Thread: News
View Single Post
Old 2008-07-14, 02:49   #32
gd_barnes
 
gd_barnes's Avatar
 
May 2007
Kansas; USA

1009910 Posts
Default Sieving k=401-1001 for n=600K-1M has started!

Bruce has kindly dedicated a good portion of his substantial computing power to sieving NPLB's next attack: k=401-1001 for n=600K-1M.

The new goal now is to complete drives 1 and 3 to n=600K by Oct. 31st and that is when Bruce is shooting to have sieving done.

With speedy 64-bit sr2sieve running on speedy 64-bit Linux machines, a very rough estimate of the optimal sieve depth is P=25T, although that has not been determined yet.

To give you an idea of the substantial resources it takes to properly sieve 300 k's at such a high n-range, I estimated that it will take 4-5 CPU years! It looks like Bruce will have no problem meeting that date himself but if we need help from the team, I'll post sieving reservations as needed.

With the newer speedy sr2sieve and 64-bit systems, we will also look to do some additional sieving on drive 3 for n=600K-1M after completing the drive 1 sieving. I'm estimating optimal to be P=20T-25T there but it is only currently sieved to P=14T, which was optimal for 32-bit systems at the time.

So come along and load up your machines as we attempt to complete drives 1 and 3 to n=600K by Oct. 31st!


Gary
gd_barnes is offline   Reply With Quote