PRP CF type

[ATH]

Why are these old PRP CF results called PRP type 1 instead of type 5 ? Was this before the different types numbers were implemented?

https://mersenne.org/M218549

1 = 64-bit residue of a^(N-1), a traditional Fermat PRP test used by most other programs
5 = 64-bit residue of a^(N*known_factors-1), same as type 1 if there are no known factors

[ewmayer]

That is odd, since the number has known factors - in that case PRP type 1 would seem to only amount to "M(p) is(not) a PRP", rather than saying anything about the status of the cofactor. Per George:

The server has two assignment types: PRP and PRP-DC.

For PRP assignments, the client can do whatever type of PRP test it wants -- preferred is type 5 if there are known factors, type 1 if there no known factors.

For PRP-DC assignments the server tells the client what type of residue is needed to match the first test. I don't think you'll get a type 5 request without known factors, but I could be wrong.

Note that some gpuowl versions produced type-4 residues. If you can't do a type-4 residue it would be best to abort the PRP-DC.

[ATH]

I assume this is just a leftover issue from the first PRP CF tests before types were implemented.
Maybe it could be fixed with a script like:
If "Number of known factors">0 and type=1 then type:=5

But then I wonder why the residues for type 1 and type 5 are different here:
https://mersenne.org/M218527
https://mersenne.org/M219851

[Uncwilly]

I have been doing some PRP on co-factors of late. I would like to get an actual report of user standings like this: https://www.mersenne.org/report_top_500_lld/
Is there a way to do that? And for ECM-F?

[ATH]

Yes:

PRP CF (1 year): https://www.mersenne.org/report_top_...=1&rank_hi=500
PRP CF (lifetime): https://www.mersenne.org/report_top_...1-01&end_date=

PRP CF DC (1 year): https://www.mersenne.org/report_top_...=1&rank_hi=500
PRP CF DC (lifetime): https://www.mersenne.org/report_top_...1-01&end_date=

ECM-F (1 year): https://www.mersenne.org/report_top_...=1&rank_hi=500
ECM-F (lifetime): https://www.mersenne.org/report_top_...1-01&end_date=

[Uncwilly]

I should have tried out the drop down.

[azhad]

Regarding PRP CF, I want to know some details:

M35596081 is one of my assigned numbers in 2007.

Someone found a factor in 2019 - "3952700815793954414143".

Mersenne.ca says "Remaining cofactor is PRP status unknown".

So I wanted to resolve that unknown; find out if the remaining cofactor is PRP or not.

Went to Manual Testing -> Assignments -> Selected First PRP tests on Mersenne cofactors" -> Enter my number only as the range -> and it gives me this assignment as:

PRP=AID_here,1,2,35596081,-1,71,0,"3952700815793954414143"

When I check the website for assignments again the line is changed:
PRP=Above_AID,1,2,35596081,-1,99,0,3,,"3952700815793954414143"

Am I doing the right test for the job (PRP Type 1 or Type 5)? If not, what needs to be changed?

[ATH]

Originally Posted by azhad

When I check the website for assignments again the line is changed:
PRP=Above_AID,1,2,35596081,-1,99,0,3,,"3952700815793954414143"

Am I doing the right test for the job (PRP Type 1 or Type 5)? If not, what needs to be changed?

That 3 after 99,0, is the base 3 of the PRP test. The next number which is not specified would be the type. I believe type 5 is default for PRP-CF tests.

[azhad]

Originally Posted by ATH

That 3 after 99,0, is the base 3 of the PRP test. The next number which is not specified would be the type. I believe type 5 is default for PRP-CF tests.

Okay, I guess I have to trust Prime95 that it is doing a Type 5 test and not a Type 1. (No indication on what type is being worked on).

[azhad]

For the record, Prime95 did select PRP Type 5 in the above scenario.