-
The number 83810205 has factors 12345 x 6789. Is there a way to know that the fourth digit of 0 in the result corresponds to 2 and 6 in the original factors? The premise of the scoring system is that there must be a way to get some improvements in there by training a solution before hand. This might be impossible, but it would be interesting to know why as it raises a lot of other issues.
2008-03-20 06:55:01 UTC by tbandrow
-
12345 x 6789.
2008-03-20 06:50:16 UTC by tbandrow
-
We prune the tree on the tail side by looking and seeing if the factored pair is larger than what we have now. Is there a way that we can skip nodes to arrive at our node?.
2008-03-20 06:49:58 UTC by tbandrow
-
The existing big int is a wrapper around the gnu bignum, which is pretty slow. It's slow partially because it is a genuinely big number, allowing for thousands or millions of digits. Can we just have a big int x86-64 assembly that does digits up a fixed number in the range we're interested in factoring. Certainly anything more than 2048 bits would be fanciful at this point.
2008-03-20 06:43:54 UTC by tbandrow
-
Why does -s sometimes make factor iteration counts go up on occasion? I would think scoring would only have a positive impact but not a negative one. Then again, an early score node could launch a search that it is exhaustive. Can you find out if a scored node would be excluded becuase a previous node was greater?.
2008-03-20 06:41:33 UTC by tbandrow
-
tbandrow committed revision 7 to the factdiv SVN repository, changing 9 files.
2008-03-20 06:37:30 UTC by tbandrow
-
tbandrow committed revision 6 to the factdiv SVN repository, changing 6 files.
2008-03-07 18:23:11 UTC by tbandrow
-
tbandrow committed revision 5 to the factdiv SVN repository, changing 2 files.
2008-03-07 18:22:07 UTC by tbandrow
-
tbandrow committed revision 4 to the factdiv SVN repository, changing 6 files.
2008-03-07 18:20:56 UTC by tbandrow
-
tbandrow committed revision 3 to the factdiv SVN repository, changing 1 files.
2008-02-22 19:44:22 UTC by tbandrow
