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)

YuL 2017-10-12 13:24

C161_3613045528091_17 (14e)
 
1 Attachment(s)
[QUOTE=YuL;469333]Reserving C161_3613045528091_17, thanks.[/QUOTE]

[URL="http://www.factordb.com/index.php?id=1100000000126228710"]C161_3613045528091_17[/URL] is factored.
[CODE]Thu Oct 12 13:15:21 2017 p70 factor: 6973758502234253884181929362856224995385676641649925515421567582226967
Thu Oct 12 13:15:21 2017 p91 factor: 2615526656926701787299805440932520462192513375052362998296857110075036721543122390667728507
[/CODE]

LA on the 4.84e6 matrix took 2.4h using 4x Dual Intel Xeon E5-2686 v4 (target density = 128).
Log attached and at [URL]https://pastebin.com/t01fTH5Z[/URL].

fivemack 2017-10-12 13:48

YuL: you have some pretty impressive compute resource there! Would you be able to run for ~100 wall-clock hours to deal with Lucas(1321) ?

VBCurtis 2017-10-12 13:53

C224_148_86 factors
 
[code]prp89 factor: 36004488451441119651214513502118979442636734789885920973089090859209192780295366731792917
prp136 factor: 1267270686372475929419863116308088251546931487306166069492255355460668707763455774860254500940743946722371347667053477745796394829918369[/code]
349M unique 32LP relations produced an 18.3M matrix at TD 134.

log at [url]https://pastebin.com/QzrLFYLq[/url]

c226_143_57 is 30% complete, about two more weeks.

YuL 2017-10-12 13:56

[QUOTE=fivemack;469691]YuL: you have some pretty impressive compute resource there! Would you be able to run for ~100 wall-clock hours to deal with Lucas(1321) ?[/QUOTE]

Amazon EC2, 100 hours of 4x r4.16xlarge would cost me ~480$ :-).

richs 2017-10-12 14:43

Reserving C192_249541_41 please.

fivemack 2017-10-12 15:48

[QUOTE=YuL;469693]Amazon EC2, 100 hours of 4x r4.16xlarge would cost me ~480$ :-).[/QUOTE]

Ah! Definitely not asking you to do that, then :)

RichD 2017-10-13 18:51

Taking C227_140_61 next.

RichD 2017-10-13 20:20

C225_140_138 factored
 
About 354 hours (power outage) to solve 19.7M matrix using -t 4, w/ TD=144.
[CODE]p81 factor: 204743040471511957974607832522577158823452984876200257817529201941362845184315697
p145 factor: 2647406304735019701172167099689199866467319081702657342977655939289431878262715543870119797539665236347622085873613993044340283536212695425604817[/CODE]
[url]https://pastebin.com/ULCw6fWK[/url]

richs 2017-10-13 21:19

C191_22926121_31 factored
 
1 Attachment(s)
[QUOTE]Reserving C191_22926121_31 please.[/QUOTE]

Finished quicker than I expected.

[CODE]p93 factor: 914270024491921792401085778556927231150530617854161585172963998054586105241044056026056902827
p98 factor: 23790037067171663886838174425797681923684271024161057724917436107609767438289181684247267006610009[/CODE]

118.7 hours on 2 threads Core i3-2310M with 4 GB memory for a 6.6M matrix at TD = 130. This matrix just fit in 4GB with 98% memory usage. Log attached and condensed version (entire log too big for Pastebin) at [URL="https://pastebin.com/mYM26G1N"]https://pastebin.com/mYM26G1N[/URL]

Factors added to Factordb.

Mini-Geek 2017-10-15 12:29

1 Attachment(s)
I'm post-processing C213_142_112 with 449159757 relations (I started it with a few relations still pending to be reported to NFS@Home). At first I tried a target_density of 120 (a standard value for me), but it failed to enter the LA step. So I tried it again without specifying a target_density and it has begun LA on 29.7 million dimensions.

Only thing is that it's estimating about 30 days on my i5-4670K to complete linear algebra. Is this to be expected, or is there something I should do to speed it up? I don't have a problem letting it run, just want to know I'm not wasting my time. :smile: I don't think I've done a job this big before.

Log attached.

swellman 2017-10-15 17:39

[QUOTE=Mini-Geek;469875]I'm post-processing C213_142_112 with 449159757 relations (I started it with a few relations still pending to be reported to NFS@Home). At first I tried a target_density of 120 (a standard value for me), but it failed to enter the LA step. So I tried it again without specifying a target_density and it has begun LA on 29.7 million dimensions.

Only thing is that it's estimating about 30 days on my i5-4670K to complete linear algebra. Is this to be expected, or is there something I should do to speed it up? I don't have a problem letting it run, just want to know I'm not wasting my time. :smile: I don't think I've done a job this big before.

Log attached.[/QUOTE]

If not specified, the default target_density (TD) used by msieve is 70. You may want to restart msieve with a TD=116, then 112, etc until msieve successfully enters LA. Or drop TD in increments of 10. It’s up to you but it should be a bit faster than TD=70.

That said, a SNFS 263 is never going to be a one week job on a single machine. Good luck!


All times are UTC. The time now is 23:07.

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