Optimization: A GAME-ENDING issue

[warrrn]

Vetex should consider making the dark sea a separate place.

[Vratien]

sigh it’s immersion.
at least this immersion is actually interesting, in the way that it shows the people are actually alive and isn’t just an annoying thing you have to manage
cough cough water poisoning, armour damage cough cough

[Scandin]

@Vratien

in the way that it shows the people are actually alive

The dead folk at Port Mistral have an aura.

---

@BlueFighter15

I personally also feel like every single humanoid having a sensable aura regardless if they can actually be fought at all or not might be harming the fps at least a tiny bit

Is, yes.
It is very likely that the game checks for an applicable model^[1] to apply the sensing to, but then you have to think; how is the sensing rendered? what’s the performance cost for having so many sensing gradients around? Too many questions, no answers.

---

1. Not exclusively humanoid, as per the Poseidon template sensing in the floor. ↩︎

[Vratien]

never mind

[BlueFighter15]

@Scandin

an applicable model

Well, it’s possible vetex hid an invisible humanoid under the floor to have something capable of speech?

@Vratien

at least this immersion is actually interesting, in the way that it shows the people are actually alive

Tucker has an aura and we’re literally able to watch him decompose in real time

[Scandin]

I was about to say “Possibly”, but then I realized we’re talking about Vetex, so yes, that is pretty much what it is then.

[tbd]

all of those are generally for server performance (ping), and script performance isn’t the issue according to vetex (you can probably verify that by going frame by frame yourself in the micro profiler)

I would highly doubt sensing is the issue, I highly doubt all of the billboardguis used to make the sensing gradients exist until you press shift + g (which is probably why literally every humanoid has one since it’s just going through everything tagged ‘humanoid’ or that has one and that’s why tucker’s body counts)

my guess is that ao’s shitty performance has something to do with loading stuff in. when im on my laptop which doesn’t have very good performance, if I stay in roughly the same area, I get ~60 fps. if I start moving or using attacks that drops fast

[Scandin]

@tbd

all of those are generally for server performance (ping), and script performance isn’t the issue according to vetex (you can probably verify that by going frame by frame yourself in the micro profiler)

From what I saw last, there isn’t profiling done in code to accurately show how long actions take, therefore I am not able to see what is bottlenecking.

@tbd

I would highly doubt sensing is the issue, I highly doubt all of the billboardguis used to make the sensing gradients exist until you press shift + g

It’s moreso hope it doesn’t exist when you aren’t sensing.

@tbd

(which is probably why literally every humanoid has one since it’s just going through everything tagged ‘humanoid’ or that has one and that’s why tucker’s body counts)

I do not believe humanoids are tagged, as there isn’t much proof of CollectionService being utilized for many systems in AO.

@tbd

my guess is that ao’s shitty performance has something to do with loading stuff in.

Very likely, as the amount of parts in the world are overloading the rendering to the point every computer is suffering.

[tbd]

actually the culprit of this update’s terrible optimization might just be corrupted dragons. the moment one loads it seems like everyone on the server’s fps just tanks no matter how good your pc is

edit : this actually isn’t true, but a group behind me were fighting an ancient (corrupted) inferno dragon and I loaded some monster island and somewhere in that time frame the server’s fps tanked globally with 10 fps drops intermittently. the update’s terrible optimization is probably still something to do with the dark sea