BOINC NFS sieving - NFS@Home - Page 142

RichD · 2014-09-10, 19:00

Finally, finishes with a nice split.

Code:

prp110 factor: 13867134919056264282298953670685030515712869911296942013531081154168321109576361091777430465710535074375777559
prp112 factor: 3840584215228170317549670273429806556050849131263726120860064074694704905554859461208897748984642493107758755817

After numerous restarts and moving to another machine it completes in about a week.

I'm guessing around 170 hours +/-10% for a 14.2M matrix using target_density=120 on Core-i5 -t 4.

The log is only for the matrix build.

jasonp · 2014-09-11, 01:09

I love this thread. I remember back in 2007 when it was a monumental achievement to finish matrices half the size of the ones that appear here, after months of work.

xilman · 2014-09-11, 08:49

Quote:

Originally Posted by RichD

Finally, finishes with a nice split.

Code:

prp110 factor: 13867134919056264282298953670685030515712869911296942013531081154168321109576361091777430465710535074375777559
prp112 factor: 3840584215228170317549670273429806556050849131263726120860064074694704905554859461208897748984642493107758755817

After numerous restarts and moving to another machine it completes in about a week.

I'm guessing around 170 hours +/-10% for a 14.2M matrix using target_density=120 on Core-i5 -t 4.

The log is only for the matrix build.

Nice split indeed!

fivemack · 2014-09-11, 09:02

Quote:

Originally Posted by jasonp

I love this thread. I remember back in 2007 when it was a monumental achievement to finish matrices half the size of the ones that appear here, after months of work.

And that's almost entirely due to your efforts - I don't think more than a factor three of the wall-time improvement is coming from Intel's extra cores and more efficient DDR3 controllers. My thanks are of course inadequate, but thanks again!

swellman · 2014-09-11, 11:53

After a few false starts, GC_12_222 has been successfully downloaded and is now in LA. Even managed to build the matrix with a target density of 128.

Should be done on Monday evening.

VictordeHolland · 2014-09-11, 20:13

I downloaded the precompiled msieve v1.52 from sourceforge, but unfortunate it was a 32 bit version, so that wasn't going to work for a 6GB job. The CUDA version (which I suppose is 64 bit) was giving me missing dll errors (the machine has a AMD graphics cards, so that was to be expected). So I reverted back to 1.52dev SVN946 compiled by Wombatman (http://mersenneforum.org/showpost.ph...7&postcount=10), which I've successfully used in the past.

I noticed on sourceforge there are numerous improvements with the official v1.52 release, anybody got a link to a precompiled v1.52 Windoze-x64 without CUDA?
BTW, I already checked: http://gilchrist.ca/jeff/factoring/index.html , but those are all 1.51 and 1.52dev versions, so that page might need an update also.

GC_2_795
Linear algebra started, ETA 110 hours

fivemack · 2014-09-12, 09:02

Reserving GC_2_803 (or C803, as I think Cullen would have called it); I'll try to start it automatically over the weekend, it'll probably take about a week.

I will also reserve L1279 (ETA evening of Monday September 22nd)

fivemack · 2014-09-13, 10:54

Code:

Sat Sep 13 11:46:46 2014  prp133 factor: 1968713615111618271539160277061755340724636268324928756535818414815976263634266978986498520053266230149202510193251991913203540371799
Sat Sep 13 11:46:46 2014  prp135 factor: 382854763035287141589388305613288247076660418667790621896348047402434425342034117309987377689458492297422709066316184572868365872512029

203 hours (not quite running 24/365) -t6 on i7/4930K for 17.9M matrix

Log at http://pastebin.com/06kiYxA0

wombatman · 2014-09-13, 16:18

I'll do GC_2_804 while waiting on my HomePrimes number to finish up.

fivemack · 2014-09-15, 08:16

Code:

Sat Sep 13 22:04:24 2014  prp63 factor: 244559704051062643105923212826547043644123384267057817678467409
Sat Sep 13 22:04:24 2014  prp166 factor: 1937096697530734735808566499420836932516732745738918202741792125888157323282536681913860846108786567798811249981761874727603629445005716053635702536673892223042015443

94.2 hours for 9.3M matrix, i7/2600K -t3

http://pastebin.com/uq98RVy8

VictordeHolland · 2014-09-15, 11:51

A new memory kit arrived, so i'm taking a shot at L1282.

GC_2_795 should be ready in a few hours.

2014-09-11, 11:53	#1556
swellman Jun 2012 11·281 Posts	GC_12_222 After a few false starts, GC_12_222 has been successfully downloaded and is now in LA. Even managed to build the matrix with a target density of 128. Should be done on Monday evening.

2014-09-11, 20:13	#1557
VictordeHolland "Victor de Hollander" Aug 2011 the Netherlands 2³×3×7² Posts	GC_2_795 I downloaded the precompiled msieve v1.52 from sourceforge, but unfortunate it was a 32 bit version, so that wasn't going to work for a 6GB job. The CUDA version (which I suppose is 64 bit) was giving me missing dll errors (the machine has a AMD graphics cards, so that was to be expected). So I reverted back to 1.52dev SVN946 compiled by Wombatman (http://mersenneforum.org/showpost.ph...7&postcount=10), which I've successfully used in the past. I noticed on sourceforge there are numerous improvements with the official v1.52 release, anybody got a link to a precompiled v1.52 Windoze-x64 without CUDA? BTW, I already checked: http://gilchrist.ca/jeff/factoring/index.html , but those are all 1.51 and 1.52dev versions, so that page might need an update also. GC_2_795 Linear algebra started, ETA 110 hours

2014-09-12, 09:02	#1558
fivemack (loop (#_fork)) Feb 2006 Cambridge, England 1918₁₆ Posts	Reserving GC_2_803 (or C803, as I think Cullen would have called it); I'll try to start it automatically over the weekend, it'll probably take about a week. I will also reserve L1279 (ETA evening of Monday September 22nd) Last fiddled with by fivemack on 2014-09-13 at 19:43

2014-09-15, 11:51	#1562
VictordeHolland "Victor de Hollander" Aug 2011 the Netherlands 2³·3·7² Posts	L1282 A new memory kit arrived, so i'm taking a shot at L1282. GC_2_795 should be ready in a few hours.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Boinc Statistics for NFS@Home borked ?	thomasn	NFS@Home	1	2013-10-02 15:31
BOINC NFS sieving - RSALS	debrouxl	NFS@Home	621	2012-12-14 23:44
BOINC?	masser	Sierpinski/Riesel Base 5	1	2009-02-09 01:10
BOINC?	KEP	Twin Prime Search	212	2007-04-25 10:29
BOINC	bebarce	Software	3	2005-12-15 18:35

2014-09-11, 01:09	#1553
jasonp Tribal Bullet Oct 2004 110111010111₂ Posts	I love this thread. I remember back in 2007 when it was a monumental achievement to finish matrices half the size of the ones that appear here, after months of work.

2014-09-13, 16:18	#1560
wombatman I moo ablest echo power! May 2013 13×137 Posts	I'll do GC_2_804 while waiting on my HomePrimes number to finish up.