking into adding 2 additional PRPnet servers: One here for TPS and one for NPLB in the future. The server machine appears to be more robust than I had previously thought when I said it was mostly maxed out.

Originally Posted by MooMoo2

Thanks, Gary! I've created a poll to determine which candidates we should test once we finish the n=480K-500K candidates to k=920000:
https://www.mersenneforum.org/showthread.php?t=28373

The poll is tied at 5 votes for n=1.7M and 5 votes for n=3.322M. Assuming current trends hold, I'll upload the quad-sieved candidates in early February.

At that time, would you be able to load n=3.322M into port 12000 and n=1.7M into the new server? However, if only port 12000 will be available, please let me know, and I'll try and make a decision as to which n-value we should work on.