Commits
Author  Commit  Message  Labels  Comments  Approvals  Date 


think I've got transition tables up





start new array based implementation to explore symmetries





add solve.bash





4ecca13
M

merge





remove distribute and gather





get rid of all the options. worry about those when the code is clean and fast





update TODO





first pass at using depth % 3 for pruning tables. only 1 table per phase. still extremely sloppy, and slower than I'd like.





solve is void, not int. cleanup





faster init_prune, start searches at max prune value instead of 0





fixed and better pruning table for 2_solve (although guess it's not a pruning table anymore is it?)





faster iterative init_prune for 2x2x2 solver





bunch of messy experiements for target length, first pass at exponential damping with small list





15d8fc8
M

merge





add one_sol, max_p1d, interval, optimal





first pass at adjusting timeout to meet target length solutions. sloppy, wrong, underdamped, get wild swings.





cleanup





add PHASE enums, move some code from solve() to begin_phase2()





update README





decrement intsead of assignment





cleanup, add option to updated printout every m lines





add SOLVE variable to easily change binary/arguments, kill all pids





add programs to makefile





add 2_all to generate all possible 2x2x2 cube states, 2_rand for random states





instead of searching with prune tables, generate full cube space and iterate across it to solve





update TODO





use 2 coordinates for all pruning tables, write/read table file, some more options





don't recross boundaries between phase 1 and 2 (misses some solutions, won't prove optimal, but faster according to kociemba)





fix max_len vs len





add some options, a little cleanup


