Author 
Message 
camel
Joined: 26 Dec 2007 Posts: 50 Location: Austria 

level 556 unsolveable? 

hi everyone,
are the crossflip levels the same for everyone? it seems as if my level 556 is not solveable or as if something rather weird going on. i tried with and without keeping blocker tiles in the 2 matrices as zero rows. i tried with 2 of my own solvers and also via a well known public Z(2) library. all the previous 555 levels are no problem ... i can solve them within seconds each.
will look into it ... but as the library comes up with a solution i guess the matrix is solveable and the problem would be on my side.
camel


Fri Oct 26, 2012 7:49 pm 


camel
Joined: 26 Dec 2007 Posts: 50 Location: Austria 



first i thought it could have to do with the square of the sides going beyond 256^2 but as i leave out the blocker tiles in my last attempts the amatrix does not reach 256^2 by 256^2 yet.
will investigate


Fri Oct 26, 2012 7:53 pm 


camel
Joined: 26 Dec 2007 Posts: 50 Location: Austria 



ah ok ... this really could be the problem ... matrix goes over 65536^2 from level 555 to 556 even without blockers ...
level=555 / board_width=256 / board_height=256
#of blockers=6
reading 65530 x 65530 matrix with at most 655300000 nonzero entries (density at most: 0.15260)
reading 65530 x 1 matrix with at most 65530 nonzero entries (density at most: 1.00000)
Starttime: 1351256590 Endtime: 1351256708
level=556 / board_width=257 / board_height=257
#of blockers=215
reading 65834 x 65834 matrix with at most 658340000 nonzero entries (density at most: 0.15190)
reading 65834 x 1 matrix with at most 65834 nonzero entries (density at most: 1.00000)
Starttime: 1351255976 Endtime: 1351256110


Fri Oct 26, 2012 8:08 pm 


camel
Joined: 26 Dec 2007 Posts: 50 Location: Austria 



turned out that the well known library really had a bug. so hacker.org even helped to make the world of math libraries a little better ))


Tue Oct 30, 2012 3:21 pm 


