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 2017-03-30 08:30

Taking C178_187039475551_19

unconnected 2017-03-30 09:12

Reserving C186_290770486991_19.

YuL 2017-03-30 16:49

[QUOTE=fivemack;455769]Good job![/QUOTE]

Thanks. I was expecting a smaller matrix though; I thought I would be able to
build a matrix with a much higher target density given the 460M+ relations for
a 32-bit job but even 132 failed, I had to lower it to 128.

VBCurtis 2017-04-02 04:31

I'll take 127^95 + 95^127 cofactor from the 14e queue.

fivemack 2017-04-03 12:49

C178_187039475551_19 done
 
1 Attachment(s)
[code]
Sat Apr 1 11:45:02 2017 p86 factor: 59969606437825471092748609807267089171442385775453880240834192297102407276550682363249
Sat Apr 1 11:45:02 2017 p92 factor: 28161118831265144175724429817967381635120122727242283196157680376868720112151713622855853951
[/code]

13 hours for 5.75M density-120 matrix on 8 threads E5-2650v3

Log at [url]https://pastebin.com/iaTHLyUC[/url] or attached

fivemack 2017-04-03 12:51

Taking C203_125_96 (ETA Monday morning, since the linear algebra should finish just after going-home time on Friday)

fivemack 2017-04-04 12:46

Taking C202_134_87 (15e)

22.8M density-120 matrix, ETA sometime in the week after Easter

richs 2017-04-04 17:29

Reserving C237_12161_59 (14e)

VBCurtis 2017-04-06 17:07

[QUOTE=VBCurtis;456003]I'll take 127^95 + 95^127 cofactor from the 14e queue.[/QUOTE]

I forgot to add target_density to the command line, and produced a 21.8M matrix at default density. I re-ran the filtering with TD = 140, and produced the identical matrix!

I realized that meant the dataset was oversieved, so I reduced to 470M relations (from initial 495M) and ran again. This produced a 20.4M matrix at density around 100; reducing to 450M rels produced 19.7M matrix at density 125. This matrix is forecast to take 80-100 hr longer than the original matrix, so I'm solving a 21.8M matrix at density 67, forecast to take 25 days or so.

Matrix-building took 8005MB, while the solving step is logged at 6500MB while top claims 7.6g used.

fivemack 2017-04-06 22:59

[QUOTE=fivemack;455171](fx: bites bullet; grimaces)

Taking C193_143_93. Anticipating it will take a couple of months.[/QUOTE]

39.06M matrix at density=128 started on 6 threads i7-5820K, ETA about six and a half weeks from now (second half of May)

fivemack 2017-04-07 16:00

C203_125_96 done
 
1 Attachment(s)
[code]
Fri Apr 7 10:30:05 2017 p60 factor: 262140789711545872971063685147422871988451144802929070574027
Fri Apr 7 10:30:05 2017 p71 factor: 56035791901525658273230054058046999996742343570248151591987937037637071
Fri Apr 7 10:30:05 2017 p73 factor: 1202663914264836729954659180323840975151179663082897780688670921683426083
[/code]

63 hours on 7 threads E5-2650v2 for 11.1M density-120 matrix (not enough relations for density-125)

Log attached and at [url]https://pastebin.com/1tVttxqP[/url]


All times are UTC. The time now is 23:12.

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