2022 - queue management for 15e_small - Page 15

VBCurtis · 2022-10-28, 02:24

So, if I want to save the queue maintainers extra thinking, I should add "lss: 0" to my job file post when I want to be weird and run SNFS jobs on -a side?

800 bits appears close to the crossover to -r being more efficient, anyway, so I likely won't have many more -a SNFS jobs. This one isn't sieving well on the -r side because the params are heavily skewed to -a sieving. Oh well, it's a small job anyway.

swellman · 2022-10-28, 08:01

Yes, the presence of lss: 0 in the job file is the only way to sieve on the -a side with NFS@Home.

swellman · 2022-11-01, 12:27

QUEUED AS 8p3_882M

8+3,882M is a SNFS 265 from the HCN project with a polynomial identified by @charybdis and improved by @jyb.

Code:

n: 24934894035265638985483448331535184812810674246049153483668376024856858817325258228172553572964578589228377014167414480967672615471169277059629331138156850452853934263918464184552586877758437154925038099574446936668517
skew: 2.4495
type: snfs
size: 265
c6: 1
c4: -36
c3: 36
c2: 324
c1: -648
c0: 216
Y1: 2667471553700568248898162945687552
Y0: -178405961588244985132525045510417509577433411
rlim: 134000000
alim: 180000000
lpbr: 32
lpba: 32
mfbr: 94
mfba: 64
rlambda: 3.55
alambda: 2.4

Results of test sieving on the rational side with Q in blocks of 1000:

Code:

MQ       Norm_yield
35          3418
50          3792
75          3206
100         3040
150         2595
200         2296

Suggesting a sieving range for Q of 35-185M to generate 450M raw relations.

swellman · 2022-11-13, 14:59

QUEUED AS 5p3_418

5+3,418 is a SNFS 265 composite from the HCN project now ready for 15e_small.

Code:

n: 9236186365208393683761891577206220903847462932013719243913478769946715059540132016907478468379121373476441255016694432836013908670016260372325346090623829357242495475759083866218436944358054267455737635013971569514493
skew: 1.0000
type: snfs
size: 265
c5: 1
c4: -1
c3: -4
c2: 3
c1: 3
c0: -1
Y1: -491436992041778317434363998472690582275390625
Y0: 132348898008484429804225753871267183929581915600724546
rlim: 134000000
alim: 180000000
lpbr: 32
lpba: 32
mfbr: 94
mfba: 64
rlambda: 3.55
alambda: 2.4

Results of test sieving on the rational side with Q in blocks of 1000:

Code:

MQ       Norm_yield
35          2444
50          2546
75          2665
100         2456
150         2188
200         2013
235         1967

Suggesting a sieving range for Q of 35-235M to generate 455M raw relations.

RichD · 2022-11-18, 22:07

QUEUED AS 250__853_29m1

C234 from the OPN t600 file.
It was a tossup until lower Q, rlambda is better at 2.5 then 2.4.

Code:

n: 251289298873670991124760163172962673503802324282396040839061659598608795764382436043062075890803561214626248248170525190331334962227124574166303095726989250069853181547918388949865893926288141837462337121003964894291874475267164831599
# 250590853^29-1, difficulty: 251.97, skewness: 25.11, alpha: 0.00
# cost: 9.64799e+18, est. time: 4594.28 GHz days (not accurate yet!)
skew: 25.109
c6: 1
c0: -250590853
Y1: -1
Y0: 988157274738653515216440335366956921869493
type: snfs
rlim: 134000000
alim: 134000000
lpbr: 31
lpba: 31
mfbr: 62
mfba: 90
rlambda: 2.5
alambda: 3.4

Trial sieving 2K blocks.

Code:

  Q  Yield N-Yld
 18M  4073  3809
 20M  3963  3833
 60M  3117  3253
100M  2888  2851
150M  2271  2413
190M  2356  2187
194M  2149  2275

kruoli · 2022-11-20, 22:11

QUEUED AS 543_3_543p1

543·3⁵⁴³+1 is a GCW number that is ready for 15e_small to be run as SNFS.

Code:

n: 11896128255503055078424583343842305801042443520132116034869213034674980908012128697474270389819432972421954545745883894586512762202294075418852373078076117332938198458994104136835177627790746899167893732984897051866159758086694067652708761397865821827634869
skew: 0.82101
# skew by cownoise
type: snfs
size: 265
c6: 181
c0: 9
Y1: -1
Y0: 26183890704263137277674192438430182020124347
rlim: 134000000
alim: 180000000
lpbr: 32
lpba: 32
mfbr: 94
mfba: 64
lss: 1
rlambda: 3.6
alambda: 2.4

Test sieving 1K blocks gave

Code:

Q	corr. yield
35M	3259
40M	3294
70M	3324
125M	2894
200M	2330

So I am suggesting sieving Q from 40M to 190M on the rational side for a bit over 460M relations.

I would like to do the LA.

swellman · 2022-11-21, 20:21

QUEUED AS 5p4_418

5+4,418 is a SNFS job of difficulty 265 (quintic) and is ready for sieving on 15e_small:

Code:

n: 693046629991041173312270599802660096269619008887483417151952764624349322477971400340703737674599164256771952827498640989334105065193120181149246798239867594832245019166159745364611616245462629928587347100512685932602279806500206963371129078792205199601
# 5^418+4^418, difficulty: 265.61, skewness: 1.00
# cost: 2.67643e+19, est. time: 12744.93 GHz days (not accurate yet!)
skew: 1.00000
type: snfs
size: 265
c5: 1
c4: -1
c3: -4
c2: 3
c1: 3
c0: -1
Y1: -27487790694400000000000000000000000000000000000000
Y0: 132348903717475198803264914964337934368510483275127121
rlim: 134000000
alim: 180000000
lpbr: 32
lpba: 32
mfbr: 94
mfba: 64
rlambda: 3.55
alambda: 2.4

Results of test sieving on the rational side with Q in blocks of 1000:

Code:

MQ       Norm_yield
35          2271
50          2546
75          2520
100         2641
140         2338
180         2203
220         2028

Suggesting a Q range of 35-225M to generate 450M raw relations.

VBCurtis · 2022-11-22, 05:59

QUEUED AS C172_3408_1726

Aliquot sequence 3408:1726 is a C172, ready for sieving on e-small:

Code:

n: 2775040077456459928050737052674465699616973825403521271814700507670506512129157839332638537165310672787033241301533616464016528283840854775687969503039500441006989029755391
skew: 6042529.411
type: gnfs
lss: 0
c0: 39331242371199782245701027366136112126135
c1: 1539779741379458345925821717502367
c2: -5083455373259077619131836683
c3: 191825071297341953837
c4: 52533906312984
c5: 4379760
Y0: -1137015351004040622991864705196668
Y1: 23696576339906948518079
# MurphyE (Bf=4.295e+09,Bg=2.147e+09,area=2.684e+15) = 7.469e-08
# cownoise says 2.786e-13
rlim: 67000000
alim: 33500000
lpbr: 32
lpba: 32
mfbr: 62
mfba: 91
rlambda: 2.45
alambda: 3.75

Test-sieving 1kQ blocks on the -a side indicates a region from 8-44M should yield 325M+ relations, which should be enough for a decent matrix:

Code:

#Q=10M	7563 rels, 0.0352 sec/rel 47 spq
#Q=15M	8917 rels, 0.0393 sec/rel 52 spq
#Q=20M	10946 rel, 0.0432 sec/rel 65 spq
#Q=25M	12504 rel, 0.0463 sec/rel 76 spq
#Q=30M	8798 rels, 0.0477 sec/rel 54 spq
#Q=40M	8928 rels, 0.0488 sec/rel 57 spq

VBCurtis · 2022-11-22, 06:56

QUEUED AS 5_2_802m1

5*2^802-1 is ready for e-small:

Code:

n: 16192653180263380311121310759949129923101439130017465752802405040046484840490884987396600462001519515171025943628208912153988265007815843450216451754922531793300738100571679841
type: snfs
skew: 0.97
c6: 5
c0: -4
Y1: -1
Y0: 21778071482940061661655974875633165533184
rlim: 44000000
alim: 100000000
lpbr: 31
lpba: 32
mfbr: 59
mfba: 61
rlambda: 2.65
alambda: 2.4

Test-sieving 1kQ blocks on the usual -r side indicates Q from 9-60M should be enough for 240M relations:

Code:

#Q=10M	5129 (0.047 sec/rel) 61 spq
#Q=20M	5826 (0.052 sec/rel) 70 spq
#Q=30M	4858 (0.060 sec/rel) 62 spq
#Q=40M	4918 (0.055 sec/rel) 62 spq
#Q=50M	4449 (0.066 sec/rel) 60 spq
#Q=60M	3630 (0.063 sec/rel) 54 spq

I'll handle the matrix.

swellman · 2022-11-28, 12:04

QUEUED AS 3p2_557

3+2,557 is a SNFS 265 difficulty composite from the HCN project now ready for sieving on 15e_small.

Code:

n: 483180001821279491965263537959768506162202836225449813054816002904842125596253688412389640222784006267309112169275501598404357737644811248770766499768331097886879174539345150912180304425610941683084240791747518321612104404765495347150011
# 3^557+2^557, difficulty: 265.76, skewness: 0.85
# cost: 2.70575e+19, est. time: 12884.51 GHz days (not accurate yet!)
skew: 0.85028
type: snfs
size: 265
c5: 9
c0: 4
Y1: -2596148429267413814265248164610048
Y0: 91297581665113611259115979754590511595360241199911147
rlim: 134000000
alim: 180000000
lpbr: 32
lpba: 32
mfbr: 94
mfba: 64
rlambda: 3.55
alambda: 2.4

Results of test sieving on the rational side with Q in blocks of 1000:

Code:

MQ       Norm_yield
35          2480
50          2539
75          2385
100         2315
150         2213
200         2088
235         1956

Suggesting a sieving range for Q of 35-235M in order to generate 450M raw relations.

swellman · 2022-12-08, 15:24

QUEUED AS 11p7_256

11+7,256 is a SNFS 266 (quintic) from the HCN project currenly finishing up ECM. It can be sieved on 15e_small.

Code:

n: 2553981413852098572466442907173552469368292614188450511004339359650468343633796951471911300802730754162273587594690111839247019022949713234925002260501298686994374468022738547775937015563607743378915495570632585191989227207056914045976465409
# 11^256+7^256, difficulty: 266.60, skewness: 0.91
# cost: 2.87817e+19, est. time: 13705.58 GHz days (not accurate yet!)
skew: 0.91357
type: snfs
size: 266
c5: 11
c0: 7
Y1: -12589255298531885026341962383987545444758743
Y0: 129129938167664848157333145589394151932880352885953011
rlim: 134000000
alim: 180000000
lpbr: 32
lpba: 32
mfbr: 94
mfba: 64
rlambda: 3.55
alambda: 2.4

Results of test sieving on the rational side with Q in blocks of 1000:

Code:

MQ       Norm_yield
35          2218
50          2424
75          2387
100         2362
150         2024
200         1962
235         1847

Suggesting a sieving range for Q of 35-235M in order to generate ~440M raw relations.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Queue management for 14e queue	VBCurtis	NFS@Home	170	2023-01-02 15:27
2022 Queue management of 15e	swellman	NFS@Home	186	2022-12-27 12:53
Queue management for 16e queue	VBCurtis	NFS@Home	154	2022-12-23 21:35
Queue management for e_small and 15e queues	VBCurtis	NFS@Home	254	2022-01-02 01:59
Improving the queue management.	debrouxl	NFS@Home	10	2018-05-06 21:05

2022-10-28, 02:24	#155
VBCurtis "Curtis" Feb 2005 Riverside, CA 160B₁₆ Posts	So, if I want to save the queue maintainers extra thinking, I should add "lss: 0" to my job file post when I want to be weird and run SNFS jobs on -a side? 800 bits appears close to the crossover to -r being more efficient, anyway, so I likely won't have many more -a SNFS jobs. This one isn't sieving well on the -r side because the params are heavily skewed to -a sieving. Oh well, it's a small job anyway.

2022-10-28, 08:01	#156
swellman Jun 2012 F0A₁₆ Posts	Yes, the presence of lss: 0 in the job file is the only way to sieve on the -a side with NFS@Home.