OK, I'll have it done overnight; looks like about a 20-hr job.

[QUOTE=VBCurtis;545748]OK, I'll have it done overnight; looks like about a 20-hr job.[/QUOTE]

Things seem to be working again now at my end; I'd already started the CADO job before the connection issues arose during poly selection, and most of the clients didn't die, so they started up again and sieving is now about 20% complete. Sorry about this!

I just finished poly select, first sieve WUs hadn't been turned in yet. No problem.
Here's my poly: [code]n: 49665846035497032723349836963420409711194081665832301543290489337278371278618849862398242745980757054061618614684403578901807721147888058272899
skew: 213811.297
c0: -209887345227259003605546539620692
c1: 260454928360311414107555995
c2: 6163263545723237497104
c3: -102846049175697625
c4: 96186887688
c5: 632880
Y0: -3414002759550885303634841155
Y1: 389509549701351678203
# MurphyE (Bf=2.147e+09,Bg=2.147e+09,area=5.369e+13) = 2.537e-06[/code]
I'd like to compare my poly select params to CADO's usual ones- can you post your poly, and I'll compare via cownoise score-evaluator?

I didn't use the default parameters either - mine were a slightly adjusted version of your c140.params from the CADO improved parameters thread, namely
[code]tasks.polyselect.degree = 5
tasks.polyselect.P = 220000
tasks.polyselect.admin = 1080
tasks.polyselect.admax = 8e4
tasks.polyselect.adrange = 600
tasks.polyselect.incr = 60
tasks.polyselect.nq = 15625
tasks.polyselect.nrkeep = 60
tasks.polyselect.ropteffort = 14[/code]

FWIW, here's my poly:
[code]n: 49665846035497032723349836963420409711194081665832301543290489337278371278618849862398242745980757054061618614684403578901807721147888058272899
skew: 924049.234
c0: -36772090407503830677708783005743168
c1: 157617098726207046453463977084
c2: 172679540766368633687860
c3: -395687871979779177
c4: -219745135620
c5: 64800
Y0: -5371503940098331992387869877
Y1: 211767968130007793723
# MurphyE (Bf=2.147e+09,Bg=1.074e+09,area=1.342e+14) = 1.219e-06
# f(x) = 64800*x^5-219745135620*x^4-395687871979779177*x^3+172679540766368633687860*x^2+157617098726207046453463977084*x-36772090407503830677708783005743168
# g(x) = 211767968130007793723*x-5371503940098331992387869877[/code]

Yours is substantially better - what were your parameters?

On to a c166 at index 2214. I'm running t45 now so best for anyone running ECM to start at least at the 50 digit level.

Looks like someone gave us a helping hand... ryanp, is that you?

[QUOTE=charybdis;545771]Yours is substantially better - what were your parameters?[/QUOTE]

Cownoise says yours is skew 1128723.48744 score 1.38713265e-11.

It rates mine as skew 286670.64786 score 7.28030446e-12.

I'm not surprised yours is better; I'm surprised it's twice as good! Yours is similar to the record for C145, while mine is just bad.

Here were my params:
[code]tasks.polyselect.degree = 5
tasks.polyselect.P = 280000
tasks.polyselect.admin = 1920
tasks.polyselect.admax = 2e5
tasks.polyselect.adrange = 960
tasks.polyselect.incr = 120
tasks.polyselect.nq = 15625
tasks.polyselect.nrkeep = 72
tasks.polyselect.ropteffort = 16
[/code]

[QUOTE=VBCurtis;545806]Cownoise says yours is skew 1128723.48744 score 1.38713265e-11.

It rates mine as skew 286670.64786 score 7.28030446e-12.

I'm not surprised yours is better; I'm surprised it's twice as good! Yours is similar to the record for C145, while mine is just bad.

Here were my params:
[code]tasks.polyselect.degree = 5
tasks.polyselect.P = 280000
tasks.polyselect.admin = 1920
tasks.polyselect.admax = 2e5
tasks.polyselect.adrange = 960
tasks.polyselect.incr = 120
tasks.polyselect.nq = 15625
tasks.polyselect.nrkeep = 72
tasks.polyselect.ropteffort = 16
[/code][/QUOTE]

I'm seeing skew 279762.81487 score 1.66392900e-11 for yours so I'm guessing you made a typo.

The main difference between our parameters seems to be the larger incr and correspondingly larger admax that you used.

The c132 is all yours if you want it btw.

My machines are busy tonight, I wouldn't finish it before going to sleep.

Thanks for catching the score typo- I checked a couple lines, and then just decided I had a bad poly. I've been testing incr of 120 vs 144 vs 210 for 140-150 digit params; nice to see there's improvement possible.

Someone cracked it, and pushed us to i2218- a C197.
I'm running curves at B1=8e7 now.

Already done 6000 curves at 15e7, with another 9000 to come. Whoever did the last few terms might well have done even more than that.