Support NeoGAF

[RoadHazard]

This was a big discussion topic before the PS5 was released. As we know, the console runs using variable CPU/GPU frequencies rather than the typical locked frequencies with variable power consumption. This means it can slightly downclock the GPU when there's not enough power to run it at max clock. We were told this would behave exactly the same on every PS5, so you would never see one PS5 performing worse than another due to slight differences in cooling efficiency or power delivery.

However, it seems like different PS5s ARE performing differently. Specifically in God of War Ragnarök, where most performance analyses swear the performance mode is 100% locked to 60fps and never drops a frame. Well, on my PS5 it does drop frames, rather consistently. It's not MUCH, but a frame here and there, mostly noticable as slight hitches or judder when panning the camera. I know what locked 60 looks like, and this is not it. Some parts of the game are a perfect 60, but most are not, and in a few environments it has been very obvious.

You can find some Reddit threads about this issue, but it doesn't seem to be very widespread. So either most people (including DF etc) are just blind to dropped frames, or most PS5s really are running this game better than mine.

So I was thinking about how this could be the case. Obviously it's not my specific copy of the game, so that made me think about the variable frequency. Were we lied to? Do some PS5s actual perform worse than others?

 

[Pelta88]

I picked up the Series X and a PS5 at launch. And for me personally, the real and tangible difference between them, when it comes to performance...

 

[RoboFu]

Are you using a user installed ssd?

 

[RoadHazard]

I picked up the Series X and a PS5 at launch. And for me personally, the real and tangible difference between them, when it comes to performance...

This isn't about PS5 vs XSX, it's about PS5 vs PS5. But yeah, the controller is pretty cool (but less ergonomic in my hands than the DS4 was).

 

[lh032]

maybe your game area has more objects or just some minor software bugs?
tbf, i believe its not possible for DF to retest the same area many times to confirm on the average frame rate, if there is a 1-5 frame difference between ps5s on the same game area, i dont see a big deal on this.

 

[RoadHazard]

Are you using a user installed ssd?

No. And this isn't loading stutter, it happens when just standing still in one place and panning the camera around.

 

[RoadHazard]

maybe your game area has more objects or just some minor software bugs?
tbf, i believe its not possible for DF to retest the same area many times to confirm on the average frame rate, if there is a 1-5 frame difference between ps5s on the same game area, i dont see a big deal on this.

I saw this right from the first area in the game. And it always happens in the same places, worse in some areas than others, which indicates it's a GPU load thing. It's not "random".

 

[lh032]

I saw this right from the first area in the game. And it always happens in the same places, worse in some areas than others, which indicates it's a GPU load thing. It's not "random".

hmm is it a launch model? im wondering maybe the different "parts" in the different version of PS5 did affect the overall performance .

 

[GymWolf]

Framerate are usually never locked on console, even when df try the first 10 hours of the game that are usually not the one with absolute hell on screen.

Their test methodology has always been flawed.

I always had hiccups on any "locked" console game, most people just don't care\notice, especially when you have guys saying that stuff like returnal is locked when the fucking game has gigantic slowdowns every 2x3 with hell on screen.

 

[buenoblue]

I mean without actually measuring the frame rate on your ps5 it's just subjective. I definitely noticed little stutters and drops on my my ps5, especially when jumping over a low wall. But I think for the reviewers and myself it's close enough. I think they are just happy it's not like GOW on PS4 pro where it never hits 60 and your ps4 pro has to be tethered to the ground to stop it taking off .

 

[RoadHazard]

hmm is it a launch model? im wondering maybe the different "parts" in the different version of PS5 did affect the overall performance .

Pretty much. Not first batch, had to wait until March of 2021 I think, but definitely the first model.

 

[RoadHazard]

Framerate are usually never locked on console, even when df try the first 10 hours of the game that are usually not the one with absolute hell on screen.

Their test methodology has always been flawed.

I always had hiccups on any "locked" console game, most people just don't care\notice, especially when you have guys saying that stuff like returnal is locked when the fucking game has gigantic slowdowns every 2x3 with hell on screen.

Yeah, but this doesn't seem to happen when there's a lot of action going on specifically, it's more about what the environment is. There can be absolutely nothing going on at all, and in some environments there will always be slightly stuttering.

And yeah, Returnal is DEFINITELY not locked either. In some biomes it's more like a 50fps game.

I mean without actually measuring the frame rate on your ps5 it's just subjective. I definitely noticed little stutters and drops on my my ps5, especially when jumping over a low wall. But I think for the reviewers and myself it's close enough. I think they are just happy it's not like GOW on PS4 pro where it never hits 60 and your ps4 pro has to be tethered to the ground to stop it taking off .

Yeah, but the people who have actually measured performance are pretty much all saying it never drops a frame. Whereas I can constantly see dropped frames. Not like it's running at 55fps, but a duplicate frame here and there, resulting in visible judder.

 

[Kilau]

I haven’t noticed that in the frame rate mode. The 40fps mode which is unlocked I think when connected to a VRR display isn’t very smooth though.

 

[TrebleShot]

No.

I had this issue too, it’s a handshake thing between the game the TV and the PS5.

When using the performance mode make sure VRR is off in the system menu and HFR is off in the game menu.

It’s nothing to do with the GPU.

 

[tommib]

My PS5 was blessed by king Jim and runs everything flawlessly.

 

[Revolutionary]

I got a brand new PS5 a few weeks ago, and am very sensitive to frame drops (primarily PC gamer)... and I didn't notice a single dropped frame on PS5 Ragnarok.

PS4 Pro on the other hand... I could actually drop it below 30 by using specific Runic attacks lol

 

[GHG]

So we have one single anecdotal example which isn't scientific in nature and this thread is worded in a way which suggests the opposite?

OP you might want to get your PS5 checked or exchanged if you're still in your warranty period.

 

[solidus12]

My PS5 was blessed by king Jim and runs everything flawlessly.

Was your PS3 blessed by Kutaragi-san?

 

[Roxkis_ii]

Man, I must be lucky, GOW:R and Returnal run smoothly on my PS5. But I'm a console gamer, so maybe I'm not not as sensitive to framerate hiccups.

Hope you can get it figured out OP.

 

[tommib]

So we have one single anecdotal example which isn't scientific in nature and this thread is worded in a way which suggests the opposite?

OP you might want to get your PS5 checked or exchanged if you're still in your warranty period.

It could be a thing, boss. Gymwolf says Returnal drops for him a lot. I have 120 hours on that game and the drops I remember were when entering biome 3 from biome 2 probably due to level loading. It’s minor differences but I wouldn’t discard the potential.

 

[GHG]

It could be a thing, boss. Gymwolf says Returnal drops for him a lot. I have 120 hours on that game and the drops I remember were when entering biome 3 from biome 2 probably due to level loading. It’s minor differences but I wouldn’t discard the potential.

How does one come to the the conclusion that framedrops = variable clocks, especially without doing a deep dive?

 

[tommib]

How does one come to the the conclusion that framedrops = variable clocks, especially without doing a deep dive?

 

[RoadHazard]

No.

I had this issue too, it’s a handshake thing between the game the TV and the PS5.

When using the performance mode make sure VRR is off in the system menu and HFR is off in the game menu.

It’s nothing to do with the GPU.

Other games are locked 60fps for me, so no, not an HDMI thing.

VRR and HFR are off (can't even enable it since I don't have a VRR capable TV).

So we have one single anecdotal example which isn't scientific in nature and this thread is worded in a way which suggests the opposite?

OP you might want to get your PS5 checked or exchanged if you're still in your warranty period.

I'm not the only one, there are videos of it. For example (not my video):



I reached this specific area a few days ago, and it was one of the most stuttery for me too.

There's no way they would accept this as a warranty issue just for a few dropped frames. I know how these things work, they would 100% ship it back to me saying nothing is wrong with it. It's not like games are crashing or anything.

 

[RoadHazard]

How does one come to the the conclusion that framedrops = variable clocks, especially without doing a deep dive?

That's just a theory that COULD explain it. It's something we know is different about the PS5 compared to all previous consoles (and the Series).

 

[tommib]

Other games are locked 60fps for me, so no, not an HDMI thing.

VRR and HFR are off (can't even enable it since I don't have a VRR capable TV).



I'm not the only one, there are videos of it. For example (not my video):



I reached this specific area a few days ago, and it was one of the most stuttery for me too.

There's no way they would accept this as a warranty issue just for a few dropped frames. I know how these things work, they would 100% ship it back to me saying nothing is wrong with it. It's not like games are crashing or anything.

That’s hilarious because I see that video and don’t see any dropped frames although I do remember them in dwarf city. The bigger question is, what does it matter?

 

[Physiognomonics]

This was a big discussion topic before the PS5 was released. As we know, the console runs using variable CPU/GPU frequencies rather than the typical locked frequencies with variable power consumption. This means it can slightly downclock the GPU when there's not enough power to run it at max clock. We were told this would behave exactly the same on every PS5, so you would never see one PS5 performing worse than another due to slight differences in cooling efficiency or power delivery.

However, it seems like different PS5s ARE performing differently. Specifically in God of War Ragnarök, where most performance analyses swear the performance mode is 100% locked to 60fps and never drops a frame. Well, on my PS5 it does drop frames, rather consistently. It's not MUCH, but a frame here and there, mostly noticable as slight hitches or judder when panning the camera. I know what locked 60 looks like, and this is not it. Some parts of the game are a perfect 60, but most are not, and in a few environments it has been very obvious.

You can find some Reddit threads about this issue, but it doesn't seem to be very widespread. So either most people (including DF etc) are just blind to dropped frames, or most PS5s really are running this game better than mine.

So I was thinking about how this could be the case. Obviously it's not my specific copy of the game, so that made me think about the variable frequency. Were we lied to? Do some PS5s actual perform worse than others?

What you should do is capture the problem (at 60fps) to see if it's an actual frame drop or an animation problem. It could be caused by an animation problem that I think I actually noticed in this game when panning the camera.

EDIT: I have seen the video you later linked without noticing any framedrops (but didn't check all the long video). You should capture the problem with specific timestamps. I think it's probably an animation hiccup (probably caused by different panning speed between frames) you are mistaking as a frame-drop.

 

[GHG]

That's just a theory that COULD explain it. It's something we know is different about the PS5 compared to all previous consoles (and the Series).

There are a whole host of variables that can lead to framedrops, doesn't make sense to even attempt to draw any conclusions without having the means to investigate beyond what you see.

 

[RoadHazard]

That’s hilarious because I see that video and don’t see any dropped frames although I do remember them in dwarf city. The bigger question is, what does it matter?

I see it clear as day there, don't know what to tell you! But yeah, the Dwarf city was also a place where it was very obvious, whereas in most places it's very slight (or doesn't happen at all).

It matters in the context of the question I'm asking in this thread! Do some PS5s run games worse than others, even though we were told that couldn't happen (barring any actual defects of course)? I paid as much as anyone else for my PS5 (well, except for the suckers who bought theirs from scalpers), I don't want one that performs worse than others.

 

[RoadHazard]

What you should do is capture the problem (at 60fps) to see if it's an actual frame drop or an animation problem. It could be caused by an animation problem that I think I actually noticed in this game when panning the camera.

Yeah, I considered that, but what makes me think that's probably not the case is that in certain environments it happens very consistently while in others it doesn't happen at all. That's what makes me think it's due to GPU load.

 

[Skifi28]

No, it's been made pretty clear how the variable clocks work.

 

[RoadHazard]

No, it's been made pretty clear how the variable clocks work.

Yes. So what's going on then?

 

[Skifi28]

Yes. So what's going on then?

Maybe the game has actual performance issues. Maybe it's just stuttering caused by your drive which might not be perfect, maybe you're just too sensitive. I haven't played the game so I can't comment on how stable the framerate actually is.

 

[Physiognomonics]

Yeah, I considered that, but what makes me think that's probably not the case is that in certain environments it happens very consistently while in others it doesn't happen at all. That's what makes me think it's due to GPU load.

I think you missed my edit. I watched the video you linked without noticing any real framedrops. I think what you see is caused by irregular panning speed (that I actually noticed in that game) that you mistake for framedrops. You must be very sensitive to it.

 

[RoadHazard]

Maybe the game has actual performance issues. Maybe it's just stuttering caused by your drive which might not be perfect, maybe you're just too sensitive. I haven't played the game so I can't comment on how stable the framerate actually is.

Drive? You mean SSD? I doubt that, it's not like streaming stutter at all, it happens when standing completely still with nothing going on, just panning the camera (depending on the environment).

 

[RoadHazard]

I think you missed my edit. I watched the video you linked without noticing any real framedrops. I think what you see is caused by irregular panning speed (that I actually noticed in that game) that you mistake for framedrops. You must be very sensitive to it.

Could be. But again, why would it happen consistently in some environments and not at all in others if that was the case?

 

[Lysandros]

There are plenty of reports of specific users reporting differing performance profiles on X and Y games on X/Y scenes on Series consoles which use "fixed clocks" very famously, do "we" jump to conclusions about Microsoft engineers and say Phil blatantly deceived us about the matter?.. There can be plenty of reasons for these. Sorry but you seem to be reaching a bit far there OP. This thread is a bit surreal.

 

[buenoblue]

Yeah, but the people who have actually measured performance are pretty much all saying it never drops a frame. Whereas I can constantly see dropped frames. Not like it's running at 55fps, but a duplicate frame here and there, resulting in visible judder.

You have to remember that the reviewers get early code, plus multiple patches over time that can change the performance or look of a game. It is what it is. I'd be happy cause sooner or later you know we are going back to 30fps games when cross gen is phased out. (I can't go back, don't make me go back )

 

[Physiognomonics]

Could be. But again, why would it happen consistently in some environments and not at all in others if that was the case?

Maybe aggravated in some environments depending of various reasons. It could be some kind of animation bug that is very hard to even spot (and so debug). Maybe you could describe your problem to Santa Monica Studio to see if they can resolve the problem.

But first you need to timestamp what you see on a 60fps video to be 100% sure it's actually an animation problem and not a framedrop.

 

[M1chl]

If you have dusty fan or heated room, then for sure. But it still will be quite a marginal difference.

 

[Elios83]

It was explained pretty clearly by Cerny that every PS5 works in the exact same way.
The system does not downclock based on operating temperatures.
Every game has a processing load and the respective power consumption is calculated based on the precise instructions the code is executing.
If the total power consumption is estimated to be over a fixed max allocated budget it can downclock. But it's a fully ripetitive process, the power budget is the same, the same code will result in the same power load estimate on every PS5.

So whatever is the issue if there is a issue, must be investigated and compared across different PS5 with some objective measurements. Things like "that reviewer said it's locked 60fps" or "I feel this or that" are not objective metrics.

 

[RoadHazard]

If you have dusty fan or heated room, then for sure. But it still will be quite a marginal difference.

But we've been told that shouldn't make any difference at all. And my room isn't very hot, especially right now.

 

[RoadHazard]

It was explained pretty clearly by Cerny that every PS5 works in the exact same way.
The system does not downclock based on operating temperatures.
Every game has a processing load and the respective power consumption is calculated based on the precise instructions the code is executing.
If the total power consumption is estimated to be over a fixed max allocated budget it can downclock. But it's a fully ripetitive process, the power budget is the same, the same code will result in the same power load estimate on every PS5.

So whatever is the issue if there is a issue, must be investigated and compared across different PS5 with some objective measurements. Things like "that reviewer said it's locked 60fps" or "I feel this or that" are not objective metrics.

Yep, I know that's what we were told.

Unfortunately I only have one PS5, lol.

 

[M1chl]

But we've been told that shouldn't make any difference at all. And my room isn't very hot, especially right now.

More toasty chip = bigger power draw, because semicoductors gets higher resistance (ohm) with higher temperature.

 

[cormack12]

I don't think they are done in real time are they? Aren't they just a profile? Otherwise we would hear our fans kicking in at different times in the same game.

 

[SLB1904]

Other games are locked 60fps for me, so no, not an HDMI thing.

VRR and HFR are off (can't even enable it since I don't have a VRR capable TV).



I'm not the only one, there are videos of it. For example (not my video):



I reached this specific area a few days ago, and it was one of the most stuttery for me too.

There's no way they would accept this as a warranty issue just for a few dropped frames. I know how these things work, they would 100% ship it back to me saying nothing is wrong with it. It's not like games are crashing or anything.

Is that callisto running on a 4090?

 

[RoboFu]

Not all cpus are made the same. That’s usually why each cpu use to have one or two cores deemed the best and those where the main turbo cores. These could be different from one cpu to the next.

You might have just got a shitty draw in the silicon lottery.

 

[Thirty7ven]

I have a launch model DE and noticed hiccups in God of War Ragnarok in just one scene in Midgard but I reloaded the checkpoint and it was perfect again.

 

[Swift_Star]

Other games are locked 60fps for me, so no, not an HDMI thing.

VRR and HFR are off (can't even enable it since I don't have a VRR capable TV).



I'm not the only one, there are videos of it. For example (not my video):



I reached this specific area a few days ago, and it was one of the most stuttery for me too.

There's no way they would accept this as a warranty issue just for a few dropped frames. I know how these things work, they would 100% ship it back to me saying nothing is wrong with it. It's not like games are crashing or anything.

This looks more like a glitch than a hardware problem.
On my console this doesn't happen in neither of the modes. not the 40fps, 60fps, HFR mode.

 

[alabtrosMyster]

I did notice slight drops in 60fps mode in god of war:r when I turned the camera in Midgard specifically.

For the other games, nothing I could notice (I would have noticed irregular frame rate, or frame rate in the 50s) even before I had a VRR TV.

 

[RoadHazard]

More toasty chip = bigger power draw, because semicoductors gets higher resistance (ohm) with higher temperature.

Again, that's NOT how Sony/Cerny have said it works. The frequency throttling/boosting should not be affected by power draw or heat at all, but ONLY by what instructions are used. So it should be 100% deterministic and not depend on how hot your room is or anything. If it does work as you say they lied to us (which is exactly what I'm asking with this thread).

Not all cpus are made the same. That’s usually why each cpu use to have one or two cores deemed the best and those where the main turbo cores. These could be different from one cpu to the next.

You might have just got a shitty draw in the silicon lottery.

Again, I don't think it's supposed to work that way. It doesn't with any other console. Consoles are made to be fixed spec, unlike PCs where you can indeed have the variability you speak of.