Renslay
Member
Finally, I finished my optimal 2x2x2 solver. It is written entirely in MATLAB.
The software can build and use an optimal table, which contains all the necessary information for solving the 2x2x2 optimally.
The interesting fact about the table is that it requires only 734832 bytes to store, which is 1.6 bit information for a cube state, and still can be used as a look-up table directly.
This little project is the result of the conversation started here:
http://www.speedsolving.com/forum/showthread.php?43928-Which-Moves-in-20-Minimum&p=904485&viewfull=1#post904485
Storing the optimal table requires 1.6 bits per state, however, the software uses more memory for generating and using the table. This is only for keeping the runtime low and the source code short, managable and understandable.
The optimal table is pre-built and stored, however, the functions provide its generation. On my laptop it takes about 7 seconds.
Many thanks for Stefan Pochmann for inspiration and Tom Rokicki for the table building strategy! (And making possible the 7 seconds runtime instead of the original 30 minutes.)
You can download the package here.
Or if that doesn't work, try here.
You can find some interesting data here, resulted by the modifications of this software.
The comments are possibly full of typos.
Any bug report or suggestion is welcomed!
Renslay
The software can build and use an optimal table, which contains all the necessary information for solving the 2x2x2 optimally.
The interesting fact about the table is that it requires only 734832 bytes to store, which is 1.6 bit information for a cube state, and still can be used as a look-up table directly.
This little project is the result of the conversation started here:
http://www.speedsolving.com/forum/showthread.php?43928-Which-Moves-in-20-Minimum&p=904485&viewfull=1#post904485
Storing the optimal table requires 1.6 bits per state, however, the software uses more memory for generating and using the table. This is only for keeping the runtime low and the source code short, managable and understandable.
The optimal table is pre-built and stored, however, the functions provide its generation. On my laptop it takes about 7 seconds.
Many thanks for Stefan Pochmann for inspiration and Tom Rokicki for the table building strategy! (And making possible the 7 seconds runtime instead of the original 30 minutes.)
You can download the package here.
Or if that doesn't work, try here.
You can find some interesting data here, resulted by the modifications of this software.
The comments are possibly full of typos.
Any bug report or suggestion is welcomed!
Renslay
Last edited: