• 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.

GhostRunner PS5/XSX/XSS DF Performance Analysis

Darsxx82

Member
The Series X version most certainly didn't outperform the PS5 version during their initial analysis at all. The photo mode comparison was, in a ways, a cool experiment, but that isn't gameplay. The Series X version would even slow down and hitch during the appearance of on-screen text, a crucial design element of Control. It was almost certainly in unplayable condition with respect to the PS5 version, and anybody who watches that video can see that. So a single system update from Microsoft, with no update from the dev, instantly changed that. That is why it got the attention it did. Such a thing has never happened in modern console gaming history. The update had something to do with the Series X virtual machine that runs with the OS.
The XSX version showed better performance in the most demanding moments and in the DF video you have an example and they show it themselves.

You even have NXG in their review stating that the XSX version is slightly more stable in the most stressful moments.

Photo mode is simple proof that XSX was performing better and to what extent, which is something you couldn't know with locked framerate.

The subsequent improvement (tools?) is relevant in that it eliminated the only reason that could make the PS5 version preferable and, of course, annihilate the theory of some according to which the stutter problem in XSX was due to the memory or slower SSD.
 

SlimySnake

Flashless at the Golden Globes
Seems like an easy 4k ultra over 60fps on my 3060ti
While the xsx is equivalent to a 2080 in standard rasterization, it is only equivalent to a 2060 super in rt due. Your 3060 ti should be way better than just 4K 60 FPS seeing as how the xsx is 4K 50 in rt mode.

It's interesting, that's for sure. Would love for developers to explain what are causing the aggressive performance dips on Series X, especially in 120 hz modes. This is becoming a recurring issue in native current gen games. First, it was DMC5 SE, then one of these 3 - Dirt 5, Destiny 2 or R6 Siege, and now this game.

Also, why isn't there an Xbox One comparison? They only did PS4 for last gen?
This is something DF needs to ask Microsoft. Microsoft did have to admit this was a problem back at launch but It’s been a year and DF needs to call up MS again. The tools have been fixed. Xsx Games are consistently outperforming the ps5 but then we get games like this. Maybe this is on the devs but both consoles are using rdna 2 GPUs, it should just run better. You shouldn’t need to code to the metal to get better performance out of Xbox.

Do we know if the engine simply favors the ps APIs? Did the ps4 pro version of out perform the x1x?

I really hope DF uses these games as an opportunity to dig deep into a discussion about whether or not these consoles with the same rdna 2.0 architectures are really substantially different because otherwise results like this no matter how rare simply don’t make much sense. A 10.6 tflops 6600xt should never outperform a 13 tflops 6700xt. Pc gamers would riot.
 
While the xsx is equivalent to a 2080 in standard rasterization, it is only equivalent to a 2060 super in rt due. Your 3060 ti should be way better than just 4K 60 FPS seeing as how the xsx is 4K 50 in rt mode.


This is something DF needs to ask Microsoft. Microsoft did have to admit this was a problem back at launch but It’s been a year and DF needs to call up MS again. The tools have been fixed. Xsx Games are consistently outperforming the ps5 but then we get games like this. Maybe this is on the devs but both consoles are using rdna 2 GPUs, it should just run better. You shouldn’t need to code to the metal to get better performance out of Xbox.

Do we know if the engine simply favors the ps APIs? Did the ps4 pro version of out perform the x1x?

I really hope DF uses these games as an opportunity to dig deep into a discussion about whether or not these consoles with the same rdna 2.0 architectures are really substantially different because otherwise results like this no matter how rare simply don’t make much sense. A 10.6 tflops 6600xt should never outperform a 13 tflops 6700xt. Pc gamers would riot.
No clue what settings they used. Very clearly not ultra.
 

onQ123

Member
Something to note is that this is one of the few games that's not built to run well on PS4 it's below 1080p & the frame rate drops but PS5 is 4K 60fps with Ray Tracing.

It's clear that something about the PS5 is allowing the game to run a lot better & it's not that they are giving PlayStation special treatment.
 

01011001

Banned
It's clear that something about the PS5 is allowing the game to run a lot better & it's not that they are giving PlayStation special treatment.

all that is clear is that this game is a mess, and if it runs good on one platform they clearly gave it more attention. anyone who played this on PC (which I am one of them) knows how absolutely dogshit this game runs in some levels.
there are literally parts in this game where every single time you move the mouse, your framerate basically drops by 50%... but ONLY if you move your mouse/turn the camera... not if you walk backwards, forwards, side to side or jump... only if you turn the camera.

when the game first launched on PS4 pro, it ran like ass and the resolution was below 900p

then the RT implementation, not only does it look like shit, it also tanks performance on PC like crazy... fucking Cyberpunk 2077 with RT turned to max runs better than Ghostrunner
 
Last edited:

01011001

Banned
Seems like an easy 4k ultra over 60fps on my 3060ti

turn on RT and you will drop to 10fps my dude. the RT implementation of this game on PC is notoriously bad. I played through it on a 3060ti and at 1440p I had to play with DLSS Balanced mode to hit 60fps with RT on.

the issue being, on PC you have no control over the RT settings outside of off and on. if you turn it on it will run fill resolution RT reflections (that still manage to look like complete ass) and it will run with RT AO, Shadows and GI

the console versions only run with RT reflections and at lower than native resolution as well. an option like this should be in the PC version tbh.
 
Last edited:
It's clear that something about the PS5 is allowing the game to run a lot better & it's not that they are giving PlayStation special treatment.
How do you know that the PS5 version wasn't given special treatment? The devs took the time to implement distinctive dual sense rumble. It was mentioned in the DF review. It wouldn't be surprising at all that they simply took more time with this particular version of the game and optimized it in ways they clearly did not on Xbox AND PC.
 

onQ123

Member
all that is clear is that this game is a mess, and if it runs good on one platform they clearly gave it more attention. anyone who played this on PC (which I am one of them) knows how absolutely dogshit this game runs in some levels.
there are literally parts in this game where every single time you move the mouse, your framerate basically drops by 50%... but ONLY if you move your mouse/turn the camera... not if you walk backwards, forwards, side to side or jump... only if you turn the camera.

when the game first launched on PS4 pro, it ran like ass and the resolution was below 900p

then the RT implementation, not only does it look like shit, it also tanks performance on PC like crazy... fucking Cyberpunk 2077 with RT turned to max runs better than Ghostrunner
And what make the PS5 so special that these devs would go out of their way to make it run better than other platforms?
 

01011001

Banned
And what make the PS5 so special that these devs would go out of their way to make it run better than other platforms?

is this a serious question?... it's clearly the system that sells the most 3rd party games right now... how else could you explain that every other version until now was a complete mess at launch and some versions still are to this day, and only the PS5 version seems to be fine?
 
Last edited:

onQ123

Member
How do you know that the PS5 version wasn't given special treatment? The devs took the time to implement distinctive dual sense rumble. It was mentioned in the DF review. It wouldn't be surprising at all that they simply took more time with this particular version of the game and optimized it in ways they clearly did not on Xbox AND PC.
DualSense rumble is simple to support
 

onQ123

Member
is this a serious question?... it's clearly the system that sells the most 3rd party games right now... how else could you explain that every other version until now was a complete mess at launch and some versions still are to this day, and only the PS5 version seems to be fine?
The PS5 version isn't fine it has a audio bug .

So these bad devs who couldn't get it right after all this time on other platforms all of a sudden figured out how to get extra performance out of the PS5?
 

01011001

Banned
The PS5 version isn't fine it has a audio bug .

So these bad devs who couldn't get it right after all this time on other platforms all of a sudden figured out how to get extra performance out of the PS5?

so every version sucks, some suck in one way, other suck in other ways.

explain how the PC version has levels where the framerate gets cut in half every time you turn the camera, literally going from 180fps down to ~90fps for me with raytracing off, and from ~70fps down to ~40fps with RT on... how do you explain that? you can't. and using a game like that to infer anything about any system on the market is simply stupidity... nothing more nothing less. it is pure stupidity.

this game is so unpolished that they didn't even properly flag every surface for RT reflections. there are objects in this game that have RT reflections in one level, but only have screen space reflections a level later... I am specifically talking about the glass on the hacking terminals here (that look exactly the same in every level), but there are also other objects that for no good reason only have screen space reflections instead of raytracing reflections.

then in the cyberspace levels if you turn on RT reflections almost all of the ground becomes non-reflective, I said almost because some parts of teh ground will have RT reflections... but with RT off the ground will be full of screen space reflections, if this is intended behaviour then wtf... if not then also WTF?

this game is a complete mess on a technical level. it is a different kind of mess depending on where you play, but it's still a mess.
 
Last edited:

onQ123

Member
so every version sucks, some suck in one way, other suck in other ways.

explain how the PC version has levels where the framerate gets cut in half every time you turn the camera, literally going from 180fps down to ~90fps for me with raytracing off, and from ~70fps down to ~40fps with RT on... how do you explain that? you can't. and using a game like that to infer anything about any system on the market is simply stupidity... nothing more nothing less. it is pure stupidity.

this game is so unpolished that they didn't even properly flag every surface for RT reflections. there are objects in this game that have RT reflections in one level, but only have screen space reflections a level later... I am specifically talking about the glass on the hacking terminals here (that look exactly the same in every level), but there are also other objects that for no good reason only have screen space reflections instead of raytracing reflections.

then in the cyberspace levels if you turn on RT reflections almost all of the ground becomes non-reflective, I said almost because some parts of teh ground will have RT reflections... but with RT off the ground will be full of screen space reflections, if this is intended behaviour then wtf... if not then also WTF?

this game is a complete mess on a technical level. it is a different kind of mess depending on where you play, but it's still a mess.
Actually I can explain it the game is running into a bottleneck on your system
 

Sosokrates

Report me if I continue to console war
While the xsx is equivalent to a 2080 in standard rasterization, it is only equivalent to a 2060 super in rt due. Your 3060 ti should be way better than just 4K 60 FPS seeing as how the xsx is 4K 50 in rt mode.


This is something DF needs to ask Microsoft. Microsoft did have to admit this was a problem back at launch but It’s been a year and DF needs to call up MS again. The tools have been fixed. Xsx Games are consistently outperforming the ps5 but then we get games like this. Maybe this is on the devs but both consoles are using rdna 2 GPUs, it should just run better. You shouldn’t need to code to the metal to get better performance out of Xbox.

Do we know if the engine simply favors the ps APIs? Did the ps4 pro version of out perform the x1x?

I really hope DF uses these games as an opportunity to dig deep into a discussion about whether or not these consoles with the same rdna 2.0 architectures are really substantially different because otherwise results like this no matter how rare simply don’t make much sense. A 10.6 tflops 6600xt should never outperform a 13 tflops 6700xt. Pc gamers would riot.

I think its just a case that more work is needed for the xbox series versions.
The xbox series is a more complecated platform then the PS5. Theres two different platform (the s + x) for xbox and the RAMs are configured differently. We also dont know when work was started for this game, pehaps work started later on xbox.

So imo I think its definitely a case of software and logistics. Devs have said that the PS environment is easier to work with.
 
Last edited:

sendit

Member
is this a serious question?... it's clearly the system that sells the most 3rd party games right now... how else could you explain that every other version until now was a complete mess at launch and some versions still are to this day, and only the PS5 version seems to be fine?

This game was released on PC first using DirectX libraries and api’s. You know what console also uses DirectX?
 
do you mind GIF by DJ Khaled
Missed opportunity to put Xbox or PS logo on the front of that cigar or whatever the hell he's smoking lol
 
You need to rewatch the video and what DF says ... DF is clear, between 1-2 fps in very limited and specific situations and being 0 fps in the rest of the game. And NO, it is not a 1-2 fps improvement over game at 30fps. It is 1-2 fps, I repeat, "in specific and limited situations" in unlocked framerate up to 60fps. The fact is that its detection was only possible thanks to the photo mode where the framerate is unlocked (the same for DMC). Otherwise it would not have been detected on RT Mode.
Yes, it is. From DF analysis. That's a 6% improvement in that scene. Not bad from a firmware update. this is not during photo mode (DMC5). The fact that it's the highest improvement in the absolute minimum framerate is also very interesting.

image0.png
 
Last edited:

DenchDeckard

Moderated wildly
Isnt this exactly what was predicted. Some games are going to flip flop. I think to get a true gauge we need newly released games with dedicated teams to each platform. Not these last gen port jobs, by small indie studios. Look at the mess that was mortal shell on ps5, that was fixed. I’m sure this will in a patch or two.

that said, mirroring what I have said earlier. The ps5 is the absolutely best place to play this game currently. It wipes the floor with the series version.
 
Last edited:

Shmunter

Member
I don´t get the difference between the 60fps-Mode and the RT-mode on PS5. Both run @4K with 60fps, one has RT, any benefits in the other mode?

Why would you not choose the RT mode, if there are no benefits in the other mode?
I noticed a bit more controller latency on the RT mode. But my tv is 30+ Ms in game mode to start with, unlikely to be noticeable on more modern lower latency sets.
 
Last edited:

Darsxx82

Member
Yes, it is. From DF analysis. That's a 6% improvement in that scene. Not bad from a firmware update. this is not during photo mode (DMC5). The fact that it's the highest improvement in the absolute minimum framerate is also very interesting.

image0.png
No matter how hard you try, the reality is that NO.
It's funny because, despite the fact that there is a video, you insist and put captures of those "very specific and limited situations" to take what they say and reality out of context.
I repeat, the reality is that there is NO 1-2 fps enhancement on a sustained basis. 99% of the gameplay (both in Control and in DMC) remains unchanged with the same fps as before the firmware update .
 

Md Ray

Member
turn on RT and you will drop to 10fps my dude. the RT implementation of this game on PC is notoriously bad. I played through it on a 3060ti and at 1440p I had to play with DLSS Balanced mode to hit 60fps with RT on.

the issue being, on PC you have no control over the RT settings outside of off and on. if you turn it on it will run fill resolution RT reflections (that still manage to look like complete ass) and it will run with RT AO, Shadows and GI

the console versions only run with RT reflections and at lower than native resolution as well. an option like this should be in the PC version tbh.
On PC, you can use the Universal Unreal Engine 4 Unlocker tool to unlock the command console for all UE4 games where you can adjust each individual RT and other settings more granularly.

Here's Alex's vid on it (timestamped):


How do you know that the PS5 version wasn't given special treatment? The devs took the time to implement distinctive dual sense rumble. It was mentioned in the DF review. It wouldn't be surprising at all that they simply took more time with this particular version of the game and optimized it in ways they clearly did not on Xbox AND PC.
Really, have you tested the PC version? See this:
Must be their PCs.

Alex covered the PC version a while ago using RTX 2080 Ti, 2060, RX 6800 XT and reported no such issues except for stutters which he suspects were due to shader compilation which frequently occurs in DX12 titles. I also tried the game myself on Steam and it was pretty good at utilizing the 3070 to its maximum judging by Nsight and its power consumption.




Timestamped:
 
Last edited:
No matter how hard you try, the reality is that NO.
It's funny because, despite the fact that there is a video, you insist and put captures of those "very specific and limited situations" to take what they say and reality out of context.
I repeat, the reality is that there is NO 1-2 fps enhancement on a sustained basis. 99% of the gameplay (both in Control and in DMC) remains unchanged with the same fps as before the firmware update .
Of course this is specific to modes and scenes, I never claimed the contrary. You claimed there wasn't any case with 1-2fps improvement at 30fps ish. I just showed you there was.
 

DenchDeckard

Moderated wildly
Game performing better on PS5 = “Xbox version is unoptimized… dev should not have time to work on it”.

Game performing better on Xbox = “Cleary Xbox is the strong console in the world”

Let see when people will realize PS5 has several advantages over Xbox that allow that new trend with better performance on PS5 happens.

bro, when mortal shell was broken on ps5 you went to twitter and confronted the devs. Can you show the same energy here?
it’s pretty obvious there is a glaring issue with the series versions. It’s dropping to like 40s to 50s are you saying the ps5 is like 20 to 30 percent more powerful than a series x? We factually know that isn’t the case. And if someone tries to tell me the magic of the io is doing some outer worldly shit on this small indie game that loads levels in UE4, then I know that some people are genuinely lacking Of any common sense.

there’s obviously an issue with the Xbox versions. It will probably be patched. this game isnt worth the fanboy madness.

just scratch it as a win for ps5 and move on lol. This isn’t evidence of anything.
 
Last edited:
And what make the PS5 so special that these devs would go out of their way to make it run better than other platforms?
It takes more effort to port a game from PC to PS5 than from PC to Xbox. The latter is basically a single mouse click (if you're lazy).

The game runs like shit on both XSX and XSS. The devs clearly didn't put any effort into the port. The game was/is buggy on PC too, with random performance drops even on high end PCs.
 

Lysandros

Member
I wonder if an agresive DRS solution or even a straight 1800P would fix those heavy dips in RT/60 FPS mode on XSX. Some tweaking also seem to be needed for the 120 FPS mode.
 
Last edited:

Md Ray

Member
The game runs like shit on both XSX and XSS. The devs clearly didn't put any effort into the port. The game was/is buggy on PC too, with random performance drops even on high end PCs.
Do you have any evidence of this? Because DF didn't talk about any "random perf drops" in their PC coverage.
 
Last edited:

Darsxx82

Member
Of course this is specific to modes and scenes, I never claimed the contrary. You claimed there wasn't any case with 1-2fps improvement at 30fps ish. I just showed you there was.
Up to 60fps not 30fps. And you have gone to put the only situation found as an example of what happens regulary when it is not so. I repeat, the improvement is never 1-2 fps extra throughout the game, only in very limited and specific situations that barely remain in time being 99% of the rest of the game 0fps improvement. That is, this is not a case of a 30fps game with moments at 28-29 fps that now runs at 30fps all times.
 
How come I or Alex didn't encounter this?
Sample size of 2 isn't indicative of much

Edit
Also it's very clear Xbox didn't receive proper optimisation, and PlayStation did. This isn't time or effort, it's that this twitch play fast paced game which basically needs 60fps isn't a lock on Xbox and it is on ps5. I don't care if the Xbox version runs at half res and low settings to get locked 60, but absolutely nothing was done to get a locked fps. The 2 versions have what looks like exact same settings, but one runs like shit. That's a lazy job done no matter how you slice it.
 
Last edited:

onQ123

Member
It takes more effort to port a game from PC to PS5 than from PC to Xbox. The latter is basically a single mouse click (if you're lazy).

The game runs like shit on both XSX and XSS. The devs clearly didn't put any effort into the port. The game was/is buggy on PC too, with random performance drops even on high end PCs.
Maybe the people who have problems on their PC'S have the same bottleneck as the Xbox Series X which isn't a problem on PS5. Every PC is not the same.
 

onQ123

Member
It takes more effort to port a game from PC to PS5 than from PC to Xbox. The latter is basically a single mouse click (if you're lazy).

The game runs like shit on both XSX and XSS. The devs clearly didn't put any effort into the port. The game was/is buggy on PC too, with random performance drops even on high end PCs.
By the way if it only take a mouse click to port games to Xbox Series X that would mean that devs have more free time for optimization so you're killing your own excuse.
 
Discussing the video that started the thread that's it's about the performance differences on a game is somehow discouraged by the mods, they even deleted posts commenting on the performance issues with one console, but changing subjects to question a firmware update is somehow allowed.


Nice to see the new narrative being spread around here this times.
Sony's machine have an inherent advantage because it takes more effort to port to it meaning that games are more optimized to run on it.
Poor Microsoft...
 

Lysandros

Member
UE4 + 60 fps on PS4 is bound to run like shit. 😂

It should NOT run like shit on Series X though, the system has both the CPU and GPU grunt to run that fine.

Anyhow I've noticed that UE is one of those Engines than often favor PS5 for some reason I don't know. This is not the only UE4 game that runs better on Sony's machine. Crash It's About Time, for instance, performs better and sometimes at a higher rez on PS5 too.

On the other hand other engines like the RE Engine usually perform better on Series X (DMC V, RE Village...).

I guess it's true that some engines favor narrow and fast and some favor wide and slow design. But even taking that into account I still think Series X is underperforming in this game. It's just too large of a difference, way larger that what we saw in Crash.
I was curious if F1 2021 was also an UE4 title, PS5 version having ~10% higher FPS at same resolution/settings in replays. But no apparently, it uses the proprietary EGO Engine 4.0. I would like to add that DMC 5 was 'very close' with the 120 FPS mode performing better on PS5 in gameplay. With the new firmware update performance might be virtually identical in RT mode, since the margin was very small to begin with, around ~4%. Yes, in the end different design philosophies favoring different engines to varying degrees.
 

Snake29

RSI Employee of the Year
Game performing better on PS5 = “Xbox version is unoptimized… dev should not have time to work on it”.

Game performing better on Xbox = “We see a trend

Let see when people will realize PS5 has several advantages over Xbox that allow that new trend with better performance on PS5 happens.

Sorry fixed that for you.

And for those that are still in denail. Thse devs created this game for PC which is using DirectX, so i'm not sure what kind of dumb excuses you guys using.

The Xbox is not some magical console that can do everything better, and that was clear from launch.
 
Last edited:

OmegaSupreme

advanced basic bitch
It takes more effort to port a game from PC to PS5 than from PC to Xbox. The latter is basically a single mouse click (if you're lazy).

The game runs like shit on both XSX and XSS. The devs clearly didn't put any effort into the port. The game was/is buggy on PC too, with random performance drops even on high end PCs.
Give us a source on any of this. Otherwise it's pure speculation. Is the ps5 cpu not able to go higher? That could account for framerate differences in some titles that are more cpu dependent.
 
Last edited:
Top Bottom