2018 15e post-processing reservations and results - Page 14

debrouxl · 2018-10-19, 10:04

The early RSALS "solution" of using a quorum of 2 for WUs might help against cheating by a single user using few computers... but not against coordinated cheating by multiple users using enough computers, and in the absence of cheating, it would waste half of the grid's computing power / double sieving time.

ISTR about a relations validator, but we're not using that in production. If done correctly, it wouldn't even necessarily be much slower than the current raw relations file production code.

VBCurtis · 2018-10-19, 15:19

Taking Fibonacci(1519).

fivemack · 2018-10-22, 15:07

Code:

p85 factor: 3726136642949946816528745367893675091194750264889581150199364793798895306815542863773
p156 factor: 756060880117253023409403445987365206687468608094473608902076696797774286172357743550738272761220360293807327855040825499161769412167531966292613480116433809

205.5 hours for 21.40M density-136 (not 138) matrix on 6 cores i7-4930K, 128-bit vectors.

Log attached and at https://pastebin.com/CT5TXFPW

fivemack · 2018-10-24, 12:03

Code:

Tue Oct 23 20:32:14 2018  p123 factor: 105066982192152272880111479167058062830872232282860973653314097665017697299850989908689096684465088038408522535114737932029
Tue Oct 23 20:32:14 2018  p145 factor: 2294535254428777564537271103515508214047202353428614676877787617011738874787436955256050625681636274611332741498436333638705278331641046191157629

About 174 hours for 26.71M density-118 (not 120) matrix on 14 cores i9-7940X (la_block=4096, VBITS=256). Log attached and at https://pastebin.com/tw6SyN33

fivemack · 2018-10-26, 08:14

Code:

Thu Oct 25 22:07:38 2018  p78 factor: 718995944250057271907045612703017337403579266807856154351702295438893093761151
Thu Oct 25 22:07:38 2018  p165 factor: 122242082107500935528242108444783884427112919378733837395987136222288003613410023065006780298446985639571061257857001268402714256520759939283085592301563987769020807

75.8 hours for 13.08M density-152 (not 154) matrix on 6 cores i7-4930K (VBITS=128). Log attached and at https://pastebin.com/iMJcescU

Batalov · 2018-10-27, 00:14

Tom, are you getting enough rels for the quartic Lucas/Aurif'n ?

Code:

L4155B = Lucas(4155) larger Aurifeuillian	Lucas	SNFS(231.56)	32
	40-200 M	100 %	0	1713	38284	303,058,498	13,583,961	Tom Womack

This one (and two more below) are listed as 32/32 jobs, but are they really 32/31's or 32/30's lopsided projects?

I will take from your result onto the next two quartics: if yours needs extra sieving, then those two do too.

fivemack · 2018-10-27, 12:28

They're not lopsided, I've pushed the extra sieving that the server was suggesting

fivemack · 2018-10-27, 12:28

Taking C220_142_118

ETA evening of Sunday 11 November

VBCurtis · 2018-10-29, 16:48

C188_150_83 is factored:

Code:

prp77 factor: 24705240030201196947941913183112341699334786370036351414881100462202441062147
prp111 factor: 750567552596149617202988865036899779170844144751752063088170326493051863535197547001241103433930996414270503517

Log at https://pastebin.com/Usu0Rc9d

481M raw relations yielded 371M unique. TD 132 produced a 19.06M matrix, which took about 10 days to solve on ten cores of Xeon-SB with all hyperthreads busy on lasieve14e.

Batalov · 2018-10-30, 17:19

Not sure what's up with the ftp server - no new data is added in >24 hours (by wget), but the stats appear to be changing. I am monitoring F2085 and L4215B.

https://escatter11.fullerton.edu/nfs/crunching_e.php

pinhodecarlos · 2018-10-30, 21:02

Quote:

Originally Posted by Batalov

Not sure what's up with the ftp server - no new data is added in >24 hours (by wget), but the stats appear to be changing. I am monitoring F2085 and L4215B.

https://escatter11.fullerton.edu/nfs/crunching_e.php

The job creator is stuck, it’s not pushing what’s left to 100 %, that’s the problem. I’ll issue an email to Greg to kick the server up.

2018-10-27, 00:14	#149
Batalov "Serge" Mar 2008 Phi(4,2^7658614+1)/2 22164₈ Posts	Tom, are you getting enough rels for the quartic Lucas/Aurif'n ? Code: L4155B = Lucas(4155) larger Aurifeuillian Lucas SNFS(231.56) 32 40-200 M 100 % 0 1713 38284 303,058,498 13,583,961 Tom Womack This one (and two more below) are listed as 32/32 jobs, but are they really 32/31's or 32/30's lopsided projects? I will take from your result onto the next two quartics: if yours needs extra sieving, then those two do too.

2018-10-27, 12:28	#151
fivemack (loop (#_fork)) Feb 2006 Cambridge, England 1100011101110₂ Posts	Taking C220_142_118 ETA evening of Sunday 11 November Last fiddled with by fivemack on 2018-11-01 at 20:35

2018-10-29, 16:48	#152
VBCurtis "Curtis" Feb 2005 Riverside, CA 5²·11·17 Posts	C188_150_83 is factored: Code: prp77 factor: 24705240030201196947941913183112341699334786370036351414881100462202441062147 prp111 factor: 750567552596149617202988865036899779170844144751752063088170326493051863535197547001241103433930996414270503517 Log at https://pastebin.com/Usu0Rc9d 481M raw relations yielded 371M unique. TD 132 produced a 19.06M matrix, which took about 10 days to solve on ten cores of Xeon-SB with all hyperthreads busy on lasieve14e.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
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
Crash doing large post-processing job	wombatman	Msieve	22	2013-12-04 01:37
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

2018-10-19, 10:04	#144
debrouxl Sep 2009 3D1₁₆ Posts	The early RSALS "solution" of using a quorum of 2 for WUs might help against cheating by a single user using few computers... but not against coordinated cheating by multiple users using enough computers, and in the absence of cheating, it would waste half of the grid's computing power / double sieving time. ISTR about a relations validator, but we're not using that in production. If done correctly, it wouldn't even necessarily be much slower than the current raw relations file production code.

2018-10-19, 15:19	#145
VBCurtis "Curtis" Feb 2005 Riverside, CA 4675₁₀ Posts	Taking Fibonacci(1519).

2018-10-27, 12:28	#150
fivemack (loop (#_fork)) Feb 2006 Cambridge, England 2×3,191 Posts	They're not lopsided, I've pushed the extra sieving that the server was suggesting

2018-10-30, 17:19	#153
Batalov "Serge" Mar 2008 Phi(4,2^7658614+1)/2 10010001110100₂ Posts	Not sure what's up with the ftp server - no new data is added in >24 hours (by wget), but the stats appear to be changing. I am monitoring F2085 and L4215B. https://escatter11.fullerton.edu/nfs/crunching_e.php