Subproject #3: 700k-800k sequences to 100 digits

[unconnected]

Originally Posted by 10metreh

BTW, how is 770k-780k getting on?

It is ready but I need time to upload data into DB and make detailed report.

[10metreh]

Originally Posted by unconnected

It is ready but I need time to upload data into DB and make detailed report.

For the milestones, have you got the dates when each 5k half (770k-775k and 775k-780k) finished?

Originally Posted by Greebley

For windows, it doesn't correctly put in a new line so you get two indices on the same line. So it does continue, but you have to go back and add the newline in or the db is missing one index. I am pretty sure this wasn't ever fixed.

This doesn't happen to me, and I'm on Windows.

[Buzzo]

done with 761478
reserving 768072

[unconnected]

Originally Posted by 10metreh

For the milestones, have you got the dates when each 5k half (770k-775k and 775k-780k) finished?

770k-775k - Jan 12
775k-780k - Jan 9

[Joshua2]

Originally Posted by 10metreh

Also, you don't need to post your .elf files. Just submit them to factordb.com by running aliqueit <the_sequence> -s 0.

i dont have wget on my windows 7.

[Mini-Geek]

Originally Posted by Joshua2

i dont have wget on my windows 7.

This is normal, since it's a GNU app normally. You can download a version of it for Windows. The setup link from http://gnuwin32.sourceforge.net/packages/wget.htm should be all you need.

[EdH]

OK, looking for some "sage" advice.

I ran 761040 for 14 hours today without any advancement. Most of the day it seemed to say "c92: running ### ECM curves..." I won't be able to run it for a full day tomorrow, if at all. I have my 24/7 machine working on the 4788 team sieve. Next time I start aliqueit, I don't want to just rerun today. Should I start aliqueit with the -e option next time to skip that portion, or should I run ggnfs on the composite and skip letting aliqueit run the steps?

[frmky]

765864, 767520, 767640, 767802, 767550 are done.
I'll take 780042, 780312, 780324, 780390, 780414, 780504, 780654, 780870, 780888, 780906, 780912.

[RichD]

Originally Posted by EdH

OK, looking for some "sage" advice.

I ran 761040 for 14 hours today without any advancement. Most of the day it seemed to say "c92: running ### ECM curves..." I won't be able to run it for a full day tomorrow, if at all. I have my 24/7 machine working on the 4788 team sieve. Next time I start aliqueit, I don't want to just rerun today. Should I start aliqueit with the -e option next time to skip that portion, or should I run ggnfs on the composite and skip letting aliqueit run the steps?

Yep, I'm in the same boat. I have a 0.8 GHz G4 Mac and would recommend running the -e option to cut my loses, but I don't have ggnfs (just yet). Only running msieve on the big boys.

[Batalov]

I'll take 781254 781302 781950 782472 782838 782880 783804 783960 783996

[mdettweiler]

Originally Posted by EdH

OK, looking for some "sage" advice.

I ran 761040 for 14 hours today without any advancement. Most of the day it seemed to say "c92: running ### ECM curves..." I won't be able to run it for a full day tomorrow, if at all. I have my 24/7 machine working on the 4788 team sieve. Next time I start aliqueit, I don't want to just rerun today. Should I start aliqueit with the -e option next time to skip that portion, or should I run ggnfs on the composite and skip letting aliqueit run the steps?

Definitely go with -e. I've noticed that aliqueit's formula for estimating GNFS time (which it uses for calculating ECM amounts) somewhat heavily overestimates on numbers smaller than C95 or so, thus leading to too much ECM. Last time a new version of aliqueit was released, the crossover to QS was around C95--and for numbers below that, aliqueit's QS estimation formula is much more accurate, whether you're actually going to do QS or GNFS with 11e. However, unfortunately there's no easy way that I know of to tell it to use the QS formula for C<95 while leaving the GNFS cuttoff at 85. Anyway, the upshot of this is that if it was running for more than 20-30 minutes or so doing ECM for a C92, it's probably plenty.