GIMPS' first Fermat factor!

[akruppa]

Saving residues every 5M should be fine. Doing the double-check in 14 pieces is probably more practical than in 68, and each one should take only a few days.

Alex

[mdettweiler]

Originally Posted by akruppa

Saving residues every 5M should be fine. Doing the double-check in 14 pieces is probably more practical than in 68, and each one should take only a few days.

Alex

Huh? You can distribute doublechecks like that? I thought in order to pick up in the middle of a test, you have to have the save file for everything up to that point, so you may as well just do the whole thing through.

[smh]

Originally Posted by mdettweiler

Huh? You can distribute doublechecks like that?

You can distribute the save file and everyone can run the 5M iterations till the next save file. If the residues match, there was no error during that part of the run. The risk is, of course, that if there is an error early in the run, a couple of doublechecks will be wasted.

[TimSorbet]

Yes, you need the save file for everything up to that point in order to resume a test in the middle, but like smh said, distributing save files at intervals and checking each section's residues works. If all sections match, then it's actually even more certain that the final result is correct than a single DC with matching residues. (since the chance of residues just happening to match goes from something like 2^64 to something like 2^64^('number of save files', which is ceiling('number of total iterations'/'number of iterations between save file copy-offs')), which from akruppa's post should be, for this run, 2^64^14)
It's not feasible to run distributed DCs in normal cases, like for normal GIMPS DCs, since you have to distribute save files that are several MB. Besides, needless to say, the nature of the massive parallelity (is that a word? you know what I mean) of GIMPS's DCs means it would provide minimal benefits.
For a DC like that of F26, though, it fits perfectly.

If the save files are put out and DCd as soon as they are saved off, then there will be very little wasted work. If we figure that the DC runs as fast as the first-run, then only the time to run 5M iterations will be wasted for each error.

[CRGreathouse]

Originally Posted by Mini-Geek

If all sections match, then it's actually even more certain that the final result is correct than a single DC with matching residues. (since the chance of residues just happening to match goes from something like 2^64 to something like 2^64^('number of save files', which is ceiling('number of total iterations'/'number of iterations between save file copy-offs')), which from akruppa's post should be, for this run, 2^64^14)

I would have thought 2^64 to 2^(64k) rather than 2^(64^k). Unless you meant (2^64)^k = 2^(64k), of course -- but I'm used to the convention a^b^c = a^(b^c).

[TimSorbet]

Originally Posted by CRGreathouse

I would have thought 2^64 to 2^(64k) rather than 2^(64^k). Unless you meant (2^64)^k = 2^(64k), of course -- but I'm used to the convention a^b^c = a^(b^c).

You're probably right.
Let's simplify it a bit to try to find out: instead of 2^64, we'll use 2^2, and k=3.
The odds of three such flips all matching up is .25*.25*.25, or .25^3, or 1/64, which is 1/(2^2)^3.
Yep, you're right, it's 2^64k.

[ATH]

The remaining number of F26 is composite:

UID: athath, F26/76861124116481 is not prime. RES64: 6C433D4E3CC9522E. Wd1: 7BD8A30F,00000000

F26/76861124116481 interim Wd1 residue AE87D4D932FC3DF3 at iteration 1000000
F26/76861124116481 interim Wd1 residue F2716A338E6E6EBC at iteration 2000000
F26/76861124116481 interim Wd1 residue 1C1CBA26FE389227 at iteration 3000000
F26/76861124116481 interim Wd1 residue 01DDE9CADFF26078 at iteration 4000000
F26/76861124116481 interim Wd1 residue 556D6405C818F46D at iteration 5000000
F26/76861124116481 interim Wd1 residue E1F551A8AD31EC12 at iteration 6000000
F26/76861124116481 interim Wd1 residue 99511915410EF3E9 at iteration 7000000
F26/76861124116481 interim Wd1 residue 1FEE46100F4BFD30 at iteration 8000000
F26/76861124116481 interim Wd1 residue DF4A9E71C5A5AB85 at iteration 9000000
F26/76861124116481 interim Wd1 residue 0B2076FE17D9257F at iteration 10000000
F26/76861124116481 interim Wd1 residue 474E35E749F20330 at iteration 11000000
F26/76861124116481 interim Wd1 residue 9A2FA0540866B29E at iteration 12000000
F26/76861124116481 interim Wd1 residue 754F4113FCB5D0D2 at iteration 13000000
F26/76861124116481 interim Wd1 residue 77A29ACA327342E0 at iteration 14000000
F26/76861124116481 interim Wd1 residue 48F56556183B315A at iteration 15000000
F26/76861124116481 interim Wd1 residue 562431741A43CECA at iteration 16000000
F26/76861124116481 interim Wd1 residue 44D2CE56A6B66DD0 at iteration 17000000
F26/76861124116481 interim Wd1 residue 956C29B9FAC5C3D5 at iteration 18000000
F26/76861124116481 interim Wd1 residue 3F884922256F9B9F at iteration 19000000
F26/76861124116481 interim Wd1 residue 0162B16F72294488 at iteration 20000000
F26/76861124116481 interim Wd1 residue 6AA9E95044D69741 at iteration 21000000
F26/76861124116481 interim Wd1 residue 487F0275DC7D334E at iteration 22000000
F26/76861124116481 interim Wd1 residue 141421DC8D617F7A at iteration 23000000
F26/76861124116481 interim Wd1 residue 3D738BCEE704A895 at iteration 24000000
F26/76861124116481 interim Wd1 residue 8B8EACA16EDCBD04 at iteration 25000000
F26/76861124116481 interim Wd1 residue 26650173A8038D01 at iteration 26000000
F26/76861124116481 interim Wd1 residue F8CDB032FA451238 at iteration 27000000
F26/76861124116481 interim Wd1 residue DD44C2B03514B2D4 at iteration 28000000
F26/76861124116481 interim Wd1 residue 36902A9AC1A30F73 at iteration 29000000
F26/76861124116481 interim Wd1 residue 95467ED0921CFAB6 at iteration 30000000
F26/76861124116481 interim Wd1 residue E14D1F2E655C9B43 at iteration 31000000
F26/76861124116481 interim Wd1 residue 76082581A0ECFB54 at iteration 32000000
F26/76861124116481 interim Wd1 residue 7F02E42FD4CE9D68 at iteration 33000000
F26/76861124116481 interim Wd1 residue 377C70DF388A58E6 at iteration 34000000
F26/76861124116481 interim Wd1 residue C25E1154C4F9FEAC at iteration 35000000
F26/76861124116481 interim Wd1 residue 1DD329E92A944406 at iteration 36000000
F26/76861124116481 interim Wd1 residue CBDED833FAA65397 at iteration 37000000
F26/76861124116481 interim Wd1 residue FF2D5E22919776EA at iteration 38000000
F26/76861124116481 interim Wd1 residue FEBDF08EAF61798A at iteration 39000000
F26/76861124116481 interim Wd1 residue A57C6E53515D9D40 at iteration 40000000
F26/76861124116481 interim Wd1 residue 8DBCF6463070CDA8 at iteration 41000000
F26/76861124116481 interim Wd1 residue 8ED2F50765E4A35C at iteration 42000000
F26/76861124116481 interim Wd1 residue B6E947745F9EC0D1 at iteration 43000000
F26/76861124116481 interim Wd1 residue 877D7B0C8853D9D1 at iteration 44000000
F26/76861124116481 interim Wd1 residue DF806EDB608211FB at iteration 45000000
F26/76861124116481 interim Wd1 residue 17795281FA1D388C at iteration 46000000
F26/76861124116481 interim Wd1 residue 7B6B8A2F7E4F3C18 at iteration 47000000
F26/76861124116481 interim Wd1 residue 9C209862DCB82278 at iteration 48000000
F26/76861124116481 interim Wd1 residue A4BC1D0F631E15E6 at iteration 49000000
F26/76861124116481 interim Wd1 residue B68B35A87E25463D at iteration 50000000
F26/76861124116481 interim Wd1 residue C166A000CD7AFEB6 at iteration 51000000
F26/76861124116481 interim Wd1 residue D42D89D4A9340795 at iteration 52000000
F26/76861124116481 interim Wd1 residue 55158051DCBCA9AD at iteration 53000000
F26/76861124116481 interim Wd1 residue 995BC4E3FE6B0899 at iteration 54000000
F26/76861124116481 interim Wd1 residue 6CC66E75DC740FC1 at iteration 55000000
F26/76861124116481 interim Wd1 residue 593F5ADA166AFB6D at iteration 56000000
F26/76861124116481 interim Wd1 residue 539EF8E111E3066C at iteration 57000000
F26/76861124116481 interim Wd1 residue E98B2676B1EE5FC9 at iteration 58000000
F26/76861124116481 interim Wd1 residue D26FAE44A5545FD9 at iteration 59000000
F26/76861124116481 interim Wd1 residue B96BE569D4406C25 at iteration 60000000
F26/76861124116481 interim Wd1 residue 74415CDDB6D5AD50 at iteration 61000000
F26/76861124116481 interim Wd1 residue AEDCFC33C6754707 at iteration 62000000
F26/76861124116481 interim Wd1 residue EB33C58290D69B8F at iteration 63000000
F26/76861124116481 interim Wd1 residue 9A069F580D23ADC2 at iteration 64000000
F26/76861124116481 interim Wd1 residue 2FE2DE89CE503DE8 at iteration 65000000
F26/76861124116481 interim Wd1 residue B763250871FC2B23 at iteration 66000000
F26/76861124116481 interim Wd1 residue 83FC06C89428A04E at iteration 67000000

[R.D. Silverman]

Originally Posted by ATH

The remaining number of F26 is composite:

Nicely done. Not a surprise, of course.

[ATH]

The remaining 40,403,531 digit factor of F27 is composite:

UID: athath, F27/151413703311361/231292694251438081 is not prime. RES64: 481F26965DE16117. Wd1: AD647FF8,00000000

I'm not going any higher :) This one took 8 months on and off, roughly 190 days cpu time on a Core2duo (Conroe) E6750 2.66 Ghz.

Residues every 1M iterations:
F27residues.txt

[Andi47]

Originally Posted by ATH

The remaining 40,403,531 digit factor of F27 is composite:

UID: athath, F27/151413703311361/231292694251438081 is not prime. RES64: 481F26965DE16117. Wd1: AD647FF8,00000000

I'm not going any higher :) This one took 8 months on and off, roughly 190 days cpu time on a Core2duo (Conroe) E6750 2.66 Ghz.

Residues every 1M iterations:
F27residues.txt

It seems that your results of proving the compositeness of F25, F26 (and now F27) did not make it to Wilfried Keller's Page. Can you please send him an email?

[ATH]

Originally Posted by Andi47

It seems that your results of proving the compositeness of F25, F26 (and now F27) did not make it to Wilfried Keller's Page. Can you please send him an email?

Yeah, I sent him an email today.