View Single Post
Old 2008-02-04, 19:04   #2
gd_barnes
 
gd_barnes's Avatar
 
May 2007
Kansas; USA

132×61 Posts
Default Status for sieveing 300<k<400 for n=260K-1M

As you may know, about a week ago I started sieving 300<k<400 for n=260K-1M. I said I may ask for some sieving help in the near future. I have some good, better, and best news:

Good news:
With 9 cores now on it, I will reach P=6T by ~2/26 (my original target being end of Feb.) so I won't need any help unless someone REALLY wants to sieve! Let me know if you enjoy sieving and we can coordinate.

I reached P=660G last Friday and have now split P=~600G ranges over the 9 cores (660G-1.2T, 1.2T-1.8T, 1.8T-2.4T, etc.). At that point, I'll break off n=260K-600K for everyone and continue sieving n=600K-1M.

Better news:
Everyone can keep their speedy machines running LLR tests on our current files. The Feb. month end target also gives us plenty of time to start finding primes for 300<k<400 starting at n=333333 before they start dropping off (probably about 3-4 months from now). This will put as much historical data as possible for this k-range in Prof. Caldwell's database.

Best news:
There will be something for everyone for 300<k<400. A few individual-k reservations, some testing starting at n=600K for those who like big primes (after sieving to n=1M is complete), and of course the usual team drive set up.

I think this works out well. My machines consistent mostly of 1.66-Ghz Dell CD's and 3-Ghz P4's, none of which are overclocked, mixed in with a couple of Athlon's. Many of you have much faster machines. I think that puting these somewhat slower machines on sieving is a better use of them because LLRing is about 95% of total time. (Of course I have to have a few doing some LLRing!)

Any thoughts or opinions are always welcome.


Gary
gd_barnes is offline   Reply With Quote