View Single Post
Old 2009-11-10, 04:56   #1
gd_barnes
 
gd_barnes's Avatar
 
May 2007
Kansas; USA

237628 Posts
Default Sieving Riesel & Sierp base 16

Hey all,

I'm currently sieving Riesel and Sierp base 16 together for the remaining unreserved k's for n=200K-500K as a slightly lower priority effort. The optimum sieve depth for breaking off n=200K-300K is P=12T. This assumes the very likely possibility that we won't complete testing both drives to n=300K within ~1 year and so only takes into account the rate of factor removal from the n=200K-300K portion of the file; not the entire file.

I'm not going to make a formal sieving drive out of it but if anyone would like to help, please post a reservation here.

The file sieved to P=8T is attached to this posting. I'll remove factors from it as we progress. It takes one 64-bit 2.4 Ghz core ~5 days to sieve a P=1T range using 64-bit sr2sieve.

When we're at P=12T, I'll remove the n=200K-300K part of the file and post some files in the Riesel base 16 drive for testing. When/if we find some more primes on Sierp base 16 before reaching n=200K, I'll remove those k's from the file. We can also continue sieving n=300K-500K as a very low priority effort if anyone wants to.

These are good bases to run the much faster-testing power-of-2 bases. Iirc, even with the new PFGW, as a general rule, I've found that they test ~2.5-3 times faster than non-power-of-2 bases for the same size k-value and # of digits. (Yes, the smaller k-values can test much faster than the larger ones, even on non-power-of-2 bases using the new PFGW.)


Reservations:
Code:
gd_barnes      P<=3T       complete
Flatlander     P=3T-4T     complete
gd_barnes      P=4T-5T     complete
Flatlander     P=5T-8T     complete
gd_barnes      P=8T-12T    complete
(Sieving is paused as we break off and test n=200K-300K.)
Gary
Attached Files
File Type: zip sieve-riesel-sierp-base16-200K-500K.zip (211.3 KB, 136 views)

Last fiddled with by gd_barnes on 2009-12-14 at 09:24 Reason: update status
gd_barnes is offline   Reply With Quote