[QUOTE=VBCurtis;485651]About 2000 posts' worth, right here in this thread, just waiting for someone to compile it in one big spreadsheet![/QUOTE]

I’ll start cracking on it. In the meantime I’m releasing the composite since it’s undersieve. With Td=100 it’s going to take at least 330 hours.

Taking C207_103xx849_11.

[QUOTE=pinhodecarlos;485658]Taking C207_103xx849_11.[/QUOTE]

Nice to see you on here again Carlos.

Taking C210_14073559_31 please

[QUOTE=Speedy51;485670]Nice to see you on here again Carlos.

Taking C210_14073559_31 please[/QUOTE]

Cheers Jarod but not for long.

[QUOTE=pinhodecarlos;485674]Cheers Jarod but not for long.[/QUOTE]
In that case enjoy yourself while you are here. Hopefully you will be back

[QUOTE=pinhodecarlos;485655]I’ll start cracking on it. In the meantime I’m releasing the composite since it’s undersieve. With Td=100 it’s going to take at least 330 hours.[/QUOTE]

The catch with this is that SNFS difficulty is not obvious, as it scales with the size of the coeffs in the poly, in a way I'm not aware of a clear formula for. If you really do tackle this, please include the largest SNFS poly coeff as one column in the data. Perhaps we can group them by coeff size (say, under 100/100-10k, 10k-1M, etc) to try to draw some conclusions.

For me, the bigger difficulty is that the polys are in one thread, while the matrix stats are in another. If you do tackle this, even for a subset of the data (say, GNFS on 15e), the data set should help us more tightly predict req'd relations.

[QUOTE=VBCurtis;485679]The catch with this is that SNFS difficulty is not obvious, as it scales with the size of the coeffs in the poly, in a way I'm not aware of a clear formula for. If you really do tackle this, please include the largest SNFS poly coeff as one column in the data. Perhaps we can group them by coeff size (say, under 100/100-10k, 10k-1M, etc) to try to draw some conclusions.

For me, the bigger difficulty is that the polys are in one thread, while the matrix stats are in another. If you do tackle this, even for a subset of the data (say, GNFS on 15e), the data set should help us more tightly predict req'd relations.[/QUOTE]

I would suggest that the Murphy E-value reported by msieve is quite a good metric for the difficulty of an SNFS polynomial; and fortunately that shows up in all the msieve log files.

C207_103xx849_11 done.

[CODE]
Thu Apr 19 20:20:59 2018 p55 factor: 4385701910982243445748053671553196903045671888640475391
Thu Apr 19 20:20:59 2018 p69 factor: 241616641062841768092055658091560761410611519688085701304353922761899
Thu Apr 19 20:20:59 2018 p84 factor: 128597668999542368610789545460494496653346003032071849959844631548891311127900241741

[/CODE]

[url]https://pastebin.com/RL00yB8v[/url]

Taking C184_91605427_29.

134^73+73^134 C177 cofactor (corrected run) factored

[code]
prp76 factor: 2383458983078941039108335285610074143703958845222726611972037485537612641639
prp101 factor: 75600103640208057105914290035659904254989498252237849586813581131726051804470711641509238102720936589
[/code]

263.7M raw / 209.2M unique

Taking C214_249541_43