• 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.

[DF] Elden Ring Beta: PS5/ Xbox Series X/S First Look - Quality vs Performance Modes Tested

O

onQ123

Member
SlimySnake said:
After a full year of ps5 winning several face offs, they need to stop being baffled. It was tools at launch, baffling today or perhaps they were wrong to question cerny and insinuate ps5 was 9 tflops.

We discussed this and it seems the combination of the ps5 io, the higher clocks and the Xsx sticking more CUs in shader arrays than normal is probably what’s contributing to ps5 keeping up with the 18% faster xsx gpu.

When xsx does win, it wins big like in hitman, metro and doom so maybe it is due to different engines preferring higher CUs vs higher clocks.
Click to expand...

I tried to give people a heads up but I'm crazy

www.neogaf.com

40 CU 9.75 TFLOPs RX 5700 XT vs 64 CU 12.6 TFLOPs RX Vega 64 ( Different architectures but probably the best way to compare PS5 & Xbox SX for now)

Why RX 5700 XT vs RX Vega 64 & not RX 590 vs RX Vega 64? My reason for this is because RX 590 is ROP limited with 32 ROPs while RX 5700 XT has the same 64 ROPs as the RX Vega 64. RX 5700 XT is a newer architecture & benefit from that but at the same time I'm comparing it to the RX Vega 64 &...
www.neogaf.com www.neogaf.com
 
LordOfChaos

LordOfChaos

Member
Wish we could hear more from the developer about some of these reasons. I.e is the PS5 being twice as fast at loading because they're starting to take of a lot of the hardware offload that a lot of games haven't touched much yet, or is it just generally less time spent optimizing the SX?

We'll probably never know, unless later updates change things, but I'm a curious boy.
 
Last edited:
Lysandros

Lysandros

Member
SlimySnake said:
Nvidia actually said that the ps3 gpu had more tflops than the 360. And the cell had 7-8 cpu cores/threads/whatever they were calling it while the xenon processor had just 3.

In reality The 250 gflops ps3 gpu was more like a 170 gflops gpu compared to the 240 gflops 360 gpu and the 360 cpu was able to do just as many calculations as the cell. I think someone ran a ballerina demo and found they were both 100 gflops.

Its been a while so my specs might be a bit off but Both Sony and nvidia had touted the cell and the rsx as more powerful while the 360 guys remained silent with big dick confidence knowing the final games would inevitably show the difference.
Click to expand...
No. This is the benchmark you are talking about, i remember it well:

mvQZSsq.jpg


The figures are for the number of dancers. As one can clearly see PS3 CELL massively outperforms X360 CPU by x3 (!) to the point that makers of the benchmark remarked (word for word) "Cell is a beast." That makes sense since it was significantly more powerful compared X360 CPU (due to SPUs mostly) having twice the computatinal capabilities (200 GFlops vs 100 GFlops). RSX was rated at 192 GFlops. Thus combined CELL+RSX exceeded X360. Now, it was a very difficult architecture to program for comparatively. And the overall picture as to game performance was much more complicated than that of course.

Edit: Again, those were vastly different architectures as to CPU, GPU and RAM not even remotely comparable to today's picture.

Edit II: Much more importantly than CELL's rated max theoretical compute figure you could potentially achieve very high saturation (up to 100%!) on the SPUs as explaned by Cerny in the DF interview about PS5, thus very high efficiency in 'real world' compute throughput. The very reason for the architectural choice of the Tempest Engine.
 
Last edited:
Plantoid

Plantoid

Member
SlimySnake said:
There are like 4 TVs with vrr and two of them are from Sony whose implementation sucks dick. The rest are expensive oled panels. Surely that’s not an alternative.

Most 60 FPS games in previous gens had drops but there is a difference between 1-2 dropped frames vs consistently dropping ten frames. All the fighting, racing and sports games I played manage this by making sacrifices. Action adventure games don’t want to do that and it shows.
Click to expand...
Sorry, you're wrong, even cheap LCD nanocell TV's from LG have vrr.
 
StateofMajora

StateofMajora

Banned
Epic Sax CEO said:
Contrary to my own predictions last year, it's talking longer for the SeX to show it's compute and CPU advantage.
Definitively not a premature ejaculator, I just hope than that when it cums it'll be worth the best tantric sex.



Demons remake surely looks better, but also runs at a much lower resolution.
Click to expand...
Nope demons souls is native 4k with 30fps and I wonder if eldin ring is using checkerboard rendering.
 
LordOfChaos

LordOfChaos

Member
Lysandros said:
No. This is the benchmark you are talking about, i remember it well:

mvQZSsq.jpg


The figures are for the number of dancers. As one can clearly see PS3 CELL massively outperforms X360 CPU by x3 (!) to the point that makers of the benchmark remarked (word for word) "Cell is a beast." That makes sense since it was significantly more powerful compared X360 CPU (due to SPUs mostly) having twice the computatinal capabilities (200 GFlops vs 100 GFlops). RSX was rated at 192 GFlops. Thus combined CELL+RSX exceeded X360. Now, it was a very difficult architecture to program for comparatively. And the overall picture as to game performance was much more complicated than that of course.

Edit: Again, those were vastly different architectures as to CPU, GPU and RAM not even remotely comparable to today's picture.
Click to expand...

This is a test of SIMD performance, for which the Cell was absolutely a beast for the time it launched. But games don't just run straight line SIMD on them all day of course. Where it suffered was branchy paths that weren't easily vectorizable (i.e you'd see framerate dips compared to 360 in areas that used a lot of CPU despite on paper the Cell boasting more flops, sometimes it was devs not putting in the work, sometimes it wasn't, but developers don't have unlimited time and budget)

It was definitely an interesting novel direction. I do wish we got to see what the Cell could do when paired with a GPU more like Xenos, where it wasn't just making up for RSX's shortfalls by the end when a number of studios were using it well. And especially with unified memory, maybe with an eDRAM to feed the lowered bandwidth.
 
Last edited:
M

MrA

Banned
SlimySnake said:
After a full year of ps5 winning several face offs, they need to stop being baffled. It was tools at launch, baffling today or perhaps they were wrong to question cerny and insinuate ps5 was 9 tflops.

We discussed this and it seems the combination of the ps5 io, the higher clocks and the Xsx sticking more CUs in shader arrays than normal is probably what’s contributing to ps5 keeping up with the 18% faster xsx gpu.

When xsx does win, it wins big like in hitman, metro and doom so maybe it is due to different engines preferring higher CUs vs higher clocks.
Click to expand...
it's going to depend on how the game is using the gpu, more parallel tasks, more cus= win, less parallel tasks, faster clocks= win
 
Lysandros

Lysandros

Member
SlimySnake said:
After a full year of ps5 winning several face offs, they need to stop being baffled. It was tools at launch, baffling today or perhaps they were wrong to question cerny and insinuate ps5 was 9 tflops.

We discussed this and it seems the combination of the ps5 io, the higher clocks and the Xsx sticking more CUs in shader arrays than normal is probably what’s contributing to ps5 keeping up with the 18% faster xsx gpu.

When xsx does win, it wins big like in hitman, metro and doom so maybe it is due to different engines preferring higher CUs vs higher clocks.
Click to expand...
XSX GPU is only 'faster' at 'some' metrics like floating point calculations or number of ray cast due to having more CUs essentially. It isn't faster 'overall', quite the contrary, every single component of its GPU is slower by 18% compared to PS5. This of course impacts GPU throughputs tied to fixed functions units and increase latency between the components.

Edit: I would also consider Ghostrunner and this one as pretty big technical 'wins' for PS5, among other very interesting cases like the Tourist.
 
Last edited:
rickybooby87

rickybooby87

Member
Interesting set of results for sure. I’d be happy with a locked 4K mode at 40FPS on 120hz displays (like insomniac’s rift apart).

The fact that the PS5 could load the game twice in the same amount of time it takes the XSX to load it once is impressive.

the biggest glaring issue for me, is the pop in. That’s gotta be fixed asap.

overall, good showing on PS5, especially considering the narrative in the year run up before launch about expected performance deltas between the two systems.
 
Snake29

Snake29

RSI Employee of the Year
rickybooby87 said:
Interesting set of results for sure. I’d be happy with a locked 4K mode at 40FPS on 120hz displays (like insomniac’s rift apart).

The fact that the PS5 could load the game twice in the same amount of time it takes the XSX to load it once is impressive.

the biggest glaring issue for me, is the pop in. That’s gotta be fixed asap.

overall, good showing on PS5, especially considering the narrative in the year run up before launch about expected performance deltas between the two systems.
Click to expand...

 
Anchovie123

Anchovie123

Member
Theres still a lot of time until release so im hopeful that they will iron out these performance problems come release but if they dont thats seriously embarrassingly pathetic on Froms part. I will play the PS4 Pro version if not fixed come launch. I also seriously doubt the RT mode will be 60 so who even gives a shit about that.
 
Snake29

Snake29

RSI Employee of the Year
SlimySnake said:
After a full year of ps5 winning several face offs, they need to stop being baffled. It was tools at launch, baffling today or perhaps they were wrong to question cerny and insinuate ps5 was 9 tflops.

We discussed this and it seems the combination of the ps5 io, the higher clocks and the Xsx sticking more CUs in shader arrays than normal is probably what’s contributing to ps5 keeping up with the 18% faster xsx gpu.

When xsx does win, it wins big like in hitman, metro and doom so maybe it is due to different engines preferring higher CUs vs higher clocks.
Click to expand...

Even when XSX has "big" wins, this console still has a lot off issues with performance like in Hitman (dropping to 40s). Not seeing a big win with Hitman in the end when the performance was nearly locked on the PS5 at 1800p, while the native 4K version had stutters and more frame drops. That only means that the XSX couldn't really handle the native 4K together with a better performance. The game should've been dropped to 1800p, same as PS5.

I even bet the next Hitman game will suddenly be the same on both consoles, that's why i think that with Hitman...it was purely a engine/optimisation thing than a hardware limitation.
 
Last edited:
lh032

lh032

I cry about Xbox and hate PlayStation.
Epic Sax CEO said:
Contrary to my own predictions last year, it's talking longer for the SeX to show it's compute and CPU advantage.
Definitively not a premature ejaculator, I just hope than that when it cums it'll be worth the best tantric sex.



Demons remake surely looks better, but also runs at a much lower resolution.
Click to expand...
demon souls looks cripsed, i dont gaf about resolution.
 
Last edited:
Banjo64

Banjo64

cumsessed
skit_data said:
Looks like PS5 is and will be the ”Souls”-machine going forward. Fine with me, its one of the devs/games that matter to me.
Click to expand...
Never in doubt considering DeS and Bloodborne and the likely hood of future From exclusives. PS has always been my go to for From games, like having them all on one profile.
 
Shmunter

Shmunter

Member
SSfox said:
Bluepoint releases games for just one console, FS for hundreds consoles. There's a difference.

This is why me and other wise gentlemen like Bryan75 are worried about Sony focusing on other than making games for PS5 and ONLY PS5, if you see what i mean.
Click to expand...
The only saving grace would be Sony studios not being burdened by cross platform development with a pure PS5 focus. A porting studio can then deal with bringing it across to other platforms. A recent purchase brings hope.
 
Shmunter

Shmunter

Member
X not looking good, but then it’s only a fraction of a fraction of the install base with S being the dominant within the Xbox family. Why optimise for an outlier.

Of course it could be that PS5 is just the most powerful next gen hardware. 🤷‍♂️
 
Last edited:
O

onQ123

Member
Shmunter said:
X not looking good, but then it’s only a fraction of a fraction of the install base with S being the dominant within the Xbox family. Why optimise for an outlier.

Of course it could be that PS5 is just the most powerful next gen hardware.
Click to expand...
Don't do this lol ,they already been running with the Xbox Series X getting less optimization narrative.
 
Last edited:
Sosokrates

Sosokrates

Report me if I continue to console war
TonyK said:
Why? What it makes a AAA game it's not the quality but the investment. Avengers is a AAA game, and Sekiro, in comparison, must be a AA game. But, obviously, quality is not related to the number of A.

I can add something it would sound even worse, Nintendo is a AA developer for me.

If Nintendo or From are AAA developers, then we need to add an extra A to Rockstar, Ubisoft, Guerrila, Naughty, etc...
Click to expand...

Maybe
Indie - $0-20million
AA 20-60million
AAA 60-150million
AAA 150+million
 
Sosokrates

Sosokrates

Report me if I continue to console war
Shmunter said:
X not looking good, but then it’s only a fraction of a fraction of the install base with S being the dominant within the Xbox family. Why optimise for an outlier.

Of course it could be that PS5 is just the most powerful next gen hardware. 🤷‍♂️
Click to expand...

Lol, if you think the X does not looking good then you must also think the PS5 is not looking good.

Its like the X is a burnt steak and the ps5 version is slightly less burnt steak.

And, why optimise the outlier?

Well why optimise the PS5 version when there loads more PS4's?

So silly...
 
O

onQ123

Member
Sosokrates said:
Lol, if you think the X does not looking good then you must also think the PS5 is not looking good.

Its like the X is a burnt steak and the ps5 version is slightly less burnt steak.

And, why optimise the outlier?

Well why optimise the PS5 version when there loads more PS4's?

So silly...
Click to expand...
He was joking and just making fun of the 'Xbox didn't get as much optimization' crowd
 
Sosokrates

Sosokrates

Report me if I continue to console war
onQ123 said:
Don't do this lol ,they already been running with the Xbox Series X getting less optimization narrative.
Click to expand...

What makes u so sure its a narritive? we dont know whats causing the inferior overall performance.
 
Last edited:
You must log in or register to reply here.

Similar threads

Digital Foundry: Stellar Blade Demo - PS5 First Impressions - Every Mode Tested!
2
62
4K
poodaddy replied
[Digital Foundry] Avatar: Frontiers of Pandora - PS5/Xbox Series X|S Tech Review - Graceful Scaling vs Maxed PC
Trailer 2 3
101
11K
JackMcGunns replied
VGTech: Assassin's Creed Mirage PS5 vs Xbox Series X|S Frame Rate Comparison
35
5K
SHA replied
Teardown - Stunning Physics & Destruction - PS5/Xbox Series X/S - DF Tech Review
Review
20
3K
touchfuzzy replied
IGN: Dragon's Dogma 2 Performance Review - PS5 vs Xbox Series X|S vs PC
2 3 4
158
13K
March Climber replied
Top Bottom