Found a factor? Post it here.

[ric]

... and I got this pretty one:

UID: RU/Nomad2C, M12103879 has a factor: 76589819640254233618412004377777346859019502009144302474620595280120616202419586537 (P-1, B1=430000, B2=8500000, E=12)

275 bits, composite (84 + 90 + 103). As a side note, had to resort to manual page submission in order to have it processed (CURL raised a timeout in automatic processing and stopped the queue).

[James Heinrich]

Originally Posted by ric

275 bits, composite (84 + 90 + 103)

That is pretty!

[Jwb52z]

P-1 found a factor in stage #2, B1=715000, B2=13048750.
UID: Jwb52z/Clay, M89067427 has a factor: 524395562032803866876089 (P-1, B1=715000, B2=13048750),

78.795 bits.

[GP2]

Originally Posted by ric

275 bits, composite (84 + 90 + 103). As a side note, had to resort to manual page submission in order to have it processed (CURL raised a timeout in automatic processing and stopped the queue).

So what would have happened if you didn't manually intervene? The factor would have been dropped and lost?

And how often has that actually happened, to other people, and no one the wiser?

[ric]

An interesting observation: here's my thinking

Originally Posted by GP2

The factor would have been dropped and lost?

No.

Originally Posted by GP2

So what would have happened if you didn't manually intervene?

I've opted to use Primenet to report results, so the large factor and all results following that one, they all stacked up into prime.spl (I saw this factor the day after its discovery, so this is a fact) - a message on screen

CURL library error: Operation timed out after 180000 milliseconds with 0 bytes received

and an equivalent entry in prime.log warned that something weird was happening at each subsequent result/communication attempt - in all likelihood, with such visible warnings, even a novice user would have come to mersenneforum to ask for advice.

Had I opted for not communicating directly to Primenet, the first subsequent manual submission would have lasted a bit longer than usual.

Originally Posted by GP2

And how often has that actually happened, to other people, and no one the wiser?

IMAO P(x)->0

Two side notes:

1. I've also pondered about asking to raise the 3 min threshold for CURL communication, but the low occurrence of these large factors IMO does not justify this measure
2. oh, and BTW I noticed that this factor does not appear in the "Recent Cleared" report, but again clipping the outliers to me is a sane measure

TL;DR: I like the way it is, and see no danger in it. Other opinions are welcome

[James Heinrich]

I suspect that the original (automatic) submission triggered an attempt to factor the factor, which is non-trivial, and the communication timed out while waiting for the result of the factor factorization. By the time the manual communication was attempted, the original factor splitting was completed and cached so the manual communication went through immediately.

On a side note, my site now displays composite factors found as the list of the prime components for greater clarity:
https://www.mersenne.ca/exponent/12103879
I also wonder why the original P-1, apparently done nearly 18 years ago, should have but didn't find two of the three factors?

[ric]

Originally Posted by James Heinrich

By the time the manual communication was attempted, the original factor splitting was completed and cached so the manual communication went through immediately.

It's likely, and my educated guess is that the "factor the factor" process was started during my initial manual submission (the large factor + subsequent no factor results), and completed - caching included - by the time I re-submitted the large factor alone (impatient? who, me? ).

This guess of mine is based on the evidence that I have two identical result lines for it in my results report (crediting half a GHz-Day each, go figure!) and on mersenne.ca as well.

Originally Posted by James Heinrich

I also wonder why the original P-1, apparently done nearly 18 years ago, should have but didn't find two of the three factors?

Not the first one of this kind, probably won't even be the last (another recent example: M12076891 - incidentally, it has a glitch right now, preventing display).

I recall a discussion somewhere else on mersenneforum, with the general consensus that re-doing older P-1 was basically futile (and however mersenne.ca maintains this list).

[James Heinrich]

Originally Posted by ric

M12076891 - incidentally, it has a glitch right now, preventing display

Already fixed

[ric]

Originally Posted by James Heinrich

Already fixed

Amazing speed! Thx

[Jwb52z]

P-1 found a factor in stage #2, B1=715000, B2=13227500.
UID: Jwb52z/Clay, M89267203 has a factor: 34165373881617506022060001 (P-1, B1=715000, B2=13227500),

84.821 bits.

[Jwb52z]

P-1 found a factor in stage #1, B1=715000.
UID: Jwb52z/Clay, M89218183 has a factor: 1578944744818761269375521 (P-1, B1=715000),

80.385 bits.