BOINC NFS sieving - RSALS - Page 41

debrouxl · 2012-08-07, 17:42

As the disk occupation ratio is now down to ~80%, I'm definitely using the clients to finish more numbers - I have started more numbers

Mathew · 2012-08-07, 19:32

GW_4_387 is complete

Code:

prp64 factor: 1515636209967489968041137518997387054449302972575057595046341897
prp82 factor: 6988427235017527575941901808277674842110407205732649461358768262340190766334140897
prp91 factor: 3630389219936118399137319889719189767005891975979850676502689317400599004722336064581555623

pinhodecarlos · 2012-08-07, 19:47

Quote:

Originally Posted by Mathew

GW_4_387 is complete

Code:

prp64 factor: 1515636209967489968041137518997387054449302972575057595046341897
prp82 factor: 6988427235017527575941901808277674842110407205732649461358768262340190766334140897
prp91 factor: 3630389219936118399137319889719189767005891975979850676502689317400599004722336064581555623

Mathew, please post the log file. Thank you.

RichD · 2012-08-08, 04:56

Quote:

Originally Posted by debrouxl

Go ahead, Rich

Whew! I think I have it running.

Code:

Tue Aug  7 23:46:49 2012  memory use: 2895.7 MB
Tue Aug  7 23:48:04 2012  linear algebra at 0.0%, ETA 139h56m

jrk · 2012-08-08, 20:33

Reserving 577_89_minus1

jrk · 2012-08-10, 00:11

Quote:

Originally Posted by jrk

Reserving 577_89_minus1

Done. Factors are in the OPN factors thread.

Reserving 4873051_31_minus1 next.

debrouxl · 2012-08-12, 11:53

The disk occupation ratio remains at ~90%, even with lengthy periods of grid starvation...
There are currently three unreserved 31-bit LPs task, among which is the hardest GNFS task ever on RSALS (a record-setting difficulty of 172 for the near-repdigit project), and two unreserved 30-bit LPs tasks.
Interested post-processers will have to manage backups of the raw relations themselves, so that I can remove the datasets from the server

BOINC's indication of the power of the grid shows that it's now around 900 GFLOPS, up from less than 300 GFLOPS most of the time. The gap between RSALS and NFS@Home is temporarily bridged by half an order of magnitude.
Therefore, if someone has hard numbers of mathematical interest (GNFS 16x or 17y, with y probably < 4; SNFS 23x-24x with a sextic polynomial), now is the time to take advantage of the RSALS end-game, as long as clients are interested

I already have an Aliquot C171 in sight (http://www.mersenneforum.org/showpos...&postcount=789), when I have made some room for new relations.

xilman · 2012-08-12, 13:30

Quote:

Originally Posted by debrouxl

Therefore, if someone has hard numbers of mathematical interest (GNFS 16x or 17y, with y probably < 4; SNFS 23x-24x with a sextic polynomial), now is the time to take advantage of the RSALS end-game, as long as clients are interested

I already have an Aliquot C171 in sight (http://www.mersenneforum.org/showpos...&postcount=789), when I have made some room for new relations.

I've quite a few SNFS and GNFS numbers from the GCW tables but, unfortunately, most haven't had a 2/9 or 1/3 level of ECM done. If you're prepared to take the risk of an ECM miss, let me know.

Paul

RichD · 2012-08-12, 14:59

I've got 33 hours left in LA for GC_10_234 and the .dat file is stored on two separate computers. I should be able to finish this one off.

Feel free to remove it from the server.

debrouxl · 2012-08-12, 15:35

Paul: sieving is currently fairly cheap at RSALS, so I'm not too afraid of ECM misses. Maybe I could make RSALS sieve GC_7_280_C185 ?

Rich: in fact, I have already removed that file yesterday or even earlier

Dmitry has sent me a polynomial, I've just started the Aliquot C171 on RSALS. It will be a 30-bit LPs task, so sieving will be performed over a fairly wide range of q values.

jrk · 2012-08-12, 15:57

Quote:

Originally Posted by jrk

Reserving 4873051_31_minus1 next.

Done. Factors are in the OPN factors thread.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
BOINC NFS sieving - NFS@Home	frmky	NFS@Home	1828	2020-11-14 04:57
BOINC	Unregistered	Information & Answers	6	2010-09-21 03:31
BOINC.BE	BATKrikke	Teams	2	2010-03-05 18:57
Boinc	Xentar	Sierpinski/Riesel Base 5	4	2009-04-25 10:26
BOINC?	masser	Sierpinski/Riesel Base 5	1	2009-02-09 01:10

2012-08-07, 17:42	#441
debrouxl Sep 2009 1721₈ Posts	As the disk occupation ratio is now down to ~80%, I'm definitely using the clients to finish more numbers - I have started more numbers

2012-08-08, 20:33	#445
jrk May 2008 3·5·73 Posts	Reserving 577_89_minus1

2012-08-12, 11:53	#447
debrouxl Sep 2009 977 Posts	The disk occupation ratio remains at ~90%, even with lengthy periods of grid starvation... There are currently three unreserved 31-bit LPs task, among which is the hardest GNFS task ever on RSALS (a record-setting difficulty of 172 for the near-repdigit project), and two unreserved 30-bit LPs tasks. Interested post-processers will have to manage backups of the raw relations themselves, so that I can remove the datasets from the server BOINC's indication of the power of the grid shows that it's now around 900 GFLOPS, up from less than 300 GFLOPS most of the time. The gap between RSALS and NFS@Home is temporarily bridged by half an order of magnitude. Therefore, if someone has hard numbers of mathematical interest (GNFS 16x or 17y, with y probably < 4; SNFS 23x-24x with a sextic polynomial), now is the time to take advantage of the RSALS end-game, as long as clients are interested I already have an Aliquot C171 in sight (http://www.mersenneforum.org/showpos...&postcount=789), when I have made some room for new relations.

2012-08-12, 14:59	#449
RichD Sep 2008 Kansas 3,389 Posts	I've got 33 hours left in LA for GC_10_234 and the .dat file is stored on two separate computers. I should be able to finish this one off. Feel free to remove it from the server.

2012-08-12, 15:35	#450
debrouxl Sep 2009 977 Posts	Paul: sieving is currently fairly cheap at RSALS, so I'm not too afraid of ECM misses. Maybe I could make RSALS sieve GC_7_280_C185 ? Rich: in fact, I have already removed that file yesterday or even earlier Dmitry has sent me a polynomial, I've just started the Aliquot C171 on RSALS. It will be a 30-bit LPs task, so sieving will be performed over a fairly wide range of q values.