Version of program to calculate factor for b^262144+1

[pepi37]

Hi!
First of all this is only idea, and since I dont know programing, I dont know is it possible or will be usable fast.
So in case of candidates b^262144+1 we all know that factors can be only prime number , multiply of number 524288 with +1
Since many people here develop OpenCL factoring program on GPU , I was wondering is is possible to be written something for this class of numbers.
For example after sieving up to some range it become less efficient and slower then PRP. But still on my case RX580 I need one hour to do PRP on this kind of candidates. So if someone here wrote program that will find factor faster then time range of 60 minutes it will new boost for this kind of candidates.

Of any program that can find factors for this kind of candidates only Prime95 with P-1 or ECM can be used. But P-1 is too slow and found only smooth factors... any other program fails to run it.

Any idea?

[paulunderwood]

Originally Posted by pepi37

Hi!
First of all this is only idea, and since I dont know programing, I dont know is it possible or will be usable fast.
So in case of candidates b^262144+1 we all know that factors can be only prime number , multiply of number 524288 with +1
Since many people here develop OpenCL factoring program on GPU , I was wondering is is possible to be written something for this class of numbers.
For example after sieving up to some range it become less efficient and slower then PRP. But still on my case RX580 I need one hour to do PRP on this kind of candidates. So if someone here wrote program that will find factor faster then time range of 60 minutes it will new boost for this kind of candidates.

Of any program that can find factors for this kind of candidates only Prime95 with P-1 or ECM can be used. But P-1 is too slow and found only smooth factors... any other program fails to run it.

Any idea?

Why don't you ask around at PrimeGrid? They search for these sorts of GFNs on a big scale.

[pepi37]

Originally Posted by paulunderwood

Why don't you ask around at PrimeGrid? They search for these sorts of GFNs on a big scale.

I ask, and got all programs they have. I run it at home and made initial sieve up to 10000P.
But after around 15000P sieve will become slower in removal rate then PRP on GPU.
So I ask in advance ...