I switched to precalculated g-values and my results aren't any different. I didn't expect them to be drastically different, but it was a good change anyways. Thanks for catching it
AppGameKit used to be able to measure in higher precision that 1 ms in v1 but not in v2. I really wish I could benchmark the code parts, but I suspect the slowdown is due to the searches in the open/closed list (yes essentially including that sorted insert as it's in the place of a search).
I'm going to attempt to create a heap in AGK. It'll require breaking Paul's rule (don't workaround to nest UDTs) but I'm doing that already. I'll discover if that's the cause of my occasional crashes too at least
It's something I can change, but the farther I go from an object-orientated code structure the more ugly and unreadable the code becomes
. Will report back
(If the heap works out I'll share it on the AppGameKit board for others to use)
Edit: Oh I suppose I could implement the heap with an array
Yep I'll do that.
Edit 2: Well that was easy
I made a Minimum Complete Binary Heap. Forgot it worked alongside arrays so nicely. I'll share it in its own thread.
Edit 3: Remembered AGk's built in sorting using Radix so this probably doesn't help me
oh well
