Best I've got so far for the C198
 

[CODE]polynomial selection complete
R0: -546268140455137701406805047345800037410
R1: 226452688334889799259
A0: 10794059562734392601889435593289839868031860972670847
A1: 501360724513596634404276748648407498857738
A2: -8581602603670133864136149879844821
A3: -91539705482567385226172
A4: 856482044142084
A5: 3024
skew 3700513720.52, size 1.518e-019, alpha -8.809, combined = 5.253e-015 rroots = 5
elapsed time 01:44:59[/CODE]

This covers an A5 search from 1 to 5000 or so and 2,400,000,000 to the max (2,405,798,807). What's interesting is that this top score is from an A5 that's not in the top 200 in terms of lowest stage 2 norm. The best from the top 200 is:
[CODE]polynomial selection complete
R0: -489092685993814624804817273139280299739
R1: 156076657625511527201
A0: -600520117635194113644227818669708485347032533709200
A1: -1602153780683263041815274765631382853486690
A2: -429595505422610212930558861226327
A3: 1378544362089548248638725
A4: 34178249697051
A5: 5256
skew 3021795735.94, size 1.382e-019, alpha -7.620, combined = 4.964e-015 rroots = 1
elapsed time 00:58:31[/CODE]

In the ranges I searched, I only ended up with two polys with a stage 2 of less than 1e26 and 64 that were less than 5e26.

Another search
 

I did another short run (~8 hours) starting with A5=100M and running to about 100,250,000. With stage 2 norm = 1e27 and stage 1 norm at default, I got only 11 hits total, and none were less than 1e26. However, a good number of the polynomials produced had e-scores greater than 5e-15. I'm including them for posterity:

[CODE]# norm 1.893964e-019 alpha -7.992029 e 5.299e-015 rroots 3
skew: 89402034.93
c0: 123921746223924742275060670812927998929662169485
c1: 17049799413036715161526011828078477731718
c2: 506769018326770074777877894993768
c3: 2180947550527859973509702
c4: -92022227268077973
c5: 100083060
Y0: -68147530827915332123779184895022449562
Y1: 8021697733956300869[/CODE]
[CODE]# norm 1.733502e-019 alpha -7.521848 e 5.040e-015 rroots 5
skew: 83571097.15
c0: -218683070026079709449356849314758577676966877763
c1: 5019551355736941177070000939047116816526
c2: 507357161658492031279123470665104
c3: -2100108213622978488866554
c4: -86005449880287573
c5: 100083060
Y0: -68147530827818882695599515402495568970
Y1: 8021697733956300869[/CODE]
[CODE]# norm 1.761543e-019 alpha -7.505234 e 5.050e-015 rroots 5
skew: 89109036.38
c0: -621726218668799833834531533416803737924109215355
c1: 7074706138633783494040330574733207905998
c2: 492243732462663421135839319024168
c3: -2802673805276135374619818
c4: -84977350649759973
c5: 100083060
Y0: -68147530827802402181778684054010615422
Y1: 8021697733956300869[/CODE]
[CODE]# norm 1.762723e-019 alpha -7.543442 e 5.091e-015 rroots 5
skew: 79844182.98
c0: -98296066074080836043452683862674060301149840903
c1: 4328182768003274669074665030106384234166
c2: 511394561215124052945774663413884
c3: -1866197593819488356358274
c4: -86345026698714573
c5: 100083060
Y0: -68147530827824326139464800808702263680
Y1: 8021697733956300869[/CODE]
[CODE]# norm 1.794033e-019 alpha -7.614441 e 5.139e-015 rroots 5
skew: 87356576.68
c0: -426749618737144673655057628306678295383995918768
c1: 4863566345882482280029346636110012195280
c2: 522553555516789088839918625127772
c3: -946447725278220200821126
c4: -87667508740680273
c5: 100083060
Y0: -68147530827845525633563470767993529941
Y1: 8021697733956300869[/CODE]
[CODE]# norm 1.763735e-019 alpha -7.571006 e 5.089e-015 rroots 5
skew: 85749920.25
c0: -326628062934777425719913708813578605737687359935
c1: 4683995311989136393957419701630738137338
c2: 524779251929316553594355748241336
c3: -611659428401645465195410
c4: -88143965657362173
c5: 100083060
Y0: -68147530827853163276475018443045824828
Y1: 8021697733956300869[/CODE]

Note that these are all from the same A5 (100083060). It also produced a bucketload of polys with scores of 4.5e-15 and up.

Can anyone help with poly selection with C165 from aliquot sequence 842592? Number is: [CODE]698533437875873623788818810532787451389939961713010404188996512533029273790663622250360518247944147564818522290209195447042954760855419302843136771291636186922582539[/CODE]

My best polys for the C198 so far:

[CODE]
# norm 2.949984e-19 alpha -8.495928 e 7.154287e-15 rroots 3
skew: 575570103.95
c0: 2910081215264507857915601041296371032525882247040
c1: -50464344069893538077082133289535935614128
c2: -151991879188447596873425329589885
c3: 165695586357399401337154
c4: -1327973791296464
c5: 240240
Y0: -227715805955647977061040399046632992289
Y1: 210854695525914833209
# norm 2.955399e-19 alpha -8.252221 e 7.152340e-15 rroots 3
skew: 451285461.55
c0: 1898416102802994181603598887301086712401782057837
c1: -8278861461548198292577984515874466835078
c2: -144468718955662954791865054393193
c3: -32287165843712910497182
c4: -1282421907719264
c5: 240240
Y0: -227715805947651949303545871734063125260
Y1: 210854695525914833209
[/CODE]The poly that sieves best depends on the range you plan to sieve. The second poly is probably the best sieving.

All polys I have with a score above 6.6e-15:

Thanks very much for those polynomials. Obviously I don't know the target range until I've done some trial sieving to determine the yield curve, but I'm running that now. Aiming for a round billion relations (will be using 3 large primes on the algebraic side - that gives a clear 15% speed-up)

We're in an area where the skew is the dominant factor for sieving time - at least, my initial results are giving an ordering by time-per-relation almost identical to the ordering by skew. I'll have more solid figures tomorrow morning.

gimarel's third-best polynomial is the one that sieves fastest for me. I still have a week or so of GPU polsel left to run, and the 15e queue isn't close to being empty, so I will let that finish, see if anything superb pops out, and then add the number. Sieving will take probably six weeks flat-out.

Anyone like to have a go at finding a poly for this C156 from Aliquot sequence 829332:3605?

[CODE]163727926668678035939618053446364480300300719840063508810834694701350128493871168693101567743639977109458829559656666787504985092716814050668990858237669441[/CODE]

I have a 6.5 day CPU poly select running, but any assistance at finding a better poly would be most appreciated!

Sure, I'll have a go starting at A1=10k.

[QUOTE=richs;425027]Anyone like to have a go at finding a poly for this C156 from Aliquot sequence 829332:3605?[/QUOTE]

I ran overnight, coeff 10k to 35k. Best poly so far is 2.62e-12, with another in the 2.50's and a bunch in the high 2.40's. I'll run it until coeff = 200k.

Thanks! My CPU search is still running....

[QUOTE=VBCurtis;425128]I ran overnight, coeff 10k to 35k. Best poly so far is 2.62e-12, with another in the 2.50's and a bunch in the high 2.40's. I'll run it until coeff = 200k.[/QUOTE]

Day 2, up to coeff 84k. Best score 2.765e-12, second best 2.72. I wouldn't bother with your CPU search unless it has already turned up a 2.50 or better! I'll get the 3 best polys so far posted tonight, so you can start test-sieving.

Still planning 2ish more GPU-days to reach coeff 200k.