mersenneforum.org 2022 - queue management for 15e_small
 Register FAQ Search Today's Posts Mark Forums Read

 2022-09-02, 03:37 #111 VBCurtis     "Curtis" Feb 2005 Riverside, CA 15AF16 Posts The catch to using Charybdis' formula is that GGNFS cannot sieve Q values below lim, so it reduces lim to Q-1 for each sieve process. If you use small starting value for Q relative to lim, lambda should be a little higher (like a tenth) than otherwise because the lim used by GGNFS can be quite small. If lambda is a bit too low, you'll miss out on some yield at small Q values but not at larger Q values- so, not too big a deal. Edit: I tried a test case, and I'm not sure the formula works (or maybe my arithmetic is bad)... If I use lim of 134M = 2^27 and LPB of 2^32, what does the formula suggest for lambda? Last fiddled with by VBCurtis on 2022-09-02 at 03:40
2022-09-02, 14:45   #112
charybdis

Apr 2020

15538 Posts

Quote:
 Originally Posted by VBCurtis Edit: I tried a test case, and I'm not sure the formula works (or maybe my arithmetic is bad)... If I use lim of 134M = 2^27 and LPB of 2^32, what does the formula suggest for lambda?
The formula uses mfb, not lpb: lambda determines which cofactors get calculated after sieving.

 2022-09-02, 17:53 #113 VBCurtis     "Curtis" Feb 2005 Riverside, CA 7×13×61 Posts Thanks for correcting my mistake. :)
 2022-09-06, 04:35 #114 richs     "Rich" Aug 2002 Benicia, California 2·17·47 Posts QUEUED AS C167_230916_6623 GNFS Aliquot C167 230916 i6623, ECM'ed to t51.4, 50 cpu thread-hours poly search, for 15e small: Code: # expecting degree 5 poly E from 3.57e-13 to > 4.10e-13 # norm 3.220878e-16 alpha -7.846393 e 5.131e-13 rroots 3 n: 68480980185965589250694154773723855827100479692664182650843065204557192011901064193836421108381794510505559718483397542729522676327402943506853535111599751262500151689 skew: 69820427.40557 c0: -1547545381426193518134376127334202552733511 c1: 210609628800511515576857126330306105 c2: -4433164644746260507425121055 c3: -378099455992993835245 c4: 1016158879626 c5: 10200 Y0: -367615617529328446474038213787280 Y1: 248517663147240641 rlim: 67000000 alim: 67000000 lpbr: 30 lpba: 30 mfbr: 60 mfba: 60 rlambda: 2.6 alambda: 2.6 lss: 0 Sieve on algebraic side Q = 10-75M. I will handle LA. Last fiddled with by swellman on 2022-09-06 at 09:51
 2022-09-06, 12:26 #115 kruoli     "Oliver" Sep 2017 Porta Westfalica, DE 1,223 Posts QUEUED AS 4_430m GCW 4,430- is an SNFS 262 which is ready for sieving. Code: n: 6894695061411733701376557556749898283209534920194547993396723084115159294060781479138699943141703038260439008560537252142483105584413625732899522723438562874394768578944986146521719768453038558103716269466523346928451842163438078096585501487785848666983647 skew: 0.37405 type: snfs size: 262 c6: 1720 c0: -1 Y0: -11150372599265311570767859136324180752990208 Y1: 1 rlim: 90000000 alim: 178000000 lpbr: 31 lpba: 31 mfbr: 91 mfba: 62 lss: 1 rlambda: 3.8 alambda: 2.7 To be sieved on the rational side. Test sieving blocks of 1k gave Code:  Q norm. yield 40M 2011 70M 1875 125M 1519 200M 1440 Therefore I suggest sieving Q 40M-200M for 260M+ relations. I will do the LA. Last fiddled with by swellman on 2022-09-06 at 12:42
 2022-09-07, 08:08 #116 bur     Aug 2020 79*6581e-4;3*2539e-3 10010110012 Posts QUEUED AS C174_1992_1695. Congratulations on new record Escore for a C174! GNFS Aliquot C174 1992 i1695, ECM'ed to t55, record size poly, for 15e small: Code: # cownoise 2.329e-13 n: 470591531631348035122835952410646926317873637893205194615682356577590330702553555628775500411575272153087457823487296743846900868386365774458010976374638888143341362679360261 skew: 1894776 Y0: -2772867613071987835179566649293458 Y1: 556525366582386152183 c0: 244760256581114438455807516175342815140 c1: 297459705653457167451974927405829 c2: -974523754475295166457644004 c3: 76633000667711098507 c4: 305333259610830 c5: 17224560 rlim: 134000000 alim: 134000000 lpbr: 31 lpba: 31 mfbr: 62 mfba: 91 rlambda: 2.7 alambda: 3.4 Algebraic side test sieving in 1k chunks: Code: MQ norm_yield 25 3647 50 3791 75 3759 100 3498 125 3366 If I'm correct, sieving Q from 25M to 100M should comfortably yield 260M raw relations. Last fiddled with by swellman on 2022-09-07 at 10:42
 2022-09-09, 21:13 #117 RichD     Sep 2008 Kansas 22×3×313 Posts QUEUED AS 34210243_31m1 C234 from the OPN t800 file. Code: n: 10512896323946525328323590051713030182707488022156443686427827035707600086867033903192945127273531284579440671455984668440538086981016196248405396732898676231093060782972818311914352957505254809179243 # 34210243^31-1, difficulty: 233.56, skewness: 0.06, alpha: 0.00 # cost: 2.30621e+18, est. time: 1098.20 GHz days (not accurate yet!) skew: 18.02 c6: 1 c0: -34210243 Y1: 46857681410903935241309012906079659443 Y0: -1 type: snfs rlim: 67000000 alim: 67000000 lpbr: 30 lpba: 30 mfbr: 60 mfba: 87 rlambda: 2.6 alambda: 3.3 Trial sieving 2K blocks. Code:  Q Yield N-Yld 18M 4381 4097 20M 4420 4275 50M 3103 3271 80M 2345 2713 100M 2563 2226 Estimated total yield from 18-100M is 130M relations. Last fiddled with by swellman on 2022-09-09 at 21:27
2022-09-09, 21:24   #118
charybdis

Apr 2020

53·7 Posts

Quote:
 Originally Posted by RichD Code: rlim: 67000000 alim: 67000000 lpbr: 30 lpba: 30 mfbr: 60 mfba: 87 rlambda: 2.6 alambda: 3.3
Sorry to be the lambda police again.
log(2^87)/log(67000000) = 3.35... try setting alambda to 3.5 or 3.6.

2022-09-10, 02:30   #119
RichD

Sep 2008
Kansas

22·3·313 Posts

Quote:
 Originally Posted by charybdis Sorry to be the lambda police again. log(2^87)/log(67000000) = 3.35... try setting alambda to 3.5 or 3.6.
No problem. lambda is the least understood parameter in my knowledge suite.

This does seem to be favorable. I won't have time tonight to fully test all the variants.

@Sean: Feel free to pull it before it goes active.

2022-09-10, 02:44   #120
swellman

Jun 2012

3,767 Posts

Quote:
 Originally Posted by RichD No problem. lambda is the least understood parameter in my knowledge suite. This does seem to be favorable. I won't have time tonight to fully test all the variants. @Sean: Feel free to pull it before it goes active.
No problem. FYI it’s a manual process by me or another mod to move a job from queued status* into sieving status, so no worries. I’ll just wait on it until you say otherwise.

*Once it goes into sieving status there’s no going back! If the parameters must be changed after sieving has begun then it must be killed by Greg, we correct the job and then resubmit it under a new but similar name, e.g. old_name_1.

 2022-09-10, 03:00 #121 RichD     Sep 2008 Kansas 22×3×313 Posts Quick update before bedtime. Checking @ 20MQ. Code: Yield N-Yld sec/rel 4420 4275 (kernel update - reboot) <= 3.3 4534 4386 0.250 <= 3.5 4536 4388 0.251 <= 3.6

 Similar Threads Thread Thread Starter Forum Replies Last Post VBCurtis NFS@Home 166 2022-11-29 10:13 swellman NFS@Home 156 2022-11-14 19:52 VBCurtis NFS@Home 149 2022-11-02 11:09 VBCurtis NFS@Home 254 2022-01-02 01:59 debrouxl NFS@Home 10 2018-05-06 21:05

All times are UTC. The time now is 19:57.

Fri Dec 2 19:57:51 UTC 2022 up 106 days, 17:26, 0 users, load averages: 1.17, 1.27, 1.23