Mining system

[median silo]

Ive made a mining system using an instance part hitbox that relies on the physics of the game n other stuff

Is it a good system or will it be a lag bomb

[hard willow]

So like you use a part and .Touched or something?

Or are you talking about voxels

[quasi tundra]

find out

[median silo]

Its like each click a 3x3x3 object spawns and when it rouches mineable it decreases health int

[quasi tundra]

just debounce and its fine

[hard willow]

I think it's fine, but I'd personally use Spatial Queries like GetPartBoundsInBox (https://create.roblox.com/docs/reference/engine/classes/WorldRoot#GetPartBoundsInBox)

Does the physics matter?

[hard willow]

@median silo

[quasi tundra]

It’s expensive

[hard willow]

I'd love to see a benchmark between the different methods

I didn't think it'd be much considering this is probably a hit every few seconds or so and not a method in a loop

[quasi tundra]

.Touched is good enough for this, I remember seeing getpartsinbounds in a production game I had was very expensive over time

[hard willow]

I've always heard Touched isn't very reliable (not that I've really had issue with it before, but just saying what I've heard)

Perhaps at one point I'll actually try and benchmark that, as it feels important for me to know

[quasi tundra]

lmk too

[paper drift]

** You are now Level 10! **

[hard willow]

How do you implement a part hitbox with Touched? I don't normally do it that way

I tried doing 100,000 calls all at once with GetPartBoundsInBox, standing in the middle of a bunch of mesh parts
100,000 GetPartBoundsInBox calls took 8.29276960005518 seconds, detecting 21600216 parts or an average of 0.0000829276960005518 seconds & 216.00216 parts per call, or 0.4975661760033109% of a 1/60 frame - Server - test:16
This could be improved with OverlapParams

It's also possible I did this totally wrong lol

With this quick code

local Players = game:GetService("Players")

Players.PlayerAdded:Connect(function(p: Player)
    p.CharacterAdded:Connect(function(c: Model)
        wait(10)

        local thr = coroutine.create(function()
            local t = os.clock()
            local parts = 0
            for i = 0, 100_000, 1 do
                parts += #workspace:GetPartBoundsInBox(c.PrimaryPart.CFrame, Vector3.one * 50)
            end
            local totalTime = os.clock() - t
            local timePerCall = totalTime / 100_000
            local secondsPerFrame = 1/60
            print(`100,000 GetPartBoundsInBox calls took {totalTime} seconds, detecting {parts} parts or an average of {timePerCall} seconds & {parts / 100_000} parts per call, or {(timePerCall / secondsPerFrame) * 100}% of a 1/60 frame`)
        end)

        coroutine.resume(thr)
    end)
end)

return {}

Hey @stone bridge, terribly sorry for the ping, but you're pretty experienced in this kind of thing, right?

I'm curious what's the best hitbox method in terms of performance?

[quasi tundra]

Nvm I just searched it up, most of them say touched is worse. I’m guessing this is a new developer so prob letting them use touched is better

GetPartsInBound for this kind of hit would be better though

[hard willow]

No totally that makes sense, I tend to do the same, trying to keep it simple for newer devs

[stone bridge]

why me

[hard willow]

Because I just think you're neat

[stone bridge]

performance is the wrong question here most of the time

[stone bridge]

the people who say .touched isn't reliable, don't know how to use .touched

which method you'd use depends on the game

performance is irrelevant

[obtuse jackal]

I've never had a problem with touched personally

At least for stationary objects