2022 Queue management of 15e - Page 17

VBCurtis · 2022-12-09, 16:51

Filter and find out? Sounds like another 8-10MQ sieving would be enough.
Is 75% uniques a reasonable guess for what would have happened originally? 75% of 450M is barely higher than the uniques count you have available now, but 10M more uniques / 20M more raw relations would surely be nice.

unconnected · 2022-12-14, 23:01

QUEUED AS C194_933436_12615

C194 from aliquot sequence 933436:i12615 is ready for 15e queue:

Code:

n: 47865687229131184317692578921068303977020064182267271352643679500427297093018908790333342047904946543962939075496676005046504518963042383403247568991757829882212233001572969185349893873168471477
# norm 7.265291e-19 alpha -7.720276 e 1.264e-14 rroots 5
skew: 32322331.46
c0: -1235667978997620363556760317293364759074225405
c1: 67395016792443133559867209567626990087
c2: 12107215710907980077439344894861
c3: -385812100351898704373027
c4: -15243862655565396
c5: 43243200
Y0: -16174143659563790702590053718286883154
Y1: 710621880250857012617
rlim: 266000000
alim: 134000000
lpbr: 33
lpba: 33
mfbr: 66
mfba: 96
rlambda: 2.4
alambda: 3.6
type: gnfs
lss: 0

Thanks Gimarel for the poly!

Suggesting sieving range is 40M-480M, this should produce 940M raw relations.

swellman · 2022-12-15, 17:08

QUEUED AS 8p7_335

8+7,335 is a SNFS 242 (quartic) job from the HCN project now ready for 15e.

Code:

n: 32276654829259799863005820044869323892783528733599738092188356613385301088083231758647709484498557100935413582866528381535912363844064879879229876964653244591875662332565057810296324688814320487977811824487767365344885926526516580327626361
skew: 1.0000
type: snfs
size: 242
c4: 1
c3: -1
c2: 1
c1: -1
c0: 1
Y1: -418377847259091645147530834859099334519176045887014771543
Y0: 3213876088517980551083924184682325205044405987565585670602752
rlim: 134000000
alim: 266000000
lpbr: 33
lpba: 32
mfbr: 96
mfba: 64
rlambda: 3.7
alambda: 2.4

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

Code:

MQ       Norm_yield
35          2012
50          2381
75          2636
100         2720
150         2682
200         2501
250         2501
300         2382
350         2226

Suggesting a sieving range for Q of 35-315M to generate 710M raw relations.

swellman · 2022-12-15, 19:43

QUEUED AS C233_149_65

C233_149_65 is a SNFS 271 (quintic) job from the XYYXF project now ready to sieve.

Code:

n: 33827771786877707998384368629844995241564446369026139635560836084670339830658077524495138575963727682982560806897427849876582683209230482064400511362995052903264712918508081295685814804022993192433559876989169295099647549838672653473
skew: 2.3045
type: snfs
size: 271
c5: 1
c0: 65
Y1: -17841094326671990120557246949
Y0: 2440061088325404525157577761293434537947177886962890625
rlim: 134000000
alim: 216000000
lpbr: 32
lpba: 32
mfbr: 94
mfba: 64
rlambda: 3.65
alambda: 2.35

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

Code:

MQ       Norm_yield
35          1921
50          2039
75          1902
100         1754
150         1718
200         1555
250         1582
300         1387

Suggesting a sieving range for Q of 35-310M to generate 455M raw relations.

kruoli · 2022-12-24, 17:58

QUEUED AS 12_239m1

12,239- is a GCW number that is ready for SNFS on 15e:

Code:

n: 590371475947183936932141481376463998055530515567659946889692583925383629509906738840163571940804130694135716394737528978868209597035932913989930194285099326417638558837704463027489797118813001623243
type: snfs
size: 261
skew: 0.39225
# skew by cownoise
lss: 1
c6: 3824
c0: -3
Y1: -1
Y0: 7348857839845432252913777775075213063487488
rlim: 134000000
alim: 266000000
lpbr: 32
lpba: 32
mfbr: 94
mfba: 64
rlambda: 3.6
alambda: 2.35

Test sieving blocks of 1kQ gave:

Code:

Q	corr. yield
30M	2480
45M	2920
70M	2816
100M	2619
130M	2663
175M	2679
250M	2162

So I am suggesting sieving on the rational side from 30MQ to 200MQ for ~460M raw relations.

Additionally, I have the GCW number 4,431+, which is also ready for SNFS on 15e:

Code:

n: 3497126816854588025004943773840081937360999065176214996317320573515625396339809862379723283740638390926305862089848211239576917354805524614801471451692044883889818634016367958216754729878463681916571730642481896188319123181967883222573996348887133
type: snfs
size: 263
skew: 0.62119
# skew by cownoise
lss: 1
c6: 431
c0: 4
Y1: -1
Y0: 22300745198530623141535718272648361505980416
rlim: 134000000
alim: 266000000
lpbr: 32
lpba: 32
mfbr: 64
mfba: 64
rlambda: 2.5
alambda: 2.35

This number and the above one did not show a clear way to go regarding using 3LP after test sieving (looking at sec/rel). So I would like to have one with and one without 3LP to see how this goes in practice. Test sieving 1kQ blocks in this case gave:

Code:

Q	corr. yield
35M	2156
40M	2391
70M	2248
125M	2206
200M	1995
300M	1803

So I would suggest sieving this number on the rational side from 35MQ to 250MQ to gather ~460M raw relations.

For both, I would like to do the LA.

kruoli · 2022-12-24, 22:24

Quote:

Originally Posted by kruoli

Additionally...

There is also a second one!

swellman · 2022-12-24, 22:37

Quote:

Originally Posted by kruoli

Additionally, I have the GCW number 4,431+, which is also ready for SNFS on 15e:

Code:

n: 3497126816854588025004943773840081937360999065176214996317320573515625396339809862379723283740638390926305862089848211239576917354805524614801471451692044883889818634016367958216754729878463681916571730642481896188319123181967883222573996348887133
type: snfs
size: 263
skew: 0.62119
# skew by cownoise
lss: 1
c6: 431
c0: 4
Y1: -1
Y0: 22300745198530623141535718272648361505980416
rlim: 134000000
alim: 266000000
lpbr: 32
lpba: 32
mfbr: 64
mfba: 64
rlambda: 2.5
alambda: 2.35

This number and the above one did not show a clear way to go regarding using 3LP after test sieving (looking at sec/rel). So I would like to have one with and one without 3LP to see how this goes in practice.

Not sure I am understanding your request. Are you asking for this job to be fully run twice, one with 3 LPs and one with 2 LPs?

Or are you also involving 12,239- in this 2 LPs vs 3 LPs request as well?

kruoli · 2022-12-24, 22:49

Quote:

Originally Posted by swellman

Not sure I am understanding your request. Are you asking for this job to be fully run twice, one with 3 LPs and one with 2 LPs?

Or are you also involving 12,239- in this 2 LPs vs 3 LPs request as well?

No, the second job is a request on its own!

swellman · 2022-12-25, 01:11

Quote:

Originally Posted by kruoli

No, the second job is a request on its own!

Got it! I will hold off with enqueuing 4,431+ for a few days to give other project(s) a chance to post. I will enqueue it on Friday, sooner if other work is posted before then.

VBCurtis · 2022-12-27, 05:46

QUEUED AS 13_2_934m1

13*2^934-1 is ready for 15e, barely:

Code:

n: 251364777953893958305437346864937186512321532050545553931322907094413853004194310233445404813220825133817964918831651082094116827715192434319246167791528636582249935929661065443238576160446734691590108930947072135695412019275157
skew: 0.83
type: snfs
c6: 13
c0: -4
Y1: -1
Y0: 91343852333181432387730302044767688728495783936
rlim: 134000000
alim: 266000000
lpbr: 33
lpba: 33
mfbr: 96
mfba: 64
rlambda: 3.6
alambda: 2.35

I gave up a little yield in favor of faster sieving, but ended up with a wide sieve range.
Testing 1kQ ranges on the -r side indicates Q from 40-475M should yield ~890M relations. I plan to sieve Q=35-40M locally with CADO for another 30M relations.

Code:

#Q=40M	3542 (0.165 sec/rel) 62 spq
#Q=70M	2780 (0.181 sec/rel) 50 spq 
#Q=100M 2781 (0.235 sec/rel) 54 spq
#Q=130M 2354 (0.227 sec/rel) 52 spq
#Q=160M 2435 (0.237 sec/rel) 55 spq
#Q=190M 2308 (0.253 sec/rel) 58 spq
#Q=220M 1970 (0.275 sec/rel) 49 spq
#Q=250M 1927 (0.298 sec/rel) 51 spq
#Q=280M 1833 (0.339 sec/rel) 50 spq
#Q=310M 1753 (0.312 sec/rel) 50 spq
#Q=340M 1765 (0.346 sec/rel) 52 spq
#Q=370M 1825 (0.322 sec/rel) 53 spq
#Q=400M 1496 (0.318 sec/rel) 45 spq
#Q=430M 1546 (0.247 sec/rel) 48 spq
#Q=460M 1461 (0.296 sec/rel) 48 spq

I would like to handle the matrix.

swellman · 2022-12-27, 12:53

QUEUED AS 3m2_635

3-2,635 is a SNFS 242 quartic from the HCN project now ready for SNFS on 15e:

Code:

n: 11931722089350811658556084807108513923688082132601172856321805520673161588114993203624756608137903876037992559263047208123948929097872425676888795237210007262472465696347726163768784492547456304341468983218049075863621
# 3^635-2^635, difficulty: 242.38, skewness: 1.00
# cost: 4.61482e+18, est. time: 2197.53 GHz days (not accurate yet!)
skew: 1.00000
type: snfs
size: 242
c4: 1
c3: 1
c2: 1
c1: 1
c0: 1
Y1: -170141183460469231731687303715884105728
Y0: 3930061525912861057173624287137506221892737197425280369698987
rlim: 134000000
alim: 266000000
lpbr: 33
lpba: 32
mfbr: 96
mfba: 64
rlambda: 3.7
alambda: 2.4

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

Code:

MQ        Norm_yield
35           1962
50           2239
75           2497
100          2451
150          2576
200          2404
250          2358
300          2249
350          2194

Suggesting a sieving range for Q of 35-335M in order to generate 710M 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 for 15e_small	swellman	NFS@Home	186	2022-12-29 17:58
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

2022-12-09, 16:51	#177
VBCurtis "Curtis" Feb 2005 Riverside, CA 53·113 Posts	Filter and find out? Sounds like another 8-10MQ sieving would be enough. Is 75% uniques a reasonable guess for what would have happened originally? 75% of 450M is barely higher than the uniques count you have available now, but 10M more uniques / 20M more raw relations would surely be nice.