I'll also take GW_11_219: results early next week (probably Monday morning)

I'll take GW_5_326.

fivemack: in fact, the input was passed through remdups4, and there were only two lines that msieve was complaining about.

For now, I'll start GW_7_270. Downloaded 7256854198 bytes in 11'22" (722s), with peaks at 19-20 MB/s.

GW_7_270 failed as well at the beginning of the matrix step, earlier than GW_2_757, with a different error which I don't have anymore (screen scrolled too far, and it's not in the log file).

jasonp mentioned at [url]http://www.mersenneforum.org/showpost.php?p=376614&postcount=14[/url] that there was something wrong with the changes in msieve SVN r966, so I've reverted to msieve SVN r965, rebuilt, and restarted filtering on GW_2_757. We'll see how filtering goes, this time :smile:

Your symptoms are exactly that of the large dataset bug. Stay tuned, I think it's about to get fixed.

I have a question,
is it normal that at Detailed status of lasievee from the number C176_3270_692 Aliquot GNFS(176) the Est. Pending Rels grow up?
The number have 329089910 Relations - but the Est. Pending Rels do not decrease. :question: Is that correct?

[URL]http://escatter11.fullerton.edu/nfs/crunching_e.php[/URL]

Regards Andi_HB

More Q-values are being queued up to sieve as results are coming back; I asked for 20-100, but it appears that 20-140 is being done.

W_2_756 done
 

[code]
prp78 factor: 374789068685526688677379478014904956820705500189408553811085313643061276770427
prp120 factor: 237877110234048839646350808833645374938138200009549725842325525328995455270196872995492313303158280189797174047653056297
[/code]

About 33 hours for 6.3M matrix on three cores i7/2600

Log at [url]http://pastebin.com/1aD5e5uU[/url]

[QUOTE=fivemack;376695]More Q-values are being queued up to sieve as results are coming back; I asked for 20-100, but it appears that 20-140 is being done.[/QUOTE]

That's intentional based on the number of relations actually returned. With 32 bit large primes, it defaults to a larger number of relations.

[QUOTE=frmky;376721]That's intentional based on the number of relations actually returned. With 32 bit large primes, it defaults to a larger number of relations.[/QUOTE]

Ah, OK. I was slightly intending to probe how few relations were required to make a usable matrix with 32-bit large primes (appreciating that I might have to do a bit of sieving myself) - it's heading for about 460M, I suspect 360M would be enough.

I'll take GW_4_378 next.


GW_5_326 eta is Friday night.