Reserved for MF - Sequence 3366

[VBCurtis]

Yea, my silicon is a bit dated: 2x 10-core Ivy Bridge xeon @2.6ghz, 1x 12-core haswell xeon @2.5ghz.

My choices for B1 across the board are based on the lowest B1 that get k=2 and are bigger than the standard B1 for a T-level:
T45 2e7
T50 6e7
T55 15e7
T60 42e7 or 65e7 (65e7 is k=3, which means B2 jumps by 50% over B1=420M)
Each of these is at least as fast to complete the related T-level as standard B1 choices, with the bonus of contributing a bit more chance to find a larger factor.
Above T60, memory restricts our choices a bit- ekeing out 5% more speed becomes secondary.

[VBCurtis]

4800 curves complete at 42e7.
I'm moving one machine to poly select for this number, while the other 20 cores continue on ECM at this size.

Poly select is on CADO, 5e6 to 5e7 with incr 4620, P=3M, and nq 15625.
A run from ~100k to 5e6 is welcome from any contributor, incr 420 is likely best down there.

[EdH]

I'm just getting my "farm" settled back in from the recent task of upgrading everything. If I can, I'll try your suggested range this evening, if no one else steps up before then.

[Gimarel]

My best polys so far. The skew could be lower.

# norm 5.124552e-19 alpha -7.806861 e 1.035e-14 rroots 3
skew: 229473397.70
c0: -28975681727567524837662736904500125083895064840
c1: 2878364533279412069293098584050068952094
c2: 12328382189869072397837639510417
c3: -134638507067018387900322
c4: -724952625833220
c5: 514800
Y0: -106316329608766223087548646624002108221
Y1: 139198582645969222983109

# norm 5.098112e-19 alpha -8.012851 e 1.030e-14 rroots 5
skew: 125001952.15
c0: -24623140536189314142763591449244095867976653480
c1: 1591604144502435503327811672315382999292
c2: 45151115187254052425427334115817
c3: -227121374773875476420597
c4: -2975093123155512
c5: 6032880
Y0: -67399847496998630851429653894088679427
Y1: 125291089571292601013129

# norm 4.973943e-19 alpha -7.015830 e 1.003e-14 rroots 5
skew: 139217445.66
c0: -7255470118724953101983117852152337087695168335
c1: 149796856433933965973020367862770013258
c2: 16841679768648646355805881003315
c3: 102010909982219747276718
c4: -911111469275220
c5: 514800
Y0: -106316339675995109699546585110006980568
Y1: 139198582645969222983109

[EdH]

I didn't turn up anything near 1e-14.

[Gimarel]

Originally Posted by EdH

I didn't turn up anything near 1e-14.

I have msieve-gpu running for four days. Maybe CADO needs a bit mor time.

[EdH]

Originally Posted by Gimarel

I have msieve-gpu running for four days. Maybe CADO needs a bit mor time.

I completed the region Curtis suggested with incr at 420. I suppose I could vary something and run it again, just for the case of experimentation.

Parameter suggestions, anyone?

Edit: I decided to try the default CADO-NFS parameters. . .

[charybdis]

ECM progress: 7200@40e7 and 26000@42e7. Together with Curtis's work shall we call that enough?

I'd help out with the polyselect but Gimarel's polys all break the old c196 record so I don't think there would be much point.

[VBCurtis]

I'm at 6000 @ 42e7, calling it quits. I agree that's enough.

I also agree that Gimarel's records plus the two CADO efforts Ed and I put in are easily enough poly select. I'll test-sieve the best CADO and Gimarel's polys in a day or two, and get the job posted to the new 16e-small queue on NFS@home by Saturday.

Ed, that gives you until Wed morning to supply a poly! If you get something within 5% of Gimarel's score, I'll test-sieve it. My poly search will finish Tuesday evening.

[EdH]

Originally Posted by VBCurtis

I'm at 6000 @ 42e7, calling it quits. I agree that's enough.

I also agree that Gimarel's records plus the two CADO efforts Ed and I put in are easily enough poly select. I'll test-sieve the best CADO and Gimarel's polys in a day or two, and get the job posted to the new 16e-small queue on NFS@home by Saturday.

Ed, that gives you until Wed morning to supply a poly! If you get something within 5% of Gimarel's score, I'll test-sieve it. My poly search will finish Tuesday evening.

My polyselect (size) is scheduled to end before 8:00 this evening, but I could break in earlier, I suppose*. How long do you think my polyselect (root) will take?

Should we see if Max would like to spin any of these?

* I might do some experimentation today while this continues to run, by trying to run a separate machine to work on what I have. But, I really haven't tried this before, so I'm not sure how how to approach it and how it may turn out. Then, again, what are the odds that the best poly would be found in the last few hours of the search?

(Since Gimarel's scores are new records, I'd hate to beat them so soon. But, then again, he does have some others.)

[VBCurtis]

I'd just let it run; root-opt will only take a couple hours, and it's not like I'll be waiting on you Wednesday in any case.