Spring RTS Engine

Deformation actually adds a lot to game play if used correctly by making the map mutable, just because there is virtually none in the games you play BasiC doesn't mean that it can't contribute. I remember when it made a difference in AA...

Well guess what, it already stores it in an array.

Well, at least now I know at least one other person thought that was a good plan

Seriously though, I guess that leaves only the number of blocks that are allowed to be queued in repathing as a possible target for this, unless somebody has a better idea. I'll try to find time to read the code here in a couple of hours or so, so that maybe I have a clue about how it's evaluating that stuff.

When stuff chain-blows like that, it must be maxing out the queue for many frames in a row- and the queue must be large enough that the processor can't actually handle that load along with other stuff. Lowering the number of blocks that may be evaluated may really help this.

Might be worth confirming that it definately is the deformation that's causing the lag/slowdown and not some other mystery process...

Deformation makes the heightmap much more complex, which causes pathfinding to require a greater load, if I recall. It's not much different from the terraform based slowdown which occurs in CA.

It's not much different from the terraform based slowdown which occurs in CA.

Does that slowdown occur because of the changes made, or because of the complexity of the paths that result, though? That'd be worth testing, imo.

This reminds me of some earlier CA versions where you could build a soft bump, that caused for me BIG lag when i made just one of them, but the normal block terraforming didnt cause lag, nobody never knew what was the reason. Now CA doesnt have soft bump anymore, i guess.

Maybe theres some bug in the code which causes it to update the paths too often?

TradeMark wrote:This reminds me of some earlier CA versions where you could build a soft bump, that caused for me BIG lag when i made just one of them, but the normal block terraforming didnt cause lag, nobody never knew what was the reason.

Sure we know the reason and it is not a pathing/terraform problem.

Also I wrote a solution for it with my huge patch, but nobody still not included it yet :<

I thought your patch was fully applied to the SVN. Oo

IIRC only 2 LUA parts not; I don't even recall seeing anything changing pathing code but it's a while ago already.

In any case, why this discussion when some with a profiler could point out exact places that cause the lag pretty fast?

Tobi wrote: In any case, why this discussion when some with a profiler could point out exact places that cause the lag pretty fast?

Queue post from BrainDamage