Did you do any optimization to get those numbers?

No, nothing special. My more ordering is something like North, South, East, West, place horizontal walls, place vertical walls. Not really anything smart.

My depth 2 on the starting position is exploring about 17k nodes taking about 46 seconds to do so :S

17k nodes is essentially the full size of the tree, which means that your alpha-beta search is not producing any beta cuts. That doesn't seem right. Regarding the nodes per second (where you are getting something like a slowdown by a factor of 500), it's a matter of picking the right language (I would argue you should use something low-levelish like C or C++ for this task) and understanding a couple of things about what operations are cheap and which ones are expensive.

By the way, my pathfinding function still needs improvement, and I expect to get more nodes per second that way. Transposition tables and some move-ordering heuristics should allow me to search deeper for a given number of nodes.

Did you do any optimization to get those numbers?

No, nothing special. My more ordering is something like North, South, East, West, place horizontal walls, place vertical walls. Not really anything smart.

My depth 2 on the starting position is exploring about 17k nodes taking about 46 seconds to do so :S

17k nodes is essentially the full size of the tree, which means that your alpha-beta search is not producing any beta cuts. That doesn't seem right. Regarding the slowdown by a factor of 500, it's a matter of picking the right language (I would argue you should use something low-levelish like C or C++ for this task) and understanding a couple of things about what operations are cheap and which ones are expensive.

By the way, my pathfinding function still needs improvement, and I expect to get more nodes per second that way. Transposition tables and some move-ordering heuristics should allow me to search deeper for a given number of nodes.