Poly search candidates - Page 2

wombatman · 2013-06-29, 04:54

Another for the C163--higher E score, but skew that's an order of magnitude higher.

Code:

polynomial selection complete
R0: -40020140479137682195488804583345
R1: 12470852768744483
A0: 1934060142089187254140727999747419918668
A1: 251989992651819359069791989348234
A2: -269320501923113497813241230
A3: -29383283268000506320
A4: 3396697929467
A5: 10416
skew 9932333.95, size 7.478e-016, alpha -6.509, combined = 9.418e-013 rroots = 5

wombatman · 2013-06-29, 05:56

Best one so far as a quintic:

Code:

polynomial selection complete
R0: -21001168756013069437856880644490
R1: 1537745313139543
A0: 1101813190447253196113041703044601808103
A1: 860667476357611390499479006857020
A2: -125729032517775400648934611
A3: -63124241910199841636
A4: 3510134989836
A5: 261744
skew 7089777.20, size 7.852e-016, alpha -6.614, combined = 9.832e-013 rroots = 5

VBCurtis · 2013-06-29, 06:04

The composite size where 6th degree can compete with 5th is thought to be around 210 digits. It may be as low as 195, but a sextic for a C155 is surely a waste of time.

The first 6-hour batch on the C155 produced 2.49e-12. While e-scores on sextics are not directly comparable, this is over triple the score of your result. Msieve (that is, Batalov's estimate code) thinks we'll find something around or over 2.6e-12.

For my own C147, I produced 500MB of stage 1 hits, which took me roughly 6 core-days to process- longer than the suggested 120 hr poly search limit. That data produced a single poly above msieve's 8.77e-12 estimate, at 9.13e-12.

For the C155, I'll have 600-700MB of stage 1 hits, still just 36hr or so on the GPU but a core-week of stage 2.
Edit: Your C163 score is already in msieve's window, after just a day! Nice find.

firejuggler · 2013-06-29, 12:18

So, my best one for the first C150 almost make the grade

Code:

 expecting poly E from 5.18e-012 to > 5.96e-012
Sat Jun 29 10:47:15 2013  R0: -35438200818517877149236947705
Sat Jun 29 10:47:15 2013  R1: 67524408571583
Sat Jun 29 10:47:15 2013  A0: 32784110189167120719008928006382776
Sat Jun 29 10:47:15 2013  A1: 188888873030153589238143265026
Sat Jun 29 10:47:15 2013  A2: -179427764511079427669989
Sat Jun 29 10:47:15 2013  A3: -4242417039255900574
Sat Jun 29 10:47:15 2013  A4: 3052432180560
Sat Jun 29 10:47:15 2013  A5: 12935160
Sat Jun 29 10:47:15 2013  skew 419146.07, size 1.295e-014, alpha -6.692, combined = 5.114e-012 rroots = 1

wombatman · 2013-06-29, 13:59

Yeah, I was mostly curious to see if there any benefit to trying 6th degrees, even with a smaller number. And based on what I've gotten thus far, it is in fact a waste of time. All the scores have been ok, but nothing spectacular. I'm working on some 5th degree ranges now. If I get anything that's better than yours, I'll put it up

Quote:

Originally Posted by VBCurtis

The composite size where 6th degree can compete with 5th is thought to be around 210 digits. It may be as low as 195, but a sextic for a C155 is surely a waste of time.

The first 6-hour batch on the C155 produced 2.49e-12. While e-scores on sextics are not directly comparable, this is over triple the score of your result. Msieve (that is, Batalov's estimate code) thinks we'll find something around or over 2.6e-12.

For my own C147, I produced 500MB of stage 1 hits, which took me roughly 6 core-days to process- longer than the suggested 120 hr poly search limit. That data produced a single poly above msieve's 8.77e-12 estimate, at 9.13e-12.

For the C155, I'll have 600-700MB of stage 1 hits, still just 36hr or so on the GPU but a core-week of stage 2.
Edit: Your C163 score is already in msieve's window, after just a day! Nice find.

firejuggler · 2013-06-29, 14:12

semi decent poly on the C155

Code:

Sat Jun 29 14:24:50 2013  expecting poly E from 2.61e-012 to > 3.00e-012
Sat Jun 29 16:06:32 2013  R0: -680628204836635134826943221897
Sat Jun 29 16:06:32 2013  R1: 44954668139851
Sat Jun 29 16:06:32 2013  A0: -631400808220929843355383580257719940
Sat Jun 29 16:06:32 2013  A1: -3931896448399836495367486860678
Sat Jun 29 16:06:32 2013  A2: 2874200947791086897271164
Sat Jun 29 16:06:32 2013  A3: 13720753381161217935
Sat Jun 29 16:06:32 2013  A4: -2389738631022
Sat Jun 29 16:06:32 2013  A5: 397320
Sat Jun 29 16:06:32 2013  skew 1189130.77, size 4.043e-015, alpha -6.493, combined = 2.567e-012 rroots = 3

wombatman · 2013-06-29, 16:32

Nice one, fire! I'm sorting through some 5th degree polynomials now to see if I can get anything better...

wombatman · 2013-06-29, 17:31

Surprisingly, I actually did come across one that's slightly better (score is good, but skew is high):

Code:

expecting poly E from 2.61e-012 to > 3.00e-012
R0: -1710798987491293177532301875105
R1: 2977085536379179
A0: -35325577353217692275205373019309099688
A1: 93960101629955806327917631955384
A2: 27232843908875185831284718
A3: -3235381678824829314
A4: 147472260525
A5: 3960
skew 11649226.70, size 4.333e-015, alpha -6.768, combined = 2.676e-012 rroots = 3

wombatman · 2013-06-30, 17:06

And here's a new best for the C155:

Code:

**Expecting 2.61e-12 to > 3.00 e-12**

R0: -428009249638926627527273633984
R1: 934477421964221
A0: 217061154789639629444011631916949755
A1: -572233306440187650641837358080
A2: -2454023742836008232872129
A3: 3244582709807438026
A4: 6367540736208
A5: 4040400
skew 606864.90, size 4.537e-015, alpha -6.117, combined = 2.782e-012 rroots = 3

And with that, I think I'm done searching on the C155 and C163 for now!

VBCurtis · 2013-07-01, 01:53

C155 poly:

Code:

# norm 6.127360e-015 alpha -7.959690 e 2.933e-012 rroots 5
skew: 1640621.52
R0: -447020206779336618711125180826
R1: 1757855198190749
c0: -2276053687049797273481909588627918189
c1: 25854343994823984053734488131562
c2: 63714329712221148759529753
c3: -832502874879406998
c4: -21370641846128
c5: 3251280

I am only 80% through the data, but this is unlikely to be beaten. I'll tackle the C163 next.

wombatman · 2013-07-01, 02:03

Wow! Nicely done!

Edit: Interesting that my search didn't pick it up. Wonder what the difference is?

2013-06-29, 06:04	#14
VBCurtis "Curtis" Feb 2005 Riverside, CA 4861₁₀ Posts	The composite size where 6th degree can compete with 5th is thought to be around 210 digits. It may be as low as 195, but a sextic for a C155 is surely a waste of time. The first 6-hour batch on the C155 produced 2.49e-12. While e-scores on sextics are not directly comparable, this is over triple the score of your result. Msieve (that is, Batalov's estimate code) thinks we'll find something around or over 2.6e-12. For my own C147, I produced 500MB of stage 1 hits, which took me roughly 6 core-days to process- longer than the suggested 120 hr poly search limit. That data produced a single poly above msieve's 8.77e-12 estimate, at 9.13e-12. For the C155, I'll have 600-700MB of stage 1 hits, still just 36hr or so on the GPU but a core-week of stage 2. Edit: Your C163 score is already in msieve's window, after just a day! Nice find. Last fiddled with by VBCurtis on 2013-06-29 at 06:38

2013-07-01, 02:03	#22
wombatman I moo ablest echo power! May 2013 29·61 Posts	Wow! Nicely done! Edit: Interesting that my search didn't pick it up. Wonder what the difference is? Last fiddled with by wombatman on 2013-07-01 at 02:11

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Poly Search vs Sieving times	EdH	Factoring	10	2013-10-14 20:00
Resume msieve poly search job?	Andi47	Msieve	1	2011-03-28 04:30
gpu poly search error	bdodson	Msieve	10	2010-11-09 19:46
Poly search for c157 from 4788:2422	henryzz	Aliquot Sequences	59	2009-07-04 06:27
Poly search for c137 from 4788:2408	axn	Aliquot Sequences	15	2009-05-28 16:50

2013-06-29, 16:32	#18
wombatman I moo ablest echo power! May 2013 29×61 Posts	Nice one, fire! I'm sorting through some 5th degree polynomials now to see if I can get anything better...