Register FAQ Search Today's Posts Mark Forums Read

2019-03-15, 01:50   #1717
Max0526

"Max"
Jun 2016
Toronto

19·37 Posts
C206 poly degree 6 -- 1.95

Quote:
 Originally Posted by Max0526 This one spins a bit higher: Code: Y0: -1538110509020104375978976968951119 Y1: 74881205760971794 c0: -535286831610710194798762751821201298078023600 c1: 5945055959053928254238681238992761412070 c2: 3244030930107740287434936150209597 c3: -2179315168806881651433735327 c4: -668653607927756486422 c5: 220538105096268 c6: 19352256 skew: 2721348.84 # size 5.852e-15, alpha -9.597, combined = 1.767e-15 rroots = 6
Actually

But it's degree 6. Will it sieve better?
Code:
Y0: -1538110509013620412372134421308659
Y1: 74881205760971794
c0: -1446560752975762872083356841587900498554577200
c1: 3766745954017790911698864365082729049350
c2: 2649277663818371317849409707330487
c3: -2394123180446139085639071247
c4: -570995136993799351822
c5: 230592376178508
c6: 19352256
skew: 2580555.41248
# lognorm 59.77, E 49.68, alpha -10.09 (proj -1.87), 6 real roots
# MurphyE(Bf=1.000e+07,Bg=5.000e+06,area=1.000e+16)=1.952e-15

Last fiddled with by Max0526 on 2019-03-15 at 02:39

2019-03-15, 02:18   #1718
Max0526

"Max"
Jun 2016
Toronto

19×37 Posts
C206 poly degree 6 -- 2.017

Quote:
 Originally Posted by VBCurtis CADO sends warm regards for our C206: Code: R0: -12601258825545276113806892086568209032311 R1: 371662080727495174910581769 A0: 874708573977327989079213908166025283601524526316032 A1: -1016897532213206546994807128393281224899768 A2: -4765540714559627430714382742228252 A3: 2352509367951779125835982 A4: 2019390882502791 A5: 705600 skew 1201838327.26, size 2.596e-20, alpha -7.662, combined = 1.786e-15 rroots = 3 Skew is super-high, but I wager it'll test-sieve better than the mid-1.6 polys we have found so far; perhaps Max's spin can halve the skew without hurting score much.
Just for VBCurtis and on the $\pi$ Day!
I probably can make it higher, it's just after 3.5 min of spin.

Code:
Y0: -12601232873195375545450587523347632672883
Y1: 743324161454990349821163538
c0: 48832161596787096493393187437766185114842126397550
c1: -205648255335205404935808363185396705274975
c2: -1052812051562592448697276219220869
c3: 1475476235275407031396575
c4: 2265743403238791
c5: 1411200
skew: 596716219.75
# size 3.095e-20, alpha -7.489, combined = 2.017e-15 rroots = 3

Last fiddled with by Max0526 on 2019-03-15 at 02:40

2019-03-15, 02:39   #1719
Max0526

"Max"
Jun 2016
Toronto

70310 Posts
C206 poly

Quote:
 Originally Posted by VBCurtis Another pretty good one from msieve: Code: # norm 2.861734e-20 alpha -8.135482 e 1.641e-15 rroots 3 skew: 337025857.62 c0: -31457341673727994357991909229335026863473566329600 c1: 1331414760494355651163380007690621767575856 c2: -13625330957293497120450703031291620 c3: -53242422612500072100095716 c4: 120731530177491365 c5: 101664900 Y0: -3159942372537343778069833854769681933927 Y1: 6310295104659930539 Continuing to 105M, then I'll jump to 180M.
This one doesn't seem to spin up.

2019-03-15, 03:06   #1720
VBCurtis

"Curtis"
Feb 2005
Riverside, CA

11×409 Posts

Quote:
 Originally Posted by Max0526 Just for VBCurtis and on the $\pi$ Day! I probably can make it higher, it's just after 3.5 min of spin. Code: Y0: -12601232873195375545450587523347632672883 Y1: 743324161454990349821163538 c0: 48832161596787096493393187437766185114842126397550 c1: -205648255335205404935808363185396705274975 c2: -1052812051562592448697276219220869 c3: 1475476235275407031396575 c4: 2265743403238791 c5: 1411200 skew: 596716219.75 # size 3.095e-20, alpha -7.489, combined = 2.017e-15 rroots = 3
As far am I'm concerned, we have a winner! I'll give CADO another couple of thread-weeks this weekend, but 2e-15 on deg 5 is higher than I thought we could achieve. Just awesome, sir!
I'll test-sieve this weekend, too; Sean mentioned frmky usually uses alim = rlim = 536M for inputs this size, so I'll try that and see what yield is like.

2019-03-15, 03:21   #1721
Max0526

"Max"
Jun 2016
Toronto

19·37 Posts
C206 poly -- 2.103 and 2.107

Quote:
 Originally Posted by VBCurtis As far am I'm concerned, we have a winner! I'll give CADO another couple of thread-weeks this weekend, but 2e-15 on deg 5 is higher than I thought we could achieve. Just awesome, sir! I'll test-sieve this weekend, too; Sean mentioned frmky usually uses alim = rlim = 536M for inputs this size, so I'll try that and see what yield is like.
Hang on! I already have 2.103 and 2.107. I'll post them later tonight, before I go to bed.

Last fiddled with by Max0526 on 2019-03-15 at 03:32

2019-03-15, 04:22   #1722
Max0526

"Max"
Jun 2016
Toronto

19·37 Posts
C206 poly

Quote:
 Originally Posted by VBCurtis CADO sends warm regards for our C206: Code: R0: -12601258825545276113806892086568209032311 R1: 371662080727495174910581769 A0: 874708573977327989079213908166025283601524526316032 A1: -1016897532213206546994807128393281224899768 A2: -4765540714559627430714382742228252 A3: 2352509367951779125835982 A4: 2019390882502791 A5: 705600 skew 1201838327.26, size 2.596e-20, alpha -7.662, combined = 1.786e-15 rroots = 3 Skew is super-high, but I wager it'll test-sieve better than the mid-1.6 polys we have found so far; perhaps Max's spin can halve the skew without hurting score much.
It was actually possible to increase the E for the same c5. But the skew is the same huge and the score is too low now.
Code:
# norm 1.563823e-24 alpha -7.718866 e 1.834e-15 rroots 3
skew: 1195708201.97422
c0: 1308013145682203235921791846993890994906709837883805
c1: -2480261159146718267773058682439963878133938
c2: -3006841586003824986868630992097940
c3: 4082857057755985761817794
c4: 2670257077726791
c5: 705600
Y0: -12601190259138135484577170468444511365284
Y1: 371662080727495174910581769

 2019-03-15, 05:07 #1723 Max0526   "Max" Jun 2016 Toronto 19·37 Posts C206 polys -- the three above 2.1 The three best polys [so far ;-) ] with the E > 2.1e-15. The first and last seem to have the same alpha and number of roots. All skews are different and were optimized by http://cownoise.com/. Code: Y0: -12601188141691177666904742646406365493031 Y1: 743324161454990349821163538 c0: 80861604392230511460818572744874760814432195796792 c1: -314265373520508376909781846026746875237031 c2: -734134308227143674162327314923583 c3: 2071969217344750528293831 c4: 2690356929862791 c5: 1411200 skew: 597950211.83 # size 3.418e-20, alpha -7.633, combined = 2.111e-15 rroots = 3 Code: Y0: -12601339721892862582748708202204842976691 Y1: 743324161454990349821163538 c0: 137590190631534878358422419258553863917875388884862 c1: 164573777586510141237335590511706400855489 c2: -1450105737514750858673830964438693 c3: 464312981092984010304351 c4: 1251482803942791 c5: 1411200 skew: 758009406.84691 # lognorm 64.15, E 56.09, alpha -8.06 (proj -1.97), 1 real root # MurphyE(Bf=1.000e+07,Bg=5.000e+06,area=1.000e+16) = 2.107e-15 Code: Y0: -12601310268654152821876683800363585596851 Y1: 743324161454990349821163538 c0: 105279833837254024353264581011870638143550621803582 c1: 52174663792680045663118941643393433715009 c2: -1382245216053833380565883756313253 c3: 684822745956980314667871 c4: 1531067490022791 c5: 1411200 skew: 649347173.63761 # lognorm 64.03, E 56.39, alpha -7.63 (proj -1.97), 3 real roots # MurphyE(Bf=1.000e+07,Bg=5.000e+06,area=1.000e+16) = 2.103e-15
 2019-03-15, 08:55 #1724 swellman     Jun 2012 24·181 Posts Nicely done! I’m stopping my search - nothing comes close to these values. Good job!
2019-03-15, 13:55   #1725
cubaq

Mar 2016

5810 Posts

Quote:
 Originally Posted by Max0526 All skews are different and were optimized by http://cownoise.com/.
Does it mean, that one may change the skew value in nfs.job and nfs.fb, and sieve will produce proper results?
What kind of 'gain' i can get?
I'm using yafu v1.34.5.

Regards, cubaq

2019-03-15, 15:32   #1726
VBCurtis

"Curtis"
Feb 2005
Riverside, CA

11×409 Posts

Quote:
 Originally Posted by cubaq Does it mean, that one may change the skew value in nfs.job and nfs.fb, and sieve will produce proper results? What kind of 'gain' i can get? I'm using yafu v1.34.5. Regards, cubaq
Yes, but usually doing so will reduce sieve speed. There is a best (fastest) skew for each polynomial, and it's sometimes not the one msieve calculates. There are a couple tools out there to double-check optimal skew. We're talking potential savings of 1-3%; not worth the manual labor on a C150 but possibly a core-year of time savings at C200+.

2019-03-15, 15:42   #1727
cubaq

Mar 2016

2·29 Posts

Quote:
 Originally Posted by VBCurtis Yes, but usually doing so will reduce sieve speed. There is a best (fastest) skew for each polynomial, and it's sometimes not the one msieve calculates. There are a couple tools out there to double-check optimal skew. We're talking potential savings of 1-3%; not worth the manual labor on a C150 but possibly a core-year of time savings at C200+.

Thanks a lot for quick reply VBCurtis. As far as I am doing not 'more' than C160 it ends a case.

cubaq

Last fiddled with by cubaq on 2019-03-15 at 15:44

 Similar Threads Thread Thread Starter Forum Replies Last Post ixfd64 mersennewiki 169 2018-09-21 05:43 carpetpool Miscellaneous Math 14 2017-02-18 19:46 cheesehead Forum Feedback 6 2009-07-28 13:02 R.D. Silverman NFSNET Discussion 13 2005-09-16 20:07 TauCeti NFSNET Discussion 0 2003-12-11 22:12

All times are UTC. The time now is 09:40.

Wed Dec 2 09:40:49 UTC 2020 up 83 days, 6:51, 1 user, load averages: 3.02, 3.05, 2.47