15:1 is what I get, so I guess 20:1 is normal.
Range 1 Range 2 Range 3 Total
Factors 40 44 68 152
Out of R. 48 59 66 173
Excluded 643 767 990 2400
Total 731 870 1124 2725
How many duplicates is normal? I'm getting a 20:1 ratio of duplicates to unique factors, with a handful out of range.
15:1 is what I get, so I guess 20:1 is normal.
Range 1 Range 2 Range 3 Total
Factors 40 44 68 152
Out of R. 48 59 66 173
Excluded 643 767 990 2400
Total 731 870 1124 2725
Yeah its like 17:1. I've only found a couple factors so there is potential for skew.
Even over a whole 1T range there can be some deviation:
908T: 18 new, 28 range, 315 dup
909T: 20 new, 29 range, 363 dup
915T: 21 new, 35 range, 404 dup
919T: 17 new, 33 range, 370 dup
926T: 23 new, 17 range, 343 dup
Quad 2.5GHz G5 PowerMac. Mmmmm.
My Current Sieve Progress: http://www.greenbank.org/cgi-bin/proth.cgi
Yeah a deviation of 6 factors over 1T is a lot. Sieving 1T is going to take me into the depths of winter so I hope to find as many factors as possible. Alternately, PRP is projected to give me only 1 or 2 proofs (30:1 speed ratio?) so I think I'm doing the right thing.
Unfortunately, one of the things I'm finding about this project is that nailing anything to within a 30:1 performance ratio is a difficult bit of math. A lot of project variables skew to a factor of 2-4 meaning that overall project performance is a grey area under a factor of 20-30. Meanwhile, sieving is considered a 30:1 improvement over PRP which is the edge of our ability to calculate. Might be why massive sieving has been such a hard sell.