GMPECM-7.1 submissions not working

[fivemack]

I tried submitting

Using B1=100000000, B2=776268975310, polynomial Dickson(30), sigma=1:2078362274
Step 1 took 408021ms
Step 2 took 1689ms
********** Factor found in step 2: 46533631553296182417142192803839571246952859587
Found prime factor of 47 digits: 46533631553296182417142192803839571246952859587
Prime cofactor 3364027155274053890624820701673288731165122378917012485751019899309153526211319048970638633629401345631554593604272199 has 118 digits

to factordb, and it told me that sigma=2078362274 did not produce a relevantly smooth curve order.

[WraithX]

That's because the "parameter" is not 0. Your sigma is 1:2078362274, which has param=1. The different param values make the ecm program use different elliptic curve equations. Since they are different equations, they will have different group orders. You can get more details in section 6 of the README included with the gmp-ecm sources.

Factordb is only setup to recognize the original ecm elliptic curve equations, ie when param=0. Syd will have to make changes to recognize/accept the different param values.

[SethTro]

Thanks for the pointer.

This is for moving the wavefront of Home Primes up to ~C160 from ~C147, something you might be interested in.

---

Ideally I'd like to be able to submit ECM info to FactorDB,

Is there data on the timing impact / success when using of -param 0 vs -param 1?
Is leaving it unset considered best practice? or am I fine to set it to `-param 0` so I can submit details to FactorDB?

Thanks for all the continued advice?

[kosta]

I've been submitting gmpecm7 findings anyway, by faking B1 and B2 values.