Polynomial Request Thread

[EdH]

Originally Posted by VBCurtis

I believe the default settings only go to 2-3M; I went for 14M to avoid any duplication.
That said, I used admin 14e6 and admax 15e6; after running all the workunits, it thinks it is 99.6% done and won't proceed to rootopt.
I've tried changing admax to a multiple of 2310, changing admax to far below 15e6 so that it is clearly more than 100% done, and tried setting admin/admax/adrange all to zero to try to force it to finish sizeopt, no joy.

So, I guess I won't be providing a poly for this after all.

I guess I'm running the same region as swellman did, then, but with CADO-NFS instead of msieve. I'll have to see if I happen on the equivalent of his. I'll make some changes with the next run.

I have noticed that both my earlier runs halted like you describe and I remember them doing so in the past and making the adjustments you describe to get them going again.

However, this time I let the server sit and tried to do some studying of the situation, including looking for processes. I have discovered that it appears that the server will not proceed until all WUs that were assigned are completed. It waits for the timedout value and then, it assigns a WU2 to another worker. It then waits for that completion, repeatedly.

I found that if I let the server just sit, it eventually comes back to life and continues on. I don't know how long that break is. (I really should have tried to find that in the log, I suppose - maybe next time.) I've had one sit at 102%. I will try to catch more of this run's activities and see if I can find out anything extra.

[VBCurtis]

Thanks, Ed- your tip worked, and I produced a poly from my search.

Score 5.753e-14, so nothing interesting.

Also, note that CADO and msieve have such different search algorithms that you shouldn't give a second thought to searching the same c5 values with the two programs.

[EdH]

If I've done this correctly and am reading things right, I may be on to something:

n: 26521232090195873108384905824300492852413283081683568418163219479089273132380406501680155963531361683795706304607082425988301635509432877463621844114521741860720947862338201013214619
skew: 54201867.872
c0: 51582596120687700276112275388206350440245099
c1: 11294221610459987563412108443101422017
c2: -117600666539497199147834931355
c3: -6080337773255668538309
c4: 48364989484644
c5: 335160
Y0: -151232835222757496322930715437493282
Y1: 268058264803156563833119
# MurphyE (Bf=4.295e+09,Bg=2.147e+09,area=2.684e+14) = 8.660e-08
# f(x) = 335160*x^5+48364989484644*x^4-6080337773255668538309*x^3-117600666539497199147834931355*x^2+11294221610459987563412108443101422017*x+51582596120687700276112275388206350440245099
# g(x) = 268058264803156563833119*x-151232835222757496322930715437493282

cownoise says:

69686041.76774 8.07782503e-14

[VBCurtis]

That'll do, sir. That'll do.

[swellman]

Originally Posted by EdH

69686041.76774 8.07782503e-14

Nicely done! Waiting for my search to finish root opt but you have nailed it Ed.

I’ll update the best poly table later today.

[EdH]

Thanks! I have mixed emotions - I don't feel I know what I'm doing, but am learning a little bit more. Thanks for inviting me to play.

The range was 1e4-2e6. I'm trying another run at 2e6-4e6, intending to include the machines that were helping with the 198 project, but oddly, many of them are crashing out and refuse to do polyselect. The WUs appear to be taking on extra charactersin their names and then can't be found. But, then again, one of the machines took on extra characters and it's running fine.

[swellman]

C191_M61_k32 from the kosta project also needs a poly. I’ve been searching for some weeks using msieve-GPU but it’s slow going.

27503817001021855161359142123867859536697596095777966914212896065652446437415997206851944428748759308483152015176751478208917632740831052389257949045686226159825984985356043711658640131389761

Currently searching 10M < c5 < 14M


Threshold e-score: 1.78e-14

Objective e-score: > 2.05e-14

Record e-score: 2.028e-14

[EdH]

What would be the e-score to trigger posting for the c191? Objective or threshold, perhaps?

[swellman]

Originally Posted by EdH

What would be the e-score to trigger posting for the c191? Objective or threshold, perhaps?

My best poly found so far is 1.54e-14, which is far from spectacular so I didn’t post it (though I did save it locally).

Anything above threshold is certainly worthy of posting I would say. Maybe even > 1.6e-14.

[EdH]

Here's one above the threshold:

n: 27503817001021855161359142123867859536697596095777966914212896065652446437415997206851944428748759308483152015176751478208917632740831052389257949045686226159825984985356043711658640131389761
skew: 52655514.429
c0: 3894101020403634460611585358413752162507886840
c1: -264898911158239952216506666695139995566
c2: -8576295453521901136589935944553
c3: 78457552052343269630821
c4: 1224390234907308
c5: -14387760
Y0: -5205781762729277297467243067909327153
Y1: 1057936587640601961509
# MurphyE (Bf=8.590e+09,Bg=4.295e+09,area=5.469e+16) = 1.102e-08
# f(x) = -14387760*x^5+1224390234907308*x^4+78457552052343269630821*x^3-8576295453521901136589935944553*x^2-264898911158239952216506666695139995566*x+3894101020403634460611585358413752162507886840
# g(x) = 1057936587640601961509*x-5205781762729277297467243067909327153

cownoise says:

64598312.74029 1.92602657e-14

I'm hoping I've got different enough settings using CADO-NFS that I'm not treading on work that you're doing. If I am (or do), don't hesitate to tell me. I'm running admin=14e6 admax=18e6 overnight. I'm not sure if the other params are CADO-NFS default or not - I've been playing with some of them.

[swellman]

Originally Posted by EdH

I'm hoping I've got different enough settings using CADO-NFS that I'm not treading on work that you're doing. If I am (or do), don't hesitate to tell me. I'm running admin=14e6 admax=18e6 overnight. I'm not sure if the other params are CADO-NFS default or not - I've been playing with some of them.

No worries Ed - I am using msieve-GPU and there’s no real chance of overlap between the two of us.

You’ve already beaten the best poly for a C192, so finding anything better going forward is gravy.