OPTIMIZATION REQUESTS

[heavy parcel]

Compare Arma Reforger and Squad graphics.

[static cosmos]

new game.

[heavy parcel]

AR performs and looks better than Squad, even the UI looks nicer.

[heavy parcel]

Squad: 2020
AR: 2022

[static cosmos]

2016 years - squad

release 2020 yeah

[heavy parcel]

That means they had even more time, yet none to show.

I know you guys protecting it, but it just simply isn't what they are saying it is.

[misty swift]

Also BI has their own engin tailored for their needs, with much more resources, and 500+,probably more experienced developers,

Not really comparable.

[heavy parcel]

Go search for a bit there is hundreds of posts, threads, videos of complaints and proof of this being discussed years ago and still no sight of any change.

[static cosmos]

in arma 3? yes

have in arma 3

many games have this

[heavy parcel]

Bro, Hell Let Loose is smaller and same release date.

[static cosmos]

not only squad

[misty swift]

I installed Squad Alpha 12, just for fun. I had 40 FPS less than I have now.

[static cosmos]

poor game.

[heavy parcel]

Better optimization, bigger updates, proper transparency.

[misty swift]

And that was before the lighting update.

[heavy parcel]

It performs and looks better in every way, sort & form.

[static cosmos]

with more dlc and stupid gameplay

[heavy parcel]

Mate, Hell Let Loose is younger and blows Squad out the water, better optimization, better visuals, active and proper transparency regarding development, animations, hit-reg, mechanics, and game systems.

[static cosmos]

I've been playing Hell let loose since beta. The game is terrible

[misty swift]

Reforger does not perform better if you compare max settings in both games.

[heavy parcel]

Squad is stagnate - slow updates and no innovation relaly.
Broken promises, poor optimization, outdated animations.

[static cosmos]

starting with the fact that there is stupid gameplay and level system

[heavy parcel]

Performs better than Squad.

[static cosmos]

Why in the game are people not allowed to repair a tank if they have not reached some level as a driver?

[misty swift]

Not on my end. If you turn everything to max I have like 30 FPS in Reforger.

[static cosmos]

Stupid and poor game

[static cosmos]

and more stupid idea from developers

[static cosmos]

what the system you have

[misty swift]

4080, 7900x3d

[heavy parcel]

Good stuff.

Recent?

[misty swift]

ddr5 6000, and to make you happy with custom timings.

[misty swift]

no

[static cosmos]

zentimings

pls screen and send

[misty swift]

Nah sorry not downloading that, I dont want FSB spy ware.

[heavy parcel]

Anyways, yeah, Squad's got to wake up and get their stuff together, that's all. It's all out there no point defending it here cus we all know it's not where it's suppost to be.

[static cosmos]

(vdd with a bug)
this is how it looks to me

[static cosmos]

lets compare, brother

[static cosmos]

I'm not Russian, I remind you

[misty swift]

i also dont want Ukrainian spy ware.

or whatever

[static cosmos]

racist

[misty swift]

sure

dont want spyware = racist

[static cosmos]

dowload on internet

dont be racist

[heavy parcel]

How and when is refusing to download software racist?

[static cosmos]

Then you are deceiving that you have custom ones. I know that DX12 can work for you.

[misty swift]

I never said that it does not work for me

[misty swift]

What die is that? M-Die?

[static cosmos]

A-die

vdd - bug in zentimings

[misty swift]

Yeah I was just about to ask you why it is that high

[static cosmos]

Originally wrote about this

zentimings confused vdd with vpp

[misty swift]

FLCK 2033
Tras 28
Trc 68
Twr 48
Refresh Intervall 50000
Trfc1 500
Trfc2 400
Frfcsb 300
Trtp 12
TrrdL 8
TrrdS 4
Tfaw 20
TwtrL 16
TwtrS 6
TrdrdScl 4
TrdrdSc 1
TwrwrScl 4
TwrwrSc 1

@static cosmos Those are my timings, should work on every DDR5 M-Die Memory

I think you can lower your Tras further as well

If that improved performance/stability in Squad, i dont know. I am probably GPU limited anyway.

[static cosmos]

agesa updated. And now Tras matters a lot.

[static cosmos]

overall ok. But you didn’t show most of the timings, you also have small flaws

[misty swift]

the other timings are default

[static cosmos]

Trfc must always be a number that can be divided by 8 for example

like a 504

Trcdwr + 12 = Tras

Refresh Intervall can make 65535

Well, if that suits you. Then use it as you want

[misty swift]

My memory cooling is not that good, because of my case and fan layout. It becomes too hot in the summe

[static cosmos]

if so then fine

[misty swift]

I dont see why. The default timing is not even devidable by 8.

[static cosmos]

This is how people accept it. And the systems themselves replace these numbers with some even ones

Yes, and AMD has such a parameter as GDM, which replaces odd numbers with even ones

[misty swift]

I think it is a myth.

[static cosmos]

GDM not myth

why do they follow the rule of division by 8, simply because the system more easily perceives numbers that are divisible by 2, 4, 8, 16

[misty swift]

For this timing I dont see a technologically necessity to be constrained to values devidable by 8

[static cosmos]

Yes. but if you make the value a little lower, you still won't gain much

it's just convenient. And there is some logic in this

[misty swift]

I find it more convenient to use 1% of the refresh interval. It comes down to personal taste i guess.

[latent shoal]

DayZ Spent 13 years in development and had half an engine rebuild, to get to where it is now. Reforger is using the full engine rebuild that took half a decade to complete, so it's expected to perform better due to DayZ being the test bed. Can expect Arma4 to run significantly better than Reforger, due to Reforger being the test bed for the complete Enfusion engine. Reforger had a very rocky start, took two years to iron out it's performance problems and even now, still has problems when players aren't spread out across the map, instead all being within 4x4km(like Squad)

Arma 3 performs well now, due to it's age. In it's hay day it was considered one of the most difficult games to run, both client and server, due to very high CPU demand. Just the nature of hardware advancement, and them upgrading from 32bit to 64bit helped too. Even so, it's still pretty performance hungry when it comes to modded content(which 90% of Arma players play).

Squad is not perfect, optimizations are being worked on. There's no point doing major content or system updates for Squad now(on UE4) as it would just require said things to be revisited in the middle of the current UE5 transition. Waste of development time & resources. I would not expect any significant updates for a while.

[steady hamlet]

Lmao comparing arma as some pinnacle of mil gaming when desync is it’s middle name

[analog egret]

wait
if squad is going to get a make over for a lot of things for the UE5 conversion
DEVS TAKE THIS OPPORTUNITY TO CODE IN A DATA ORIENTED WAY

[steady hamlet]

...those words don't really mean much

but yes, they're already working on improving the code systems.

[slow pendant]

And make sure to use singletons for more angular bandwidth 🙄!

[analog egret]

With the amount of data required to run each match
I feel like doing data oriented code will see significant performance improvements
Imagine instead of each 30+ players near you shooting
Instead of your cpu going through each memory address to run the fire function from each player class
It just goes through one actor class which takes an array of players

[analog egret]

Ngl
I feel like their netcode for bandwidth is really good already
Considering how large each match is

But yeah
I just hope they take this opportunity to really rework as many things from the ground up

[misty wind]

Went to try the Nvidia Override in the Nvidia App and it looks like Squad isn't supported in the app anymore.

[fleet fable]

i use a cloud Computer. Shadow(.)tech

[vague mulch]

But grass is only a thing up to 300m

[vague mulch]

Mabe in squad it’s also 300m but squad handles it a lot better

[static cosmos]

in the squad the lighting does not disappear after 300 meters and the grass

[deep flint]

Does anyone know of any solutions for the "Server Closed connection" issue? Im on ethernet and have good pc specs but theres gotta be some kind of fix?

[thorn lion]

#support-discussion

[broken compass]

if youd actually read you would know that’s exactly what the guide does. thanks though

[silk merlin]

this guide does absolutely nothing in reality too

maybe it does something on low end hardware

[broken compass]

helps many people lol

[silk merlin]

actual benchmarks or didnt happen

[thorn lion]

This is a solid point

[silk merlin]

Yes all the guides on internet claim magic boost to fps yet rarely anyone can back them up with actual benchmarks

[thorn lion]

There is one guide that a old comp team used that was reqired by the team and I was told it works but some stuff was like "turn on Rebar" and other stuff you should already be doing

[silk merlin]

Rebar is hit and miss, in delta force nvidia enabled it and it started causing massive fps drops

HAGS might work in squad but it has to be tested case by case basis

[thorn lion]

I would guess thats a delta force issue

[silk merlin]

not really, AMD heavily relies on Rebar for performance nvidia on the other hand is hit and miss

some games will give you lower frames with rebar on

[thorn lion]

Where are these benches from?

[silk merlin]

techspot

[silk merlin]

https://www.youtube.com/watch?v=uXlLfHDGgro

in these tests it faired a bit better

not many people benchmark it anymore

[austere tree]

What do you think?😁

[rotund echo]

POV : You want to build an AM5 PC but all X3D chips are out of stocks

[silk merlin]

They’re in stock actually

[rotund echo]

For less than 600 quids ?

Cause scalpers don't count

You can't find an X3D chips on AM5 for less than 515 € in france for exemple

That includes the 7800X3D

[silk merlin]

How is euro to gbp conversion? In uk they’re available

[rotund echo]

Same here, just very expensive

And it's about the same price with conversion i would say

[silk merlin]

Same price as 7800x3d when they launched

[rotund echo]

Yea, too much 😦

Although it was month after its release, I think I got my 5800X3D for like 330

[static cosmos]

for m-die - ok

[static cosmos]

now by the way 7800x3d is not relevant for purchase. It would be better to buy 9700x where the parity will be 1440p, and it costs less than 3dcash

[static cosmos]

although I managed to do it that way.

[static cosmos]

It costs me 750 euros

[silk merlin]

Oof

[tiny plank]

pretty sure Squad massively benefits from X3D Cache, no? The more unoptimized the game is, the more benefit there is in-general

And yeah, Squad is very memory-sensitive, it can literally double ur FPS

[static cosmos]

3dcache performance drops at high resolutions, so at 1440p the 9700x and 7800x3d have parity in performance. Even at 1080p the difference between them is not very significant

[tiny plank]

According to whom, and what data? Sounds like you're heavily over-generalizing

X3D users have highest FPS in Squad, I don't remember a 9700x being that good in it

Don't generalize a 10% FPS difference in average values for all games, most of which are optimized, to the steaming pile of shit that Squad is lol

[static cosmos]

7800x3d is outdated and there are more profitable processors to buy. hardware has such a capability as to become outdated to buy

[tiny plank]

So no data, all clear

And if CTRL+F FPS X3D ull find plenty of it showing users with 200+ FPS

In this Discord server

[static cosmos]

they have the same fps in this game

in Tarkov 9700x and 7800x3dv are the same in performance, for example

in Tarkov the optimization is crap, and the processors are the same in terms of performance

just don't waste your money on expensive outdated hardware, buy 9800x3d or 9700x

[tiny plank]

based on what data?

[static cosmos]

from personal experience, also from tests of respected people

[tiny plank]

I have a 9700x, and I get around 130fps driving around Yehorivka, ~80-100fps in very saturated moments, all at 1080p medium/high settings. I definitely wish the performance was a bit better though, it would be nice to always be at over 120fps

Here is public info

[static cosmos]

give me his nickname, I will help him with the setup

[tiny plank]

a 5800x3d gets similar performance at 4K:

O M G

4k, everything max, 120 FPS with DLSS on quality. Sometimes it Drops to 100 or 80, but still very smooth. But it got hot. Now i have both watercooled and GPU is most of the time at only 30% with 50°C```

so ur idea that at 4k, x3d doesn't help is ridicilous, zero data nothing just "my friend says so"

[static cosmos]

Say nickname your friend

With 9700x

there after we will see how 7800x3d is currently expensive and useless to buy

[rotund echo]

I wouldnt say useless to buy, but expensive yes

It's there when 9800X3D is out of stock

Although I wouldn't put 500 in a CPU just cause its successor is out of stock. Im in no rush

I missed the 1080 Ti back in the day but I didn't miss the gigachad 5800X3D which is so incredibly good you can skip an entire Ram generation with it

All the way to DDR6

[misty swift]

Internet Rule Nr.1 - Never trust some random guys claims without proof.

just saying

[static cosmos]

we have 7800x3d and 9800x3d costing 600 and 800 euros. Now 9700x is very popular because it is not expensive and has performance at the level of 7800x3d

9700x costing 336 euro

We got to this earlier than you, apparently. Because 9700x is very popular in the post-Soviet countries. Simply because our prices are twice as expensive

but I don't want to prove my words with the help of YouTube videos

Well, under the same conditions, the performance is almost the same

in a game with very poor optimization

Escape from Tarkov

it would be much more interesting if the person above introduced me to a guy with 9700x. And we would go straight into the game in one session to test the 7800x3d and 9700x to confirm my words.

@tiny plank ?

[tiny plank]

Cmon man.

1. That's not Squad.

2. That's an expensive 7600C34 RAM, even if you spend the money to get RAM tested at that speed, it's not even a given that your CPU or MOBO will handle it.

So yeah, no shit that with optimized RAM the importance of X3D diminishes.

But most people buy X3D so the don't have to overspend on RAM, and potentially brick their system just to get performancce slightly below 7800X3D.

Do a 6000C30 review for Squad, the 9700x won't be anywhere close the 7800X3D

[static cosmos]

It's easier for us, you spend 330 euros and 5 euros for setup. And you sit and enjoy it because you didn't pay 600 euros

[tiny plank]

That's like saying 13900k > 7800X3D with full tuned RAM. It is true, but also you spend like $1k just on RAM.

[tiny plank]

9700X for 360 euros + expensive RAM + risk of bricking ur system < 7800X3D for 488 euros + standard RAM

[static cosmos]

7600 1/2 is comparable in performance to 6000 cl30. that's not the point. In the video they tried to make the same conditions with the Intel processors, which are also present there. I just removed them

[tiny plank]

That's not true, and 7600 has higher risk

Obviously 7600C34 doesnt tell u full info, if tFAW and the related tRRDS is optimal, then 7600 is way better than 6000, especially on CPU without X3D

[static cosmos]

in processors there is such a thing as Gear 1 and Gear 2

[tiny plank]

doesn't matter, tFAW more relevant

gear 1 v gear 2 barely any perf difference

meanwhile tFAW optimal vs suboptimal is 30% FPS diff

[static cosmos]

many years have passed since the release of DDR5, and a lot has changed

[tiny plank]

DDR is still DDR

tFAW 32 for DDR5 is optimal

u put 48 u will lose massive FPS

if u dont even know how DDR works then doesnt matter to have this argument

[static cosmos]

7800 cl34 is comparable in performance to 6200 cl32 1/1

[tiny plank]

it's not

tFAW 32 7800 is way faster than 6200 tFAW 32

maybe tFAW 48 7800 is same as tFAW 32 6200, but then u have the same RAM kit anyway

[static cosmos]

https://docs.google.com/spreadsheets/d/1dsu9K1Nt_7apHBdiy0MWVPcYjf6nOlr9CtkkfN78tSo/edit?gid=1677627148#gid=1677627148

[tiny plank]

either way 9700x is more cost-efficient for Squad, but don't claim it has same FPS as 7800X3D, it doesn't, it's slower by some tens of %

[static cosmos]

let's compare now

[tiny plank]

Again, there's no point in arguing with u i already know u know nothing about how memory works

ur basic understanding boils to single read, and since latency is similar for single-read for 6200 c32 vs 7800 c34, u think perf is the same

[static cosmos]

bro, i just don't know why you need 1/2 on amd. If amd processors are limited by infinity fabric

[tiny plank]

but no game or appliation does single read and optimal instruction order is needed to have high perf

[static cosmos]

yours are more suitable for intel processors

[tiny plank]

Educate yourself: "DDR4 timings explained: tRRD & tFAW // THE MOST IMPORTANT MEMORY TIMINGS"

[static cosmos]

because Intel has no bandwidth limitations

[tiny plank]

Author makes good visual explanation how DDR ram works, including DDR5

believe it or not, CL doesn't matter, u can set CL to 7800CL44 and it'll perform similar to 7800CL34, as long as tFAW is optimal, so long reads are optimally buffered

the memory speed with optimal tFAW and it's related settings is most important, then the refresh latency etc etc

CL on the end of the list

[static cosmos]

and you didn't think that the complex is important there, and not just spinning individual timings?

[tiny plank]

do u know what CL and tFAW does or not lol?

i am playing on 10600k with 90 FPS average on DDR4 because I do

[static cosmos]

roughly speaking is responsible for time. each reduced tcl +200 mhzц

tfaw is responsible for a number of lines

but yes I agree, tcl on intel is not very important. You can sit at 6600 cl32

[static cosmos]

often put according to the formula = trrds multiply by 4

[tiny plank]

ahahaha lol

yeah u gotta learn more

[static cosmos]

well yes, in general I also heard a lot of good things

[magic elbow]

I hope ue5 has good optimization

[misty wind]

Will the same devs who are looking at Squad's optimization also look at Squad 44's performance? Mercury Arts would probably need that info too

[modest oriole]

Mercury (f)Arts only proved with the last update they are amateur modders who can't even deliver a well performing map which is a bottom line for the modders. How to even ask for optimization, bug fixing, good layer and gameplay design when they are unable to do what they are known for?

At this point it's time to hire new devs with proper skills if they want to save this game.

[last needle]

Fix the always resetting settings please its a pain for me

[mystic kindle]

Waste of money the game is dead and ain't coming back

It's unfortunate because it's better than hll

[floral mango]

I wish they just took Squad in its current state and reskinned it into WW2. Squad 44 feels outdated and needs so much work and polish to catch up. They've done some good work tho.

[strange jewel]

Can we change the DLSS scale factors from uneven to even, so for quality a scale factor of 0.75 would fix the temporal smearing abit via DLSS. Along with upgrading to DLSS 4 would be amazing.

[latent shoal]

You can force DLSS 4 through the Nvidia app override feature.
https://youtu.be/v7l086qvqEM

[strange jewel]

Not preset K, which looks better than J. Plus using DLSS Tweaks to view the DLSS HUD shows that when you force it through the app, it doesn't always work, and sometimes alters the preset randomly. You can use a power shell script online to force it driver level for all games. But it would be a lot more user friendly if we could just choose it in the game. DLSS 4 is available for all RTX cards, so upgrading the dll would be good.

[modest oriole]

Game is there so it's kind of a waste. They should take over, bring it to squat level, and develop together, maybe as someone suggested in the same launcher.

[mystic kindle]

Sunken cost fallacy, the game is so far behind squad it would take forever to bring it up to squads level of development. May aswell just let the modders do whatever they want to it

[modest oriole]

I guess you're right. Shame...

[frigid gale]

any update on optimization or nah

[thorn lion]

Wait till UE5, according to roadmap there should be a public playtest soon but I imagine that that will be when its ready

[frigid gale]

So when is ue5 expected

[thorn lion]

When its ready

[fluid frost]

3080 ti 16 ram intel i7 (getting 4,5gz)
running everything on low-medium
overclock is on
everything is x,ed out, any random applications
still running at 40-90 frames? like are we kidding rn how bad is it to optimize this game

[violet tendon]

oh its bad

[fluid frost]

wdym

[violet tendon]

.

[fluid frost]

yeah idk man

it drastically change dperforamnce once 8,0 came out

[violet tendon]

it is what it is and its not top priority

didnt play for months now bc of that

if they dont care i couldnt care less

[fluid frost]

i power through it cuz the games fun, you jus gotta dlss it

[violet tendon]

dlss my ass

[fluid frost]

i lowk forgot idk if i use dlss or not

the game looks horrid without it

[violet tendon]

not the fan of the scope mush in dlss

[fluid frost]

i think you put some setting wrong, i play dlss and the scopes look fine

[violet tendon]

u cant tick prioritize clearity in scopes when dlss is on

[fluid frost]

Okay then i might not be using dlss

i havent played in a bit so

spare me 💔

[violet tendon]

do urself a favor play something else

[fluid frost]

why games fun

[violet tendon]

i have a rog strix 4090 and a 9800x3d and get around 100 fps with micro stuttering (64gb of ram)

[fluid frost]

okay

are we deduzz

you are doing something wrong if you have that monster of a pc 🙏

[violet tendon]

stutters in the worst moments

[fluid frost]

thats ture

i stutter as soon as i get peaked

[violet tendon]

all other games run fine i tried evrything for a month to mitigate the problem its just the game

im 100% im not doing anything wrong

[fluid frost]

i will say, micro stutters are apart of squad so

[violet tendon]

gorss

on shadows high settings. indoors the gpu usage spikes to 95% on shadows middle or low indoors is broken the edges kinda glow

of the room

[fluid frost]

lowk just play on medium

[violet tendon]

i do it looks ugly tho and the micro stutters are still there

the game doesnt look remotly that good to be using this mutch resources

[fluid frost]

that is true, but i like the game so i have to power through

just ghot to hope

ue5

changes somethin

[violet tendon]

😂

hope dies last

[fluid frost]

not if ur hope is infinite

🛋️

[frigid gale]

this was my game around january-february last year

actually playable

but before that it was like this

so at one point they actually did something to fix the issue but its back to stuttering again I guess

[violet tendon]

crazy

[fluid frost]

Yeah that’s what my settings look like rn

[violet tendon]

they should fix their product

[frigid gale]

and this might be personal preference but I find DX11 to be more stable than DX12

[violet tendon]

with dx11 u loose 20-25 fps when using a scope

[frigid gale]

even if DX12 provides more performance I found it to stutter more and reduce overall quality of the game to the point where its almost impossible to see enemies

[fluid frost]

is that jsut for u or

[frigid gale]

I played at the start of this year for a bit I encountered stutters but no where extreme as the ones in that video, fps was around 30-50, I have a 1660 super amd ryzen 7 5800x and 32gb of ram, game is on ssd

[fluid frost]

its like even low settings doesnt barley give uany fps

[violet tendon]

we as consumers shouldnt have to go fix their product and search for fixes and tweak our systems for weeks just for the soul purpose of running squad, and then not even able to run it decently

[fluid frost]

that is true

[frigid gale]

If you go into nvidia control panel you can adjust video card settings, that helped me quite a bit

[fluid frost]

for what setting

[frigid gale]

let me find it hold on

so if you go under program settings you can adjust a few settings that will help with FPS

when you hover over the settings it shows what it will do

I found this to help out a lot

its worth giving it a try I did find an extra 20ish frames

[fluid frost]

okay ill try

ty

[violet tendon]

gl

[severe glacier]

99% of these don't do anything at all. They are there for backwards compatibility, and they only affect dx8 and dx9 titles. Dx11, dx12, and vulkan are unaffected

[frigid gale]

Well they seemed to work for me

[fluid frost]

dx 11 or dx 12?

[violet tendon]

U can try all the settings but in 95% of the cases the problem lies with squad‘s spaghetti code

[frigid gale]

DX11

but thats just personal preference

[lilac coral]

https://www.youtube.com/watch?v=yD_IWYpr3Bk

Video gives some really good feedback

[steel walrus]

Youre doing something wrong, i have 100 fps with a 2070 super and a 7800x3d, 32gb of ddr5 6000mhz ram
micro stutters could be drivers or shadowplay being on

[violet tendon]

Shadowplay is off

I get over 120

The stuttering is awefull

[violet tendon]

Like i said its the only game that has problems on my pc. I play a lot of different stuff

[static cosmos]

L3 cache is full

https://zentimings.com/

share a picture from here

sometimes the motherboard can set terrible timings

4090 and 14700k the same worked quietly in 4k 150 frames without stutters. You may also have

[violet tendon]

what u wana see

[static cosmos]

trfc high very

and another timings

[violet tendon]

ok?

[static cosmos]

here are your problems. I told you why you have problems

[violet tendon]

bc trfc is set too high?`

[static cosmos]

If you want, I can help you set it up

[violet tendon]

do i have to fiddel in the bios

[static cosmos]

yes

[violet tendon]

i mean we can try i dont mind

[static cosmos]

good

I'll help another person now

I'll write to you too

[violet tendon]

ok

funny how other games on my pc dont suffer from this...

[static cosmos]

there will be growth everywhere in general after this

[violet tendon]

[gritty fable]

Hey so here’s a good idea

Add an option to remove plants swaying in the wind

[lean hedge]

I have an i7 8700 gtx 1660 super 6gb 16gb ram i can run on high 60 fps but the only problem is the scope its veryyyy lagyy pls fix it

[steel walrus]

set your scope resolution to 75 and limit the refresh rate on the scopes to 60, it'll make it more stable

[lean hedge]

Ok ill try it

[stark parrot]

I just wana leave a message here for the devs to please optimise it somehow.

[midnight apex]

Specific to the C7A2/C8A3 rifles & C79A2 optical sight

The C7A2 selector lever is stuck on repetition, it visible switches to automatic but then the lever goes back to repetition. Switching from automatic to repetition is the same animation, switching from repetition to automatic.

The C8A3 selector lever is stuck on safe, exact same problem as will the C7, but just on safe

This problem seems like something that should be in squad already, but perhaps that's too critical on my part.

The C79A2 allows zeroing to 100m when the lowest possible zeroring with the C79 is 200m - The 800m max is correct. As well, the model of the sight system shows that 200m is the minimum. A final small point on the C79A2 is the range selector doesn't move when switching zeroing distance. Not integral to the playability of the game but would be cool the see the weapons functioning in full.

[stone rapids]

Why do i get 50-60 fps with my 3050, i know its not great but

should be able to get a little more on all low settings

[thorn lion]

Cpu? Mem speed?

[stone rapids]

i5 12400f and your talkign about ram speed?

36gb at 3600

[stone rapids]

.

[thorn lion]

Yea ram speed

[stone rapids]

36gb at 3600

[thorn lion]

I am not super versed in Specs but 12400f is a laptop model right?

[stone rapids]

no

its just old

[thorn lion]

Ah yea F means no integrated graphics

[stone rapids]

oh

i didnt even know tbg

tbh

what are your reccomendations

[static cosmos]

I think you have gear 2 mode

[stone rapids]

explain?

[static cosmos]

all intel processors without index k - have a frequency limit of 3466

[stone rapids]

i only have 8gb of vram which might be the big problem

[static cosmos]

and if it exceeds, then xmp sets you to Gear 2, which divides your memory controller into two

[static cosmos]

It’s not the problem lol

[stone rapids]

so what is the solver

[static cosmos]

Open task manager and say your frequency

[stone rapids]

3.97

[static cosmos]

Memory speed

[stone rapids]

ah

3600 mt/s

[static cosmos]

It’s 100% gear 2

[stone rapids]

how do i fix

[static cosmos]

What the motherboard

[stone rapids]

b 550 m - c

[static cosmos]

660?

[stone rapids]

let me chaeck

[static cosmos]

Win+r

[stone rapids]

yes 660

B660M-C

[static cosmos]

Open bios

And change dram frequency 3600 to 3466

And make gear 1

[stone rapids]

Alright

[static cosmos]

this is a very common problem for 12400f users, their memory is always running in Gear 2

[stone rapids]

I’m in here

How do I find what I’m looking for

[fluid frost]

Hello, I have a question. For Alienware bios settings, should I have my xmp profile on one or 2

[static cosmos]

Open tweaker in bios

[stone rapids]

I found it

It’s at 3466

Now

[static cosmos]

And gear 1

[stone rapids]

Yup

What about xm

Camp

[static cosmos]

usually you should always put the first one, the second one can be with gear 2

[static cosmos]

Xmp enabled

Change frequency

And make gear 1

And done

[stone rapids]

Alright it’s done

Now load up squad

?

[static cosmos]

Yeah you can

you will see some growth already

[fluid frost]

Would u further explain on gear 2? Gear one is the best for performance etc, right?

[static cosmos]

You can also set up secondary ones if you need more frames

[static cosmos]

Gear 1 - the best for DDR4, yeah

[stone rapids]

Should I use dose

Dlss

[static cosmos]

Gear 2 is needed to increase memory bandwidth, but this will result in high latency and low fps in games

[static cosmos]

it's not really necessary

[stone rapids]

Alright

Is there any other sort of important settings

Task manager currently

While loading In

[static cosmos]

Make low-medium settings

[stone rapids]

Got ya

Definitely much better

At 90 rn when I’d normally be at 50-60

Should I use msi afterburner

[fluid frost]

Thanks!

[static cosmos]

who set you 3600 to 12400f?

[stone rapids]

Prebuilt

Had 16 gb originally and I upgraded to 32

[static cosmos]

don't buy from them anymore, they are terrible people

[fresh terrace]

PLEASE

[stone rapids]

Mine was a pretty good deal

Like 750

[lean hedge]

Put everything on high and u will have more fps than on low

Thats how it works for me

[tiny plank]

Disable hyper-threading, it's wild that the game devs still haven't fixed WindowsDefender throttling and still haven't disabled hyper-threading of the app.

You just need to either do it in your BIOS or just set core affinity to even or odd: [0, 2, 4, 6, 8] or [1, 3, 5, 7, 9]. The amount of numbers depends on your cores.

[static cosmos]

Cpu bottleneck

[tiny plank]

V-Sync + G-Sync/Freesync + FPS-cap = (Screen Refresh Rate - 3)

G-Sync/Freesync is amazing for masking the stutters when you have FPS drop

[static cosmos]

we tried a lot with him. And some things helped him. I am more inclined to think that the bios version and legacy on the Hero cause these problems

because the Hero series of motherboards is problematic

disabling hyperthreading is a dubious idea

[tiny plank]

Try it first and then talk, man

Squad has horrible hyper-threading implementation

That makes the performance worse

[static cosmos]

That's what I did, dude. And it actually made it worse.

disabling cores is only relevant if the processor has two chiplets, for example 9900 or 9950

This is already the 3rd or 4th wave of the obsession with disabling multithreading

[tiny plank]

Who said anything about disabling cores?

[tiny plank]

What's your processor and what setting in BIOS did you change?

[static cosmos]

People ran to Zen+ to prove the efficiency and boost, how they won 3 frames
Now here too they show results at the level of error

[static cosmos]

There is an application called process lasso for this, through it you set priorities on the performance of the kernel

[tiny plank]

Why is it everytime we talk you just have like the lowest level of knowledge but such a high level of confidence...

Process lasso can't disable cores cause of the anti-cheat, dude

It shows visually but it doesn't work

[static cosmos]

I say “priorities”

Lol

[tiny plank]

You need to do it in BIOS or use a scripting language like Python to run a mother process with less cores

[tiny plank]

Yeah I know, that doesn't work

The anti-cheat that Squad uses disables it, you have the same problem in other games with the same anti-cheat

[static cosmos]

It already seems to me that you yourself suffer from pride.

[static cosmos]

I don't have such a problem

[tiny plank]

Buddy, you literally didn't disable anything and you're like "yeah I tried it it made things worse" lol

[tiny plank]

Because you're too stupid and arrogant to see that you do have the problem

https://letmegooglethat.com/?q=easy+anti-cheat+hyper-threading+process+lasso

[static cosmos]

I don't give stupid advice like "disable multithreading" if you suggest something like that then you should be responsible for your words

[tiny plank]

Buddy you never even tried it but you claimed you did

Also it's not that hard to do it through a Python script just for Squad

[static cosmos]

I asked you several times to go and confirm my words, because there is such an opportunity. But you always refused.

[tiny plank]

I did idiot

I have +50% FPS boost from no hyperthreading

Both BIOS and Python script work, Process Lasso does not

Stop coping

Last time we talked you were talking a lot about RAM and turns out you don't even know tFAW is and you think CL matters. You're technically illiterate.

And that's ok, 99% of people are, but you giving wrong advice for whatever reasons you have is just shitty thing to do

[static cosmos]

if I could talk calmly I would explain it to you, but the translator translated it terribly

[tiny plank]

How to improve FPS in Squad, it's simple. Disable all exploit protection settings for SquadGame.exe, and use a Python script to disable hyperthreading.

[tiny plank]

Bro there's nothing to explain you think Process Lasso does anything to a game with Easy Anti-Cheat enabled

You just talk, 0 experience

[static cosmos]

I have screenshots, just wait

[tiny plank]

It doesn't matter what Process Lasso or Task Managers shows you, you can even try to use RegEdit to change the core affinity...

It doesn't work with Easy Anti-Cheat, you need to launch the launcher itself with less affinity, which will launch Easy Anti-Cheat with less affinity, which will finally launch SquadGame.exe with less core affinity

Do you have Python on your system? I can write the script for you and you can test it

[static cosmos]

Its process lasso

He’s working, brother

[tiny plank]

Wdym it's working? You didn't explain your low res pictures

[static cosmos]

there, if anything, you can see how priorities were set for the threads

[tiny plank]

1. You can't, it's a low res picture. 2) It doesn't matter, it's visual only.

import subprocess

p = psutil.Process()
p.cpu_affinity([0, 1, 2, 3, 4, 5]) # Enable CPU affinities for 0, 1, 2, 3, 4, 5

print(f"Launching game launcher as child process and exitting..")
p = subprocess.Popen("squad_launcher.exe", start_new_session=True)```

There you go, Python script that actually changes core affinity

Try it and see your FPS go up and less stutter

...\steamapps\common\Squad\squad_launcher.exe <---- This is the launcher you need to put the script next to it

[static cosmos]

and dude if you even understood you would just tell the person to turn off SMT instead of making it 100 times more complicated

[static cosmos]

when I reassemble the computer I'll do a test, okay

[tiny plank]

Because the Windows Defender part is important as well for stutter

[static cosmos]

I simply don't have stutters and the fps is high without this

[tiny plank]

There's no point in trying to increase FPS when Defender throttles it

[tiny plank]

Good for you

I have older CPU and average 100+ FPS limited by my shitty GPU only at 2K resolution, a lot of people here have top-gen CPU and get lower FPS at 1080p lol with better GPUs

[static cosmos]

160 fps - 7800x3d and 150-160 on 14600k

I recently found out

without any settings there fps 100-130

[tiny plank]

Read through this thread, very few people even get 100 xd

That's just Squad devs being lazy af like disabling hyperthreading and fixing the Windows Defender throttling would be easy thing to do, literally a few weeks of work to integrate it into their launcher

But they already listened and in new UE5 versiont they fixed the excessive regedit read/writes, so it should lower the Defender throttling

[static cosmos]

many people are sitting without xmp, or their memory in Gear 2 is DDR4. the problem here is that people just don't understand this

here the user above got normal fps just by putting Gear 1 on xmp

[static cosmos]

I generally hope they fix the multithreading and fix the lags due to synchronization with the main threads

[tiny plank]

Let's see, I think some other milsim will come out to take the cake before they do that

[severe glacier]

note that this will only work for AMD cpus, for intel cpus you need to set it to odd numbers only

[tiny plank]

Cause of the whole p-core thing?

[severe glacier]

Kind of, AMD lists all the cores first, then it lists all the threads after. Intel does the opposite and lists each core and it's threads next to each other.

So on amd CPUs, node 0 is core 0 , and node 1 is core 1. But on intel, node 0 is core 0, but node 1 is also core 0. Hence why on intel you need to use every second node to keep your performance

[full harness]

Is asynchronous texture streeaming implemented ?

Is foliage procedural generation implemented ?

Is virtual geometry implemented ?

Is peer-to-peer replication implemented ?

Is adaptive tick-rate implemented ?

Is spacial partitioning for physics implemented ?

[frigid gale]

https://tenor.com/view/peace-dueces-fadeaway-gif-14327700

[severe glacier]

Always has been

[severe glacier]

That's an artist tool, nothing to do with a packaged game

[severe glacier]

Absolutely horrible idea for any non coop multiplayer (think how bad GTA v can get for network attacks)

[severe glacier]

Unreal engine does this by default

[full harness]

Could prove useful for camo instead of pre-baked textures

= less textures in VRAM pool

[full harness]

Additionally i was just probing at what was implemented and peer-to-peer replication could prove useful only for graphical effects

[severe glacier]

Then there is the tradeoff of vram Vs performance, sure it means less textures in the pool, but then the shaders need to calculate the same results all over again each frame

[severe glacier]

I don't believe you can limit P2P replication to certain actors, I think it is an all or nothing affair

[full harness]

well... i suppose that this could be coded

[severe glacier]

Ehh, you probably could in an engine like unity, but with unreal engine, you really want to stick to the builtin tools, especially for networking.

[full harness]

Wait

[full harness]

I think there is a way this can be implemented

You can use builtin replication system

Just mark graphical effects as replicated and use remote procedure calls

bReplicateMovement is a flag for affecting gameplay, this needs to be disabled and DONE

Wait... This could just be the best option... gameplay calculations for server, every other effect could be managed by p2p

[severe glacier]

I think you're misunderstanding how RPCs work. Graphical effects are already replicating (otherwise nobody could see the effects), and replication in unreal engine is network authoritative, the server tells the clients what to do, not the other way around (not technically a limitation of the engine, as client only code can call to a server only event, but that is widely considered bad practice)

Also there is only 1 way to do P2P in unreal engine, and that is with a listen server where 1 player is also the server. There is no way for one player to talk directly to another player, it must go through the server.

[full harness]

darn

okay... one idea to the trash bin

[sleek zenith]

Disable contact shadows on low settings

It's an enabled thing that wastes a lot of performance

U can disable contact but only on high settings not low for some reason

[steady badge]

what settings on lossless scaling?

[velvet niche]

arent they disabled by default?

[sleek zenith]

i pretty sure its enabled by default and your cannot disable it intil you have high or epic settings

from my experince

[latent shoal]

Contact shadows are relitively cheap and shouldn't be causing significant performance drops. It's more so higher resolution shadows and distant shadows that cause the performance problems due to the CPU hit.

[junior laurel]

When owi make video comparison about fps ue 4 vs ue 5? Cause im curious bout that

[tiny plank]

Does Squad suffer from VRAM though? AFAIK these days games aren't really limited by VRAM with all these 16 GB and 24 GB models coming up.

I just once had an extensive look into procedural textures and the line between worth-it vs not is not that clear really

[sleek zenith]

Would be nice tho to add that back as a returning option cause it used to be a thing u could change

https://tenor.com/view/squid-game-seong-gi-hun-locked-in-gif-18145163737889558922

Funny squid game number

[wispy kestrel]

It is hoped that abuse of OP privileges and electronic attacks on other servers in China will be regulated within the Chinese server. By the way, a large number of Chinese players will attack each other on the Chinese community 贴吧 (post bar).

[noble flame]

ASUS-TUF17 laptop
Processor AMD Ryzen 7 5800H with Radeon Graphics 3.20 GHz
RTX 3060 (6GB)
Installed RAM 16.0 GB

[static cosmos]

Laptop have weak memory subsystem

and the squad loves fast memory bandwidth

[noble flame]

So the frame drop because of RAM not GPU ??

[static cosmos]

Yes

Unfortunately, nothing can be done on a laptop

[noble flame]

I At the main base and when the server seeding the frames reach 120fps but the frames drop when I reach the enemy zone

[noble flame]

Enamy zone= in invasion

[static cosmos]

Well, you see that you are running into memory bandwidth

when nothing happens then fps is high

there are laptops with 96 MB l3 cache. Which compensates for this disadvantage a little. but at the same time these laptops cost a lot and the fps there will not be more than 7600x

that 7800x3d with 4800 memory - becomes very weak, because the cache cannot compensate for the RAM bandwidth

[vague mulch]

here you go, finally 45fps on al basrash

[latent shoal]

The main bulk of optimizations haven't occured, especially for Al Basrah, which only recently became finalized in terms of it's layout, building placements etc Still has a lot of optimization work to do, especially when it comes to dialing in setting scalability, which Slay plays on the higher end of settings.

One of the current focuses is dialing in PiP performance in UE5 testing. The UE5 migration is still very early days, all the foundational aspects are still being worked on.

[thorn lion]

if we assume that Slays specs are the specs in the video then they should be this

To be fair thats not the best CPU in the world so we might not be cooked

But it could shift the game to GPU heavy and in that case it could be not great because 90FPS on a 3090 in 1080p is kinda sad right?

But his system is most likely CPU bottlenecked hard so the FPS hopefully is not representative for a 3090 @ 1080p

Also OBS is running so that is also a performance hit too.

Idk if we can look at the FPS numbers in the video and make judgements on what the performance will be

[misty swift]

We dont know which graphics settings he is using.

[thorn lion]

Also Fair

Wait no

[thorn lion]

Its in the picture, Medium Quality

He mustve hit the Medium preset or we have to assume

[misty swift]

Well yeah but we still lack important details. For example he may be running DLSS DLAA or DLSS performance, we dont know.

[thorn lion]

Also fair point

[fluid frost]

is that your rifle or nah because it looks fucking sexy

[strange axle]

Not sure if I should post this here since I also made a bug report about it but here it is anyways.
As you can see in the top right corner of the video, when using a 1.0 sensitivity, the game runs perfectly. However, when using a 0.10 sens, the game stutters like crazy when trying to turn. Both tests were done on the same DPI but I've tested with many different DPIs and the problem always comes back to the low in-game sens values.

[latent shoal]

It was one of my service rifles, yes. Has since been replaced with an MRAD .338

[fluid frost]

Hey man, I remember asking you a question and i forgot what to do, it was something about xmp profiles, if you maty answer my question again thatll be nice. xmp 1 profile or 2 is the best? or is it option

[fluid frost]

yummy

[static cosmos]

Xmp 1

[fluid frost]

Okay, Thanks!

[static cosmos]

the second one is usually in gear 2 on ddr4

[fluid frost]

i have ddr4 im pretty sure

so, still xmp1 correct?

[static cosmos]

Yeah

[fluid frost]

Okay, thank you very much.

[crimson valley]

when ue 5 to squad

[steady badge]

Lenovo Legion 5 Laptop Ryzen 5 4600H GTX 1660TI 32 GB RAM, getting around 100 FPS easily on full 100 players servers using LSFG 3.0 on lossless scaling app thanks bye

[tiny plank]

copium

[tiny plank]

10900K is not that bad, it's decent. I run Squad at like ~90 FPS usually at 10600k. It depends on your RAM though, the 2x32GB DDR4 is 100% not good ram though.

https://www.youtube.com/watch?v=2UUzW5hgiFU

Look, FPS diff between 10900k and 14900k not that huge.

CPUs have stagnated a lot since a couple of years ago so that's why, especially when testing on a game like Squad where memory is so unoptimized so you'll have a lot of cache misses

It'll be the raw GHz that is the deciding factor then with the RAM latency

UE5 is probably going to be worse in GPU frametimes, but better in CPU frametimes, but let's see

[misty swift]

The most relevant factor for Squad is cache size, not GHz.

[thorn lion]

I don't really trust no name benchmark channels as they have no tangible accountability or reliability Consider looking at https://gamersnexus.net/ https://hardwareunboxed.com/

https://gamersnexus.net/cpus/amd-ryzen-9-9950x3d-cpu-review-benchmarks-vs-9800x3d-285k-9950x-more#9950x3d-conclusion

[static cosmos]

I now have 12400f in performance as 5700x3d and 5800x3d and 7600x.
It's not about the cache, but about the memory bandwidth of your processor

Well, 10600k(12mb) and 5600x(32mb) have the same performance in our game

AMD generally feels bad in a squad, simply because of its weak infinity fabric. Of course, you can compensate for this with X3D cache and tuned memory, which is what AMD does. But the cache cannot fully compensate for the weak processor throughput

it's like the bus width in nvidia 4000 video cards. They cut the bus width from 256 to 192 and tried to compensate with a larger cache, and now the video card has to access the VRAM several times more, instead of getting all the data at once

give us an open world, with high visibility, density, detail and textures of 3D objects and that's it, 3D cache is far from the solution

[latent shoal]

No, it's a fact.

[tiny plank]

Yeah, which is exactly why the GHz matters, buddy. How do you think cache works in a CPU?

The point is is that the newer 14900k isn't much faster than 10900k exactly because most games aren't pipeline optimized, so the memory latency + cache + GHz will matter more than the more modern pipeline

[tiny plank]

Yeah, but this game isn't Cyberpunk, Cyberpunk is optimized for CPU and CPU-heavy. Squad is CPU-light and unoptimized for CPU.

Squad is basically a simple game failing to do something simple efficiently, Cyberpunk is more like a complex game optimized to do something complex.

[tiny plank]

Nope, OWI devs don't need UE5 to improve performance, they just don't care about it enough so the performance will be bad in UE5 as well because they will just overload the game until it runs like shit again

[misty swift]

This statement shows that you do not understand how CPUs work in practice. If you have a CPU heavy game like Squad which needs to access a lot of different data on the memory, and which also suffers from an obsolete engine and unoptimized code, then the processor speed does not really matter, because it cannot get the necessary data fast enough anyway, because RAM in general is too slow. What matters then is the cache size because more cache means that less memory bandwidth is required. This is why X3D CPUs perform really well in this game.

[misty swift]

And this statement shows that you dont know the technical details of Squad. running a 100p multiplayer game is not "CPU-light"

[rotund echo]

"Nvidia DLSS Antialiasing", it's written

[misty swift]

This is not specific. It can mean native rendering using DLAA, or rendering at like 1/4 of the resolution using ultra performance. The performance results vary depending on the exact setting. Unless we know the exact settings, drawing exact conclusions about the performance on the video is not possible.

[rotund echo]

I need a TLDR

So, basicaly, after years cooking on that UE5 update, we finaly got a benchmark from some random youtube channel, displaying a pathetic 45 fps in Al Basrah.

The official response seems to be "We're still in early stage of optimization", like optimization wasn't the main reason for this upgrade in the first place.

And in the end, we will have to wait for another long time and we can already somewhat know that the result is expexted to be mild. Is that correct ?

[latent shoal]

Moving to UE5 is not only about improving performance, but ridding of a lot of spaghetti code, that is tied to left over legacy code and redundant no longer supported UE4 systems and tech. I give more detail here.
#squad-general message

Tldr; Future proofing squad opening the door to other planned features, improving optimization not only in terms of performance, but server side, codebase and many other aspects of the game under the hood. Which I can tell you(As out of me and you, I'm the only one who has access to Squad on UE5) is being worked towards with relative success so far.

[static cosmos]

They always talk about cache here. But I would rather have a normal architecture and memory controller so that DDR5 works perfectly. X3D has a problem with 0.1 and 1% fps when the processor relies only on it

as for me, DDR5 is worth switching to, at least because of the principle of its operation (multi-channel/multi-threading)
and the big memory bandwidth is already a bonus

[static cosmos]

zen6 processors will be like this, I think. They want to redesign the whole memory management

[rotund echo]

I hope you are right about that. Because the definition of "sucessfull" might defer beetwin devs and players.

I do wish for the game to be good, it's just all I saw happening for the past 2 years was adding more content, more bugs, and not fixing any of the olds bugs nor new.

So, assuming thoses bugs are due to spaghetti code and will be gone with UE5, great, but if that happens in 2 years that will be a very long time untill we see the fruits of that labor

There are some bugs that should be easy fixes, yet they stays for months. So it's hard to believe in anything OWI or associates says untill we see the things ourselfs

[frigid gale]

with that being said I'm now fully convinced that the ghost shooting is just a feature to make yourself believe you have schizophrenia

cause that has been in the game for like 3 years now

[tiny plank]

Does X3D really have problem with 0.1% and 1% FPS though? A lot of benchmarks look good relative to 14900k, but of course it's vs normal RAM not some super tuned $2k setup

[tiny plank]

don't make me hope D:

[static cosmos]

yes, there is usually nothing like that in benchmarks, you need to test this and that yourself. and you will see how the fps drops momentarily when the 96 MB cache is filled in the game. and it is constantly filled in games like Tarkov, Squad, Arma 3 and in general in many games

[static cosmos]

The average FPS of the X3D is high. But at the same time 0.1% and 1% are 30% worse. This creates a not very pleasant picture.
when other processors have 0.1% and 1%, it's only 10% worse, which feels much better than the fps from x3d.

It’s OC

For me, 0.1% and 1% are much better than just useless huge numbers in fps

9800x3d also suffers from this

[misty swift]

The X3d has the higherst 1% and 0.1% FPS in those graphes

What exactly do you mean?

[static cosmos]

It's big. but the difference with the average fps is 20-30%, and this is his problem.

[misty swift]

This is not an issue. By locking the max FPS the difference between average and 0.1% FPS can be reduced to the levels of the other CPUs, while still offering better low FPS.

[static cosmos]

the smaller the difference between the average fps and 0.1% and 1%, the more stable the picture is. I don't care that the average fps is high, it feels significantly worse

[static cosmos]

set fps limiter like 9700x?

[misty swift]

Yes for example.

Should work in theory.

[static cosmos]

I would rather buy 9700x and wait for Zen 6. I don't like these X3d processors at all

there infinity fabric will be reworked and maybe the problem with low 0.1% and 1% will be fixed in X3D processors

[static cosmos]

here's one from fairly recently

[tiny plank]

I am confused, isn't the 7800X3D having the highest 0.1% FPS in this graph?

Or you mean the relative drop between AVG fps and 0.1% FPS?

[static cosmos]

Yes, I'm talking about fps Avg and 1% 0.1%

the translator apparently failed to convey the idea completely

for example, for 7800x3d the difference in avg and 1% is 30 frames

while for regular processors the difference is 15-20 frames

You can already see from 0.1% and 1% that X3D has problems.
and it was done in a sterile environment with a clean system and so on.
And one can only imagine that some user might have at least 40-50 frames difference there. And an average fps of 160 frames will not help to make a good smooth picture.

[static cosmos]

X3D people look at the average fps, not even paying attention to 0.1% and 1%. The average is high for them, of course, but the values ​​of 0.1% and 1% remain the same as for a processor without cache. that's why the picture is more jerky

[tiny plank]

I mean fair enough, but the 0.1% FPS is still higher and you can always cap your FPS to be close to 0.1% FPS to make the change lower

So like if 7800X3D is 0.1% FPS of 100, then you cap it to 100 and then your avg FPS is like 100, and 0.1% FPS is like 95

Without the X3D cache you'd need to cap it to like 90 FPS and have 0.1% of 85 for example

[static cosmos]

Yes. But then I don't see the point of X3D, they were advertised as processors capable of producing many frames. But in reality they are just processors where the bandwidth is compensated by the cache.

And normal performance becomes only slightly better than processors without cache

I'm just looking. And the real increase from X3D can only be counted in a couple of games. And there are the main ones CS2 and Pubg and other eSports disciplines, where for some reason people lack 400 frames

[static cosmos]

Like this

in other games the cache is completely clogged and gives out a terrible 1% and 0.1%,

[steel walrus]

4096w limit

least power hungry 14900kf

[severe glacier]

What do you mean by the cache getting clogged? That is the caches sole job. If you cache isnt full, then you have a defective CPU and need to return it

[static cosmos]

It’s normal. He using maximum 100 watt in game. This made for stable frequency

[static cosmos]

cache - buffer with fast access to data, and it is able to fully rest on its volume "96 MB" in games

because of its small volume it is not able to transfer all the data at once, which is why it has to transfer the data two or three times. When RAM is able to bring everything at once

or you run twice for separate bottles of 1.5, although very quickly (GB/s of each chip), and I, slower (GB/s of each chip), but at one time I bring 2 bottles of 1.5 at the same time
and, as a result, we have the same 3 liters (total PSP), only someone screwed up and got tired of running back and forth, one by one, and someone, without stress, brought everything at once

here for example I will show 5700x3d, where the cache is not full and where it is completely full

200 fps

120 fps

It’s 5600 with 100 fps. 5700x3d performance is only slightly better than 5600

here is already 5700x3d with 150 fps configured memory and bandwidth of 56 GB instead of 48 with lower latencies

[static cosmos]

More better screen 5600

The cache in Squad also has a bad effect on 0.1% and 1%, although it gives a good average fps

[severe glacier]

I know how cpu cache works, and it doesn't affect programs in the way you describe.

Cache is required because ram is not fast enough to keep up with the CPU, so the cpu chokes with a memory bandwidth bottleneck.

Programs are 1 of 2 types, they are either cache friendly, or they are cache hostile. Cache friendly applications are able to keep the same data in cache for a long time, cache hostile programs are not.

Whether your program is friendly or not is determined by 2 things:

• A) how much data you are working with
  as when you are working with less data than the cache size, you can be confident that the data is always in the cache.
• or B) how you are accessing the data.
  The CPU preloads the cache by looking at how your program is using memory, and if you are using it in a predictable way (linear access is a good start) then the CPU can continually prefetch data as you use it so the program is never waiting.

Games are particularly hostile, because they have both, a lot of data to work through every tick, and they access memory in unpredictable ways in non linear places.

X3D chips help games as they give them more cache, which means that there is less swapping required to work through a particular size of memory. For example:
If a CPU has 8MiB of cache and the program needs to use 20MiB of data, the CPU must do at least 12MiB of data transfers. However, if the CPU instead had 15MIB of cache, then it only needs to transfer at least 5MiB of data.

There is no situation possible, where an OS that uses more than the CPU's cache size of memory, where the cache of the CPU is not full. It isn't possible to encounter this, the cache is always full.

You also misunderstood the sheer difference in speeds of the cache and the ram. In your example you list cache as 1x1.5gbps and ram as 2x1.5gbps, not only is this a bad example, as the CPU does not work directly off of ram (although there are instructions to force it, they are privileged and are disabled on all 3 top OS's) but the scales are also horribly misleading. System ram runs at 2x your ram speed, so 2x3600=7200MTs for ddr4 and likely 2x6000=12,000MTs for ddr5. Multiply these by how much data is provided per transfer, and you get about 100gbps for ddr4, and 200gbps for ddr5. This is also ignoring the latency, where cache has 1-5 nanoseconds, and ram has 100-200 nanoseconds

However if you look at a 5800x3d, it has 8 cores that run at 4.2Ghz, which works out at 33,600MHz. Far more than what its ddr4 ram can provide, and far more than ddr5 could've given it.

L3 Cache however, runs at speeds at or around 4-600Gbps, 4-6x ddr4 speed (for the 5800x3d) and 2-3x for ddr5 (7800x3d and above)

In the game examples you show above, the X3D chip is 20% faster than the 5600, this is exactly the performance benefit it advertises so I'm not sure what you are complaining about there,

0.1% low metrics are almost always entirely useless, as they have a rolling time of 15 minutes, where you are far more likely to have a disk lookup which causes the drop.

The reason you can see ever so slightly higher 1% lows on the non X3D chips is absolutely nothing to do with the cache, and it is everything to do with the fact that X3D chips cannot boost as high, e.g. the 5800x3d can boost to 4.2 all cores and 4.45 with fewer, whereas the 5800x can do 4.4 all cores and 4.7 with fewer.

Please look up how cpu cache works, and how it is used before you start talking about how it affects games. There is an article here by OSdev that explains everything.

[static cosmos]

Okay, let me start trying to argue in my own language:

Кэш - у Амд компенсирует слабую пропускную способность памяти, я буквально так и сказал. Но оно никак не может полностью заменить кольцевую шину у процессора. И здесь буквально сложилась такая ситуация- что у процессоров AMD очень медленная связь между ядрами и оперативной памятью. И кэш дает лишь средний фпс, а все другое уже не способен он должным образом сделать. Пропускная способность у процессоров AM5 не способна давать больше 64-70 гб на одном чиплете, ибо ограничено слабой кольцевой шиной. Когда у того же Интела пропускная способность 100+ гб на каждом процессоре, и эта особенность во многих сценариях позволяет иметь прирост в тех местах, где X3D имеет плохой фпс что как бы фактом является.
Поэтому я не люблю X3D нынешние, потому что у них INFINITY FABRIC слабая и они просто повышают тебе средний фпс, а все другое на ужасном уровне. Ибо кэшем нельзя полностью заменить кольцевую шину которая за раз способна больше сделать чем кэш за две-три ходки.

Здесь буквально ситуация как с видеокартами. Видеокартам Nvidia урезали шину с 384 до 256 и добавили много кэша чтобы компенсировать утерю эту. Но теперь предтоповые видеокарты не способны получать большую производительность в разрешении 2160, и кэш вообще никак не помогает

wrote in Russian. Because I see that the translator cannot convey my thought and you constantly tell me that I wrote literally

I don't see any sense in the current X3D processors. It's better to wait for ZEN 6, where AMD will rework their infinity fabric and the throughput will be at the level of Intel. Then the sense of 96 MB cache will be significant.

tuned intel with a frequency of 8000 and the most select and overclocked to the very limit 9800x3d with a frequency of 8800 and fabric 2233

https://youtu.be/MWWf0ZmC8CM?si=zsgWUqwry_jdMDvA

Here with a friend we tested my 12400f with DDR5 and 5700x3d, I got 10 frames more under the same conditions

18 MB cache beat a glorious 96 MB cache, oh my god

[slow pendant]

All this talk about the cache and high FPS vs 0.1% lows got me thinking about how I always felt like squad has a lot of hitching on my 5800x3d & 3080 combo. I thought it was normal (everyone gets it) but is that not the case? Are there people playing with well-setup machines that never have frametimes above 20ms? My averages are pretty much always at or above 100fps, but the stuttering is painful once every minute or two in the midst of battle.

[rotund echo]

5800X3D, 7900XTX, 0 stutter

It's apparently very random

Ask gammapheonix about it

Dude spent 4k to have a monstruous gig with a 4090 in it and he literaly rage quitted the game because the stutters were horrendous

I have personnaly no idea why my rig runs squad well and his dont, but it does

UE5 update this decade

[misty swift]

People who rage quit because of "stutters" are often just blaime something which is not themselves for loosing.

[misty swift]

Is this from a live game or local?

This is with a 4080 + 7900X3D, Fallujah 60-70 players. The spikes are caused by Hellcanon impacts lol.

[slow pendant]

5 rounds on different maps in a server that was at 50v50 each time I took the 10 minute benchmark.

Thanks for sharing! If you happen to be playing on a fully populated (100 players) Sumari / Mutaha and remember, would love to see if your frame times (stuttering) are much better than mine.

On one hand I hope that it's something that I could fix, and on the other hand I don't feel like buying new ram etc so I want to just blame the engine and pretend it'll magically get better with ue5 release without me having to do anything 😅

[slow pendant]

Would you mind sharing a capture when in a 50v50 round, ideally invasion? I would love to know the variances that you are getting since you have the same cpu, to know what I should be shooting for 🙏

[misty swift]

The server was actually almost full during this recording, I replaced the original picture. I can try to get some measurements for those other maps, yeah. But Fallujah should be one of the most performance heavy maps, the others should run better. It would be good if someone could test non-X3D CPUs as well.

[slow pendant]

Ah ok, thanks for the info!

[rotund echo]

All i have is tank gameplay 1km away from objs lmao. But yea i have a lot of records and i dont experience stutters at all. Dont have records with metrics on tho

[static cosmos]

I'm trying to record fragments now. But I'll have to compare only with 5800x3d, because 12400f can't compete with 7900x3d anymore

[static cosmos]

@slow pendant

playing on full server

Harju was the hardest. There was a game against the Serbs and they used Grads and a lot of different artillery

7900X3D has 1% and 0.1% worse, but the average fps is higher🤔

I hope AMD finally makes new Infinity fabric. Otherwise, this is complete nonsense.

[static cosmos]

You don't have to buy new RAM. Any memory on DDR4 can be configured. If anything, write to me in PM

[restive mantle]

i have problem with all 4x scope every time i open the scope fps drop to 30

any tips ?

[misty swift]

We would have to test under equal conditions tho. Same map, same graphics settings and resolutions, with same memory bandwidth. Your RAM is 400 MT/s faster than mine.

[static cosmos]

we can arrange this if you want

Because I'm wondering what the real difference is. I am 100% sure that my processor is superior to 5700x3d and 5800x3d with 7500f-7600x. And then I saw that 7900X3D was worse than me in these indicators.

and as I recall, is your system configured the same as mine?

[static cosmos]

you have a memory of 6000 - gear 1. I have a 6400 - gear 2. These 6400 mhz will not be equal to my 6400

[steel walrus]

set your scope refresh rate to 60

it helps minimize some of the performance hit when adsing but can introduce some screen tear

[half folio]

thought this thread would be dead because of UE5

when we go to UE5 there WILL be more stutters, the game WILL be laggier and there WILL be more artifacts

say goodbye to game's performance
we had no devs or anything assure these wouldn't happen and the UE5 "upgrade" isn't just a backward step disguised as an upgrade

[steel walrus]

yeah no

[modest oriole]

And after the UE5 release they will send an army of shills on steam to attack everyone who criticizes and to write fake positive reviews to bait as many sales as possible. OWI/Tencent classics 🤣

[latent shoal]

Keep out of here, thanks.

[modest oriole]

I describe the history of what happened. You should remember that hardcore optimization was promised since V2.12 in 2022 and now they will push people with older hardware away from the game or force them to buy new one.

So who's the propagandist here?😁

[latent shoal]

You spent half a decade carrying out a campaign across steam, reddit and discord, claiming OWI is being paid off by Intel, to make AMD FX CPU's run terrible. You then spent years claiming SQUAD performance is intentionally bad, to boost Nvidia/Intel sales....I highly suggest you follow any direction given by staff here since your account is on it's last strike.

[rotund echo]

It kinda is true though that OWI made a very bad move of promising screenshots leaks of new content in trade of posivitives reviews, unlike Steam TOS indicates, and that a campaign to saves the game's Steam review from going overwhelmingly negative happened when the ICO review bombing happenned

So while his whining is annoying, it's not entirely baseless

And it's also true that despite the recent interview regarding UE5 beeing very interesting, the FPS counter indicated very low numbers when you take into account the rendering resolution and the test rig used

We're talking sub 100 fps average on a 3090 and it craters to 50 when the guy uses his ACOG

On a 1080p resolution

Also, unrelated, but the server broswer is still laughably bad, and it has absolutly nothing to do with UE5

It's just on low priority because it sometimes bugs out terribly, but works most of the time, so it's just not important

[static cosmos]

I still don't understand why they can't just log in via IP. As if that would solve every problem

[modest oriole]

Wrong, I never said they favour Intel CPUs over AMD, so you made that up, or you mistaken me with someone else. Also. Couldn't spend half of a decade because copr you represent immediately banned me on every possible platform except here for asking questions about poor optimization and informing people about it. 😉

I claim that OWI/Tencent corp (like many other companies) being paid for intentional de-optimizaion of a game from 2018, fully released in 2020 to force people buy new hardware. The best evidence is hardware requirements on the steam page. Also they should've make Squat 2 instead of alienating lower hardware players.

I also claim that OWI/Tencent corp use shills on different forums and steam to attack everyone who criticizes their decisions and to write fake reviews.

I think I'm not the only one 😁

[floral mango]

mayb you should've made Squat 2.

[modest oriole]

Classic answer, like a baby. I work in different branch of engineering, ok?😁 Should I ask OWI/Tencent corp to build a power plant? What logic is this?

[floral mango]

it was just a mayb

[latent shoal]

We log every warning and negative interaction, this is only two of many examples. Now for the 100th time and final time, avoid tinfoil hat posting and spreading rumors, thanks.

[tiny plank]

I do think what ur saying is true in the sense of games being shit, but i dont think it's a conspiracy. Like just look at the Weather app in Windows, it's literally Edge in the background taking 1 GB of ram to show you if its gonna rain or not

it's just developer skill issue, all the good devs go work for fintech not for making games or apps

UE4 is suboptimal in many ways just to be like usable for average devs

[modest oriole]

It's the same statement, and part about favoring Intel over AMD is made up😁.

I don't spread a rumors and stop repeating tin foil like this is the only term you know instead of real counter arguments. OWI/Tencent corp will again de-optimize Squat for low end PCs.

I wonder whether you store false statements of OWI devs about optimizations when ones for 3 years they wrote something on Reddit or steam.

[modest oriole]

Why do we need to call it conspiracy? It's just a simple business. Less hardware being sold because of the prices, people play old games so why not to increse hardware demands for the same game.

They will basically cut out older GPUs.

[floral mango]

Only as long as you spell it as Squat from now on.

[tiny plank]

Cause what prevents u from opening ur own game studio and releasing well-optimized games?

so people suddenly dont have to buy newer PCs, u have this market inefficiency u can abuse

[half folio]

how dare you point out an obvious flaw with a product???
why wont u go fix it yourself instead of brining light to the issue?

[misty swift]

He is not pointing out flaws he is spreading conspiracy theories without providing any evidence or reason.

[half folio]

what conspiracy theories?
im saying EU5 will perform worse and it's based on other games released on the same engine and dev's history of prioritizing performance of this game

that's not a conspiracy theory that's just how things have been so far

[misty swift]

Like claiming that OWI is paid by Intel, Nvidia and Tencent to bully people with low spec PCs so that they are forced to buy new hardware. Also how do you know how UE5 will perform on release, do you have a cristal ball for that?

[half folio]

what are u on my man?
when did i say that?

the other guy said why wont u go make ur own game studio and release well optimized games :/
i answered to that

[misty swift]

Not you, the other guy.

[half folio]

then why u replying to me?

[misty swift]

Seemed like your comment was related to the other guys statement.

Anyway, nevermind then.

[rotund echo]

"How do you know how UE5 will perform on release ?"

Probably the FPS counter at the top right and the specs described in the video description

https://m.youtube.com/watch?v=Cv83hfaH8PA&pp=ygURT3dpIHVlNSBpbnRlcnZpZXfSBwkJ2ACjtWo3m0M%3D

Now sure, optimisation comes last and it can get better before release

[thorn lion]

He was using OBS during the gameplay so FPS takes a hit, also his system is a huge bottleneck

[rotund echo]

Bro

10900k

Since when is that considered garbage

[thorn lion]

With a 3090

[rotund echo]

Well yea

[thorn lion]

Bottleneck not garbage

[rotund echo]

So you need a 500 dollars 9800X3D otherwise you can't run the game ?

[thorn lion]

Squad UE4 is CPU bound so if your system is cpu bottlenecks its not going to be representative of the 3090 perf

[thorn lion]

Never said to buy a 9800x3d

[rotund echo]

Bro

[thorn lion]

We will see how CPU bound UE5 is

[rotund echo]

10900k might be old but that still is a high end cpu that was sold about 700 dollars

[thorn lion]

Maybe just wait 24 hours to make judgments

[rotund echo]

Why ? Does it release tommorow ?

[thorn lion]

I dont see how that’s relevant to the cpu bottlenecking

[thorn lion]

Yes

#squad-announcements message

[rotund echo]

Why are you even talking about a bottleneck in the first place ?

Also, thanks, didn't know it would release tommorow

[latent shoal]

What is in Slays video and will be in the public test build, will not be reflective of Squad UE5 release. UE5 still has a lot of hurdles to be bypassed when it comes to system revisions and just overall optimizations. It's what testing is for, to help dial in a lot of aspects and test differing iterations of things.

Do not treat any tests like it's a finalized complete product, because it is not.

[rotund echo]

Btw, is the playtest branch as heavy as the public branch ?

[thorn lion]

Define heavy?

[rotund echo]

Storage wise

Currently Squad eats about more than 100 GB on my PC

[thorn lion]

Someone said in the testing discord its about the same

[latent shoal]

It's around the same, current build I have is slightly less. But this will vary dependent on test build.

[thorn lion]

So expect having 2 squads installed

[rotund echo]

Oof

Well

I'll give it a fair try

But Stephano

I still have no clue what your point was btw, or why you even intervened

Like the 10900k is a beaffy, but old CPU, and I personnaly play Squad with OBS running all the time

So i have absolutly no idea what you were trying to say here

Also, the metrics i use are far more complete than a simple fps counter. I use afterburner like everyone does

They could have used that in the video also, but nevermind

[static cosmos]

in stock 10900k gives almost the same amount of frames now. I wonder what if it will be configured, because this person clearly did not have this

[magic elbow]

i have r7 3700x and 2060s, ATM im with ~70fps in game, im ready to see how many fps i have on UE5

[half folio]

"Do not treat any tests like it's a finalized complete product, because it is not."

yeah i forgot about ICO playtests and v2.12 playtests

[floral mango]

I just love Squat and the playtests have always been good and changes were made during them and on release.

[floral mango]

From my point of view the scale of this UE5 transition is so big that its a no brainer the release version will be more polished than what we play in testing.

[frigid gale]

ima pray to god for 60 fps and smooth gameplay

[topaz blaze]

60fps but no stutters and good looking scopes would be dream

[fluid frost]

How much you get

[magic elbow]

what?

[fluid frost]

How much fps did you get or you didnt playtest

[magic elbow]

The playtest opens tomorrow, I haven't tested yet but I think it's going to be pretty bad.

[fluid frost]

Im assuming its horrible, and we probably cant do anything because its just squad devs

[magic elbow]

¯_(ツ)_/¯

I already know what to expect, I just hope I'm wrong. I love this game and I'd hate to stop playing it because of an optimization update that makes it less optimized than before.

[fluid frost]

exactly, i have high hopes that im wrong about what im thinking'

37 ps on harju and game still looking like shit awesome

[magic elbow]

what is your spec?

[fluid frost]

3060 ti, 16g ram, and an gen 11, intel i7-117k

everything is mostly on low

medium*(

[magic elbow]

try to turn on XMP

[fluid frost]

XMP is on I think

i think its on profile one]

[steel walrus]

i have a 2070 super and i get anywhere from 70 to 100fps on ue5, just gotta fiddle with your settings

🤷‍♂️ i think people are going to overact no matter how good this play test goes, they could literally just change the lighting and use ue4 and claim it as ue5 with a 3fps drop and people will go "Wahhh i lost x fps this game is shit and dying!!'

we also arent in 2013 anymore, upgrading to semi decent hardware isnt out of the picture and sometimes you need to cut the 14 year old dual core cpu's off

[magic elbow]

Optimization update = more FPS or stability. Or just as much for a more complex game.
I'm not sure how long you've had the game, but the latest updates to “optimize” the game have indeed drastically reduced performance for all users for a slight graphical improvement that nobody asked for. So no, a 2018/2019 PC is normally capable of running this game without too much difficulty.

[misty swift]

There is something wrong when you get that bad performance with those specs.

[steel walrus]

I will rebuild my old system just to play on this playtest for you to give you an idea of how optimized they have made it 😄

No one is claiming youll be getting 400fps, but dont expect your fps to go to 30

also i was there for a few of the optimization updates that tanked fps, went from 100+ to 60 and sometimes even lower on a r5 3600, with my 7800x3d i get 150+ and @static cosmos helped me with my timings on my ram which brought me closer to the 170 range

[magic elbow]

Don't bother too much just for me I could test with my config tomorrow x). As I say, I just hope that with this update I won't have less FPS than I do now. Otherwise there's no point in calling it optimization.

[steel walrus]

Your rig is what I would say is somewhat lower end for current squad, I’m unsure how it would fare for ue5 but I doubt you’ll see a total loss in fps

[magic elbow]

OS: Windows 10 (x64)
Processor: Intel Core i or AMD Ryzen with 6 physical cores.
Memory: 16 GB RAM.
Graphics: Nvidia GTX 1060 or AMD Radeon 570 with at least 6GB of VRAM.
DirectX: Version 12.
Network: Broadband Internet connection.
Storage: 80 GB available space.
Additional Notes: A microphone.

recommanded perf

i have above that dude x)

[fluid frost]

What would be wrong then, how would I figure that out

[magic elbow]

i turn starcitizen with my 3700x and my little 2060s
so I'm hoping for more FPS on squad than on starcitizen 😂

[steel walrus]

These also havent been changed in years

[fluid frost]

anyone would have any ideas on what could increase fps? running average 40 fps on most of the maps at medium graphics

[static cosmos]

on a 1060 6 GB video card and a 3600 processor I managed to get 60 frames on average

[magic elbow]

ue5 ?

[static cosmos]

UE4

[fluid frost]

did u try it with ue5?

[static cosmos]

Not

Today want

[fluid frost]

alright

does anyone know if dx12 or 11 gives more performance? was loooking to switch to dx12 to see if it has any benefits

[austere tree]

dx12 is better overall

[magic elbow]

I have little lag spike on dx12. Dx11 is more stable and i have better AVG fps

[lean hedge]

Dx11 is way better for performance

[rotund echo]

I assume something doesn't work when you just copy files I guess

Trust me, it's so much worse in game that on the screen shot

[fluid frost]

ue5?

[rotund echo]

yes

[fluid frost]

why the fuck i tloook like that wahts ur settings gang

[rotund echo]

Previous game it was fairly normal, restarted after turning off the FSR that set itself on by itself, and now it's like this

[fluid frost]

turn on fsr lmao

[rotund echo]

It was also normal on native without AA at all

No idea what triggered it

Alright, something is messed up

I started normal, public branch Squad

Every settings has been reset

Random question, we 100% sure there is no link or dependances beetwin Squad and Squad-uat, right ?

Cause I didn't mess with the files in the public branch yet they all got reset

got a backup, and i'm happy about it

[static cosmos]

if it weren't for the 100% load on the video card, there would be 180 frames here

the game has become friends with the processor cores

before this there were 130 frames here

Owi please optimize game

[austere tree]

Wow such a ps1 vibes

[misty wind]

I'm curious if CPU utilisation is any different?

[teal wing]

9070 xt with 5700x3d and 3600mhz 32gb of ram
performance was 40-70 on max possible settings, performance boost was negligable by turning graphics down and the experience was overall a very stuttery mess

on the ue5 playtest

even when playing with above 60fps it stuttered and felt very sluggish

[drowsy flame]

you need at least TSR. If you don't want upscaling just put resolution to 100%. But you need either TSR, DLSS, FSR, or the like. Otherwise you don't have any anti-aliasing and that is what's causing you're issue.

[wraith compass]

ok so.... been a while since i played this game cus i gave up trying because of my setup..... can i play this game with a gtx1070 and i7700k cpu when it transitions to unreal 5 engine ? is the game going to become better optimised ?

[wraith compass]

update. iv just watched the interview with the tech... manager.. of overworld.. he says the unreal engine 5 should improve optimisation and that performance is a focus of them.. especially the hitching problem.

[tiny plank]

Is it really bottlenecking though? I play on 10600K and average 100 frames in Squad
You just need good RAM on those modules, except for the X3D, CPUs have stagnated in performance since the 10th gen
The diff between 10900k and 14900k isn't that big

UE5 running like shit has been confirmed already even on better CPUs, and it makes sense, if ur gonna use Nanite ur gonna tank performance, it's a lazy implementation that causes a lot of overdraws

[tiny plank]

I hope that's image compression and not in-game blur like that

[static cosmos]

I couldn't get the low settings to not be too loud for the video card. So I used an upscaler.

[tiny plank]

Shouldn't add blur though right? Like if anything upscaling should remove blur, but maybe ur sharpness is way down

[static cosmos]

without Gpu bottleneck I have 170-180 fps on average

[tiny plank]

That's nice, that means the game was optimized on the CPU side, but now cause of Nanite it's really shit GPU-wise

How many players was it in-game though?

[static cosmos]

90 players in fight 160-170 frames

Without Gpu bottleneck

[tiny plank]

Yeah I get it, basically on low settings, low res + upscaler?

[static cosmos]

fps drops less from the interactive map. also during art I have almost no drops

[static cosmos]

Yeah

If I do the same on UE4, then my fps there is on average 120

because cpu bottleneck

[tiny plank]

What's ur FPS like with medium settings, and same res as ur monitor, and no DLSS/FSR?

[static cosmos]

1440p preset low - 80-100 frames

[tiny plank]

that's not bad, is it not the same in UE4?

[static cosmos]

in UE4 I could at least set the medium-high preset

[tiny plank]

I see, so game is very graphics unoptimized then

Probably Nanite fault

[latent shoal]

UE5 is more GPU dependent in comparison to UE4, not specifically due to nanite, per se, just an accumilation of multiple aspects under the hood. It's why the recommended min spec is changing to a 1060 6GB Minimum.

In saying that, UE5 for Squad is still early days in it's migration, still a lot of systems being revised, nanite is only partially implemented, as Al Basrah is the only map currently with nearly all of it's planned nanite assets(still be optimized), all other maps don't, yet. Settings scalability, Global Illumination, PiP, Shadows, Dynamic skies etc etc all still need to be iterated on and dialed in, which is what the playtests are for, to get that profiling.

[ancient jay]

привет, какая у тебя видео карта?

[static cosmos]

4070

[ancient jay]

какие настройки у тебя стояли на скрине?

[static cosmos]

Низкие и апскейлер

[ancient jay]

эти 155 кадров везде?

[static cosmos]

Без этого 100 кадров

[static cosmos]

Стабильно 170 кадров если нету упора в видюху

[ancient jay]

еще вопрос!
фпс у тебя выше на движке Ue4 или Ue5?

[static cosmos]

В UE4 у меня максимальный фпс в катке был 140 кадров, здесь же вижу 180 кадров и 170 стабильных

Которых нету на UE4

Проблема UE5 в том, что видюха нагружена

[ancient jay]

интересно, сколько фепесов будет у моей 4060

[ancient jay]

я слышал что вчера у всех фпс упал на 20%

[static cosmos]

У всех это кто?

[ancient jay]

таких же игроко на 4070

я вчера мониторил стримы

[static cosmos]

У всех 4070?

[ancient jay]

те кто делали стримы - да

у всех

ну если на 4070 сейчас норм, выходит на 4060 тоже будет

[ancient jay]

а просадки во время боя которые были на Ue4, они все еще есть?

[static cosmos]

Открываю карту - просадок нету как на UE4
Во время арты нету также просадков

[ancient jay]

и речь не про 1-4 фпс который проседает, а про 30+ фпс

[static cosmos]

Игра стала более стабильной

[ancient jay]

это хорошо

надо будет чекнуть потом, но лучше я это сделаю после обнвки проца

не думаю что 10100ф норм)

спасибо за ответы

[austere tree]

Этого не может быть...

[ancient jay]

что именно?

[fluid frost]

so

does anyone wanna sum up for me

ue4 better or ue5

[misty swift]

UE5 is already better, and it is not even close to being release ready.

[fluid frost]

alr

[rotund echo]

No, it's not. Had you read properly you would have known that without anti aliasing the game still ran properly for a moment before rebooting it.

However, if the change from FSR to none only applied with the reboot and hence you are right, that would be so much worse. Like how could not running anti aliasing at all end up in such horrors

[austere tree]

Не бери в голову

[ancient jay]

бери в рот )

[austere tree]

Спасибо

[wraith compass]

anybody playing squad in unreal engine 5 with gtx 1070 ad i7700k cpu or similar. any better performance?

[crimson valley]

ue5 is having now much less fps than before dont worry guys

they said it will be better (i hope)

on 7800 4070s i have like 40-50 fps less

[rotund echo]

UE5 does solve a lot of problems that makes the life of vehicle players a pain, so i'm very happy about it

It also marks the end of buying an Nvidia card just to cheese the fog. Cause the fog only exists to hide what you can't render in UE4

Tbh, if most of the layers are still foggy after UE5 that would be annoying

[full rivet]

have u check the graphic resolution is 50% or 100%?

cauz they set the default to 50%

and even if u restore it to 100, it's still lookes blurry

[scenic parrot]

what fps did you get before?

[jolly sleet]

do you know how to fix this graphics i really tried everything

[static cosmos]

4gb? Use dx11

[verbal topaz]

Will fsr3-4 be added to the game?

[hexed tapir]

Operating System
Windows 10 Pro 64-bit
CPU
AMD FX-8350 53 °C
Vishera 32nm Technology
RAM
32.0GB Dual-Channel DDR3 @ 803MHz (11-11-11-28)
Motherboard
Gigabyte Technology Co., Ltd. 970A-DS3P (CPU 1) 26 °C
Graphics
LG ULTRAGEAR (1920x1080@60Hz)
4095MB NVIDIA GeForce GTX 1070 (EVGA) 47 °C
Storage
500GB Western Digital WDC WD6004FZWX-00BKVA0 (SATA ) 38 °C
5Tb Western Digital WDC WDS500G2B0A-00SM50 (SATA (SSD)) 30 °C
Optical Drives
ASUS DRW-1814BLT
Audio
NVIDIA High Definition Audio

Game locks up for anywhere from 20 seconds to 1.5 minutes at random, temps are stable, running stock settings NO over or under-clocking. All drivers are updated. Happens on BOTH DirectX versions.

Ironically happens LESS with mods active.

Upgrading hardware is NOT an option

[austere tree]

RIP buddy

[rotund echo]

MF runs a 13 yo rig and says upgrading is not an option

well, good luck bro

Ironicaly this must have been a very expensive build at its time

But 13 years is incredibly long in tech

Props though, you're a true believer

You bought an AMD CPU back when they were shit, I respect that

Apparently it's not even bad, nice. didn't know

I must say though, if, by some miracle, your DDR3 motherboard has an NVME slot, might be worth putting an SSD on it.

Probably the only thing you can try to do to make that antic works

Consider the following :
https://www.amazon.fr/Adaptateur-support-montage-extension-disque/dp/B0BVMCT67W/ref=sr_1_7?__mk_fr_FR=ÅMÅŽÕÑ&crid=3SA2DABPCQXV7&dib=eyJ2IjoiMSJ9.u9Pffi7QQVOXTbhAB8gQ7CO1h47vjMFkE_A8keNsR9nKVI0tbzBLlIkmOgJD2utuAToxn6nUnNrDEiHbzuCBCJElUblubvreB6PVl9E-QXiklPjHGzlOJXhI3Vvy9PJHc0pHMhOMGewvtTpAamkKnTV3RmYBTSxNNN9P3mj-_5HByj79fYMIlHpFGq8NVqTAb5kUL7xWhotRZ2oy1WKcMrnLK8b41tevx1p7ZEUr8N2ZCTtk4ioJPLO6B5f_rmaiWqZ7lJQTIAdfzFa98qRwQfFXsLXp2OPRYwucR3Z6qys.y8wZVwCT2hrCey0-gnd8ThhffwxhhO4Dt32f7Cm1kNQ&dib_tag=se&keywords=NVMe+M.2+SSD+to+PCIe+Card&qid=1745360023&sprefix=nvme+m.2+ssd+to+pcie+card%2Caps%2C74&sr=8-7

Then you put an NVME SSD

Even low tier cheap ass NVME nowadays are 10 times faster than what SATA can do.

So if you buy a disk that you will be able to reuse elsewhere, and put Squad on it, it might help

Maybe

Alright, forget everything I just wrote

PCIE gen 2 is so god danm slow according to wikipedia a gen 2 size 1 slot will have around 500 Mo/s transfer, which is already the speed of your SATA SSD

So just bear with it or play games that are more frendly towards decade old hardware I guess

[rotund echo]

TLDR

[steel walrus]

1. that rig can almost drive in some states
2. whats your gpu's max temp
3. cpu might be an issue

[steel walrus]

I think they said it was being looked into

[austere tree]

5Tb Western Digital WDC WDS500G2B0A-00SM50 (SATA (SSD)) 30 °C
Wtf actually

[rotund echo]

That thing is the saddest piece of hardware I have ever seen

That 5TB Sata SSD was probably bought 12 years ago with the rest of the build. At the time it probably cost around 600 bucks

12 years ago SSDs were incredibly expensive, dude legit bought it for 600 or more

And next thing you know, it aged like milk. Because while this disk is probably immortal and will most likely keep working for another 10 years no issues, NVME drives have since completly overshadowed SATA SSD because they are incomparably faster and have also become even cheaper than their SATA counterpart for the same capacity

[austere tree]

IMO, the only thing you can do is to overclock NB and memory, but I doubt it will make a significant difference.

[rotund echo]

Overclock a dead ass DDR3 memory

Imo that guy could buy a low end used laptop and still get better performance

[austere tree]

Yeah but FXs are part-time 8 cores, they are highly dependent on memory.

[rotund echo]

800 Mhz btw

[austere tree]

It's 1600 because double data rate

[rotund echo]

But so is the DDR4 3200 modules then

No, wait

Are you sure about what you say ?.

Cause according to wikipedia DDR3 can range from 800 to 2133 Mhz

Maybe that guy got some REAL 800 Mhz here

[austere tree]

Nah, timings are too high

[rotund echo]

Won't change the fact his PC belong in a museum

It had a long and fullfilling life, but now it's about time for it to retire to grandma's so she can have a free machine from his loved grandson

This PC can still perfectly be used for daily use, but for gaming, yea na