• Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.

The PlayStation 5 and Xbox Series X/S versions of Squadrons have a critical performance difference.

Status
Not open for further replies.

Vasto

Member
Duplicate: https://www.neogaf.com/threads/star...-on-xbox-series-x-s-with-120fps-modes.1575227

According to EA, Squadrons will have two different graphics settings on Xbox Series S and Series X, akin to the performance modes found in something like Team Ninja's Nioh franchise. One, aimed at better visual quality, will improve the game's lighting model and pushes the game up to 1440p at 60 fps on the Xbox Series S and up to 4K at 60 fps on the Xbox Series X. Meanwhile, the second graphics setting that's geared toward better performance gets Squadrons up to 1440p at 120 fps on Series S and up to 4K at 120 fps on Series X.

Conversely, Squadrons on PlayStation 5 will run at the same resolution and frame rate as the PlayStation 4 Pro version, except with improved lighting. It's been reported that the aerial dogfighter hits 1440p and 1800p at 60 FPS on the more-powerful PlayStation 4, so that's about what you can expect here.

This means Star Wars: Squadrons on Xbox Series S / Series X is capable of things that aren't possible in the PS5 version: 4K and 120fps. That's a stark contrast, and it's unclear why this is the case or if a future update might change things.

 
Last edited by a moderator:

StreetsofBeige

Gold Member
So PS5 is basically running at the same res and fps as PS4 Pro or Series S. Ugh.

Someone in the link's comment section said it might have to do with PSVR BC.
 
Last edited:

FlyLo

Member
It is completely bullshit and frankly yes this is lazy on their part. You’re telling me this cannot run massively better on the PS5?!!!

Bullllshit.
lol for sure. But w/e, its some mid size studio making right? Not one of the big AAA ones.
 

Quantum253

Member
I find it tough to believe that the PS5 is incapable of running this game any better than a PS4 Pro.
I would too, unless it's just a straight port from pro. It could be unkown sjnce Sony is still seems quiet about backwards compatible performances or boosts.
 

Smoke6

Member
No. Stop. This is because they're not bothering to patch the ps5 version for some reason.

Hopefully gamers speak with their wallets and don't support EA, scum of the industry.
Yeah like someone said earlier it’s could be the PSVR that’s the cause as going higher frame and res might cause issues with the headset maybe
 
No, it's a pretty big margin actually.

"Sony has opted for a custom AMD RDNA 2-based GPU inside the PS5, which provides 10.28 teraflops of power with 36 compute units running at 2.23GHz each. Microsoft has picked a custom AMD RDNA 2-based GPU for the Xbox Series X, but it can hit 12 teraflops of power with 52 compute units at 1.825GHz each"

1.72tflops is large in comparison to this gen (I believe the OG PS4 is ~1.7)... but this is a smaller margin of power by % (unless my tired math is off) than the Pro vs. the One X.
 
D

Deleted member 801069

Unconfirmed Member
tenor.gif
 

SoraNoKuni

Member
Oh no, preorder cancelled.

Seriously now, it's just ea being lazy or some weird bug with the bc.
Also first time I hear this game.
 

Neo_game

Member
This must to be one of those rare title that his hitting both the resolution and fps targets for both SS and SX.
 
Let's break this up for you guys, because it seems many can't read, nor is the 'journalist" any good at breaking down information:

Modes with improved lighting model:
  • Series S: up to 1440p at 60 fps
  • Series X: up to 4K at 60 fps
  • PS5: Reported to hit 1440p and 1800p at 60 FPS on PS4 PRO

120 FPS mode (basic lighting model):
  • Series S: up to 1440p at 120 fps
  • Series X: up to 4K at 120 fps
  • PS5: n/a

The wording is very manipulative, on one side he mentions the top resolutions, on the other he mentions both the upper and lower limit (as something "reported"). I'd say just from the writing style that the person is manipulative and maybe even deceptive.

Anyway, 4K vs 1800p is about what we should expect in terms of resolution difference when the PS5 and series X when they have their GPUs maxed out. Series S seems to pull off some pretty good numbers all things considered.
 
Last edited:

DForce

NaughtyDog Defense Force
There are people who actually think this is this result of the PS5 not being able to handle higher res and framerate. Lol
 

rnlval

Member
Given that their internal AMD architectures are similar (with power tipping to the Series X, but not by a *huge* margin) - this seems odd.
APU size

PS5 = ~309 mm2

XSX = 360.45 mm2 which includes 76 MB SRAM on-chip data storage. Remove CPU's 4MB L2 and 8MB L3 cache from APU's 76 MB SRAM, it results in 64 MB SRAM for the GPU. Remove GPU's 5MB L2 cache from 64MB results in 59 MB SRAM.

I could use RDNA NAVI 10's CU model to work out XSX GPU's extra SRAM storage.
 
APU size

PS5 = ~309 mm2

XSX = 360.45 mm2 which includes 76 MB SRAM on-chip data storage. Remove CPU's 4MB L2 and 8MB L3 cache from APU's 76 MB SRAM, it results in 64 MB SRAM for the GPU. Remove GPU's 5MB L2 cache from 64MB results in 59 MB SRAM.

I could use RDNA NAVI 10's CU model to work out XSX GPU's extra SRAM storage.

So you’re saying that difference in memory infrastructure and design can impact performance in such a way? That’s shocking, if that’s the reason here. This seems like some sort of software or developer thing.
 

Spokker

Member
I'm pretty sure this is because XSS and XSX are running an auto-improved XboxOne version and PS5 is running PS4-pro profile (like the GitHub leak suggested)
This isn't it. The PS5 version will have improved lighting according to the article. That can't be done if it's running it in the PS4 Pro profile.
 

rnlval

Member
"Sony has opted for a custom AMD RDNA 2-based GPU inside the PS5, which provides 10.28 teraflops of power with 36 compute units running at 2.23GHz each. Microsoft has picked a custom AMD RDNA 2-based GPU for the Xbox Series X, but it can hit 12 teraflops of power with 52 compute units at 1.825GHz each"

1.72tflops is large in comparison to this gen (I believe the OG PS4 is ~1.7)... but this is a smaller margin of power by % (unless my tired math is off) than the Pro vs. the One X.
TFLOPS is meaningless without memory bandwidth bound factors i.e. external memory bandwidth and on-chip cache size.
 

Romulus

Member
According to EA, Squadrons will have two different graphics settings on Xbox Series S and Series X, akin to the performance modes found in something like Team Ninja's Nioh franchise. One, aimed at better visual quality, will improve the game's lighting model and pushes the game up to 1440p at 60 fps on the Xbox Series S and up to 4K at 60 fps on the Xbox Series X. Meanwhile, the second graphics setting that's geared toward better performance gets Squadrons up to 1440p at 120 fps on Series S and up to 4K at 120 fps on Series X.

Conversely, Squadrons on PlayStation 5 will run at the same resolution and frame rate as the PlayStation 4 Pro version, except with improved lighting. It's been reported that the aerial dogfighter hits 1440p and 1800p at 60 FPS on the more-powerful PlayStation 4, so that's about what you can expect here.

This means Star Wars: Squadrons on Xbox Series S / Series X is capable of things that aren't possible in the PS5 version: 4K and 120fps. That's a stark contrast, and it's unclear why this is the case or if a future update might change things.



You're trolling. Theres no way you think a game that looks average and runs at 1440p 60fps on pro but can't do 4k 60fps on ps5. Easily.
 
Last edited:
APU size

PS5 = ~309 mm2

XSX = 360.45 mm2 which includes 76 MB SRAM on-chip data storage. Remove CPU's 4MB L2 and 8MB L3 cache from APU's 76 MB SRAM, it results in 64 MB SRAM for the GPU. Remove GPU's 5MB L2 cache from 64MB results in 59 MB SRAM.

I could use RDNA NAVI 10's CU model to work out XSX GPU's extra SRAM storage.

It was 76MB of SRAM on the entire APU.

There are SRAM arrays everywhere, even in the PHYs for the memory controllers. You cannot thus infer how much SRAM is cache on the GPU from this number. The number itself (76MB) is pretty meaningless.
 

rnlval

Member
So you’re saying that difference in memory infrastructure and design can impact performance in such a way? That’s shocking, if that’s the reason here. This seems like some sort of software or developer thing.
Both companies are not fully transparent with RDNA 2 request changes. MS's transparency is better than Sony's, but there's are more unknowns.

We do know APU die sizes for both XSX and PS5.
 

rnlval

Member
It was 76MB of SRAM on the entire APU.

There are SRAM arrays everywhere, even in the PHYs for the memory controllers. You cannot thus infer how much SRAM is cache on the GPU from this number. The number itself (76MB) is pretty meaningless.
It's pretty easy to remove CPU's SRAM from the total 76 MB SRAM.

XSX GPU's 5 MB L2 cache and known SRAM (e.g. L0, LDS, L1, vector, and scalar register files, crossbar cache) locations from NAVI 10 model can be removed from the remaining 64 MB SRAM.
 
Last edited:
Status
Not open for further replies.
Top Bottom