New mersenne factor found, what to do with it?

[mersenneNoob]

I found a new mersenne factor:
17120076576497274889 divides M9999971879
I found it using factor5. How do I send it to the Mersenne.ca server?
If I did not do any miscalculations this is the 6th factor of M9999971879.

[mersenneNoob]

Do I send an email or what?

[axn]

You can try submitting it via https://www.mersenne.ca/bulk-factors.php

But it looks like it is set to accept mfaktc result format.

I'll move this thread to mersenne.ca subforum. Maybe James will see it.

[mersenneNoob]

thanks, axn!

[James Heinrich]

Originally Posted by axn

You can try submitting it via https://www.mersenne.ca/bulk-factors.php

That would be the correct way to submit results. Since mfaktx doesn't go above 2³² exponents the result file should only include factor lines in this format:

UID: mersenneNoob/myComputerName, M9999971879 has a factor: 17120076576497274889

The "UID" section is required, and nofactor results in the >4G range are not accepted.

[mersenneNoob]

Originally Posted by James Heinrich

Since mfaktx doesn't go above 2³² exponents the result file should only include factor lines in this format:

But I used factor5 not mfaktc. Anyway I already submitted it using the line you provided. :)

[kriesel]

Originally Posted by mersenneNoob

But I used factor5 not mfaktc.

Please help with GIMPS wavefront TF with a GPU, or supplement the 100Mdigit TF with a GPU. Trial factoring is not effective use of a CPU when there's a GPU alternative.
If you feel compelled to work above the range of mfaktx (centuries ahead of GIMPS TF need), there's the possibility of mmff for double mersennes on GPU. Or for mersennes above the 2³² of mfaktx, there's Ernst's Mfactor program, several times the speed of factor5.

Please educate yourself on the available resources and run time scaling.

[ET_]

Originally Posted by kriesel

Please help with GIMPS wavefront TF with a GPU, or supplement the 100Mdigit TF with a GPU. Trial factoring is not effective use of a CPU when there's a GPU alternative.
If you feel compelled to work above the range of mfaktx (centuries ahead of GIMPS TF need), there's the possibility of mmff for double mersennes on GPU. Or for mersennes above the 2³² of mfaktx, there's Ernst's Mfactor program, several times the speed of factor5.

Please educate yourself on the available resources and run time scaling.

Is there a version of Mfactor for 64-bit ARM CPUs (like Raspberry)?

[kriesel]

Originally Posted by ET_

Is there a version of Mfactor for 64-bit ARM CPUs (like Raspberry)?

No idea if it will compile for that. It seems possible, since Mlucas is. I imagine an answer could come from

1) inspecting the source code
2) attempting a compile
3) a reply by Ernst


A quick look into source module factor.c for "arm" finds

#if (TRYQ > MAX_TRYQ) && !defined(USE_ARM_V8_SIMD) // Ignore on ARMv8, since no TF support there anyway

[birtwistlecaleb]

Good job! You can say that you can multiply 17120076576497274889 enough times to get M9999971879!

[James Heinrich]

Originally Posted by birtwistlecaleb

Good job! You can now say that you proved that M9999971879 is composite!

Not really, 5 other smaller factors were known for 3 years before this latest factor.