My original benchmarks were for Eight Queens. I ran some more tests with
larger N and found that at N=14, the Tarantella version is the clear
performance winner on my machine.
> On Jul 30, 2019, at 7:03 PM, Mark Engelberg wrote:
>
> Thanks for writing the n-queens code and the blog post.
Thanks for writing the n-queens code and the blog post. It's great to see
tarantella performing well.
There's a certain amount of overhead associated with setting up the dancing
links data structure, so I would conjecture that as your problem gets more
complicated (e.g., increasing n), you'd see
Thanks for updated Tarantella. I also enjoyed re-watching your talk.
I just wrote a blog post to cover a simple solution to the Eight Queens problem
using Tarantella.
http://conjobble.velisco.com/2019/07/30/tarantella-queens.html
This is really nice talk that's given me a bunch to think about. Thanks!
On Mon, Jul 29, 2019 at 4:23 AM Mark Engelberg
wrote:
> "You won't believe this one weird trick for solving Sudokus and other
> puzzles."
>
> Tarantella is an implementation of Knuth's Dancing Links algorithm. I
>
"You won't believe this one weird trick for solving Sudokus and other
puzzles."
Tarantella is an implementation of Knuth's Dancing Links algorithm. I
demonstrated in my 2017 Clojure Conj talk how it can be used to solve an
assortment of puzzles (https://youtu.be/TA9DBG8x-ys) and some of you may