Linear algebra reservations, progress and results - Page 32

Dubslow · 2015-09-10, 06:51

I'll try 11611_109.

Could someone estimate how much memory various 15e jobs would take? Am I right in guessing that (say) the GNFS194 from 4788 would be more than 16 GiB?

fivemack · 2015-09-10, 09:24

None of the 15e jobs will fit in 8GB, and with 32-bit large primes the filtering on the bigger jobs gets tight-to-impossible in 16G, but the matrix itself isn't too bad until you're into the high-C180s.

The C182 from 147,125 took 14GB for matrix-building; the C257 SNFS 135,82 took 11GB for matrix building.

C193_2340_742 was less over-sieved and took less memory to build the matrix, but the matrix was huge (31.5M) and would not fit in 16G.

Note that the log file understates memory use for the matrix phase; for example, a matrix where the log says 'memory use: 4730.8MB' is currently running with virt=6588M res=6.1G

VictordeHolland · 2015-09-10, 12:13

GW_9_257

Code:

prp82 factor: 2657134845708193116575756441180816679272820461238700993416806516117935617001467469
prp97 factor: 5636962207860684217535820039843158929868095238470807539851438174522620002736180244701000991814209

target_density=115 (120 failed filtering), 12.6M matrix, ~120h on 4 cores i7 3770k
log attached

frmky · 2015-09-10, 22:26

Quote:

Originally Posted by frmky

I've taken W_795 and GW_3_502. Both should be done tomorrow.

W_795

Code:

p104 factor: 54874034920931135861719895187381751989817503532077501887202266914532304010850653889600111682921895593127
p105 factor: 355018650276589253569732059273899112648065410049958016080785072067440080051354861668797071294222218684689

GW_3_502

Code:

p96 factor: 122838606011459566941521222717596412576527332246352792452595461733479143231281284137259355896787
p146 factor: 46114840318482652720226564339831810352662035775416203849592174391227241727948892725892538747025784365157459587577397810713036812808162989587158579

Jarod · 2015-09-11, 03:50

If there are any more numbers with 29 bits like 55559_219 if nobody wants them (1) I would be like to see if I can complete the LA phase. As I have done these numbers years ago and they were fairly easy. I have 16 gig of RAM.

pinhodecarlos · 2015-09-11, 04:02

Quote:

Originally Posted by Speedy51

If there are any more numbers with 29 bits like 55559_219 if nobody wants them (1) I would be like to see if I can complete the LA phase. As I have done these numbers years ago and they were fairly easy. I have 16 gig of RAM.

You can try 14441_218 or 61111_221. I'll go for the first one.

Jarod · 2015-09-11, 06:01

Quote:

Originally Posted by pinhodecarlos

You can try 14441_218 or 61111_221. I'll go for the first one.

I will take they 61111_221 thanks so much

xilman · 2015-09-11, 08:13

Quote:

Originally Posted by frmky

W_795

Code:

p104 factor: 54874034920931135861719895187381751989817503532077501887202266914532304010850653889600111682921895593127
p105 factor: 355018650276589253569732059273899112648065410049958016080785072067440080051354861668797071294222218684689

GW_3_502

Code:

p96 factor: 122838606011459566941521222717596412576527332246352792452595461733479143231281284137259355896787
p146 factor: 46114840318482652720226564339831810352662035775416203849592174391227241727948892725892538747025784365157459587577397810713036812808162989587158579

Nice! Emphatically not ECM misses.

Paul

fivemack · 2015-09-11, 09:03

Code:

Thu Sep 10 19:32:57 2015  p67 factor: 4305484965211660320499368589393459040079707968434477586884591804617
Thu Sep 10 19:32:57 2015  p173 factor: 14741232394200649623358498484902297359802972545902775872664720200270136843033002261414940248655497471086834979835615230603232298339578937981741217011036131394287636501736993

52.8 hours for 10.94M target-density-120 matrix on E5-2650v2

Log attached

pinhodecarlos · 2015-09-11, 09:39

Sorry, I didn't mean I was going to post-process 14441_218. I was advising J to to 14441_218.

Jarod · 2015-09-11, 21:42

Quote:

Originally Posted by pinhodecarlos

Sorry, I didn't mean I was going to post-process 14441_218. I was advising J to to 14441_218.

If you don't want 14441_218 & the owner isn't in a hurry for the result (couple of weeks tops) I can take this number as well as 61111_221? Out of interest who do I send these results to? I just had a look at the numbers and saw that 14441_218 is reserved to Lionel Debroux I am more than happy if you want to keep this number

2015-09-11, 03:50	#346
Jarod AKA Speedy51 Oct 2012 New Zealand 11100011₂ Posts	29 Bits numbers If there are any more numbers with 29 bits like 55559_219 if nobody wants them (1) I would be like to see if I can complete the LA phase. As I have done these numbers years ago and they were fairly easy. I have 16 gig of RAM.

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

2015-09-10, 06:51	#342
Dubslow Basketry That Evening! "Bunslow the Bold" Jun 2011 40<A<43 -89<O<-88 3·29·83 Posts	I'll try 11611_109. Could someone estimate how much memory various 15e jobs would take? Am I right in guessing that (say) the GNFS194 from 4788 would be more than 16 GiB?

2015-09-10, 09:24	#343
fivemack (loop (#_fork)) Feb 2006 Cambridge, England 1100100011000₂ Posts	None of the 15e jobs will fit in 8GB, and with 32-bit large primes the filtering on the bigger jobs gets tight-to-impossible in 16G, but the matrix itself isn't too bad until you're into the high-C180s. The C182 from 147,125 took 14GB for matrix-building; the C257 SNFS 135,82 took 11GB for matrix building. C193_2340_742 was less over-sieved and took less memory to build the matrix, but the matrix was huge (31.5M) and would not fit in 16G. Note that the log file understates memory use for the matrix phase; for example, a matrix where the log says 'memory use: 4730.8MB' is currently running with virt=6588M res=6.1G

2015-09-11, 09:39	#351
pinhodecarlos "Carlos Pinho" Oct 2011 Milton Keynes, UK 11523₈ Posts	Sorry, I didn't mean I was going to post-process 14441_218. I was advising J to to 14441_218.