mersenneforum.org When will it stop?
 Register FAQ Search Today's Posts Mark Forums Read

 2010-05-17, 02:10 #1 ThomRuley     May 2003 3638 Posts When will it stop? Been running msieve for about a week on the same number, and noticing some confusing behaviour: Total yield: 573661 0/0 mpqs failures, 368057/167535 vain mpqs milliseconds total: Sieve 6153294 Sched 0 medsched 991531 TD 2859366 (Init 95818, MPQS 554544) Sieve-Change 2967952 TD side 0: init/small/medium/large/search: 25369 260872 71926 215193 631399 sieve: init/small/medium/large/search: 124365 1070449 109121 1595205 500229 TD side 1: init/small/medium/large/search: 19996 220143 73040 218601 1025779 sieve: init/small/medium/large/search: 65963 692984 110585 1622884 261509 appending spairs.out.T0 to spairs.out appending spairs.out to 4-27-10.dat Found 18965932 relations, 113.4% of the estimated minimum (16720000). -> msieve -s example\4-27-10.dat -l example\4-27-10.log -i example\4-27-10.ini - nf example\4-27-10.fb -t 1 -nc1 113.4%? Shouldn't it be doing the linear algebra step now? Why has it kept going?
 2010-05-17, 02:58 #2 Batalov     "Serge" Mar 2008 Phi(3,3^1118781+1)/3 2·23·197 Posts That's "... 113.4% of the estimated minimum ..." (this minimum is necessary but not sufficient). Minimum is estimated from past experience. Before minimum is collected, there is no use to even present the data to msieve's filtering attempt (which precedes algebra). The specific amount needed for your project will be controlled by the script. Just leave it running; everything looks ok.
 2010-05-17, 03:13 #3 ThomRuley     May 2003 3638 Posts Thanks. Guess I'm just used to DC projects where the number of steps needed is exactly known.
 2010-05-17, 03:30 #4 EdH     "Ed Hall" Dec 2009 Adirondack Mtns 2×3×72×11 Posts It has stopped (sieving), as shown by: Code: -> msieve -s example\4-27-10.dat -l example\4-27-10.log -i example\4-27-10.ini - nf example\4-27-10.fb -t 1 -nc1 It made it to 113.4% because it sieves in batches and for the last batch, it started below 100% and by the time it finished the batch, it was at 113.4% It should next move to -nc2 and then -nc3 Last fiddled with by EdH on 2010-05-17 at 03:32
2010-05-17, 13:36   #5
ThomRuley

May 2003

35 Posts

Quote:
 Originally Posted by EdH It has stopped (sieving), as shown by: Code: -> msieve -s example\4-27-10.dat -l example\4-27-10.log -i example\4-27-10.ini - nf example\4-27-10.fb -t 1 -nc1 It made it to 113.4% because it sieves in batches and for the last batch, it started below 100% and by the time it finished the batch, it was at 113.4% It should next move to -nc2 and then -nc3
Actually, what happened was that the batch that went over went to 102%, then 106%, then 110% then 113%. Should I be concerned about that?

 2010-05-17, 14:33 #6 fivemack (loop (#_fork))     Feb 2006 Cambridge, England 143168 Posts Have a look at the file example\4-27-10.log if it has a note at the bottom about 'needs 1000000 more relations', just let everything carry on; this just means that the minimum has been under-estimated and it needs to find more relations. If not, post the bottom thirty or so lines here and we'll tell you what to do.
 2010-05-17, 20:59 #7 ThomRuley     May 2003 24310 Posts I haven't found a log file, but there is a HUGE text file (now around 3GB). Is that it? I would look at it but I don't have enough RAM.
 2010-05-17, 21:20 #8 kar_bon     Mar 2006 Germany B0A16 Posts You can download a small Lister program which supports big files, too. Try the 32Bit-version. Perhaps it helps. But I think the 3GB file is the relations found, the LOG file is more ggnfs.log or msieve.log.
 2010-05-17, 23:01 #9 EdH     "Ed Hall" Dec 2009 Adirondack Mtns 2·3·72·11 Posts I agree that the 3GB file is probably the relations, but shouldn't the log file be the one in the call line for msieve (-l example\4-27-10.log) and found in the example sub-directory in whatever directory msieve is run from? How large is the composite? 16.7M+ is a large number of relations. A c106 was fine with well under 5M for me recently. In fact, I was having trouble with too many relations for awhile. How long does the filtering step (-nc1) take based on relations? I don't remember it ever taking very long, but I haven't been working with composites much over 110 digits any time recent.
 2010-05-18, 00:56 #10 Batalov     "Serge" Mar 2008 Phi(3,3^1118781+1)/3 2·23·197 Posts Apparently, Thom is factoring a gnfs-127 from the OP thread. He will be done soon.
 2010-05-18, 01:01 #11 ThomRuley     May 2003 35 Posts Can't seem to find that kind of file. Is that going to be a problem?

 Similar Threads Thread Thread Starter Forum Replies Last Post SaneMur Riesel Prime Search 4 2011-07-18 17:04 Joshua2 Software 6 2005-05-16 16:36 jocelynl 15k Search 2 2004-07-10 13:31 ebx 3*2^n-1 Search 9 2004-07-06 15:30 paulunderwood 3*2^n-1 Search 8 2004-05-19 07:10

All times are UTC. The time now is 03:05.

Wed Jul 15 03:05:42 UTC 2020 up 112 days, 38 mins, 0 users, load averages: 1.87, 1.71, 1.66