2019 14e post processing reservations and results - Page 35

unconnected · 2019-05-29, 08:14

Taking 11p7_253.

fivemack · 2019-05-29, 09:12

Taking 105__363_5m1

RichD · 2019-05-29, 12:16

Code:

p58 factor: 2022260770073844209836360706501022480849432133178426577181
p124 factor: 2044676412643726232258552936114391493264604182716836080609049288471058758434284082305287748163282927150523141307401269738081

212M unique relations built a 10.8M matrix using TD=124. (128 failed)
Solve time about 56 hours. (-t 4)

Log at: https://pastebin.com/iqk0ewxd

richs · 2019-05-29, 19:11

Taking 890__871_5m1

fivemack · 2019-05-30, 09:13

Taking 12p11_253 in a day or so when the sieving is finished

jyb · 2019-05-30, 17:31

Quote:

Originally Posted by fivemack

Taking 12m5_253 in a day or so when the sieving is finished

Speaking of finishing sieving, what exactly is going on with 12m5_253? I'm used to seeing work units not quite add up for a while, but by the end of the job they usually get accounted for, with maybe one or two going missing (e.g. see 105__363_5m1, which seems to be short by one).

But 12m5_253 has over 500 work units that seem to have disappeared. Does anyone know how the server's accounting works, and what may have gone wrong here? Given the high duplicate rate I've been seeing for recent jobs, that missing work could really matter for this number.

fivemack · 2019-05-30, 17:48

Code:

Thu May 30 18:45:08 2019  p73 factor: 6608194267370904331301449200719209762795704309049411404730589661235454681
Thu May 30 18:45:08 2019  p112 factor: 1792468618927064721502055756964562839368545738202015959700853297472663392247222270367962097294647521714779068591

Three hours on 10 cores Xeon Silver 4114 VBITS=256 for 4.02M matrix at density 136 (not 138) - it took longer to run the filtering on a rather busy machine than to solve the matrix.

Log at https://pastebin.com/c24ky587

Code:

Thu May 30 12:12:10 2019  begin with 237674088 relations and 175684940 unique ideals

is an absurd level of oversieving for numbers which are this easy, the clique removal is taking half as long as the whole linear algebra job.

RichD · 2019-05-31, 03:50

Quote:

Originally Posted by fivemack

is an absurd level of oversieving for numbers which are this easy, the clique removal is taking half as long as the whole linear algebra job.

I re-calculated the estimated yield and came up with similar numbers as NFS@home posted. I must have made a math error along the way. Yes, it was well over-sieved.

I am happy if I can solve a 31-bit job in under a week using my Sandy Bridge core-i5 boxes. Or 32-bit jobs in under 2.5 weeks. I would like to upgrade to xxxDECA-core system but I'm not that familiar with liquid cooled systems.

fivemack · 2019-05-31, 08:49

Quote:

Originally Posted by RichD

I re-calculated the estimated yield and came up with similar numbers as NFS@home posted. I must have made a math error along the way. Yes, it was well over-sieved.

I am happy if I can solve a 31-bit job in under a week using my Sandy Bridge core-i5 boxes. Or 32-bit jobs in under 2.5 weeks. I would like to upgrade to xxxDECA-core system but I'm not that familiar with liquid cooled systems.

I used liquid cooling on my i9-7940X for a while, but basically it's not integrated enough with the motherboard to make sense (that is, if the liquid cooling fails you don't get anything resembling useful alerts in system logs, you just see a temperature increase and mysterious power-downs); I've switched to a Noctua NH-U9S now and it's been fine for a year.

fivemack · 2019-05-31, 08:53

Quote:

Originally Posted by jyb

Speaking of finishing sieving, what exactly is going on with 12m5_253? I'm used to seeing work units not quite add up for a while, but by the end of the job they usually get accounted for, with maybe one or two going missing (e.g. see 105__363_5m1, which seems to be short by one).

But 12m5_253 has over 500 work units that seem to have disappeared. Does anyone know how the server's accounting works, and what may have gone wrong here? Given the high duplicate rate I've been seeing for recent jobs, that missing work could really matter for this number.

I am very confused - my post https://mersenneforum.org/showpost.p...&postcount=379 says I'm taking 12p11_253, but your quote of it https://mersenneforum.org/showpost.p...&postcount=380 says 12m5_253. I am half inclined to take 12m5_253 too, and half inclined to see your post as a reservation of 12m5_253 in rather peculiar form. Please advise.

RichD · 2019-05-31, 13:52

Code:

p76 factor: 9654015162709013215641338252766919080261591243535477939503567221352050022127
p147 factor: 597881515215062318815670814440778211054642428667061738438953843564857073865271578816912414953710336219533816167946203481602901748167118857178594989

208M unique relations built an 11.1M matrix using TD=124. (128 failed)
Solve time about 78.5 hours. (-t 4)

Log at: https://pastebin.com/HVYmYFaY

2019-05-29, 12:16	#377
RichD Sep 2008 Kansas 2⁶×53 Posts	7p4_319 factored Code: p58 factor: 2022260770073844209836360706501022480849432133178426577181 p124 factor: 2044676412643726232258552936114391493264604182716836080609049288471058758434284082305287748163282927150523141307401269738081 212M unique relations built a 10.8M matrix using TD=124. (128 failed) Solve time about 56 hours. (-t 4) Log at: https://pastebin.com/iqk0ewxd

2019-05-30, 17:48	#381
fivemack (loop (#_fork)) Feb 2006 Cambridge, England 2³×11×73 Posts	105__363_5m1 done Code: Thu May 30 18:45:08 2019 p73 factor: 6608194267370904331301449200719209762795704309049411404730589661235454681 Thu May 30 18:45:08 2019 p112 factor: 1792468618927064721502055756964562839368545738202015959700853297472663392247222270367962097294647521714779068591 Three hours on 10 cores Xeon Silver 4114 VBITS=256 for 4.02M matrix at density 136 (not 138) - it took longer to run the filtering on a rather busy machine than to solve the matrix. Log at https://pastebin.com/c24ky587 Code: Thu May 30 12:12:10 2019 begin with 237674088 relations and 175684940 unique ideals is an absurd level of oversieving for numbers which are this easy, the clique removal is taking half as long as the whole linear algebra job. Last fiddled with by fivemack on 2019-05-30 at 17:52

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
2018 15e post-processing reservations and results	fivemack	NFS@Home	221	2019-01-04 13:08
2018 14e post processing reservations and results	pinhodecarlos	NFS@Home	551	2019-01-04 13:06
16e Post Processing Progress	pinhodecarlos	NFS@Home	8	2018-11-28 13:45
Update on 7^254+1 post processing	dleclair	NFSNET Discussion	4	2005-04-05 09:51
Post processing for 2,757-	xilman	NFSNET Discussion	3	2003-11-06 14:23

2019-05-29, 08:14	#375
unconnected May 2009 Russia, Moscow 2,593 Posts	Taking 11p7_253.

2019-05-29, 09:12	#376
fivemack (loop (#_fork)) Feb 2006 Cambridge, England 2³·11·73 Posts	Taking 105__363_5m1

2019-05-29, 19:11	#378
richs "Rich" Aug 2002 Benicia, California 2³·3·5·11 Posts	Taking 890__871_5m1

2019-05-30, 09:13	#379
fivemack (loop (#_fork)) Feb 2006 Cambridge, England 1100100011000₂ Posts	Taking 12p11_253 in a day or so when the sieving is finished