[QUOTE=bchaffin;242048]I'm running on a 12-core Westmere machine @ 2.93GHz. 11 cores are running ecm, and one is running msieve (no polys found yet).



A single curve at 110M takes 521 seconds. Based on that, 5300 curves would take about 3 days. So far I've done ~480. I'm fine with leaving it running for a few more days.[/QUOTE]

500@11e7, no factor.

I have run msieve on the GPU overnight, this is the best poly I have found:

[code]N 201662644625726223965504252321039229795541434802496056546534987125943330407091389007249091194783132977342010346522074086728121107031532400663175295605381304978591
SKEW 98978961.30
R0 -59786668246412028563961686958249
R1 209832325935724097
A0 1410324506049558171491985943536919366368608
A1 -315860711479235872142973805177138848
A2 1649030326880651398237137078
A3 345764896891481418099
A4 -39599847386
A5 264
# size 6.248e-016, alpha -7.470, combined = 7.955e-013 rroots = 3[/code]

The poly posted by jrk yields ~1.4 relation per q value over a range of 1000 q values starting at (r&a)lim/2.
For a 30-bit large primes task, on average, RSALS clients can sieve the bulk of a range of 30-40M q values, split in WUs of 10K q values, in 3-4 days.

[QUOTE=debrouxl;242436]The poly posted by jrk yields ~1.4 relation per q value over a range of 1000 q values starting at (r&a)lim/2.
For a 30-bit large primes task, on average, RSALS clients can sieve the bulk of a range of 30-40M q values, split in WUs of 10K q values, in 3-4 days.[/QUOTE]Will you also be able to handle the post-processing?

I've completed about 5300 curves at 110M now. I've found 3 polys but the best is norm 5.531429e-16 so not as good as Andi47's. I'll leave it running until I see that sieving has started, but I'll probably sit this one out since I'll be traveling for the next couple weeks.

I think the poly from jrk is the better poly - isn`t ist?
[URL]http://www.mersenneforum.org/showpost.php?p=241828&postcount=1071[/URL]

schickel: RSALS relies on external contributors to do the post-processing. Is there somebody interested in Aliquot sequences willing to do the job ?

AFAICS, the best poly posted here so far was by jrk, indeed.

Ok, count me in for finishing.

Reserving 50M-80M for the RSALS grid :smile:
At the current rate of crunching, I'll start "C162_4788_2602" by Wednesday.

Pls send me a PM when needed, because I may not notice/follow specific threads. (But I'll receive a PM for sure.)

I've just added "C162_4788_2602" to the RSALS work generation infrastructure. For now, WUs will be generated from q=50M to 80M.

At the time of this writing, nearly 5/6 of the 50M-80M range have been distributed to clients, with more than 1/5 of the corresponding results returned, yielding ~6M raw relations.
I'm therefore expanding the RSALS reservation to 80M-90M :smile: