Planning & Coordination for 2,2330L - Page 8

vebis · 2019-04-28, 18:55

Do we settle with one of the found polys?

VBCurtis · 2019-04-28, 21:22

The msieve folks haven't had very much time; for a job that takes ~50 CPU-years to sieve, a GPU-year of poly select is a reasonable target. I'm sure we don't have 12 GPUs searching, so a month isn't long enough.

I haven't gotten around to any CADO search myself, either. There is much left to do- ECM pretesting isn't yet complete, either.

vebis · 2019-04-29, 19:43

I put 1200 core-days into the polysearch, evenly splitted between deg-5 and deg-6. I don‘t know how this converts to gpu years.

Max0526 · 2019-04-30, 23:45

@vebis
My bet is we will not get anything better than 1.763e-15 for degree 5 poly. Your original poly after 600 core-days was pretty strong and we got very lucky with the spin (+13.5%).
In degree 6 that probably means E > 2.1e-15, and I have a hard time crossing a 2-handle line with all your great degree 6 candidates.
Patiently waiting for the test-sieve results...

RichD · 2019-05-01, 00:49

My best score so far is 1.3 with degree 5. Continuing with 33M with msieve-GPU.

VBCurtis · 2019-05-01, 02:18

Quote:

Originally Posted by Max0526

@vebis
My bet is we will not get anything better than 1.763e-15 for degree 5 poly. Your original poly after 600 core-days was pretty strong and we got very lucky with the spin (+13.5%).
In degree 6 that probably means E > 2.1e-15, and I have a hard time crossing a 2-handle line with all your great degree 6 candidates.
Patiently waiting for the test-sieve results...

I don't think CADO's default poly select settings are very good, and have two ideas I wish to try for improvement. I won't put nearly the resources in that Vebis used, but I want to see how competitive these ideas are. I'm in midterms right now; give me a week or so to get at least one of them tested, and then I'll also have time for test-sieving. Hopefully this 1.76 deg 5 will sieve 10% or more faster than the deg 6 tests I did, leading us below 50 core-years of expected sieving.

Gimarel · 2019-05-02, 15:05

I've run GPU polyselect for 3 days now. I have nothing competitive to 1.7e-15 yet but 4 polys above 1.38e-15. Maybe one of them is spinable.

Code:

# norm 2.268219e-20 alpha -9.159842 e 1.406896e-15 rroots 5
skew: 537137182.72
c0: 201136016414882404020745322408968491381394661407872
c1: 18443567562394707400723332615212125571164956
c2: -126808754424960017662126588751357314
c3: -63508439009354339473675934
c4: 402712014124143779
c5: 200840640
Y0: -4408360209867911475011006261136065163329
Y1: 208056570822173619719

# norm 2.250907e-20 alpha -8.254714 e 1.398482e-15 rroots 5
skew: 203737086.95
c0: -39023279351822263757649392196439290158044668030800
c1: 234293248880316954011172219691584603724020
c2: 20374363588896664401734214803072256
c3: 49887809355389561953368353
c4: -565682664212116068
c5: 201801600
Y0: -4404153751761832881906341865269260577461
Y1: 100036617348797298821

# norm 2.233059e-20 alpha -9.090323 e 1.387992e-15 rroots 5
 skew: 781890408.98
c0: 6632916566231902971415991145436233511882533517078304
c1: 35063000149988900390567210762645058368136750
c2: -140156630347139255242856341696624655
c3: -71050822348447487543054190
c4: 243561083025356676
c5: 100540440
Y0: -5062665536312230975791209695650674433165
Y1: 367519913313526297889

# norm 2.169643e-20 alpha -7.356187 e 1.384190e-15 rroots 5
skew: 256204300.06
c0: -18923777149059534087726928025111736809726687968080
c1: 1055338864201618793242848967990140672188697
c2: -3663854191651275199601669633022108
c3: -53329572334584775812110477
c4: 34960671080784883
c5: 201201000
Y0: -4406779963316103452680276509973239820313
Y1: 210111803588673209159

yoyo · 2019-05-03, 06:09

Quote:

Originally Posted by VBCurtis

There is much left to do- ECM pretesting isn't yet complete, either.

yoyo@home was voted for the Pentathlon (https://www.seti-germany.de/boinc_pentathlon/).
Start: 06 May 2019 00:00 UTC
End: 11 May 2019 00:00 UTC

I assume that the remaining ECM curves will be processed during this race.

yoyo

Max0526 · 2019-05-04, 00:28

Quote:

Originally Posted by Gimarel

I've run GPU polyselect for 3 days now. I have nothing competitive to 1.7e-15 yet but 4 polys above 1.38e-15. Maybe one of them is spinable.

Code:

# norm 2.268219e-20 alpha -9.159842 e 1.406896e-15 rroots 5
...
# norm 2.250907e-20 alpha -8.254714 e 1.398482e-15 rroots 5
...
# norm 2.233059e-20 alpha -9.090323 e 1.387992e-15 rroots 5
...
# norm 2.169643e-20 alpha -7.356187 e 1.384190e-15 rroots 5

The top two quick-spin into

Code:

Y0: -4408360209869090258211521458702876425563
Y1: 416113141644347239438
c0: 11572744639136622516898971054433671073624146401130
c1: 4968519235395260371890442335727074610993419
c2: -62826050370389033853709695230989919
c3: -72570528450051070906131110
c4: 794045028225497158
c5: 803362560
skew: 278646086.03
# size 1.878e-20, alpha -9.160, combined = 1.417e-15 rroots = 5

Code:

Y0: -4404153751761648772515040791744564432641
Y1: 200073234697594597642
c0: -4815345740956093702696476648186899246607549975000
c1: 77445214858216490262158987142423472342665
c2: 10319161718573132419107345056292918
c3: 45730269914283618960294113
c4: -1127651331417512136
c5: 807206400
skew: 106594367.02
# size 1.863e-20, alpha -8.255, combined = 1.408e-15 rroots = 5

vebis · 2019-05-05, 19:20

VBCurtis gave me 2 parameter files for polyselection. This is the top poly of the first file from cado and as an attachment the other top80.

Code:

n: 334377437706404684733884220732190564550147039308686365251932207592010868660163280944362994773042289837865752943086095333357861608005044927356022538834108845220872215597190097852981295940487497385774178731621
skew: 457883977.784
c0: -134290292880687501661256572740494690644774086231345
c1: -12796351748956634711765230605498525115012
c2: 7475858736173889682062354826325821
c3: -2446745267681131688984128
c4: 6942146306860384
c5: 9527280
Y0: -8110610416048511788085122669244949919058
Y1: 8488432067100201108680911
# MurphyE (Bf=3.436e+10,Bg=1.718e+10,area=5.469e+16) = 1.54e-08
# found by revision 9436c2ff8
# f(x) = 9527280*x^5+6942146306860384*x^4-2446745267681131688984128*x^3+7475858736173889682062354826325821*x^2-12796351748956634711765230605498525115012*x-134290292880687501661256572740494690644774086231345
# g(x) = 8488432067100201108680911*x-8110610416048511788085122669244949919058

VBCurtis · 2019-05-06, 00:54

cownoise tells me the score for this one is 1.463e-15.

No better than the one your run found, but Max may have a look-and-spin.

Thanks for trying!

2019-04-29, 19:43	#80
vebis Oct 2015 2²·17 Posts	I put 1200 core-days into the polysearch, evenly splitted between deg-5 and deg-6. I don‘t know how this converts to gpu years. Last fiddled with by vebis on 2019-04-29 at 19:45

2019-04-30, 23:45	#81
Max0526 "Max" Jun 2016 Toronto 19×47 Posts	C207 poly score @vebis My bet is we will not get anything better than 1.763e-15 for degree 5 poly. Your original poly after 600 core-days was pretty strong and we got very lucky with the spin (+13.5%). In degree 6 that probably means E > 2.1e-15, and I have a hard time crossing a 2-handle line with all your great degree 6 candidates. Patiently waiting for the test-sieve results...

2019-05-01, 00:49	#82
RichD Sep 2008 Kansas 17×199 Posts	C207 Poly My best score so far is 1.3 with degree 5. Continuing with 33M with msieve-GPU.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Coordination thread for redoing P-1 factoring	ixfd64	Lone Mersenne Hunters	81	2021-04-17 20:47
big job planning	henryzz	Cunningham Tables	16	2010-08-07 05:08
Sieving reservations and coordination	gd_barnes	No Prime Left Behind	2	2008-02-16 03:28
Sieved files/sieving coordination	gd_barnes	Conjectures 'R Us	32	2008-01-22 03:09
Special Project Planning	wblipp	ElevenSmooth	2	2004-02-19 05:25

2019-04-28, 18:55	#78
vebis Oct 2015 2²×17 Posts	Do we settle with one of the found polys?

2019-04-28, 21:22	#79
VBCurtis "Curtis" Feb 2005 Riverside, CA 11400₈ Posts	The msieve folks haven't had very much time; for a job that takes ~50 CPU-years to sieve, a GPU-year of poly select is a reasonable target. I'm sure we don't have 12 GPUs searching, so a month isn't long enough. I haven't gotten around to any CADO search myself, either. There is much left to do- ECM pretesting isn't yet complete, either.

2019-05-06, 00:54	#88
VBCurtis "Curtis" Feb 2005 Riverside, CA 2⁸·19 Posts	cownoise tells me the score for this one is 1.463e-15. No better than the one your run found, but Max may have a look-and-spin. Thanks for trying!