no, it's hosted at the wrong end of my insufficiently reliable home ADSL connection

[quote=smh;214516]
[quote=debrouxl;214499]... I was crossing fingers that we'd see only single ECM misses (since a single factor in the 45-60 digits range wouldn't have changed the difficulty of the factorization, SNFS would have remained easier than GNFS)... failed..[/quote]
Unless that single factor would have left a prime cofactor of course.[/quote]
There was a fairly painful to watch [URL="http://homepage2.nifty.com/m_kamada/math/c/11117.htm#N228_C201"]point in case[/URL] two months earlier, but nobody mentioned it, so I thought that was taken as a lesson (from mother nature!), but instead now saw this c250 and my heart sunk again. 2/9 for SNFS, 2/7 for ECM is a good rule of thumb (some people would say - that's even too low).

[QUOTE=Batalov;214552]There was a fairly painful to watch [URL="http://homepage2.nifty.com/m_kamada/math/c/11117.htm#N228_C201"]point in case[/URL] two months earlier, but nobody mentioned it[/QUOTE]

[url]http://www.mersenneforum.org/showpost.php?p=207824&postcount=71[/url]

A reminder of how awesome three-channel RAM is
 

I moved one of the linalg jobs to my i7/920 machine.

Previously, running four threads on the dual-2.5GHz-quad-opteron was going at about 90% the speed of four threads on a Q6600 - I imagine this is the overhead of having two memory controllers, as well as the matrix being slightly bigger.

On the i7/920, it runs at 210% the speed of the Q6600; should be done by 27 May, at which point I'll move the job on the Q6600 to the i7 to finish.

I have successfully performed the filtering of "31_149_minus1", and started the LA step. But it requires too much RAM for my computers (during the loading phase, it peaks at 2.7 GB of RAM...). I have therefore uploaded 31_149_minus1-compressed-with-pbzip2.dat.{cyc,mat,chk} files to the server, so that someone else can finish the LA, and provide me with the resulting .dep file :smile:

I've got another quad-core back in commission, and am at a lull with large GNFS jobs from the aliquot sequences, so I'll do this one too.

(1330: job running, ETA 97 hours from now; I downloaded the .dat so am running all three steps. Should have a factor for you on Thursday evening)

[url]http://www.fivemack.org/gnfs-completion/logs/31_149.mlog[/url] (P59 * P155)

Thanks :smile:
Considering that even a full ECM run to t55 (i.e. beyond ~2/9 of the SNFS difficulty) was rather unlikely to find a p59, this is probably not an ECM miss. Good.

Activity seems to have picked up a bit between yesterday and today: the number of WUs in progress has raised from 2000-2500 to nearly 4000. Since we're soon going to start sieving a couple 31-bit-primes tasks for OddPerfect, this is good.

[url]http://www.chiark.greenend.org.uk/~twomack/89999-243.mlog[/url] (P77*P160)

Thanks Tom, now reported :smile:

A nice big job. Any idea why there were relation reading errors reported when the linear algebra built the matrix? One would think all the bad relations would have been thrown away by then, since they would never have contributed to matrix columns, but the number of dependencies found doesn't seem to be affected.