Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[blue sam3]

Also, back when there was the economy, 2-tile rooms were the best possible option, which is impossible without walls (at least that's how I played).

Actually, those rooms were 2 walkable tiles + 9 wall tiles + 1 door.  If you hit - you could actually shrink the room farther (and have a 1 walkable tile bedroom with only a bed in it).

And if you really wanted to make it cheap, it was better to put a whole load of beds in a room together, all overlapping each other. Bottomed the value right down.

[oneguy]

Without walls, it's not a fort, it's a settlement

My understanding was that the more paths a given dwarf has between point A and B, the longer the pathing algorithm takes. So the ideal fort in terms of fps efficiency would have exactly one path from any one single spot to any other.

ya, but that would mean that every dwarf would then bump into every other dwarf regardless of were they are going

[Ruirize]

Without walls, it's not a fort, it's a settlement

My understanding was that the more paths a given dwarf has between point A and B, the longer the pathing algorithm takes. So the ideal fort in terms of fps efficiency would have exactly one path from any one single spot to any other.

In some cases, this is true.

It is A* pathfinding nature to head directly to a target, and only spread out the search when an obstical has been hit. (Not the way it's programmed, I know)
There can be any number of paths, but A* will always choose the shortest. At no speed cost. The speed cost comes with more complex, less direct paths.

[Karnewarrior]

Without walls, it's not a fort, it's a settlement

My understanding was that the more paths a given dwarf has between point A and B, the longer the pathing algorithm takes. So the ideal fort in terms of fps efficiency would have exactly one path from any one single spot to any other.

In some cases, this is true.

It is A* pathfinding nature to head directly to a target, and only spread out the search when an obstical has been hit. (Not the way it's programmed, I know)
There can be any number of paths, but A* will always choose the shortest. At no speed cost. The speed cost comes with more complex, less direct paths.

That's why the main halls in my fort are designated to be High-traffic areas. The Algorithm actually checks the path multiple times if it's normal, even more for low, and a whopping 25 times if it's restricted. High-traffic areas need only a viable path be checked once before the dwarf will take it, meaning it keeps the majority of the traffic in my fortress in the halls and the pathing quick. I get >100 FPS.

Of course it doesn't hurt that my computer is a total beast.

[Draco18s]

That's why the main halls in my fort are designated to be High-traffic areas. The Algorithm actually checks the path multiple times if it's normal, even more for low, and a whopping 25 times if it's restricted. High-traffic areas need only a viable path be checked once before the dwarf will take it, meaning it keeps the majority of the traffic in my fortress in the halls and the pathing quick. I get >100 FPS.

Of course it doesn't hurt that my computer is a total beast.

You don't understand what traffic settings do, do you?

They set the path weight of the tile, such that the "calculated distance" a tile has is increased by the weight (in addition to the "calculated distance" so far).  What a path weight of 2 means is that the A* will check 2 "high traffic" tiles before 1 normal.  And 12 or 13 normal tiles before checking a single restricted tile.

It's not that the algorithm runs more, it's that it is weighted to prefer certain tiles.

Here's a basic lecture:
http://www.cs.cmu.edu/~motionplanning/lecture/Asearch_v8.pdf