Reserved for MF - Sequence 276

[unconnected]

BTW, sequence 3366 also needs some hard ECM on C195@i2180. I did full t55 (18000@11e7) but have limited resources to take it further.

[yoyo]

Would it be possible to run stage 1 for the 10000 curves with B1=29e8 on GPU?
And can somebody do it?
I would run stage 2 afterwards.

[swellman]

I dropped a note to Ryan yesterday alerting him to the existence of this thread in case he had interest. This morning he sent me this

Input number is
146958403500264369785547548773534064181201262260218085297440392878383892186824414643378590631107799775986185342359664860673182664805915775593953028576817423653521299222191420905092930824557639076598363
(201 digits)
Using MODMULN [mulredc:0, sqrredc:2]
Using B1=7600000000, B2=324909696561468, polynomial Dickson(30), sigma=814068971
dF=1048576, k=25, d=11741730, d2=19, i0=629
Expected number of curves to find a factor of n digits:
35      40      45      50      55      60      65      70      75      80
8       21      60      190     660     2483    9979    42978   194971  938688
Step 1 took 39791886ms
Using 24 small primes for NTT
Estimated memory usage: 4777M
Initializing tables of differences for F took 4615ms
Computing roots of F took 149378ms
Building F from its roots took 81163ms
Computing 1/F took 29062ms
Initializing table of differences for G took 1134ms
Computing roots of G took 112613ms
Building G from its roots took 85745ms
Computing roots of G took 113386ms
Building G from its roots took 88387ms
Computing G * H took 14793ms
Reducing  G * H mod F took 15953ms
Computing roots of G took 117551ms
Building G from its roots took 86185ms
Computing G * H took 15073ms
Reducing  G * H mod F took 15664ms
Computing roots of G took 113218ms
Building G from its roots took 86291ms
Computing G * H took 15155ms
Reducing  G * H mod F took 15685ms
Computing roots of G took 113035ms
Building G from its roots took 89617ms
Computing G * H took 15002ms
Reducing  G * H mod F took 16119ms
Computing roots of G took 115256ms
Building G from its roots took 88074ms
Computing G * H took 14599ms
Reducing  G * H mod F took 15358ms
Computing roots of G took 111318ms
Building G from its roots took 87578ms
Computing G * H took 14781ms
Reducing  G * H mod F took 15865ms
Computing roots of G took 114522ms
Building G from its roots took 87471ms
Computing G * H took 14960ms
Reducing  G * H mod F took 16199ms
Computing roots of G took 114005ms
Building G from its roots took 87993ms
Computing G * H took 15252ms
Reducing  G * H mod F took 15668ms
Computing roots of G took 116615ms
Building G from its roots took 90888ms
Computing G * H took 16013ms
Reducing  G * H mod F took 15646ms
Computing roots of G took 110877ms
Building G from its roots took 87894ms
Computing G * H took 14924ms
Reducing  G * H mod F took 15510ms
Computing roots of G took 113445ms
Building G from its roots took 88641ms
Computing G * H took 15545ms
Reducing  G * H mod F took 15649ms
Computing roots of G took 111441ms
Building G from its roots took 87906ms
Computing G * H took 15017ms
Reducing  G * H mod F took 15154ms
Computing roots of G took 111244ms
Building G from its roots took 88618ms
Computing G * H took 15138ms
Reducing  G * H mod F took 16981ms
Computing roots of G took 113272ms
Building G from its roots took 86699ms
Computing G * H took 14530ms
Reducing  G * H mod F took 17007ms
Computing roots of G took 113789ms
Building G from its roots took 88423ms
Computing G * H took 15182ms
Reducing  G * H mod F took 15466ms
Computing roots of G took 114517ms
Building G from its roots took 91294ms
Computing G * H took 16384ms
Reducing  G * H mod F took 16707ms
Computing roots of G took 112979ms
Building G from its roots took 87971ms
Computing G * H took 16115ms
Reducing  G * H mod F took 15482ms
Computing roots of G took 119187ms
Building G from its roots took 91174ms
Computing G * H took 15448ms
Reducing  G * H mod F took 16214ms
Computing roots of G took 119303ms
Building G from its roots took 89950ms
Computing G * H took 15699ms
Reducing  G * H mod F took 16519ms
Computing roots of G took 120544ms
Building G from its roots took 91226ms
Computing G * H took 16179ms
Reducing  G * H mod F took 16001ms
Computing roots of G took 121619ms
Building G from its roots took 91794ms
Computing G * H took 15702ms
Reducing  G * H mod F took 16251ms
Computing roots of G took 126970ms
Building G from its roots took 93380ms
Computing G * H took 15486ms
Reducing  G * H mod F took 16862ms
Computing roots of G took 122080ms
Building G from its roots took 91765ms
Computing G * H took 17270ms
Reducing  G * H mod F took 16071ms
Computing roots of G took 130862ms
Building G from its roots took 96044ms
Computing G * H took 17502ms
Reducing  G * H mod F took 18512ms
Computing polyeval(F,G) took 168438ms
Computing product of all F(g_i) took 586ms
Step 2 took 6337368ms
********** Factor found in step 2:
3135163172555178875064536378046330234984169364877281361211958842231
Found probable prime factor of 67 digits:
3135163172555178875064536378046330234984169364877281361211958842231
Probable prime cofactor
46874243990462637310356201495732887005986266709475578981920765898213102249350955733998141848135127026706804456577752515184428860000573
has 134 digits

A very lucky hit even for Ryan, though I fear toes may have been inadvertently stepped on.

[VBCurtis]

I considered running some curves, and I certainly wouldn't think anything but "hellyea!" if I'd found that factor. Yay Ryan!

Only a C191 next, within 15e NFS@home range. Extra Yay!

[Dubslow]

That's extraordinary. Paul Zimmerman drops a sequence he's been working on for more than a decade, due to no ECM hits on a very large composite, meanwhile you email Ryan and within 24 hours he has a hit at 7 digits above the previously tested level?

That's some baseball-type weirdness-luckyness right there. I'm astonished. I wonder what PZimmerman thinks. (Has anyone alerted him?)

By the way I guess we're working on this sequence for now. yoyo, could you (or ChristianB) update the rkn reservations accordingly please?

[LaurV]

Yeah, my thinking too, haha, mother life is a bitch, isn't her?

[yoyo]

Sequence is not reserved here https://www.rechenkraft.net/aliquot/AllSeq.html.
If someone wants to work on it she/he can reserve it in the reservation thread.

Does somebody work on it?

[VBCurtis]

The sequence should be reserved by/to the forum, and this thread will track results. I'm a fan of using NFS@home when we can, but Ryan may wish to push a few terms on his own- I think that's his right, since he found the factor!

[ryanp]

I don't really have the time to work on this sequence at the moment, so I'm happy to let the forum crack away at it for now.

[swellman]

Well folks, what's next? Anybody want to start ECM on the C191 at say the t50 level?

I can help once it goes to the t55 level in a few days. Or is yoyo better suited for most/all of the ECM for this series?

[Dubslow]

I'm nearly done with my far-too-monstrous reservation of sequences, after which I'll likely contribute here. But it's only a Sandy Bridge, more than an order of magnitude slower than recent CPUs.