What's up with 2233183?

[tha]

Alpertron found a 39 bit factor of 2233183 recently. Wasn't that factor supposed to have been found by the server itself, or how did it work in those days?

[manfred4]

He found 2 64-bit factors and that 39 bit factor was already known before. Nothing special I can see.

[Madpoo]

Originally Posted by manfred4

He found 2 64-bit factors and that 39 bit factor was already known before. Nothing special I can see.

Yeah, it's just a case of the historical records for smaller exponents, and smaller "easy to find" factors, not being recorded at all as far as who/when found them.

[tha]

Originally Posted by manfred4

He found 2 64-bit factors and that 39 bit factor was already known before. Nothing special I can see.

Ah yeah, When I saw the same exponent listed three times at http://www.mersenne.org/report_expon...exp_hi=&full=1 with the same composite factor each time I assumed he found three previous unknown factors.

[alpertron]

The last time my computer found 3 factors in a row using P-1 algorithm was in January 27, when it found the product of 3 prime factors of M2017627.

[Madpoo]

Originally Posted by tha

Ah yeah, When I saw the same exponent listed three times at http://www.mersenne.org/report_expon...exp_hi=&full=1 with the same composite factor each time I assumed he found three previous unknown factors.

I'm not sure why, but in this case he did return 3 results for the same composite factor:
M2233183 has a factor: 163762219018395829535684102277067383089 (P-1, B1=500000, B2=15000000, E=12)

Just like the "history" section shows.

Not sure why it reported it 3 times within milliseconds of each other... bug in the client software?

When factors are checked in, Primenet does indeed verify that it's actually a factor, and also checks if the factor itself is composite (only prime factors are recorded of course). So it makes no sense to check the same thing in multiple times... just weird. And why wouldn't the software check on it's own if it's a composite factor? Doubly weird... maybe just laziness on the part of the software.

[alpertron]

In general, PrimeNet factors the composite number delivered by Prime95 and then it shows the same composite number sent two or three times, depending on the number of prime factors of that composite, so in the results page you can see composite factors in the same second.

In this particular case, there was a timeout when I sent the results, so I sent them again, so the composite number 163762219018395829535684102277067383089 appears thrice instead of twice.

[alpertron]

My computer found three new known prime factors of another Mersenne number in a row: M2293559

[Batalov]

Quoting from memory from a certain classic novel: "The man is dead. Paralysis of the heart. I can tell even without my stethoscope." (O.Bender, not to be confused with the cartoon series)

Without my stethoscope, I'd venture to guess that events like these are well in the Agresti-Coull confidence interval, given the ranges and number of candidates tested.

It would be cool if you'd find another factor for M7508981 or M9100919. That would be cool...

[alpertron]

Originally Posted by Batalov

Quoting from memory from a certain classic novel: "The man is dead. Paralysis of the heart. I can tell even without my stethoscope." (O.Bender, not to be confused with the cartoon series)

Without my stethoscope, I'd venture to guess that events like these are well in the Agresti-Coull confidence interval, given the ranges and number of candidates tested.

It would be cool if you'd find another factor for M7508981 or M9100919. That would be cool...

The event is not so common because all 3 prime factors must be greater than 3*10¹⁸. This lower bound is being improved by TJAOI, so in the future this event will be even more difficult.