mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   NFS@Home (https://www.mersenneforum.org/forumdisplay.php?f=98)
-   -   Linear algebra reservations, progress and results (https://www.mersenneforum.org/showthread.php?t=20023)

fivemack 2017-11-15 15:55

[QUOTE=Dubslow;471849]Other than the divisor of the inner term being wrong according to [URL="http://factordb.com/index.php?id=1100000000128028191"]FDB (should be 371330, not 79)[/URL][/quote]

The divisor is wrong, but not for the reason you state. It should be 71, not 79 - that's a typo - and the extra factor 5230 comes from the definition of phi_n.

Dubslow 2017-11-15 16:21

[QUOTE=fivemack;471876]The divisor is wrong, but not for the reason you state. It should be 71, not 79 - that's a typo - and the extra factor 5230 comes from the definition of phi_n.[/QUOTE]

Aha, obviously I'm not very well versed in cyclotomics. Forgot about the algebraic factor (mostly because it's more convenient when working with FDB and yafu). Thanks.

fivemack 2017-11-15 16:48

L3295A done
 
1 Attachment(s)
[code]
Wed Nov 15 16:11:47 2017 p88 factor: 2523736458627921023145795717278312510332675044476120748935257857252531967979710404049931
Wed Nov 15 16:11:47 2017 p93 factor: 117529146260985677311078059731021197946847596724602818549439979645777947375020581782484159121
[/code]

243.5 hours on 7 cores E5/2650v2 for 20.31M density-130 (not 132) matrix

Log attached and at [url]https://pastebin.com/V570tfkz[/url]

fivemack 2017-11-15 17:06

Taking C230_142_86

swellman 2017-11-17 20:39

C223_129_100 Factored
 
1 Attachment(s)
C223_129_100

[code]
p66 factor: 300634338355028902107770297818841985362732708412225230534671189001
p158 factor: 13710890265852876939918234253281665271763816874772099511164109834698568681020540058631256270440730166652606033577790433814449742651202525507914405580763759981
[/code]

461048659 raw relations / 345308985 unique relations

log attached

VBCurtis 2017-11-17 21:27

3408_1671 factors
 
[code]prp88 factor: 1417911815027846802053291284074202683296044263447677506019370900162722120468816094645719
prp93 factor: 145373073512535632032343452896302114311357961994569163931012272061735693010559257271548488289[/code]
log at [url]https://pastebin.com/QipkwDme[/url]

411M raw relations, 344M unique. Matrix TD 132 produced 14.6M matrix at density 96.

unconnected 2017-11-18 18:34

Reserving C167_4788_12505.

RichD 2017-11-19 13:19

Taking 8081^67-1 (15e).

RichD 2017-11-19 13:31

C227_135_76 factored
 
408M unique relations to build a 20.0M matrix using TD=128. (TD=132 failed).
[CODE]p77 factor: 15964490763617942821554929254720694835604074989287437741802172054773394694877
p151 factor: 4460802349198982408280765869545747805981738719596425412499688908971784347693405040128665910961202600700378834301226877059280416661164739559148356741843[/CODE]
[url]https://pastebin.com/WQQF7FwS[/url]

VBCurtis 2017-11-19 18:35

13*2^828-1 factors
 
Log at [url]https://pastebin.com/WdPhpLZE[/url]
[code]prp62 factor: 31218770086328113069106512645320557869405043516087342127128143
prp134 factor: 21628664499104777790648131454693068351775431323246347241538024533780289511774766756972222779871836576200837163337851485290669686178819
[/code]
This was a 14e/33 job. 650M raw relations, 536M unique was oversieved; TD = 132 produced a density 88 (sparse part density 79) matrix of size 16.5M.

I would have requested 410M relations as 32LP; 650M at 33LP is faster to find (in seconds and in Q-range) than 410M at 32, and 650M was too many.

fivemack 2017-11-20 22:39

C196_135_124 done
 
1 Attachment(s)
[code]
Sat Nov 18 19:59:10 2017 p75 factor: 493256075733235828719603706694310826387263994582477684236086531883097166311
Sat Nov 18 19:59:10 2017 p121 factor: 9672920637877273831373571056996202576954575529763025727282423366872027650358677672581556097977772035594575812103918000419
[/code]

Uneventful except for the time taken:

[code]
Thu Aug 31 10:27:56 2017 using block size 8192 and superblock size 1474560 for processor cache size 15360 kB
Thu Aug 31 10:33:54 2017 commencing Lanczos iteration (6 threads)
Thu Aug 31 10:33:54 2017 memory use: 20910.7 MB
Thu Aug 31 10:38:17 2017 linear algebra at 0.0%, ETA 2030h49m
Thu Aug 31 10:39:39 2017 checkpointing every 30000 dimensions
Sat Nov 18 00:12:52 2017 lanczos halted after 702030 iterations (dim = 44393476)
Sat Nov 18 00:13:30 2017 recovered 31 nontrivial dependencies
Sat Nov 18 00:13:34 2017 BLanczosTime: 6798080
Sat Nov 18 00:13:34 2017 elapsed time 1888:21:21
[/code]

The power is reasonably reliable here. The machine is an i7-5820K with 32G memory, the job used all six cores and somewhere in the high twenties gigabytes of memory, for two and a half months.

Log attached and at [url]https://pastebin.com/n7CxLdX8[/url]


All times are UTC. The time now is 23:04.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2021, Jelsoft Enterprises Ltd.