ctak - An implementation of Tak and a computer opponent in C

Age	Commit message (Collapse)	Author
40 hours	experimenting with a lazy queue implementationlazy_queue	tslil
	presently it's not correct
40 hours	More towel wringing: re-implemented check_road_colour	tslil
	Previously check_win would call check_road_colour once for each road colour, and check_road_colour would call a depth-first search (DFS) for each of the two axes. This meant that we were doing (up to) four depth-first searches for each call of check_win. I have replaced both axial DFS with the world's worst TM implementation of a connected component generation algorithm backed by the least guaranteed disjoint set data structure. Essentially doing anything about union find correctly is slower than just ... not doing it. Although we lose the asymptotic complexity, in practice we're doing this millions of times per turn, for a fixed board size and that's what matters. All in all, it appears that i've managed to shave about 69ns off check_win, per call -- nice! This amounts to 50ms or so saved at depth 5 per engine move, in one of my test games. Unfortunately nearly 99% of the time is still taken by evaluating the convolutional neural network. It's slow.
40 hours	Trying to make things faster	tslil
	I tried the following, but they all made things worse: - moving away from the singly-linked (tail tracking) list for actions by: + using an array zipper for a deque + using an array to poorly hold a floating deque - caching the results of generating move lists in the transposition table and then + copying the resulting list/zip/deque instead of generating it + applying the move-to-front without copying, but this made the search order worse. Presumably in this case shallower nodes were messing up the search tree with garbage moves? I think some of this is not supposed to happen, but i have just the right combination of poor evaluation function and naively ordered and cheap move generation that i'm in a local minimum here.
40 hours	Fix copyright notice in files, and small preemptive optimisation	tslil clingman
	Eventually there'll be a more complicated data generation step than the one we're presently using, so having it in-lined in the loop is wasteful. Ideally also this would be update per ply and we could avoid recalculating it entirely for every query -- though it's probably ``fast enough'' for now. Also, caching is WIP.
40 hours	Rename ct_k -> ct, IANAL but ...	tslil

40 hours	TEI interface working!	tslil clingman

40 hours	Tried some naive iterative deepening. Work on TEI interface next	tslil clingman
	If TEI is implemented, then i could make use of Morten's racetrack (https://github.com/MortenLohne/racetrack) and develop a quantitative measure of the bot's performance. This is the current priority.
40 hours	Just some #weightgoals ;)	tslil clingman
	It turns out that while i was training on a 0/1 classification problem, i was using 2*eval - 1. Training using this function instead, and on bot-dominated game choices (chosen_player in extract.sh) seems to have given a better evaluation function. At the least, Morten's swindle doesn't work anymore.
40 hours	Fairly important bug fixes to lcdlib, LCD now echoes input!	tslil clingman
	Input polling without line-buffering is done using ncurses, so the buildroot configuration had to change accordingly to include that library. The Makefile changed to accommodate stand-alone building of ct1986 and to include -lcurses where appropriate. There were also some typos about copying ct1986 and ctaklm to the correct directories.
40 hours	Try to squeeze out a little more performance	tslil clingman
	``Common wisdom'' dictates that placements are often better than stack moves, so we bias the generated move list in this fashion. Seems to be a little faster.
40 hours	Purged uninteresting statistics	tslil

40 hours	Added license information!	tslil clingman

40 hours	Print progress before recursion rather than after	tslil clingman

40 hours	Typo	tslil clingman

40 hours	Removed treap in favour of linked-list chained hash table	tslil clingman

40 hours	It would appear that any function call whatsoever is slower :/	tslil clingman
	For now we'll stay with directly recomputing it at each non-terminal node
40 hours	I don't have the presence of mind to debug this right now	tslil clingman

40 hours	Somehting along these lines, i'm tired	tslil clingman

40 hours	Small changes to build ct1986	tslil clingman

40 hours	Storing best moves!	tslil clingman

40 hours	This matches alpha-beta!	tslil clingman

40 hours	Stable negamax-alpha-beta fail-soft	tslil clingman

40 hours	Once again, adding TT changes the outcome	tslil clingman

40 hours	Actually it seems before i was mis-counting	tslil clingman

40 hours	There is still a bug, it doesn't appear to be checking enough	tslil clingman

40 hours	Stripping debug stuff	tslil clingman

40 hours	It was a silly typo! Hoorah!	tslil clingman

40 hours	Still bugs...	tslil clingman

40 hours	Still some bugs, standing stone becomes flat at depth4 self-play??	tslil clingman

40 hours	There's still something wrong	tslil clingman

40 hours	This looks better to me and confirms scribbles on paper	tslil clingman

40 hours	I don't understand why action_list.c:306 != 259	tslil clingman

40 hours	There are still some bugs in the undo almost surely...	tslil clingman

40 hours	Lots of bugfixes, mostly uint vs int. Still weirdness in game	tslil clingman

40 hours	This is the basic idea, there's ≥ 1 bug (generates illegals...)	tslil clingman

40 hours	Working on action lists to refactor	tslil clingman

40 hours	Still trying	tslil clingman
	\|
40 hours	Hash collisions	tslil clingman

40 hours	Fixed some important bugs	tslil clingman

40 hours	Attempting Zobrist hashing	tslil clingman

40 hours	Not correct usage, need to try hashing	tslil clingman

40 hours	no speed increase ...	tslil clingman

40 hours	Trying caching to speed things up	tslil clingman

40 hours	Thanks compile warnings!	tslil clingman

40 hours	Starting work on LCD library	tslil clingman

40 hours	Finally (?) fixed winning avoidance	tslil clingman

40 hours	Renaming	tslil clingman

40 hours	Many fixes, i think this is actually correct	tslil clingman

40 hours	Experimenting	tslil clingman

40 hours	Merge branch 'pi'	tslil clingman