The "Hey YOU" Thread

[wblipp]

This thread is for attempts to contact and/or inform other factordb users. Be aware the other users may not read this forum, or even know of it's existence.

[kar_bon]

To the guy who's doing n^m-1 (thounsands of small composites with n ~ 6- to 15-digit numbers and m ~ 5- to 17-digit exponent number):

Example:
The C79=4545240977...71<79> is the composite part of (30488817170364239412727^5-1)/5796234755344305420995084262210186.

The quotient 5796234755344305420995084262210186 was not fully factored in FactorDB (just done it) so the 30488817170364239412727^5-1 wasn't, too.

The problem:
If the C79 factors to Px * Py the n^5-1 remains not fully factored in the FactorDB and someone (if >= 70 digits) has to do this factorisation although it could have been done if the factors were inserted to the n^5-1 and not to (n^5-1)/x!

This x sometimes contains greater factors which are not known to FactorDB to the originally n^5-1 value.

Please insert all new factors found to the n^5-1 value instead of (n^5-1)/x!


Another Example:
C79 is part of 14103553^17-1, this is a C122 with a C96!
So there's a small chance, that someone is doing the C79 while someone else tries factoring the C96 at the same time!

[chris2be8]

I'm running scripts on two of my older systems to factor small numbers in factordb. They download random_composites.txt and factor everything they can from it (using SNFS if possible). But they just return:
(line from random_composites.txt)=p1*p2*p3 etc.
Is that going to cause this problem? If so what should the script do? I'm factoring hundreds of numbers per day so manual intervention is impossible.

The jobs are usually quick enough that collisions with someone else doing the number at the same time are rare.

Chris

[wblipp]

Originally Posted by kar_bon

To the guy who's doing n^m-1

That's probably the oddperfect project - at least when n and m are both prime. "Usually" (p^q-1), (p^q-1)/(p-1) and (p^q-1)/((p-1)*a) are in synch, all showing the same final composite. Once in synch, they stay in synch with any new factor showing up in all three forms. From time to time I push factors into factordb to increase the likelyhood of synchronization. It sounds like I need to do that for the batch of numbers that we are currently checking to see if additional factors are known.

[wblipp]

To the guy who is entering lots of Homogenous Cunninghams (a^n+b^n, a, b, n small) - are you aware that when n has an odd divisor k, (a^(n/k)+b^(n/k)) is an algebraic divisor of your number?

The small end of the unfactored composites list is currently swamped with your numbers, many of which could be reduced under the automatic 70 digit limit by informing the factordb of these algebraic factors. For example, I just cleared 651^28+127^28 by telling the factordb about 651^4+127^4. You can load thousands of these factors quickly using the Report Factors page with lines like

651^28+127^28 = 651^4+127^4
(651^33+74^33) = (651^11+74^11) (651^3+74^3)

[BudgieJane]

Originally Posted by wblipp

To the guy who is entering lots of Homogenous Cunninghams (a^n+b^n, a, b, n small) - are you aware that when n has an odd divisor k, (a^(n/k)+b^(n/k)) is an algebraic divisor of your number?

Yes, thanks.

The small end of the unfactored composites list is currently swamped with your numbers, many of which could be reduced under the automatic 70 digit limit by informing the factordb of these algebraic factors. For example, I just cleared 651^28+127^28 by telling the factordb about 651^4+127^4. You can load thousands of these factors quickly using the Report Factors page with lines like

651^28+127^28 = 651^4+127^4
(651^33+74^33) = (651^11+74^11) (651^3+74^3)

Sorry about that. I'm working on (a) stopping so many getting on there at any one time, and (b) clearing them up.

Are there any instructions for how to use FactorDB and how FactorDB works? I've had a good look around and can't find any.

[henryzz]

Originally Posted by BudgieJane

Are there any instructions for how to use FactorDB and how FactorDB works? I've had a good look around and can't find any.

Not really. We are available for questions though.

[cubaq]

I am surely newby, and has some simple question:

- what will happen after my number receive C* from factordb.com/search.

[wblipp]

Originally Posted by cubaq

what will happen after my number receive C* from factordb.com/search.

factordb is a resource to collect known factorizations and primality proofs.

If your composite is under 70 digits, the factordb workers will factor it. This usually happens in seconds to minutes.

Above 70 digits it will sit there until somebody factors it. There are people that donate their resources to factor the smallest numbers. This usually factors numbers into the low 90 digits, and has occasionally factored everything into the high 11x level. You can see how this stands by going to the bottom of the status page and clicking on "Distribution of composite numbers."

Beyond that, your composite will sit there until someone gets interested enough to devote the resources to factor it. If you are the only person interested, that mean until you factor it.

[PawnProver44]

Originally Posted by wblipp

factordb is a resource to collect known factorizations and primality proofs.

If your composite is under 70 digits, the factordb workers will factor it. This usually happens in seconds to minutes.

Above 70 digits it will sit there until somebody factors it. There are people that donate their resources to factor the smallest numbers. This usually factors numbers into the low 90 digits, and has occasionally factored everything into the high 11x level. You can see how this stands by going to the bottom of the status page and clicking on "Distribution of composite numbers."

Beyond that, your composite will sit there until someone gets interested enough to devote the resources to factor it. If you are the only person interested, that mean until you factor it.

It is very useful for submitting primes about 5,000 digits. Is there any way i can "Organize a list" of random long primes of requested size?

[wblipp]

Originally Posted by PawnProver44

Is there any way i can "Organize a list" of random long primes of requested size?

You can use this page to get a list of primes starting at any particular size. You will need to do the trimming to exactly that size yourself and supply your own randomization.