• 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.

Digital Foundry: Warhammer 40K: Space Marine 2 - PS5/Xbox Series X|S/PC Tech Review - Is 60FPS Viable on Consoles?

kevboard

Member
Both them seem genuinely perplexed by the frame rate advantage for XSX

it is an atypically large difference.
the only game that I know with a similarly large advantage for Series X is Apex Legends in the 120fps mode. where the PS5 version at first ran at half the resolution on average, and after multiple patches now tries to stay at the same or similar resolutions, at the cost of framerate stability. this leads to the PS5 regularly dropping below 90fps in scenes where the Series X sticks close to 120fps.
 

Xyphie

Member
The only real difference between the XSX/XSS and PS5 CPU beyond negligble clock speed speed differences is the FPU, which has reduced performance in PS5 to save area. So hypothetically if the game really used AVX that could explain the difference beyond just software stuff.

Wild seeing the 7800X3D do 2x the frame rate of a 3600. This is like to most CPU limited game ever.
 

Jesb

Gold Member
This further cements my thoughts that just sticking with GFN is the best route than a console.
 

saintjules

Gold Member
- Current gen only
- PS5/SX have 2 modes, "Speed" and "Quality"
- Series S only has a singular mode

Visuals:
- Speed mode targets 60 FPS with DRS 720p to 1080p which upscales to 4K with FSR2
- Quality mode targets 30 FPS with DRS from 1080p to 1440p which upscales to 4K with FSR2
- Series S 720p to 1080p DRS with FSR2 upscale to 1440p
- Visual quality and features broadly the same across all consoles and modes
- Quality mode has marginal improvement in shadow quality but it's nothing too big
- Performance mode can look 'fuzzy'

- Quality mode's 30 FPS is pretty tightly locked with a single drop here and there and during camera cuts in cut-scenes
- Series S has a few more blips comparatively but is otherwise just as equally well locked to its only 30 FPS mode

Performance:
- PS5 Performance mode targets 60 but has frequent drops to 50's and mid 40's (even in 30's)
- As more enemies come on screen, the performance drops more. When sparsely populated, the game can hit 60 FPS
- One segment with a flame thrower and lots of enemies on screen dips to the mid 30's during combat

- Series X is a marked improvement over PS5 with 10 to 15 FPS higher in same situations
- At worse the flame thrower segment which drops to mid 30's on PS5 drops to mid 40's on SX
- Cut-scenes can run 21% faster on SX compared to PS5
- SX's broader VRR window additionally helps smooth out the drops where PS5 is limited to 48hz (the game does not have native 120hz support)

PC Shader Compilation:
- Alex does not recommend using the TAA option cause it uses FSR 1
- PC version has a robust DRS toggle that helps keep performance on target
- The game is more CPU limited
- Launch version still has shader comp stutters

Curious as I've seen you do this before, why not post all this at the top? Thanks for the summary though of course.
 
Really not much to read into the frame rate advantage as this is an outlier result.

Such an outlier that DF themselves are surprised by it.

As much as people wish there were “patterns” - the only pattern going on the last 4 years is PS5 and Series X being the closest competing consoles of all time when it comes to performance in games.
 
Last edited:

RoboFu

One of the green rats
Oh no...
oh-really-david.gif



fb12M0H.png
 

Gaiff

SBI’s Resident Gaslighter
Yep, this is quite insane but at the same time most games have that 60fps options and options are great, consoles didn't have them before 2016.
In this case we have so, so 60fps with bad IQ vs. stable, properly frame paced 30fps with decent IQ. Not that bad.

Game is clearly CPU limited:

rJ10ABC.jpeg


So DF (and me) wonder why XSX version performs much better. XSX version is much better with higher FPS and much higher VRR range.



It stutters when some specific effects appear on the screen for the first time, so maybe every 20, 30 minutes (probably more at the start)?
What GPU has he got paired with the 7800X3D? I don't think he specifies it, but I assume the 4070 he has shown in the rest of the video.
 

T4keD0wN

Member
What GPU has he got paired with the 7800X3D? I don't think he specifies it, but I assume the 4070 he has shown in the rest of the video.
Hard to say, could be anything, probably 4090, i dont think it matters since he is testing how the CPU performs there
 
This game's probably got one of, if not the, biggest gaps in multi-platform performance I've seen in a long while, since the days of Bayonetta PS3.

Za6eYxG.png
Posting the most cherry-picked screenshot you could have possibly found in the video to make this post is actually hilarious but not surprising.
 
I haven't really noticed any shader stutter in the PC version, but I play at a locked 60 fps, so maybe they resolve within 16.66 ms and will only be visible at higher frame rates. This has happened for other games in the past. I have a beefy 12 core CPU that handles this game's swarms beautifully.
 

adamsapple

Or is it just one of Phil's balls in my throat?
Posting the most cherry-picked screenshot you could have possibly found in the video to make this post is actually hilarious but not surprising.

Er .. Yes ? It's to show the difference in the lows. When nothing's happening on screen, both games are at 60, no difference in situations like that, so kinda pointless to highlight.
 
Last edited:

Rosoboy19

Member
Both them seem genuinely perplexed by the frame rate advantage for XSX
Yeah, it’s weird. I mean, at some point the higher CU count and faster CPU clock in XSX should make a difference, right? Especially if a game is pushing the hardware to the max.
 

intbal

Member
this has to be the biggest framerate advantage we've seen on Series X so far, only Apex Legends' 120fps mode shows a similar difference.
Banishers had a large gap. But DF never tested it and those results were from youtubers with unproven reliability.
 

Topher

Identifies as young
Yeah, it’s weird. I mean, at some point the higher CU count and faster CPU clock in XSX should make a difference, right? Especially if a game is pushing the hardware to the max.

That's been the theory all generation, but rarely has it been a factor. Guess this is just one of those rare times.
 
Er .. Yes ? It's to show the difference in the lows. When nothing's happening on screen, both games are at 60, no difference in situations like that, so kinda pointless to highlight.
The screenshot you posted does not “show the difference in the lows”.

The screenshot you posted is an example of the absolute worst case scenario performance delta that only lasts a brief second before returning to a smaller delta.

The actual average delta of performance - at least according to the footage we have in the video - between the 2 consoles is not represented by the screenshot you posted and you know that, lol.
 
Last edited:

Spyxos

Member
This is insane. We're frequently back to PS360-era resolutions. We're a massive step back from the X1X/PS4 Pro in terms of IQ.
I was just thinking the other day how sharp the games on the ps4 pro and xbox one x were, last few games i played on consoles very extremly blurry.
 

adamsapple

Or is it just one of Phil's balls in my throat?
The screenshot you posted does not “show the difference in the lows”.

The screenshot you posted is an example of the absolute worst case scenario performance delta that only last a brief second before returning to a smaller delta.

The actual average delta of performance between the 2 consoles is not representative of the screenshot you posted and you know that, lol.

There are entire stretches of combat segment(s) in the video that have a 10 or thereabout FPS gap, look at the graph line in this picture.

Like I already said above, when nothing's happening both consoles are at 60, but that's not really stressing either console either so posting pictures of empty spaces with both showing 60 is kinda pointless.

PH4yNh7.png






Banishers had a large gap. But DF never tested it and those results were from youtubers with unproven reliability.


That was a great game. Deserved more love and coverage across the board.
 
Last edited:

TrueLegend

Member
Yep, this is quite insane but at the same time most games have that 60fps options and options are great, consoles didn't have them before 2016.
In this case we have so, so 60fps with bad IQ vs. stable, properly frame paced 30fps with decent IQ. Not that bad.

Game is clearly CPU limited:

rJ10ABC.jpeg


So DF (and me) wonder why XSX version performs much better. XSX version is much better with higher FPS and much higher VRR range.



It stutters when some specific effects appear on the screen for the first time, so maybe every 20, 30 minutes (probably more at the start)?
I don't know what you are on but PS1 had more 60fps games than PS5
 
Last edited:
There are entire stretches of combat segment(s) in the video that have a 10 or thereabout FPS gap, look at the graph line in this picture.

Like I already said above, when nothing's happening both consoles are at 60, but that's not really stressing either console either so posting pictures of empty spaces with both showing 60 is kinda pointless.

PH4yNh7.png
I’m not talking about the times when both consoles, or even 1 console is at 60.

When both consoles are under 60, the average delta of performance is nowhere near the 16fps difference in the screenshot you posted, hence me pointing out the clownishness of cherry-picking that particular screenshot.

It would be like me selectively screenshotting a frame in the video where the consoles are only within 2 or 3fps of each other, and accompanying it with a post saying “not really that much of a difference, nothing to see here”.

That would be equally farther from the truth as what you posted.
 
Last edited:

intbal

Member
The strange thing is that the cutscene is also showing a large performance gap. There should be a significant reduction in CPU load for a cutscene, since it's a fixed scene and the game isn't calculating any enemy AI or world physics.
Perhaps the PS5 code has a bug.
 
The only real difference between the XSX/XSS and PS5 CPU beyond negligble clock speed speed differences is the FPU, which has reduced performance in PS5 to save area. So hypothetically if the game really used AVX that could explain the difference beyond just software stuff.

Wild seeing the 7800X3D do 2x the frame rate of a 3600. This is like to most CPU limited game ever.
Focus Home, which is publisher there games are more PC side (DX Pro) just like CD Project, which can be reason PS5 having issues because Focus Home does not know how to optimize the API. Same happened with CD Red with The Witcher 3 and Cyberpunk both ran worse on PS4.
 
Last edited:

adamsapple

Or is it just one of Phil's balls in my throat?
It would be like me selectively screenshotting a frame in the video where the consoles are only within 2 or 3fps of each other, and accompanying it with a post saying “not really that much of a difference, nothing to see here”.

You're welcome to do that. Not sure why you're getting so worked up over a screenshot clearly showing the worst case scenario.


The strange thing is that the cutscene is also showing a large performance gap. There should be a significant reduction in CPU load for a cutscene, since it's a fixed scene and the game isn't calculating any enemy AI or world physics.
Perhaps the PS5 code has a bug.

The game needing to render at the low resolutions that it does, with DRS, probably means consoles have GPU restrictions to work around too.
 
It's very rarely the case where both SX and PS5 are within a 1-2 frames of each other in "speed" mode in gameplay. Most of the time, it's 8-10 frames, going as high a ~16 fps difference.
Still think something else is going on here besides just being CPU limited
 
You're welcome to do that. Not sure why you're getting so worked up over a screenshot clearly showing the worst case scenario.
Not worked up at all actually.

This game's probably got one of, if not the, biggest gaps in multi-platform performance I've seen in a long while, since the days of Bayonetta PS3.

Za6eYxG.png


Awkward Jay Z GIF by Complex
This is just a disingenuous post and I decided to point it out, that’s all.

Pretty sure we’ve seen at least a few games in the last decade with an ~7-10fps average difference in performance. Highly doubt you’ll need to go back to the PS3 days to find an outlier like this.
 
Seriously People are underestimating DX12 advantage, MS on 2015 highlighted that optimizing game on PC will help developers to easily port game on Xbox as well as both API are same.,

Maybe the developers mainly optimized the game for PC just like Black Myth Wong (not releasing maybe due to Xbox Series S or other reason) and had issues on PS5 same like this game.
 
Last edited:

Bojji

Member
The only real difference between the XSX/XSS and PS5 CPU beyond negligble clock speed speed differences is the FPU, which has reduced performance in PS5 to save area. So hypothetically if the game really used AVX that could explain the difference beyond just software stuff.

Wild seeing the 7800X3D do 2x the frame rate of a 3600. This is like to most CPU limited game ever.

I like your way of thinking but this would make 3600 also perform much better than PS5.

For CPU limited games, try Hogwarts Legacy or Jedi Survivor with RT, this game is not on the same level haha.

What GPU has he got paired with the 7800X3D? I don't think he specifies it, but I assume the 4070 he has shown in the rest of the video.

Probably 4090, just testing CPUs.

I don't know what you are on but PS1 had more 60fps games than PS5
PS1? Maybe PS2, and I don’t think that’s true either. Majority of PS5 games have a 60fps mode.

Yep, PS5 has like ~90% of games with performance options, most of them hit that target. PS2 had probably around ~50% or less and PS1? 10%?
 

saintjules

Gold Member
Since you'll probably get a PS5 Pro, it was probably the right call right?

I am thinking to get the PS5 version myself but I don't think we should expect all devs to start making updates for the Pro just yet.

If what was said about the mandate for games to be patched to play on the Pro to be true on or after Sept. 16, then this game is released a week prior to that. Maybe they don't optimize it further for the Pro. I'm sure they would consider it but the Series X working a bit better here makes me want to try it there.
 
I bought for PC. But, having said that, this being the largest gap that I've seen in a direct head-to-head, kind of funny that no one is like "Probably a bug/further optimization needed" which I know why none of that was mentioned, it's so that Alex and Co. can feel validated about their stance on the XSX in the run up to the launch. Ladies and gents, VELOCITY ARCHITECTURE HAS ENTERED THE CHAT.
 
Top Bottom