OFFICIAL "SERVER PROBLEMS" THREAD

[Madpoo]

Originally Posted by Uncwilly

That has been noted before, but no word from the powers that be.

Fixed. James brought it to my attention. The hourly stats that run was throwing an error thanks to a couple of machines (from an anonymous user) that had a rolling average of zero. No idea why. Anyway, it threw some unexpected divide-by-zero errors that are now being avoided.

[LaurV]

Originally Posted by Madpoo

Anyway, it threw some unexpected divide-by-zero errors

They did it on purpose!

[tha]

[Comm thread Jun 4 14:19] Updating computer information on the server
[Comm thread Jun 4 14:19] Sending result to server: UID: -*****-, M5000473 has a factor: 209330976360004991784025594343086488165867134495557474039114517517233285534935270447 (P-1, B1=600000, B2=12000000, E=12)
[Comm thread Jun 4 14:19] 
[Comm thread Jun 4 14:22] CURL library error: Operation timed out after 180001 milliseconds with 0 bytes received
[Comm thread Jun 4 14:22] CURL library error: Operation timed out after 180001 milliseconds with 0 bytes received
[Comm thread Jun 4 14:22] Visit http://mersenneforum.org for help.
[Comm thread Jun 4 14:22] Will try contacting server again in 70 minutes.

I get this result consistently for over a day now. Can the server handle this size a factor?

Also, two dozen other results with much smaller factors reported were acknowledged but then ignored by the server since the factor(s) were already known. How does one get the P-1 factoring new B1/B2 values to be accepted?

[James Heinrich]

Originally Posted by tha

M5000473 has a factor: 209330976360004991784025594343086488165867134495557474039114517517233285534935270447 (P-1, B1=600000, B2=12000000, E=12)
Can the server handle this size a factor?

Interesting question, and the answer may appear to be "no". At least, it appears the server is hanging up on splitting a composite factor of that size (and I'm not entirely sure why, since it's an easy factor to split, being the composite of the 6 known small factors for that exponent).

I'll take a look, and failing everything else will ask George to manually insert your result.

[kruoli]

Originally Posted by tha

M5000473 has a factor: 209330976360004991784025594343086488165867134495557474039114517517233285534935270447

Actually, your factor is highly composite and is the product of the following factors:
7760734097,
33273147343,
998404440127,
236614851647663,
49063162886176471,
69941149552602344167.
All of these factors were known! Maybe the 276 bits are too much for the server.

Ignoring the B-S-extension of P-1, you'll never had found that factor if it wasn't composite, because
\[2 \cdot kp + 1 = 209330976360004991784025594343086488165867134495557474039114517517233285534935270447\]
with
\[k = 3^3 \cdot 7 \cdot 113 \cdot 673 \cdot 1456253911501367260035342432108965873325923379588222150307120692354091,\]
so one would need B1 = 673 and B2 = 1456253911501367260035342432108965873325923379588222150307120692354091 to find that factor.

You could have added the known factors to your worktodo line, e.g. Pminus1=N/A,1,2,exponent,-1,B1,B2,bitsFactored,"knownFactor1,knownFactor2" etc.

[kruoli]

Originally Posted by tha

How does one get the P-1 factoring new B1/B2 values to be accepted?

I'll always submit the JSON manually. It seems to be necessary when any factors were previously known. But I make sure to exclude all known factors for the computation (like above).

[James Heinrich]

Originally Posted by kruoli

I'll always submit the JSON manually. It seems to be necessary when any factors were previously known. But I make sure to exclude all known factors for the computation (like above).

Submitting JSON is good, and preferred. But we're working on fixing the server problem that's causing this particular hangup, so it should be able to accept this result properly in the near future (few hours?).

[kruoli]

My point was that submitting P-1 results via Prime95 on exponents with already known factors will always fail, even if I include all known factors in the worktodo line.

Sending result to server: UID: manoli94/Threadripper-PC, Mxxxxxxx completed P-1, B1=4000000, B2=100000000, E=12, Wh4: XXXXXXXX

PrimeNet error 40: No assignment
P-1 result for Mxxxxxxx was not needed

If I then submit the JSON line manually, it works, but gives no credit (I'm fine with that).

[James Heinrich]

Originally Posted by James Heinrich

it appears the server is hanging up on splitting a composite factor of that size (and I'm not entirely sure why, since it's an easy factor to split, being the composite of the 6 known small factors for that exponent).

Sorry it's taken so long but (after much effort) we have finally figured out what was causing the hangup in splitting large factors, so submitting large composite factors should no longer be any problem. Using your example, you would now get something like this:

Found 1 lines to process.
Splitting composite factor 209330976360004991784025594343086488165867134495557474039114517517233285534935270447 into:
* 49063162886176471
* 33273147343
* 236614851647663
* 69941149552602344167
* 7760734097
* 998404440127
processing: P-1 factor 49063162886176471 for M5000473 (B1=600,000, B2=12,000,000, E=12)
Error code: 40, error text: Factoring result for M5000473 was not needed
processing: P-1 factor 33273147343 for M5000473 (B1=600,000, B2=12,000,000, E=12)
Error code: 40, error text: Factoring result for M5000473 was not needed
processing: P-1 factor 236614851647663 for M5000473 (B1=600,000, B2=12,000,000, E=12)
Error code: 40, error text: Factoring result for M5000473 was not needed
processing: P-1 factor 69941149552602344167 for M5000473 (B1=600,000, B2=12,000,000, E=12)
Error code: 40, error text: Factoring result for M5000473 was not needed
processing: P-1 factor 7760734097 for M5000473 (B1=600,000, B2=12,000,000, E=12)
Error code: 40, error text: Factoring result for M5000473 was not needed
processing: P-1 factor 998404440127 for M5000473 (B1=600,000, B2=12,000,000, E=12)
Error code: 40, error text: Factoring result for M5000473 was not needed

[tha]

Originally Posted by James Heinrich

....we have finally figured out what was causing the hangup in splitting large factors, so submitting large composite factors should no longer be any problem. ...

Fantastic, the server not hanging up on it is great progress. So, what I want now is the server to accept the new B1/B2 values which are part of the submitted results. Can that be done?

Although the two issues are not directly related, the server also ignores the new B1/B2 values if just one small factor was already known, these two issues showed up in the same attempt to report results.

[James Heinrich]

Originally Posted by tha

what I want now is the server to accept the new B1/B2 values which are part of the submitted results. Can that be done?
Although the two issues are not directly related, the server also ignores the new B1/B2 values if just one small factor was already known

I believe the two issue are directly related, but I'll need George's input on how to resolve it. I think it should be a simple fix.