Poly search candidates

schickel · 2013-06-27, 13:22

Quote:

Originally Posted by VBCurtis

I think we both would benefit from playing with settings and finding polys for a couple of smaller numbers. Perhaps some of the heavy hitters in this forum would like to supply you and I a couple of C155-180s to poly search? We can discuss settings, try to learn what stage1 bound produces the largest rate of useful nps hits per hour of gpu time, etc.

I have some numbers from my aliquot sequences queued up that would benefit from some heavy poly searching:

Code:

a115302.i5902 c150: 722986402529387761531523245174740725218763586252727774665788018930931517968238781361941678683265501221456903023062936317170941945069518708941274228073
a7044.i3420 c163:   1069286439460530805668612534096557556319118307440678075882272080432198210140482055955419148293522070111238366691021950622658624812365704607046438999695718693112229
a572000.i3128 c150: 122376126144695612022009890341256236363641981363125998702726545592271314432434257780727895029949816564645168871916657082807007147983169173961146078109
a611156.i7542 c155: 58035019786194098807181925479740537352988162044968814759257077868985812298144342954852184074381182357631459257468349391638272898828522380388658996312387583

They're currently in ECM testing, but I expect to have to put (most of) them through GNFS.....

For something to compare to, here is a poly I found for a c161 from a363270.i1775:

Code:

n: 91572763939319037745727514758662679674088061621685557085664194173403532319145745043107884001905080870498879201294878091215721815926596888093059278470362580371507
# norm 1.095460e-015 alpha -7.627276 e 1.156e-012 rroots 5
skew: 33361457.56
c0:  383355381299806903665668126264661160636400
c1:  4045284341356807274846442992265900
c2: -2318642919279642643732932976
c3: -4805046335712948621
c4:  2748629879684
c5:  5124
Y0: -28211802837843590414718838382517
Y1:  363459609449559037

I'm going to start GNFS on this one when I hit a hard number on my currently active sequence.

wombatman · 2013-06-27, 18:32

I'd be happy to work on the C155. I'll start the poly searching today.

firejuggler · 2013-06-27, 22:35

a very quick search on the first c150

Code:

# norm 1.609539e-014 alpha -8.035676 e 5.060e-012 rroots 5
skew: 55072249.88
c0: 1819977917444241850538479411223658896160
c1: 1771423220507065762039012633780080
c2: 61832710225179014067573822
c3: -4804805510554134361
c4: -15968892176
c5: 120
Y0: -359740968067837562238081234761
Y1: 3807520958703259

i'll let it run for a few hour

VBCurtis · 2013-06-28, 03:56

I'll do a GPU-day on the C155 also, to compare results, and then start on the C163. I'll keep doing the 3 steps separately, to see how results vary vs the np1-nps combo.

-Curtis

firejuggler · 2013-06-28, 08:27

Eight hours later, no improvement on the C150 poly

firejuggler · 2013-06-28, 16:37

a light poly for the 163
Fri Jun 28 18:35:35 2013 R0: -85947222884751164428165834069570
Fri Jun 28 18:35:35 2013 R1: 11198382273354161
Fri Jun 28 18:35:35 2013 A0: -90871099326788541532725912746222919655521
Fri Jun 28 18:35:35 2013 A1: 1792856973270648979649109055958235
Fri Jun 28 18:35:35 2013 A2: 462294595261376402965340374
Fri Jun 28 18:35:35 2013 A3: 6142380049087671731
Fri Jun 28 18:35:35 2013 A4: -119766512679
Fri Jun 28 18:35:35 2013 A5: 228
Fri Jun 28 18:35:35 2013 skew 63700732.89, size 7.874e-016, alpha -5.911, combined = 9.716e-013 rroots = 5

henryzz · 2013-06-28, 21:02

Looking at your polys, firejuggler, they all start with small leading coefficients and as a consequence have huge skew. It might be worth you searching at a higher leading coefficient.

firejuggler · 2013-06-28, 21:24

As mentionned earlier, those are *quick-and-dirty* search, just done to give an idea of the poly score.

firejuggler · 2013-06-28, 22:59

here is a more serious search for the first C 150

Code:

# norm 1.444667e-014 alpha -6.917632 e 4.814e-012 rroots 3
skew: 841145.75
c0: -325738126959923032706969414442066400
c1: -1840003628971626993662283648330
c2: 225769551869279160391393
c3: 3762824820552507006
c4: -2143214131752
c5: 1686960
Y0: -53260479116005806504982958319
Y1: 13372535599069

searched poly with leading coef from 1 to 3e6

wombatman · 2013-06-29, 01:51

Here's an initial result for the C163:

Code:

expecting poly E from 9.45e-013 to > 1.09e-012

polynomial selection complete
R0: -6390374448189173278546531251014
R1: 19504335320833379
A0: -20025904729622787109637795434613157531
A1: 413767607592671960996692844369211
A2: 895005471926494085299577289
A3: -374505915870943397111
A4: -430367904368930
A5: 100337328
skew 1369087.07, size 6.093e-016, alpha -7.290, combined = 8.247e-013 rroots = 5

Still running some, but it's a pretty good start. Doing the -npr step on the C155--searched for sextics, and it should hopefully finish up soon.

Edit: Here's one with a slightly better score with a much lower skew:

Code:

polynomial selection complete
R0: -6389956263867362886877922677130
R1: 33280324830544727
A0: 767797939339909723668396375755528757
A1: 33215971061213227743193900791371
A2: -370232822547281454298146599
A3: -315552628829057509909
A4: 646291445176044
A5: 100370160
skew 709419.34, size 6.326e-016, alpha -6.967, combined = 8.418e-013 rroots = 5

wombatman · 2013-06-29, 03:47

And here's what I've gotten from ~1 day of GPU searching on the C155. Not great, but a good start.

Code:

expecting poly E from 2.61e-012 to > 3.00e-012
polynomial selection complete
R0: -6119479740188029422377425
R1: 2635387682623
A0: 42492734092992596689899104401432272
A1: -276970245278370518463806817308
A2: -54986121077830867971702912
A3: 21087170634838748433
A4: 3474626431755520
A5: -129010006404
A6: 1105104
skew 65069.71, size 3.420e-011, alpha -8.637, combined = 7.509e-013 rroots = 4

2013-06-28, 08:27	#5
firejuggler "Vincent" Apr 2010 Over the rainbow 101101101011₂ Posts	Eight hours later, no improvement on the C150 poly Last fiddled with by firejuggler on 2013-06-28 at 08:27

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

2013-06-27, 18:32	#2
wombatman I moo ablest echo power! May 2013 744₁₆ Posts	I'd be happy to work on the C155. I'll start the poly searching today.

2013-06-28, 03:56	#4
VBCurtis "Curtis" Feb 2005 Riverside, CA 53·113 Posts	I'll do a GPU-day on the C155 also, to compare results, and then start on the C163. I'll keep doing the 3 steps separately, to see how results vary vs the np1-nps combo. -Curtis

2013-06-28, 16:37	#6
firejuggler "Vincent" Apr 2010 Over the rainbow 5553₈ Posts	a light poly for the 163 Fri Jun 28 18:35:35 2013 R0: -85947222884751164428165834069570 Fri Jun 28 18:35:35 2013 R1: 11198382273354161 Fri Jun 28 18:35:35 2013 A0: -90871099326788541532725912746222919655521 Fri Jun 28 18:35:35 2013 A1: 1792856973270648979649109055958235 Fri Jun 28 18:35:35 2013 A2: 462294595261376402965340374 Fri Jun 28 18:35:35 2013 A3: 6142380049087671731 Fri Jun 28 18:35:35 2013 A4: -119766512679 Fri Jun 28 18:35:35 2013 A5: 228 Fri Jun 28 18:35:35 2013 skew 63700732.89, size 7.874e-016, alpha -5.911, combined = 9.716e-013 rroots = 5

2013-06-28, 21:02	#7
henryzz Just call me Henry "David" Sep 2007 Liverpool (GMT/BST) 1100001000100₂ Posts	Looking at your polys, firejuggler, they all start with small leading coefficients and as a consequence have huge skew. It might be worth you searching at a higher leading coefficient.

2013-06-28, 21:24	#8
firejuggler "Vincent" Apr 2010 Over the rainbow 37·79 Posts	As mentionned earlier, those are quick-and-dirty search, just done to give an idea of the poly score.