2022 - queue management for 15e_small - Page 10

swellman · 2022-08-09, 16:30

I’ve enqueued this as posted (with some slight reordering of the lines), but are you sure about 3 LPs on the rational side while sieving on the algebraic side? It can be changed until it is thrown into sieving status.

If it test sieved best using these parameters then so be it but it looked odd to me. But we don’t see many factorizations of pure Cullen numbers anymore, maybe they’re a different beast?

kruoli · 2022-08-09, 16:50

Yes! This was my raw data (rational/algebraic/sieve side):

Code:

22A: total yield: 2312, q=100001029 (0.21729 sec/rel)
     58 Special q, 396 reduction iterations
22R: total yield: 2104, q=100001029 (0.21938 sec/rel)
     54 Special q, 384 reduction iterations
23A: total yield: 2882, q=100001029 (0.20435 sec/rel)
     58 Special q, 396 reduction iterations
23R: total yield: 2746, q=100001029 (0.23284 sec/rel)
     54 Special q, 384 reduction iterations
32A: total yield: 3017, q=100001029 (0.22544 sec/rel)
     58 Special q, 396 reduction iterations
32R: total yield: 2614, q=100001029 (0.21088 sec/rel)
     54 Special q, 384 reduction iterations
33A: total yield: 3755, q=100001029 (0.99578 sec/rel)
     58 Special q, 396 reduction iterations
33R: total yield: 3414, q=100001029 (1.01115 sec/rel)
     54 Special q, 384 reduction iterations

Edit, normalized:

Code:

swellman · 2022-08-09, 16:59

Quote:

Originally Posted by kruoli

Yes! This was my raw data (rational/algebraic/sieve side):

Code:

22A: total yield: 2312, q=100001029 (0.21729 sec/rel)
     58 Special q, 396 reduction iterations
22R: total yield: 2104, q=100001029 (0.21938 sec/rel)
     54 Special q, 384 reduction iterations
23A: total yield: 2882, q=100001029 (0.20435 sec/rel)
     58 Special q, 396 reduction iterations
23R: total yield: 2746, q=100001029 (0.23284 sec/rel)
     54 Special q, 384 reduction iterations
32A: total yield: 3017, q=100001029 (0.22544 sec/rel)
     58 Special q, 396 reduction iterations
32R: total yield: 2614, q=100001029 (0.21088 sec/rel)
     54 Special q, 384 reduction iterations
33A: total yield: 3755, q=100001029 (0.99578 sec/rel)
     58 Special q, 396 reduction iterations
33R: total yield: 3414, q=100001029 (1.01115 sec/rel)
     54 Special q, 384 reduction iterations

Ok, thank you for the confirmation! It’s already queued and I’ll move it to sieving status later tonight.

charybdis · 2022-08-09, 17:24

Quote:

Originally Posted by swellman

I’ve enqueued this as posted (with some slight reordering of the lines), but are you sure about 3 LPs on the rational side while sieving on the algebraic side? It can be changed until it is thrown into sieving status.

This does make sense for numbers where the algebraic and rational norms are very similar in size. The special-q essentially gives you a large prime for free on the side that you're sieving on, so this is morally 3 large primes on each side.

swellman · 2022-08-11, 15:11

QUEUED AS 11p8_254

11+8,254 is a SNFS 264 composite from the HCN project ready for sieving on 15e_small.

Code:

n: 124770258398178107242131430473369370375510834286483169003880299586574202593247463580711485332956285974313251948273561177793178361754355619422761510174288177249352768474487586257335461086084765626108729190558791341
type: snfs
size: 264
skew: 0.8993
c6: 121
c0: 64
Y1: -85070591730234615865843651857942052864
Y0: 54763699237492901685126120802225273763666521
rlim: 134000000
alim: 134000000
lpbr: 32
lpba: 32
mfbr: 94
mfba: 64
rlambda: 3.5
alambda: 2.8

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

Code:

MQ       Norm_yield      Speed (sec/rel)
35          3381              0.151
50          3291              0.163
75          2879              0.194
100         2654              0.211
140         2493              0.241
180         2183              0.273
205         1909              0.310

Suggesting a sieving range for Q of 35-205M to generate 445M raw relations.

Aiming for fewer raw relations (old target was ~460M) due to observations by @VBCurtis. I’ll wait for Yoyo to wrap it up before enqueuing the job.

RichD · 2022-08-22, 15:14

QUEUED AS 51113_250

C251 from the Near-Repdigit project, 51113_250.
The SNFS difficulty is 250.71-digit.

Code:

n: 51111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111113
m: 1000000000000000000000000000000000000000000
c6: 23
c0: 850
skew: 1.83
type: snfs
rlim: 134000000
alim: 134000000
lpbr: 31
lpba: 31
mfbr: 62
mfba: 62
rlambda: 2.7
alambda: 2.7

Trial sieving 2K blocks.

Code:

  Q  Yield N-Yld
 20M  5472  5293
 60M  4298  4486
100M  4100  4047
140M  3390  3651

RichD · 2022-08-24, 12:04

QUEUED AS 21111_249

C250 from the Near-Repdigit project, 21111_249.
The SNFS difficulty is 250.28-digit.

Code:

n: 2111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111
m: 100000000000000000000000000000000000000000
c6: 19000
c0: -1
skew: 0.19
type: snfs
rlim: 134000000
alim: 134000000
lpbr: 31
lpba: 31
mfbr: 62
mfba: 62
rlambda: 2.7
alambda: 2.7

Trial sieving 2K blocks.

Code:

  Q  Yield N-Yld
 20M  6456  6245
 60M  4887  5100
100M  4462  4404
124M  4422  4274

swellman · 2022-08-26, 11:48

QUEUED AS 11p3_254

11+3,254 is a SNFS 264 from the HCN project now ready for sieving on 15e_small.

Code:

n: 67424057153850060120677779598359258132056887148905399271362551582217879519699108422784725600826084662672009166900497025803379070453570749977082733500637054989796931680909432338535168344460673746876997654857923342528013856780638780077430294609703515749
skew: 0.6485
type: snfs
size: 264
c6: 121
c0: 9
Y1: -109418989131512359209
Y0: 54763699237492901685126120802225273763666521
rlim: 134000000
alim: 266000000
lpbr: 32
lpba: 32
mfbr: 94
mfba: 64
rlambda: 3.5
alambda: 2.8

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

Code:

MQ       Norm_yield      Speed (sec/rel)
35          4104              0.226
50          4045              0.235
75          3616              0.265
100         3421              0.307
150         3092              0.339
200         2549              0.388

Suggesting a sieving range for Q of 35-165M in order to generate 455M raw relations.

bur · 2022-09-01, 07:33

edit: I looked up this old factorisation, so I'll just use those parameters.

I am looking into test sieving for a C174 of aliquot 1992. Can I just use the CADO parameters for GGNFS?

CADO has:

Code:

tasks.I = 14 tasks.qmin = 39200000 tasks.lim0 = 48100000 tasks.lim1 = 67600000 tasks.lpb0 = 30 tasks.lpb1 = 31 tasks.sieve.mfb0 = 60 tasks.sieve.mfb1 = 90 tasks.sieve.ncurves0 = 15 tasks.sieve.ncurves1 = 8

It seems GGNFS needs lambda specified, is it like that and if so, which value would be good?

swellman · 2022-09-01, 11:30

Quote:

Originally Posted by bur

edit: I looked up this old factorisation, so I'll just use those parameters.

I am looking into test sieving for a C174 of aliquot 1992. Can I just use the CADO parameters for GGNFS?

CADO has:

Code:

tasks.I = 14 tasks.qmin = 39200000 tasks.lim0 = 48100000 tasks.lim1 = 67600000 tasks.lpb0 = 30 tasks.lpb1 = 31 tasks.sieve.mfb0 = 60 tasks.sieve.mfb1 = 90 tasks.sieve.ncurves0 = 15 tasks.sieve.ncurves1 = 8

It seems GGNFS needs lambda specified, is it like that and if so, which value would be good?

I use lambda values based strictly on lpb, namely (lpb,lambda) pairs are as follows:

Code:

(29,2.5)
(30,2.6)
(31,2.7)
(32,2.8)
(33,3.0)
(34,3.2)

Not terribly insightful but it seems to work. The old GGNFS documentation lacked a lot of detail, though it’s been years since I read it.

charybdis · 2022-09-01, 14:05

Quote:

Originally Posted by bur

It seems GGNFS needs lambda specified, is it like that and if so, which value would be good?

Lambda should be set to log(2^mfb)/log(lim) + 0.2-0.3. It doesn't matter what base you take the log to, it'll give the same result.

2022-08-09, 16:30	#100
swellman Jun 2012 2⁴×3×5×17 Posts	Cullen 849 I’ve enqueued this as posted (with some slight reordering of the lines), but are you sure about 3 LPs on the rational side while sieving on the algebraic side? It can be changed until it is thrown into sieving status. If it test sieved best using these parameters then so be it but it looked odd to me. But we don’t see many factorizations of pure Cullen numbers anymore, maybe they’re a different beast? Last fiddled with by swellman on 2022-08-09 at 16:31

2022-09-01, 07:33	#108
bur Aug 2020 796581e-4;32539e-3 733 Posts	edit: I looked up this old factorisation, so I'll just use those parameters. I am looking into test sieving for a C174 of aliquot 1992. Can I just use the CADO parameters for GGNFS? CADO has: Code: tasks.I = 14 tasks.qmin = 39200000 tasks.lim0 = 48100000 tasks.lim1 = 67600000 tasks.lpb0 = 30 tasks.lpb1 = 31 tasks.sieve.mfb0 = 60 tasks.sieve.mfb1 = 90 tasks.sieve.ncurves0 = 15 tasks.sieve.ncurves1 = 8 It seems GGNFS needs lambda specified, is it like that and if so, which value would be good? Last fiddled with by bur on 2022-09-01 at 07:59

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