Fast Breeding (guru management) - Page 96

Max0526 · 2017-05-20, 03:07

@unconnected
Once again, CADO rocks! Every time you need a better one, just throw a PM at me. Will try to help.

swellman · 2017-05-23, 21:10

14e queue is running dry again. Here is a rollup of existing proposed/unqueued candidates from earlier in this thread, linked for convenience.

QUEUED C195_148_98 xyyx post 1020

QUEUED C196_122_115 xyyx post 1020

QUEUED 104281^47m OPN post 1025

QUEUED C193_131_102 (15e) xyyx post 1026

QUEUED C213_142_74 xyyx post 1034

QUEUED C206_129_95 xyyx post 1035

QUEUED C207_127_103 xyyx post 1035

QUEUED C205_125_117 xyyx post 1038

QUEUED C208_138_73 xyyx post 1038

QUEUED 25125200742702138527^13m OPN post 1044

QUEUED C165_11040_10029 Aliquot post 1045

VBCurtis · 2017-05-30, 17:23

QUEUED 1/6 1335 13*2^799-1:

Code:

# 13*2^799-1 difficulty: 241 
n: 232389108699618321557459290694950351218738646825678975874486530700896401553189617686479734599832733812040057417227643521832109054078090754653623125431200337186463151290106481591370414553915983374424635623
m: 10889035741470030830827987437816582766592
type: snfs
skew: 0.61
c6: 26
c0: -1
rlim: 110000000
alim: 110000000
lpbr: 32
lpba: 32
mfbr: 64
mfba: 64
rlambda: 2.7
alambda: 2.7

ECM'ed to half a t55.
Q of 15M to 110M should be sufficient for ~360M raw relations.
I'd like to do the post-processing when it's ready.

13*2^798 was slightly undersieved with q-range 20 to 110M; I sieved 10M relations locally to build a matrix at TD = 110. I attempted to extrapolate from previous factorizations, and only test-sieved at a single q; laziness doesn't pay, I suppose.
Test-sieved 1k blocks:

Code:

Q   yield   sec/rel
15M   4732   0.0384
30M   4727   0.0448
60M   3840   0.0545
90M   3512   0.0635
120M  3426   0.0735

swellman · 2017-05-31, 20:49

Two more candidates for 14e now ready for sieving. Both survived full t55 by yoyo@Home and a bit more @B1=3e8 by me.

QUEUED 1/6 1330 C214_123_104

Code:

n: 5166719264740945896517016191148357776632291080534010469449788562081667109333054962658097843673856429378080020663924737928712208787524079208840757268034774812593545689245880357917064575607533579431569927489851960447
# 123^104+104^123, difficulty: 253.18, anorm: 3.26e+040, rnorm: -4.44e+046
# scaled difficulty: 254.20, suggest sieving rational side
# size = 6.624e-013, alpha = 0.000, combined = 9.240e-014, rroots = 0
type: snfs
size: 253
skew: 1.0253
c6: 15129
c0: 17576
Y1: -43822462860668387010961525553205767831552
Y0: 337587917446653715596592958817679803
rlim: 240000000
alim: 240000000
lpbr: 32
lpba: 32
mfbr: 64
mfba: 64
rlambda: 2.8
alambda: 2.8

Test sieving in blocks of 5000 Q

Code:

Q=20M  total yield: 10115, q=20005003 (0.51952 sec/rel)
Q=100M total yield: 7939, q=100005001 (0.70206 sec/rel)
Q=200M total yield: 6510, q=200005007 (1.09317 sec/rel)
Q-300M total yield: 5758, q=300005039 (1.33745 sec/rel)

suggesting a range of 20-390M for Q

QUEUED 1/6 1330 C254_133_85

Code:

n: 24371997703997701483819053358015776510449740371883940079796043643360158564268707877768888155207525820198986127371298406856290019651385818690083714485756294147753130840309706077562173253165803041474008079912351622609418032599444972105863647091703823200949
# 133^85+85^133, difficulty: 258.54, anorm: 2.13e+038, rnorm: 2.70e+048
# scaled difficulty: 260.23, suggest sieving rational side
# size = 5.590e-013, alpha = 0.000, combined = 7.863e-014, rroots = 0
type: snfs
size: 258
skew: 1.0775
c6: 85
c0: 133
Y1: -541904769658563069794308330729
Y0: 2800376120856162211833149645328521728515625
rlim: 268000000
alim: 268000000
lpbr: 32
lpba: 32
mfbr: 64
mfba: 64
rlambda: 2.8
alambda: 2.8

Test sieving in blocks of 5000 Q

Code:

Q=20M  total yield: 8755,  q=20005003 (0.76666 sec/rel) 
Q=100M total yield: 7158, q=100005001 (0.99961 sec/rel) 
Q=200M total yield: 5977, q=200005007 (1.37363 sec/rel) 
Q-300M total yield: 5631, q=300005039 (1.57613 sec/rel)

suggesting a range of 20-430M for Q

unconnected · 2017-06-01, 08:48

C165_11040_10029 is 31bit task, so please add 40MQ. Also I'll reserve it for postprocessing.

RichD · 2017-06-08, 00:53

A C164 from the OPN project is ready.
I tried a/rlim of 134M but it was worse. 20-25% slower while gaining <10% more yield.
Maybe values of 80M or 100M might be better which I didn't try because this looks pretty good.

Code:

n: 48617051659411879901148500183191173740459719818818239572906825331510589697540739558361422095833883018398491446348980638157678408112092901699165057691650922870794961
# 12400411646533^17-1, GNFS-164, sieve on algebraic side
lss: 0
Y0: -39473793098425407586702362013059
Y1: 261219411908353
c0: 36414268145373788770115624216770967668
c1: 42969519334008936189506779266008
c2: -125646566348823834900729739
c3: -10224912539480172989
c4: 5332389399891
c5: 507276
skew: 4174573.51659
rlim: 67000000
alim: 67000000
lpbr: 31
lpba: 31
mfbr: 62
mfba: 62
rlambda: 2.6
alambda: 2.6

Trial sieving 5K blocks.

Code:

  Q  Yield
 20M 12853
 60M 14816
100M 12710

QUEUED 8/6 1300

RichD · 2017-06-09, 14:35

SNFS-252 job is ready from OPN.

Code:

n: 2628443707681354167068187357307406276224836247570952408476253944812890672190460092817686395342995854169493921586010588285093342807630218122037987668109546335074209003822352304893720480661143
# 179743^47-1, difficulty: 252.22, skewness: 7.51, alpha: 0.00
# cost: 9.83625e+18, est. time: 4683.93 GHz days (not accurate yet!)
skew: 7.512
c6: 1
c0: -179743
Y1: -1
Y0: 1089471535698978337428627067875381019180801
m: 1089471535698978337428627067875381019180801
type: snfs
alim: 268000000
rlim: 268000000
lpbr: 32
lpba: 32
mfbr: 64
mfba: 64
rlambda: 2.8
alambda: 2.8

Trial sieving 5K blocks.

Code:

swellman · 2017-06-14, 09:15

Two more candidates for 14e

C215_145_52

Code:

n: 28517792151961169713334701529683994469920634832298395618828402880211937617714358390875895038921867809408730936364349284979531990120923644890094872165509528815503746898447469462582487316122189734747938969600125124951
# 145^52+52^145, difficulty: 248.82, anorm: 2.90e+032, rnorm: 2.15e+055
# scaled difficulty: 252.63, suggest sieving rational side
# size = 2.395e-017, alpha = 0.000, combined = 9.207e-014, rroots = 1
type: snfs
size: 248
skew: 7.3206
c5: 1
c0: 21025
Y1: -4108469075197275390625
Y0: 58089409991159212052332855232011755388709594202112
rlim: 240000000
alim: 240000000
lpbr: 32
lpba: 32
mfbr: 64
mfba: 64
rlambda: 2.8
alambda: 2.8

Test sieving with Q in blocks of 5000

Code:

Q=20M 6501 rels
Q=100M 8277 rels
Q=200M 8326 rels
Q=300M 7750 rels
Q=400M 6879 rels

suggesting a sieving range of 20-340M for Q

C216_143_53 - note that test sieving indicates sieving on the -a side is best

Code:

n: 314653115809307951898713853390593679365304026849878683717647306448540598024840725423941604678327491979309970558045182675162419704208615919288865579283623418810465733214078622746767013840787080771222275807872799719543
# 143^53+53^143, difficulty: 249.41, anorm: 2.11e+040, rnorm: 7.90e+046
# scaled difficulty: 250.51, suggest sieving algebraic side
# size = 6.352e-013, alpha = 0.000, combined = 8.929e-014, rroots = 0
type: snfs
size: 249
skew: 9.3229
c6: 13
c0: 8535703
Y1: -2273168619161481613
Y0: 241335311011519234780052665404754645838881
rlim: 240000000
alim: 240000000
lpbr: 32
lpba: 32
mfbr: 64
mfba: 64
rlambda: 2.8
alambda: 2.8

Test sieving with Q in blocks of 5000

Code:

Q=20M 12022 rels
Q=100M 10368 rels
Q=200M 7240 rels
Q=300M 5871 rels
Q=400M 6022 rels

suggesting a sieving range of 20-325M for Q

VBCurtis · 2017-06-21, 05:43

13*2^800-1 has been ECMed to ~0.6t55, and is ready for 14e:

Code:

#13*2^800-1 difficulty: 242
n:  93279338268454023866245440870308237806520338071962741869336774244320823422643775297036986460002650705701592500634862028121070426777291354105093480717237604707298542950947863172273537595119
m: 10889035741470030830827987437816582766592
type: snfs
skew: 0.53
c6: 52
c0: -1
rlim: 110000000
alim: 110000000
lpbr: 32
lpba: 32
mfbr: 64
mfba: 64
rlambda: 2.7
alambda: 2.7

I didn't test sieve, but extrapolating from 13_m_798 and 13_m_799 suggests a Q range of 15 to 120M should be sufficient. I'd like to do the LA.

chris2be8 · 2017-06-22, 16:11

If I was doing 13*2^800-1 I'd test sieve a degree 5 poly like the following to see it the lower coefficients beat the better degree.

Code:

#13*2^800-1 difficulty: 242
n: 75140219816134685225301475917061127924425548477298465135977357790861112720590532493747026331535493816353613052814667153290078550537040542030860937761702125774722744583825144915771505144999160066933624361096003812212466492784977
# m = 2^160
m: 1461501637330902918203684832716283019655932542976
type: snfs
c5: 13
c0: -1

NB. Are you sure the n in the poly you posted is right? My checking script says the poly does not evaluate to a multiple of n.

Chris

VBCurtis · 2017-06-22, 17:12

No, not certain; I may have miscopied something. I'll do some test-sieving this evening and confirm settings/Q-ranges.
I've been factoring these numbers since about 2^600; deg 6 became faster than deg 5 around 2^660.

2017-05-23, 21:10	#1047
swellman Jun 2012 11×281 Posts	14e queue is running dry again. Here is a rollup of existing proposed/unqueued candidates from earlier in this thread, linked for convenience. QUEUED C195_148_98 xyyx post 1020 QUEUED C196_122_115 xyyx post 1020 QUEUED 104281^47m OPN post 1025 QUEUED C193_131_102 (15e) xyyx post 1026 QUEUED C213_142_74 xyyx post 1034 QUEUED C206_129_95 xyyx post 1035 QUEUED C207_127_103 xyyx post 1035 QUEUED C205_125_117 xyyx post 1038 QUEUED C208_138_73 xyyx post 1038 QUEUED 25125200742702138527^13m OPN post 1044 QUEUED C165_11040_10029 Aliquot post 1045 Last fiddled with by fivemack on 2017-06-08 at 12:01 Reason: all queued now (lunchtime 8th June 2017)

2017-05-30, 17:23	#1048
VBCurtis "Curtis" Feb 2005 Riverside, CA 11×443 Posts	QUEUED 1/6 1335 132^799-1: Code: # 132^799-1 difficulty: 241 n: 232389108699618321557459290694950351218738646825678975874486530700896401553189617686479734599832733812040057417227643521832109054078090754653623125431200337186463151290106481591370414553915983374424635623 m: 10889035741470030830827987437816582766592 type: snfs skew: 0.61 c6: 26 c0: -1 rlim: 110000000 alim: 110000000 lpbr: 32 lpba: 32 mfbr: 64 mfba: 64 rlambda: 2.7 alambda: 2.7 ECM'ed to half a t55. Q of 15M to 110M should be sufficient for ~360M raw relations. I'd like to do the post-processing when it's ready. 132^798 was slightly undersieved with q-range 20 to 110M; I sieved 10M relations locally to build a matrix at TD = 110. I attempted to extrapolate from previous factorizations, and only test-sieved at a single q; laziness doesn't pay, I suppose. Test-sieved 1k blocks: Code: Q yield sec/rel 15M 4732 0.0384 30M 4727 0.0448 60M 3840 0.0545 90M 3512 0.0635 120M 3426 0.0735 Last fiddled with by fivemack on 2017-06-01 at 12:32*

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
System management notes	kriesel	kriesel	7	2020-10-21 18:52
Improving the queue management.	debrouxl	NFS@Home	10	2018-05-06 21:05
Script-based Primenet assignment management	ewmayer	Software	3	2017-05-25 04:02
Do normal adults give themselves an allowance? (...to fast or not to fast - there is no question!)	jasong	jasong	35	2016-12-11 00:57
Power Management settings	PrimeCroat	Hardware	3	2004-02-17 19:11

2017-05-20, 03:07	#1046
Max0526 "Max" Jun 2016 Toronto 893₁₀ Posts	@unconnected Once again, CADO rocks! Every time you need a better one, just throw a PM at me. Will try to help.

2017-06-01, 08:48	#1050
unconnected May 2009 Russia, Moscow 5041₈ Posts	C165_11040_10029 is 31bit task, so please add 40MQ. Also I'll reserve it for postprocessing.

2017-06-22, 17:12	#1056
VBCurtis "Curtis" Feb 2005 Riverside, CA 11×443 Posts	No, not certain; I may have miscopied something. I'll do some test-sieving this evening and confirm settings/Q-ranges. I've been factoring these numbers since about 2^600; deg 6 became faster than deg 5 around 2^660.