mersenneforum.org  

Go Back   mersenneforum.org > Factoring Projects > Cunningham Tables

Reply
 
Thread Tools
Old 2021-12-07, 09:15   #12
Gimarel
 
Apr 2010

19410 Posts
Default

Quote:
Originally Posted by swellman View Post
Done searching. Can’t find a poly even close to Gimarel’s gem.
I think that you giving up too fast. I have two more polys with a similar score and I'm optimistic to find one with a better score.
Code:
# norm 7.375882e-16 alpha -12.011597 e 3.578e-16 rroots 4
skew: 58305240.91
c0: 1333072541380389597470354173207214344325876193954300
c1: 249877985192636981198595328795551397207435728
c2: 15277361638643693781836775615903933945
c3: 37382780628139609655397222494
c4: -8175707456683442327327
c5: 24504116869820
c6: 352800
Y0: -1113549152873829308361401995126669672
Y1: 271159787779363909556551

# norm 7.085980e-16 alpha -11.225581 e 3.489e-16 rroots 6
skew: 17789353.14
c0: -34119248366663846185278942437796618591687616006920
c1: 10291106877280642918720056960799132150546209
c2: 384150905485505421860272295941740301
c3: -161742382206765912260333396451
c4: -4141445927376482706709
c5: 316761446612130
c6: 378000
Y0: -1113550171027765748936524242536827922
Y1: 15182916390265622764861
Gimarel is offline   Reply With Quote
Old 2021-12-08, 00:56   #13
Max0526
 
"Max"
Jun 2016
Toronto

919 Posts
Default

I am trying to spin all three now.
Max0526 is offline   Reply With Quote
Old 2021-12-08, 01:36   #14
swellman
 
swellman's Avatar
 
Jun 2012

1101000111102 Posts
Default

Quote:
Originally Posted by Gimarel View Post
I think that you giving up too fast. I have two more polys with a similar score and I'm optimistic to find one with a better score.
My search results were pretty disappointing. I ran CADO with incr=420 and adrange=1680 but found nothing worth mentioning. Tried 4620/18480 still nothing above 2.3.

Even ran both msieve-GPU and CADO centered around a6 = 25200. Three days of optimization gave me an e-score with a 1.6 handle! Just couldn’t find a search methodology with any apparent promise.

I may try one more time, using 4620/18480 with an a6 range of 10-50M, in blocks of 5M.

Whatever your modified code does it appears to spin straw into gold. Kudos!
swellman is online now   Reply With Quote
Old 2021-12-08, 02:26   #15
VBCurtis
 
VBCurtis's Avatar
 
"Curtis"
Feb 2005
Riverside, CA

2×2,579 Posts
Default

Please post here which specific ranges you ran with CADO, so I don't duplicate work when I give this a try in a couple weeks. I'm busy with ECM pretesting on a couple SNFS jobs in preparation for one of those "boinc challenges" that aim a ton of clients at NFS@home, but once those are queued I'll put a thread-year or so into this search- likely starting around xmas.
VBCurtis is offline   Reply With Quote
Old 2021-12-08, 03:03   #16
swellman
 
swellman's Avatar
 
Jun 2012

2×23×73 Posts
Default

Quote:
Originally Posted by VBCurtis View Post
Please post here which specific ranges you ran with CADO, so I don't duplicate work when I give this a try in a couple weeks. I'm busy with ECM pretesting on a couple SNFS jobs in preparation for one of those "boinc challenges" that aim a ton of clients at NFS@home, but once those are queued I'll put a thread-year or so into this search- likely starting around xmas.
Not sure if you’re asking me, but I searched using CADO over a6: 0-1M and 1.7-2M. And it took two+ weeks with the top find having a 2.3 handle.

I may now start a6 up even higher than your planned range with incr = 4620, or perhaps an even greater value. Cover more ground with less depth? My trial burrowing for days around a known “good a6” gave very poor results. I had hoped to duplicate Gimarel’s result but no.
swellman is online now   Reply With Quote
Old 2021-12-08, 03:11   #17
Max0526
 
"Max"
Jun 2016
Toronto

11100101112 Posts
Default

Nothing better yet. Some spinoffs of Gimarel's polys with a 3-handle for EdH's CADO spinner:
Code:
Y0: -556774576856127686087597601737762682
Y1: 271159787779363909556551
c0: 333075016176933526796868868454780289178872124848275
c1: 124891755530710265955345151175634145711461700
c2: 15277014406982003106713252768111351633
c3: 74967800240925120461782079260
c4: -32704344958945097026108
c5: 195980573793760
c6: 5644800
skew: 29712531.04
# size 5.675e-16, alpha -12.010, combined = 3.265e-16 rroots = 4
Code:
Y0: -1113552721143769423765907856623629681
Y1: 2868039601322175325006
c0: -6282102854980536317153525502211690842928383870
c1: -11686012871431240158579355681910076732003
c2: 36463498823184902691751367577951338
c3: -427099224754354481707443407
c4: -15588651231096741648862
c5: -181182646433064
c6: 201600
skew: 3756945.75
# size 5.426e-16, alpha -9.518, combined = 3.144e-16 rroots = 6
Code:
Y0: -1113550175170075278943523304605559969
Y1: 30365832780531245529722
c0: -4611885609149496144154873452859766517988321983060
c1: 2511430227447958572479504111033456556138914
c2: 257310037972570815054523237014681683
c3: -156987162634503515722170703379
c4: -9146258518655766468518
c5: 1264570699904520
c6: 3024000
skew: 9174416.21
# size 5.106e-16, alpha -10.994, combined = 3.093e-16 rroots = 6
Max0526 is offline   Reply With Quote
Old 2021-12-08, 05:48   #18
VBCurtis
 
VBCurtis's Avatar
 
"Curtis"
Feb 2005
Riverside, CA

2·2,579 Posts
Default

Quote:
Originally Posted by swellman View Post
Not sure if you’re asking me, but I searched using CADO over a6: 0-1M and 1.7-2M. And it took two+ weeks with the top find having a 2.3 handle.

I may now start a6 up even higher than your planned range with incr = 4620, or perhaps an even greater value. Cover more ground with less depth? My trial burrowing for days around a known “good a6” gave very poor results. I had hoped to duplicate Gimarel’s result but no.
I'm thinking I'll try admin of 100M or so, with incr of 110880 (selected by the CADO group for the RSA-240 poly search). I can be flexible- whomever chooses the lowest incr value should have priority in an a6 range, so I'll shift upward if someone searches into the high 8-digits.
VBCurtis is offline   Reply With Quote
Old 2021-12-08, 12:53   #19
swellman
 
swellman's Avatar
 
Jun 2012

64368 Posts
Default

Quote:
Originally Posted by VBCurtis View Post
I'm thinking I'll try admin of 100M or so, with incr of 110880 (selected by the CADO group for the RSA-240 poly search). I can be flexible- whomever chooses the lowest incr value should have priority in an a6 range, so I'll shift upward if someone searches into the high 8-digits.
In that case, I’ll sign up for admin=10M, incr of 30030, admax=50M. Will restart my search this evening.

That leaves the range 50-100M open for searching.
swellman is online now   Reply With Quote
Old 2021-12-10, 15:32   #20
swellman
 
swellman's Avatar
 
Jun 2012

2·23·73 Posts
Default

My first section (10-15M) is done, with 2.44e-16 the best result.

I am rerunning this with ropteffort=140 in lieu of 100 used in this first run. I expect no improvement but it seems worth a try.

Another machine is running the top end of my search range 40-50M (with ropteffort=100).

These runs are faster than I expected, so if no one wants it I can search 50-100M in a week or so.
swellman is online now   Reply With Quote
Old 2021-12-11, 14:16   #21
Gimarel
 
Apr 2010

3028 Posts
Default

I think, that the parameters for cado have to be reconsidered. For about 16h I run
Code:
polyselect -d 6 -n 24023387191766184217094927222587025087468868508838209688355658860897738047963529318736265313199625192712082581481439831563094403250831892547179972207284600405234081309183045488404642574713121441527058807512178264038863337 -P 5000000 -admin 1260 -admax 126001 -incr 420 -sopteffort 1 -t 32 -v -nq 279936
This isn't complete yet.

Running msieve rootsieve on the resulting size-optimized polys gives a best score 3.097e-16. Cados ropt with ropteffort 1 gets a score 2.875e-16 on the same size-optimized poly but ropteffort 10 gets only 2.760e-16.
The size-optimized poly is:
Code:
n: 24023387191766184217094927222587025087468868508838209688355658860897738047963529318736265313199625192712082581481439831563094403250831892547179972207284600405234081309183045488404642574713121441527058807512178264038863337
Y1: 806008873539147657940235233
Y0: -879851971181155162208432563347393133
c6: -1016400
c5: -15235385963040
c4: -6294711379783144129999
c3: 5466480360897072625699033948
c2: 493105692357882839776420919503677508
c1: -100163327848708707409468990789835415938548
c0: -2088003479658792575985631978579826901018379459729
# exp_E 53.50, lognorm 62.92, skew 8018993.10, 4 rroots
msieves rootsieve:
Code:
# norm 6.084435e-16 alpha -10.618832 e 3.097e-16 rroots 2
skew: 17808880.34
c0: -42002780267256221187650180356142750792071754151336
c1: 651262519104063232720115627970796415197105
c2: 1326111862738691484929056117647791495
c3: 27473728438213413248959196895
c4: -6237993554117149363999
c5: -9683067191040
c6: -1016400
Y0: -880585805990113246889342540216454148
Y1: 806008873539147657940235233
cados ropt:
Code:
n: 24023387191766184217094927222587025087468868508838209688355658860897738047963529318736265313199625192712082581481439831563094403250831892547179972207284600405234081309183045488404642574713121441527058807512178264038863337
Y0: -879988034345106179263717112397311096
Y1: 806008873539147657940235233
c0: -1067923738334962179029304141722744115795509256660
c1: -127025912283259850282903340420626685825756
c2: 489133544750679016918999325219311391
c3: 9712702591017452259191128304
c4: -6282286343689740768799
c5: -14205908960640
c6: -1016400
skew: 8653593.983
# lognorm 62.98, E 53.26, alpha -9.73 (proj -2.12), 4 real roots
# MurphyE(Bf=1.000e+07,Bg=5.000e+06,area=1.000e+16)=2.875e-16
### Best MurphyE so far is 2.875e-16, av. exp_E 53.50, av. E 53.26
Gimarel is offline   Reply With Quote
Old 2021-12-12, 20:49   #22
swellman
 
swellman's Avatar
 
Jun 2012

D1E16 Posts
Default

Quote:
Originally Posted by swellman View Post
My first section (10-15M) is done, with 2.44e-16 the best result.

I am rerunning this with ropteffort=140 in lieu of 100 used in this first run. I expect no improvement but it seems worth a try.

Another machine is running the top end of my search range 40-50M (with ropteffort=100).

These runs are faster than I expected, so if no one wants it I can search 50-100M in a week or so.
I reran my search 10-15M, with ropteffort=140. Result is unchanged, i.e. escore of 2.44e-16.

40-50M is in root optimization now on a separate machine, and I’m starting the 15-40M search on the above machine now. Should take about a 7-8 days.
swellman is online now   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
Polynomial selection Max0526 NFS@Home 9 2017-05-20 08:57
Best way to scale polynomial selection pastcow Msieve 6 2013-05-08 09:01
Updated polynomial selection jasonp Msieve 65 2011-05-01 19:06
2^877-1 polynomial selection fivemack Factoring 47 2009-06-16 00:24
Polynomial selection CRGreathouse Factoring 2 2009-05-25 07:55

All times are UTC. The time now is 02:28.


Sat Jan 22 02:28:27 UTC 2022 up 182 days, 20:57, 0 users, load averages: 1.64, 1.43, 1.48

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2022, Jelsoft Enterprises Ltd.

This forum has received and complied with 0 (zero) government requests for information.

Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.
A copy of the license is included in the FAQ.

≠ ± ∓ ÷ × · − √ ‰ ⊗ ⊕ ⊖ ⊘ ⊙ ≤ ≥ ≦ ≧ ≨ ≩ ≺ ≻ ≼ ≽ ⊏ ⊐ ⊑ ⊒ ² ³ °
∠ ∟ ° ≅ ~ ‖ ⟂ ⫛
≡ ≜ ≈ ∝ ∞ ≪ ≫ ⌊⌋ ⌈⌉ ∘ ∏ ∐ ∑ ∧ ∨ ∩ ∪ ⨀ ⊕ ⊗ 𝖕 𝖖 𝖗 ⊲ ⊳
∅ ∖ ∁ ↦ ↣ ∩ ∪ ⊆ ⊂ ⊄ ⊊ ⊇ ⊃ ⊅ ⊋ ⊖ ∈ ∉ ∋ ∌ ℕ ℤ ℚ ℝ ℂ ℵ ℶ ℷ ℸ 𝓟
¬ ∨ ∧ ⊕ → ← ⇒ ⇐ ⇔ ∀ ∃ ∄ ∴ ∵ ⊤ ⊥ ⊢ ⊨ ⫤ ⊣ … ⋯ ⋮ ⋰ ⋱
∫ ∬ ∭ ∮ ∯ ∰ ∇ ∆ δ ∂ ℱ ℒ ℓ
𝛢𝛼 𝛣𝛽 𝛤𝛾 𝛥𝛿 𝛦𝜀𝜖 𝛧𝜁 𝛨𝜂 𝛩𝜃𝜗 𝛪𝜄 𝛫𝜅 𝛬𝜆 𝛭𝜇 𝛮𝜈 𝛯𝜉 𝛰𝜊 𝛱𝜋 𝛲𝜌 𝛴𝜎 𝛵𝜏 𝛶𝜐 𝛷𝜙𝜑 𝛸𝜒 𝛹𝜓 𝛺𝜔