Linear algebra reservations, progress and results - Page 47

VictordeHolland · 2015-12-05, 18:52

C153_P233_plus_1

Code:

prp57 factor: 976091661440498273311588215450771879582220118969346171503
prp96 factor: 430335904840578306669039776242416581629866475129170214872930601654724618834763764783224209970307

matrix is 4140851 x 4141076 (1976.1 MB)
------------------------------------------------------------------------
C154_P182_plus_1

Code:

prp71 factor: 27277576813606975571844943309761673123428776562927433726921949754679613
prp84 factor: 171066199323584838258211112753849127051765672129864072245708787815508766170527186951

matrix is 4358959 x 4359184 (2091.8 MB)
------------------------------------------------------------------------

VictordeHolland · 2015-12-05, 18:53

C155_P209_plus_1

Code:

prp74 factor: 10346196175428152102773520718167922375362025229971136961917614228903776537
prp82 factor: 4105695312892115636553366234774522477274854964663968305029273169198918335942867279

matrix is 4502743 x 4502968 (2160.1 MB)
------------------------------------------------------------------------

All filtered with TD=130, LA took 10-12h each, logs attached (couldn't add 3 attachments to one post).

frmky · 2015-12-06, 01:49

For the 14e numbers, I've added a pastebin field for the unique part of a pastebin url containing the log. If that is populated, on the status page the comment becomes a link to the corresponding pastebin.

Edit: Now 15e has been given the same treatment.

Xyzzy · 2015-12-06, 01:51

pinhodecarlos · 2015-12-06, 08:59

My msieve logs for pastebin url.

VictordeHolland · 2015-12-06, 15:38

Much better and cleaner now.
Thanks Greg!

fivemack · 2015-12-07, 09:53

Code:

Sat Dec  5 09:09:56 2015  p68 factor: 99713931517486444010746977398434147387312495439028057685457930885581
Sat Dec  5 09:09:56 2015  p115 factor: 6337366248063930626919828317773198377075808129453921717685844042320912353140083589239167509613707066272579659320439

34.4 hours on 7 threads E5-2650v2 for 9.04M matrix at density 120. Log attached.

fivemack · 2015-12-07, 09:58

Code:

Sat Dec  5 17:44:54 2015  p91 factor: 2132511512443414333678564691107777973475861603076069009971937493950396472549072567490874627
Sat Dec  5 17:44:54 2015  p129 factor: 532873974251185874870260239618395036338610187500129178927301700770853208890035150980092485967757603488674407284211962191930161857

8.3 hours for 4.63M density-120 matrix on 7 threads E5-2650v2.

Log attached.

Xyzzy · 2015-12-07, 16:18

The best TD for L1397 we can get is 108. Doesn't that mean the job is going to run significantly longer?

Code:

Sat Dec  5 11:35:46 2015  setting target matrix density to 140.0
Sat Dec  5 16:00:06 2015  setting target matrix density to 130.0
Sat Dec  5 20:01:24 2015  setting target matrix density to 125.0
Sun Dec  6 09:13:49 2015  setting target matrix density to 120.0
Sun Dec  6 13:49:42 2015  setting target matrix density to 116.0
Sun Dec  6 20:20:04 2015  setting target matrix density to 112.0
Mon Dec  7 06:17:51 2015  setting target matrix density to 108.0

(We are still running 115!+1, which ended up with a TD of 116.)

fivemack · 2015-12-07, 16:50

Quote:

Originally Posted by Xyzzy

The best TD for L1397 we can get is 108. Doesn't that mean the job is going to run significantly longer?

Longer, but not significantly longer (in particular, definitely not enough longer to make up for the humans-editing-files cost involved in re-opening the sieving even if it doesn't end up right at the back of the queue again)

Have you got an ETA for 115!+1 ?

Xyzzy · 2015-12-07, 19:28

Quote:

Originally Posted by fivemack

Have you got an ETA for 115!+1 ?

It should be done on or about December 14th.

We have found that our system's memory bandwidth limits us to using only three threads for optimal performance.

(Hopefully we are not holding anything up with our glacial progress!)

2015-12-06, 01:49	#509
frmky Jul 2003 So Cal 4120₈ Posts	For the 14e numbers, I've added a pastebin field for the unique part of a pastebin url containing the log. If that is populated, on the status page the comment becomes a link to the corresponding pastebin. Edit: Now 15e has been given the same treatment. Last fiddled with by frmky on 2015-12-06 at 07:59 Reason: Ditto.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
restarting nfs linear algebra	cubaq	YAFU	2	2017-04-02 11:35
Linear algebra at 600%	CRGreathouse	Msieve	8	2009-08-05 07:25
Linear algebra crashes	10metreh	Msieve	3	2009-02-02 08:34
Linear algebra proof	Damian	Math	8	2007-02-12 22:25
Linear algebra in MPQS	R1zZ1	Factoring	2	2007-02-02 06:45

2015-12-06, 01:51	#510
Xyzzy "Mike" Aug 2002 2035₁₆ Posts

2015-12-06, 15:38	#512
VictordeHolland "Victor de Hollander" Aug 2011 the Netherlands 2³·3·7² Posts	Much better and cleaner now. Thanks Greg!