I test sieve everything, but it was late last year and I don't remember. If YAFU did suggest 3 ALP I may have changed it back to 2 (and then test sieved). I don't have much experience with using 3 LPs and have no instincts for that method. I'll check YAFU in the morning.


ETA: from personal experience I know YAFU occasionally outputs 3 LPs.

HP2(4496)
 

If there's an available spot, I'd like to request a GNFS C190 to continue Home Primes Base 2 (4496). It's been tested up to a t60 (BIG thanks to yoyo for providing help with that).

The number is:
[CODE]2463736601122447266257298700393603699043981908096866290769923015377025843584630424369597406016232631749751217215397861267748677508553789649900663934956675572965438394247075872260218511247217[/CODE]

[CODE]polynomial selection complete
R0: -6122675429695590794594870091405716508
R1: 811165482525010679
A0: -1094002446677973934687677554322517275899294365
A1: 213239803679241390276248022925209305657
A2: -12858299808518702215380662301317
A3: -131441698551532940592641
A4: 487261347251442
A5: 286344
skew 164406308.24, size 1.321e-018, alpha -7.507, combined = 2.024e-014 rroots = 3[/CODE]

This is the best polynomial I've found so far, though I can continue hunting a bit and see if anything better pops up if the test sieving for it proves unimpressive.

[QUOTE=swellman;425144]I test sieve everything, but it was late last year and I don't remember. If YAFU did suggest 3 ALP I may have changed it back to 2 (and then test sieved). I don't have much experience with using 3 LPs and have no instincts for that method. I'll check YAFU in the morning.


ETA: from personal experience I know YAFU occasionally outputs 3 LPs.[/QUOTE]

My test-sieving result is
[code]
2A2R total yield: 6225, q=240010031 (0.96452 sec/rel)
3A2R total yield: 8688, q=240010031 (0.84621 sec/rel)
2A3R total yield: 8167, q=240010031 (0.86090 sec/rel)
[/code]

so fairly clear preference for 3A

[QUOTE=fivemack;425180]My test-sieving result is
[code]
2A2R total yield: 6225, q=240010031 (0.96452 sec/rel)
3A2R total yield: 8688, q=240010031 (0.84621 sec/rel)
2A3R total yield: 8167, q=240010031 (0.86090 sec/rel)
[/code]

so fairly clear preference for 3A[/QUOTE]

How about 3A3R? Any reason to exclude it?

[QUOTE=axn;425182]How about 3A3R? Any reason to exclude it?[/QUOTE]
It breaks lasieve5. Not sure why.

[QUOTE=axn;425182]How about 3A3R? Any reason to exclude it?[/QUOTE]

I've never found a case in which it's other than incredibly slow (4x worse than the other permutations) - my guess is that raising the mpf threshold to 96 on both sides increases enormously the amount of mostly-unproductive mpqs that runs.

[QUOTE=swellman;423888]Yoyo completed 30k curves at B1=260M on [url=http://www.factordb.com/index.php?id=1100000000661978232]this C181[/url] of aliquot sequence 3366.

[url=http://www.mersenneforum.org/showthread.php?t=18449&page=15]
Several GNFS polys were found using GPU msieve as well as CADO, starting with post 162[/url].[/QUOTE]

Here is a summary of the best polys found by several individuals for aliquot C181_3366_i2124. It is unknown if the same version of msieve was used to generate these polys and their respective Murphy e-scores. As such I have listed them in order of decreasing test sieve performance, but all are close.


[code]
expecting poly E from 7.47e-014 to > 8.59e-014
polynomial selection complete
skew: 27455029.29
c0: -29103488581110291761470265464081501440963200
c1: 2993874746608901791828362969242757760
c2: 755756990971875242506513249346
c3: -2594415869184082734595
c4: -688441371089106
c5: 9815400
Y0: -43609006827127633229968650474586587
Y1: 30695282756017099
# size 1.065e-017, alpha -7.794, combined = 7.362e-014 rroots = 3
elapsed time 19:28:48
[/code]


[code]
# norm 1.275569e-017 alpha -8.987163 e 7.265e-014 rroots 3
skew: 91826954.20
c0: -5010984928556546834369183181893308978133981616
c1: 107493796305280197439323537463109388072
c2: 1449080018344543787393819255354
c3: -64041478374807114794477
c4: -104963846484778
c5: 3723240
Y0: -52938738751154714424688482444638645
Y1: 33993486786774233
[/code]


[code]
# norm 1.255195e-17 alpha -7.418710 e 7.273e-14 rroots 1
skew: 12833137.64
c0: -1651844500202603092273209171676707935633280
c1: -666189676666650339051903353897671332
c2: -22882374219520520514775531444
c3: 19138754816351501187919
c4: 118556687648450
c5: 11772024
Y0: -42052088395839133578660941153950487
Y1: 3988882332181554479
[/code]


[code]
# norm 1.410097e-17 alpha -6.283760 e 7.775168e-14 rroots 5
skew: 55099868.10
c0: -26923371051647154482176838989390697776905419
c1: 270802559180583462999025156019328254
c2: 79868109989680201381603449209
c3: -598976376776587014086
c4: -36162563490522
c5: 13860
Y0: -162033630917183209833724693720279332
Y1: 9661052375909867827
[/code]


As a point of interest, one poly was generated by CADO. Test sieving showed its speed to be less than the above polys (but with better yield), though speed may perform better at larger ranges. I could not recall how to force msieve to provide an e-score on an existing poly and a search of this forum provided no clues.

[code]
skew: 28549120.0
c0: 31695133714159212627102779580357485588240640
c1: -7140263814077019281755963680358035664
c2: -138241522083103402667415875346
c3: 17364522990413160912171
c4: 91780801517494
c5: 209520
Y0: -93609149816794578396382623932473153
Y1: 5902366602845958906530717
[/code]

If you put the CADO poly into a poly file as you quoted it and run [c]yafu "nfs($(cat num))" -ns[/c], it should auto fill some parameters, and I'm hoping as part of that it will also score the poly and print the score. Hopefully. I'm not really sure to be honest.

[QUOTE=swellman;425399]I could not recall how to force msieve to provide an e-score on an existing poly and a search of this forum provided no clues.[/quote]

Just put the polynomial in msieve format into an msieve.fb file and run msieve -nc1 with no msieve.dat file present

[code]skew 28549120.00, size 8.396e-18, alpha -6.740, combined = 6.137e-14 rroots = 3[/code]

I have queued the c5=9815400 polynomial. The 15e queue is probably now a couple of months long (we are getting about 25 million relations a day on the C198, which needs about a billion relations, and the SNFS difficulty-280 is also not going to be straightforward).

Let's see if we can get a few more 18x-digit GNFS jobs ready - I'd say most of them deserve on the order of a month of GPU polynomial selection.

C187_4788_i5243 is a prime candidate, though it probably needs a bit more ECM (ECM complete so far: at least 4K curves at 110e6, though possibly more, and 3680 curves at 260e6).