Factoring for a publication - Page 36

Max0526 · 2021-06-25, 15:02

Quote:

Originally Posted by EdH

If you're not in a hurry for it, post an SNFS poly for (12, 5) and I'll play with it. But it might take me a little while to clear some other things.

Two best and very close spun polys, +22.8%, test-sieving may help to decide.

Code:

Line 1 : {'c4': '4', 'c3': '8', 'c2': '14', 'c1': '16', 'c0': '7', 'Y1': '178512009857421326929091882365072481276865298148585', 'Y0': '-21953952420563086883720242097847447142404785093424'}

n: 107934724803710842815193466239223994044210187743786463450800508445929080930258030578089590916696032770283244667212623716102585976501017568820072004971606073
# e = 2.19511122e-12
skew: 0.65268
type: snfs
c4: 4
c3: 8
c2: 14
c1: 16
c0: 7
Y1: 178512009857421326929091882365072481276865298148585
Y0: -21953952420563086883720242097847447142404785093424

Line 2 : {'c4': '1', 'c3': '0', 'c2': '8', 'c1': '-12', 'c0': '7', 'Y1': '200465962277984413812812124462919928419270083242009', 'Y0': '-21953952420563086883720242097847447142404785093424'}

n: 107934724803710842815193466239223994044210187743786463450800508445929080930258030578089590916696032770283244667212623716102585976501017568820072004971606073
# e = 2.19391411e-12
skew: 0.75986
type: snfs
c4: 1
c3: 0
c2: 8
c1: -12
c0: 7
Y1: 200465962277984413812812124462919928419270083242009
Y0: -21953952420563086883720242097847447142404785093424

Best original poly:

Code:

Line 1 : {'c4': '1', 'c3': '-12', 'c2': '62', 'c1': '-168', 'c0': '196', 'y1': '178512009857421326929091882365072481276865298148585', 'y0': '-313116114873716480090743280534450068268921026110322'}
Skew, Murphy E:
2.92361 1.78703639e-12

swishzzz · 2021-06-25, 20:18

Quote:

Originally Posted by charybdis

After 12700 CPU-hours of sieving, (6,-8) c179/snfs229 is in linear algebra:

Code:

linear algebra completed 34422 of 13311918 dimensions (0.3%, ETA 74h21m)

This was a bit harder than I expected, mainly because I underestimated how many relations would be needed. SNFS still looks to have been the right choice, just about: I reckon this was similar in difficulty to 178-digit GNFS. (and it's possible that better parameters would bring this down another digit)

If it ever comes to it, (8,-7) c184/snfs239 and c186/snfs239 will require careful test-sieving. My hunch is that the c184 will be best done by GNFS and the c186 by SNFS.
The c179 and c182 from (7,-8) are both clearly GNFS jobs.

Just finished t55 ECM on (9, -6) c178, no factors. Unspun poly is also SNFS 229, so this one seems to be right on the edge.

Max0526 · 2021-06-27, 16:57

Two best, +12.8%, very close E scores, test-sieving recommended:

Code:

Line 1 : {'c4': '1', 'c3': '0', 'c2': '8', 'c1': '12', 'c0': '7', 'Y1': '260857381135525029949243719294318313710433146912', 'Y0': '-364767459095540964152363616829471937271602132297'}

n: 170495892879689667889057667069421358402225903909290658832736196272405169062702473702500004756386265341260162168796213412508909744072928218898504600918747000881143901037569470047
# e = 8.71649946e-12
skew: 1.46824
type: snfs
c4: 1
c3: 0
c2: 8
c1: 12
c0: 7
Y1: 260857381135525029949243719294318313710433146912
Y0: -364767459095540964152363616829471937271602132297

Line 2 : {'c4': '1', 'c3': '4', 'c2': '14', 'c1': '32', 'c0': '28', 'Y1': '260857381135525029949243719294318313710433146912', 'Y0': '-103910077960015934203119897535153623561168985385'}

n: 170495892879689667889057667069421358402225903909290658832736196272405169062702473702500004756386265341260162168796213412508909744072928218898504600918747000881143901037569470047
# e = 8.70209692e-12
skew: 1.38062
type: snfs
c4: 1
c3: 4
c2: 14
c1: 32
c0: 28
Y1: 260857381135525029949243719294318313710433146912
Y0: -103910077960015934203119897535153623561168985385

Best original:

Code:

Line 1 : {'c4': '1', 'c3': '-12', 'c2': '62', 'c1': '-168', 'c0': '196', 'y1': '260857381135525029949243719294318313710433146912', 'y0': '-417804684311034125695367541053483003859697308439'}
Skew, Murphy E:
2.8701 7.72405909e-12

Max0526 · 2021-06-27, 18:13

Four best spun polys, +16.6%, close E scores, test-sieving recommended:

Code:

Line 1 : {'c4': '2', 'c3': '6', 'c2': '7', 'c1': '0', 'c0': '5', 'Y1': '260857381135525029949243719294318313710433146912', 'Y0': '-364767459095540964152363616829471937271602132297'}

n: 25388891615954670461360948076859077726826994389098365522047648957045208436622151964350726034407506280027329769873882728638251806526870392543793301719418076557471047891086997
# e = 9.48944743e-12
skew: 1.35581
type: snfs
c4: 2
c3: 6
c2: 7
c1: 0
c0: 5
Y1: 260857381135525029949243719294318313710433146912
Y0: -364767459095540964152363616829471937271602132297

Line 2 : {'c4': '2', 'c3': '14', 'c2': '37', 'c1': '40', 'c0': '20', 'Y1': '260857381135525029949243719294318313710433146912', 'Y0': '-103910077960015934203119897535153623561168985385'}

n: 25388891615954670461360948076859077726826994389098365522047648957045208436622151964350726034407506280027329769873882728638251806526870392543793301719418076557471047891086997
# e = 9.4562857e-12
skew: 1.18773
type: snfs
c4: 2
c3: 14
c2: 37
c1: 40
c0: 20
Y1: 260857381135525029949243719294318313710433146912
Y0: -103910077960015934203119897535153623561168985385

Line 3 : {'c4': '1', 'c3': '6', 'c2': '14', 'c1': '0', 'c0': '40', 'Y1': '130428690567762514974621859647159156855216573456', 'Y0': '-364767459095540964152363616829471937271602132297'}

n: 25388891615954670461360948076859077726826994389098365522047648957045208436622151964350726034407506280027329769873882728638251806526870392543793301719418076557471047891086997
# e = 9.39525925e-12
skew: 2.71153
type: snfs
c4: 1
c3: 6
c2: 14
c1: 0
c0: 40
Y1: 130428690567762514974621859647159156855216573456
Y0: -364767459095540964152363616829471937271602132297

Line 4 : {'c4': '1', 'c3': '14', 'c2': '74', 'c1': '160', 'c0': '160', 'Y1': '130428690567762514974621859647159156855216573456', 'Y0': '-103910077960015934203119897535153623561168985385'}

n: 25388891615954670461360948076859077726826994389098365522047648957045208436622151964350726034407506280027329769873882728638251806526870392543793301719418076557471047891086997
# e = 9.36040422e-12
skew: 2.37987
type: snfs
c4: 1
c3: 14
c2: 74
c1: 160
c0: 160
Y1: 130428690567762514974621859647159156855216573456
Y0: -103910077960015934203119897535153623561168985385

Best original:

Code:

Line 1 : {'c4': '2', 'c3': '-30', 'c2': '169', 'c1': '-420', 'c0': '392', 'y1': '260857381135525029949243719294318313710433146912', 'y0': '-417804684311034125695367541053483003859697308439'}
Skew, Murphy E:
2.89571 8.14036696e-12

charybdis · 2021-06-28, 16:58

Quote:

Originally Posted by charybdis

After 12700 CPU-hours of sieving, (6,-8) c179/snfs229 is in linear algebra:

(6,-8) c179 is done:

Code:

p74 factor: 86022890209360816301938530821988654991039504448418863958026499303508753681
p105 factor: 576932364411697609643265415249894142913435989897979132448369192413060252968415853013591704634388357802399

richs · 2021-06-28, 21:50

Quote:

Originally Posted by richs

(5, -9) C251 ecm'ed to t45. Continuing to t50.

t50 complete, no factors.

EdH · 2021-06-29, 16:07

Well, I suppose it's because of my ignorance in being able to choose all the right parameters, but the c156 I'm working on via SNFS took four days to make it to the square root failure point. I have Msieve doing LA right now with an ETA of 7.5 hours. Then we can see what Msieve SR thinks. . .

I'm certain, I can GNFS a c160 in less than a day. Could the parameters have made this run that bad?

Max0526 · 2021-06-29, 17:29

Quote:

Originally Posted by charybdis

(6,-8) c179 is done:

Code:

p74 factor: 86022890209360816301938530821988654991039504448418863958026499303508753681
p105 factor: 576932364411697609643265415249894142913435989897979132448369192413060252968415853013591704634388357802399

Thank you so much, charybdis!
I marked the sheet ~2 pm yesterday and forgot to answer here.

Max0526 · 2021-06-29, 17:31

Quote:

Originally Posted by richs

t50 complete, no factors.

Thank you, richs!
swishzzz updated the sheet yesterday.

Max0526 · 2021-06-29, 17:48

Quote:

Originally Posted by EdH

Well, I suppose it's because of my ignorance in being able to choose all the right parameters, but the c156 I'm working on via SNFS took four days to make it to the square root failure point. I have Msieve doing LA right now with an ETA of 7.5 hours. Then we can see what Msieve SR thinks. . .

I'm certain, I can GNFS a c160 in less than a day. Could the parameters have made this run that bad?

I think this post from charybdis would be the most important to consider before starting SNFS: https://www.mersenneforum.org/showpo...&postcount=109
The manual parameter selection I posted here https://www.mersenneforum.org/showpo...&postcount=104 is taken from Makoto Kamada's STDKMD site.
See the top chart here https://stdkmd.net/nrr/wanted.htm for a 100 small SNFS polys, e.g., https://stdkmd.net/nrr/c.cgi?q=32227_191 at the very top (copy-pasted below).

Code:

These parameters were not fully adjusted. The approximate expressions which were used for making the parameters are as follows.
(with a strike through) deg: exponent≤105 ? 4 : exponent≤210 ? 5 : 6 or exponent≤144 ? 4 : 6 (with a strike through)
d: log10(cdeg)+deg×log10(m) [digits]
time: 10^(d/30-4) [hours]
skew: |c0/cdeg|^(1/deg)
rlim: round(7×10^(d/60+3))
lpbr: floor(d/25+21)
mfbr: floor(d/8+31)
rlambda: floor(d/25+18)/10

EdH · 2021-06-29, 18:09

Thanks! I'm wondering what I did here. Looking at the SNFS params file, it's all messed up. There's no side choice, the values are all wrong and even the siever was wrong (I =12). I suppose I'm lucky it found relations. I'm going to go back to sleep for a while. I'll post the factors for the c156 later, but then I think I'll take a break for a bit.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
factoring 2ⁿ-2 equivalent to factoring 2ⁿ-1(I think)	baih	Miscellaneous Math	9	2020-09-21 07:11
OpenCL GPU P-1 Factoring and ECM Factoring	xx005fs	GPU Computing	3	2018-10-27 14:49

2021-06-29, 16:07	#392
EdH "Ed Hall" Dec 2009 Adirondack Mtns 2²×1,303 Posts	Well, I suppose it's because of my ignorance in being able to choose all the right parameters, but the c156 I'm working on via SNFS took four days to make it to the square root failure point. I have Msieve doing LA right now with an ETA of 7.5 hours. Then we can see what Msieve SR thinks. . . I'm certain, I can GNFS a c160 in less than a day. Could the parameters have made this run that bad?

2021-06-29, 18:09	#396
EdH "Ed Hall" Dec 2009 Adirondack Mtns 5212₁₀ Posts	Thanks! I'm wondering what I did here. Looking at the SNFS params file, it's all messed up. There's no side choice, the values are all wrong and even the siever was wrong (I =12). I suppose I'm lucky it found relations. I'm going to go back to sleep for a while. I'll post the factors for the c156 later, but then I think I'll take a break for a bit.