BOINC NFS sieving - NFS@Home - Page 106

debrouxl · 2014-02-19, 21:09

The sextic difficulty 242 has basically the same ~1 relation per q value yield at q0=45M, but it's slower than the sextic posted by Sean.
The quintic difficulty 240 has pretty much the same yield as the others in the same range, faster than your sextic but slightly slower than the sextic posted by Sean.

swellman · 2014-02-20, 00:29

Quote:

Originally Posted by debrouxl

Sean: I tried your yafu-generated poly for the XYYXF C176_118_93 number

snip

but it doesn't sieve much better than the polys generated by snfspoly: ~1 relation per q value at q0=45M (rlim /2 = alim / 2) using 31-bit LPs.

That number proves surprisingly hard for a SNFS difficulty 23x (depending on the polynomial) number. Even though 14e could probably do it with a lengthy sieving, 15e would probably be more efficient.

Yes, my test sieving used 15e. Also, I ran it on the algebraic side FWIW.

Really appreciate you guys taking the time to even look at this composite. I've likely got tunnel vision on it, as I have spent a lot of effort trying to factor it.

RichD · 2014-02-20, 02:40

C160_3408_1385 splits as:

Code:

prp68 factor: 17880278619984695184050502790206873556361981045516886688928661227471
prp93 factor: 228502311121984348922806887427415202001523194440641388076613271354581292876973003527160307423

fivemack · 2014-02-20, 09:07

Code:

Wed Feb 19 22:53:55 2014  prp79 factor: 6544139735540280558492681564220189691920039926567620314258950958984293109132781
Wed Feb 19 22:53:55 2014  prp141 factor: 356041230470821300547265003988221353676049541264454212033019140637311164592494814068768166474884500439741810418568259479009834952510802545219

Log at http://pastebin.com/wcG0pwQG

7.4M matrix, 57 hours on i7/2600

fivemack · 2014-02-20, 09:09

Taking GC_4_391, eta Monday morning

Also, ETA for 2340_723 is Friday morning - I suppose I could have asked for slightly less sieving.

RichD · 2014-02-20, 22:39

The C160_3408_1385 built a 6.0M matrix using target_density=124 and took just over 22 hrs (in BL) on a Core-i5 (-t 4).

fivemack · 2014-02-21, 08:05

Code:

Fri Feb 21 04:01:00 2014  prp68 factor: 33764306304594609841151385700590125608455755423072146929348187732319
Fri Feb 21 04:01:00 2014  prp101 factor: 34171287558191566663884446485221802740049394413628828438507905979286574134540823657330555119715424329

19.5 hours on i7/4770 for a 5.9M matrix

fivemack · 2014-02-21, 08:08

Code:

Fri Feb 21 06:55:18 2014  prp56 factor: 34470240407647113273727147928606646498352887484685412491
Fri Feb 21 06:55:18 2014  prp118 factor: 2146601110002558873853257238895849150974507931390059443029499513822227376252498348631767281157367204096404394646303011

93 hours on i5/750 -t4 for 8.7M matrix

swellman · 2014-02-21, 11:38

C_2_784 splits

Code:

prp95 factor: 62989812257204905134330822366504114809484465259945825102840522386263274321822248099151904049177
prp110 factor: 22394703467804797846381410780003403139074909488534959121267242416187245444804045804712301688742956615022179337

Mini-Geek · 2014-02-23, 20:27

GC_6_302's c160 = p68*p93. It took my i5-750 about 240 calendar hours for the 12.9M matrix, but even with -t 4 the CPU usage averaged around 69% (~2.76 cores, so ~662 CPU hours for the LA). All of the logs from my harrowing experience are attached, in case anyone is silly enough to want to pore through them.

Code:

prp68 factor: 10922077385007413926959844864921688969621938072657126375168970438021
prp93 factor: 396325903304263476304132664503070294343931044028469032434386038612698624514895837180351478017

Oddly enough, my other NFS@Home post-proc, C161_4788_5193, also split as p68*p93. I had to double-take, and for a brief moment thought I had somehow rerun the sqrt of that Aliquot number instead of GC_6_302.

fivemack · 2014-02-23, 20:43

Taking GW_5_337 (eta Friday morning)

2014-02-20, 02:40	#1158
RichD Sep 2008 Kansas 2⁶·53 Posts	C160_3408_1385 splits as: Code: prp68 factor: 17880278619984695184050502790206873556361981045516886688928661227471 prp93 factor: 228502311121984348922806887427415202001523194440641388076613271354581292876973003527160307423

2014-02-20, 09:09	#1160
fivemack (loop (#_fork)) Feb 2006 Cambridge, England 1918₁₆ Posts	Taking GC_4_391, eta Monday morning Also, ETA for 2340_723 is Friday morning - I suppose I could have asked for slightly less sieving. Last fiddled with by fivemack on 2014-02-20 at 12:03

2014-02-20, 22:39	#1161
RichD Sep 2008 Kansas 6500₈ Posts	Data Point The C160_3408_1385 built a 6.0M matrix using target_density=124 and took just over 22 hrs (in BL) on a Core-i5 (-t 4).

2014-02-23, 20:43	#1166
fivemack (loop (#_fork)) Feb 2006 Cambridge, England 2³·11·73 Posts	Taking GW_5_337 (eta Friday morning) Last fiddled with by fivemack on 2014-02-27 at 10:05

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

2014-02-19, 21:09	#1156
debrouxl Sep 2009 977 Posts	The sextic difficulty 242 has basically the same ~1 relation per q value yield at q0=45M, but it's slower than the sextic posted by Sean. The quintic difficulty 240 has pretty much the same yield as the others in the same range, faster than your sextic but slightly slower than the sextic posted by Sean.