By using this site, you agree to our Privacy Policy and our Terms of Use. Close

Forums - Nintendo - How Will be Switch 2 Performance Wise?

 

Switch 2 is out! How you classify?

Terribly outdated! 3 5.26%
 
Outdated 1 1.75%
 
Slightly outdated 14 24.56%
 
On point 31 54.39%
 
High tech! 7 12.28%
 
A mixed bag 1 1.75%
 
Total:57
sc94597 said:
HoloDust said:

That's the thing with custom ports for custom hardware - if you do some napkin math, RTX 3060 is to Switch 2 (both Ampere) what something like RX 6600XT is to Steam Deck (both RDNA2) - but that is only if you let GPU inside Steam Deck run at its boosted clock (1600MHz) - like CP2077 does. And if you look up Outlaws benchmarks for those two GPUs, 3060 is just slightly above 6600XT.

I don't know at what GPU frequency that guy is running Steam Deck at when testing Outlaws, not that it matters much. Outlaws on Switch 2 (just like in case of CP2077) has lot of custom tweaks and cutbacks that help it run the way it runs, and those tweaks are not available on PC, even with lowest settings, so direct comparison is not really possible.

I find Outlaws to be best showcase of what Switch 2 can do, so far. But I'm not convinced that custom port of Outlaws would not perform fairly well on Deck (though not look as good, due to lack of DLSS/FSR4), given how both architectures behave in it when compared in their beefier iterations, instead of that stuttery mess you have in that video (is that even running of NVMe?).

I am not sure if we could use the RTX 3060 : SW2 RX 6600XT : Steam Deck relationship to understand much, given that RTGI could be a bottleneck at lower compute/render-load for RDNA2 in a way it isn't for Ampere, but not much of a bottleneck when running the game at 1080p Ultra on these beefier twins because other aspects of the game's render pipeline might predominate after getting past that initial "minimum requirement" for the most basic RTGI. 

The fact that the game doesn't scale much in terms of performance with different input resolution hints that the bottleneck for Steam Deck isn't necessarily pure general render-load and cutting out some geometry probably would only marginally improve performance on Steam Deck. Especially when many of these cutbacks seem to be to improve a CPU-bottleneck on SW2 more than a GPU-bottleneck (entire assets are removed, object culling is changed, lower NPC density, etc) and Steam Deck has a significantly more powerful CPU than SW2 that shouldn't be too much of a bottleneck in the first place. I'd suspect a "Steam Deck" optimized version would instead cut into the RT pretty heavily, given that - that is the only modification that seems to drastically improve performance (unless one considers adding FG to "improve performance.") Just porting a SW2-esque version that mostly is cutting to save CPU resources and scaling by resolution to Steam Deck probably won't make it playable in the same way reducing render resolution doesn't already. 

And of course adding new lighting solutions is more difficult than cutting assets, from a development perspective. I do agree though, that it is hard to compare like-to-like here. But that is true of every PC vs. Console comparison and really isn't that important when talking about the end results for a given price -- which are what actually matter. Every platform, including Switch 2 (as HW-accelerated RT wasn't fully taken advantage of, as an example) could benefit from more optimization and more development resources. 

Not sure I follow that bit about scaling with different input resolutions...but as I said, I'd love to see at what clock GPU runs in Deck in Outlaws, if it's at its maximum allowed (like CP2077 does) or not.

Anyway, Ubisoft games tend to really like AMD's GPUs - Outlaws is no different. While this SW2 port might be custom solution that is best suited to run on this specific nVidia low powered SoC, I'm fairly certain that, if Deck was actual handheld console and given the same treatment the way Switch 2's port has been, we would see Outlaws running on it without problems that running off the shelf PC version has.

But yeah, as just PC in the handheld form, actual 1 to 1 comparisons is not really possible, and as always you need beefier PC to match optimized console, and that of course means higher price.



Around the Network
HoloDust said:

Not sure I follow that bit about scaling with different input resolutions...but as I said, I'd love to see at what clock GPU runs in Deck in Outlaws, if it's at its maximum allowed (like CP2077 does) or not.

Anyway, Ubisoft games tend to really like AMD's GPUs - Outlaws is no different. While this SW2 port might be custom solution that is best suited to run on this specific nVidia low powered SoC, I'm fairly certain that, if Deck was actual handheld console and given the same treatment the way Switch 2's port has been, we would see Outlaws running on it without problems that running off the shelf PC version has.

But yeah, as just PC in the handheld form, actual 1 to 1 comparisons is not really possible, and as always you need beefier PC to match optimized console, and that of course means higher price.

The game seems to run at a variable 18-30 fps with significant stuttering (very frequent drops to 12-15 fps) regardless if you use FSR 3 ultra performance (internally ~260p) or performance mode (internally ~ 400p.)* Ultra performance is a bit more stable, but you'd expect a lot more of a performance gain from reducing the resolution to 45% of what it was in performance mode. Something else is limiting its performance on Steam Deck. And given what happens when you remove RT, in addition to the fact that the Series S and Switch 2 versions have reduced and/or absent RT effects in certain scenes compared to PS5/Series X/PC. That seems to be the culprit here. 

Some have argued it is Proton, but we see similar results with RDNA2 (and to a lesser extent - RDNA 3) APUs on Windows. 

Having said that, I do think the game could run on Steam Deck if it were a dedicated platform (or even if they just made an optimized Steam Deck setting like with The Witcher 3 remaster or Cyberpunk 2077), but that likely would require heavy reworking of the lighting system even more so than what was done on Switch 2 and Series S. Or more extreme compromises overall. And in that same vein of thinking, the SW2 version could run even better than it does, maybe with higher quality RT, if they utilized the RT cores more than they do in Outlaws' implementation of RTGI (which seems to use the CPU and non-RT GPU cores for a lot of what the RT cores would do in an NVIDIA-targeted game.) There is rarely enough optimization, even on closed-platforms, to fully utilize the hardware, and SW2's hardware certainly isn't being fully utilized by a port of a game that targeted PS5/Series X in its first development phase. 

*Meanwhile the Switch 2 version is internally running at 540p in handheld-mode and 720p in docked-mode according to Digital Foundry and is locked at 30fps. So it isn't just DLSS making up for the image-quality difference here, but the actual internal resolutions are higher too. 



FF7 Remake director says the game will run at "2K" (so 1440p?) on Switch 2:

“Regarding DLSS, we are using it this time. As I mentioned earlier, we wanted to provide the same experience for the user both in docked mode and handheld mode. So TV mode stably renders in 2K resolution. In handheld mode the hardware performance is slightly reduced. As a result, rendering at 2k resolution becomes difficult, so we lower it by using DLSS upscaling to output at 2k. So we can achieve 2k output even in handheld and docked mode, and the user doesn’t experience much loss.”

“What determines the graphic quality of the characters is the lighting processing, but if we change the formula of the lighting, it would probably affect the impression that the characters give to the users. So in relation to the lighting of the characters, we kept it the same as on other consoles. We changed things like fog, and post-effects. The lowest priority elements, rather than the lighting, have all been rewritten and optimized for the Switch 2. So visually, in terms of the user experience, it’s something users might not even notice. Even with the Switch 2 specs, it allows you to reach 30 frames stably.”

https://www.gonintendo.com/contents/53356-final-fantasy-vii-remake-series-director-details-switch-2-version-s-specs-defends



sc94597 said:
HoloDust said:

Not sure I follow that bit about scaling with different input resolutions...but as I said, I'd love to see at what clock GPU runs in Deck in Outlaws, if it's at its maximum allowed (like CP2077 does) or not.

Anyway, Ubisoft games tend to really like AMD's GPUs - Outlaws is no different. While this SW2 port might be custom solution that is best suited to run on this specific nVidia low powered SoC, I'm fairly certain that, if Deck was actual handheld console and given the same treatment the way Switch 2's port has been, we would see Outlaws running on it without problems that running off the shelf PC version has.

But yeah, as just PC in the handheld form, actual 1 to 1 comparisons is not really possible, and as always you need beefier PC to match optimized console, and that of course means higher price.

The game seems to run at a variable 18-30 fps with significant stuttering (very frequent drops to 12-15 fps) regardless if you use FSR 3 ultra performance (internally ~260p) or performance mode (internally ~ 400p.)* Ultra performance is a bit more stable, but you'd expect a lot more of a performance gain from reducing the resolution to 45% of what it was in performance mode. Something else is limiting its performance on Steam Deck. And given what happens when you remove RT, in addition to the fact that the Series S and Switch 2 versions have reduced and/or absent RT effects in certain scenes compared to PS5/Series X/PC. That seems to be the culprit here. 

Some have argued it is Proton, but we see similar results with RDNA2 (and to a lesser extent - RDNA 3) APUs on Windows. 

Having said that, I do think the game could run on Steam Deck if it were a dedicated platform (or even if they just made an optimized Steam Deck setting like with The Witcher 3 remaster or Cyberpunk 2077), but that likely would require heavy reworking of the lighting system even more so than what was done on Switch 2 and Series S. Or more extreme compromises overall. And in that same vein of thinking, the SW2 version could run even better than it does, maybe with higher quality RT, if they utilized the RT cores more than they do in Outlaws' implementation of RTGI (which seems to use the CPU and non-RT GPU cores for a lot of what the RT cores would do in an NVIDIA-targeted game.) There is rarely enough optimization, even on closed-platforms, to fully utilize the hardware, and SW2's hardware certainly isn't being fully utilized by a port of a game that targeted PS5/Series X in its first development phase. 

*Meanwhile the Switch 2 version is internally running at 540p in handheld-mode and 720p in docked-mode according to Digital Foundry and is locked at 30fps. So it isn't just DLSS making up for the image-quality difference here, but the actual internal resolutions are higher too. 

Yeah...and that was initial premise of response to your "...having said that, a lot of 9th Generation feature heavy titles run poorer on it than Switch 2" - that if they were dedicated ports for Steam Deck, using its strengths, instead of just running general purpose PC version, things would look different...well, just like with any console, really.
(I was just thinking yesterday "really doubt that RX 6700 (PS5 equivalent) will run Ghost of Yotei 4K native, as it runs on PS5" 



HoloDust said:

Yeah...and that was initial premise of response to your "...having said that, a lot of 9th Generation feature heavy titles run poorer on it than Switch 2" - that if they were dedicated ports for Steam Deck, using its strengths, instead of just running general purpose PC version, things would look different...well, just like with any console, really.
(I was just thinking yesterday "really doubt that RX 6700 (PS5 equivalent) will run Ghost of Yotei 4K native, as it runs on PS5" 

Of course, if you downgrade the game by removing much of the 9th Generation features (replacing RT with baked lighting or SSR) then you can probably run the game at a locked 30fps on Steam Deck like SW2 and Series S. The whole point of my original response was that the SW2 is handling the 9th Gen features better than Steam Deck so far. The fact that the SW2 version has missing assets and reduced asset density compared to the PC version, seemingly to save on CPU resources, doesn't contradict this. 

Isn't Ghost of Yotei on PS5 going to run at 30fps @ 4k and 60fps @ a dynamic 1080p - 1440p? That doesn't sound impossible for an RX 6700, especially if not maxing the PC settings (which most console versions aren't equivalent of.) Most PC gamers just don't play games at 4k 30fps.

Last edited by sc94597 - on 28 September 2025

Around the Network
sc94597 said:
HoloDust said:

Yeah...and that was initial premise of response to your "...having said that, a lot of 9th Generation feature heavy titles run poorer on it than Switch 2" - that if they were dedicated ports for Steam Deck, using its strengths, instead of just running general purpose PC version, things would look different...well, just like with any console, really.
(I was just thinking yesterday "really doubt that RX 6700 (PS5 equivalent) will run Ghost of Yotei 4K native, as it runs on PS5" 

Of course, if you downgrade the game by removing much of the 9th Generation features (replacing RT with baked lighting or SSR) then you can probably run the game at a locked 30fps on Steam Deck like SW2 and Series S. The whole point of my original response was that the SW2 is handling the 9th Gen features better than Steam Deck so far. The fact that the SW2 version has missing assets and reduced asset density compared to the PC version, seemingly to save on CPU resources, doesn't contradict this. 

Isn't Ghost of Yotei on PS5 going to run at 30fps @ 4k and 60fps @ a dynamic 1080p - 1440p? That doesn't sound impossible for an RX 6700, especially if not maxing the PC settings (which most console versions aren't equivalent of.) Most PC gamers just don't play games at 4k 30fps.

I don't know, not convinced that sort of rework would be needed - Deck is same architecture as XS/PS5, and game already uses hybrid approach for some things. What I think that's lacking is fine tuned port, not Deck's lack of capabilities per se.

As for Yotei, I guess we'll have to wait and see - I think we're finally starting to see games that are, due to fixed hardware and APIs that can go "deeper" than DX12, pushing more than what equivalent PC GPUs can with DX12.



HoloDust said:

I don't know, not convinced that sort of rework would be needed - Deck is same architecture as XS/PS5, and game already uses hybrid approach for some things. What I think that's lacking is fine tuned port, not Deck's lack of capabilities per se.

As for Yotei, I guess we'll have to wait and see - I think we're finally starting to see games that are, due to fixed hardware and APIs that can go "deeper" than DX12, pushing more than what equivalent PC GPUs can with DX12.

One of the reasons why I think there probably does need to be reworking (rather than just minor optimization) is the fact that an RX 6400 runs the game similar to a Series S despite being on paper somewhat weaker than a Series S. Basically runs the game at low settings with an internal 720p resolution at 30fps. So the PC version isn't that poorly optimized when theoretically worse hardware is achieving the same as the console version (which has additional optimizations.) 

An RX 6400 is about 90% of a Series S in theoretical performance. The Steam Deck is about 45% of an RX 6400 in terms of theoretical performance.

Ostensibly you could make that difference by scaling internal resolution to something like 45% of 720p, but we don't see that. Even at internally 260p the game isn't running at 30fps. Scaling isn't linear on the same architecture here.

So something in the render pipeline is bottlenecking performance on the Steam Deck. RT seems to be the natural answer of what that might be. The Steam Deck's CPU certainly isn't the issue because CPU utilization remains pretty low when playing the game and it has a decent CPU.



sc94597 said:
HoloDust said:

I don't know, not convinced that sort of rework would be needed - Deck is same architecture as XS/PS5, and game already uses hybrid approach for some things. What I think that's lacking is fine tuned port, not Deck's lack of capabilities per se.

As for Yotei, I guess we'll have to wait and see - I think we're finally starting to see games that are, due to fixed hardware and APIs that can go "deeper" than DX12, pushing more than what equivalent PC GPUs can with DX12.

One of the reasons why I think there probably does need to be reworking (rather than just minor optimization) is the fact that an RX 6400 runs the game similar to a Series S despite being on paper somewhat weaker than a Series S. Basically runs the game at low settings with an internal 720p resolution at 30fps. So the PC version isn't that poorly optimized when theoretically worse hardware is achieving the same as the console version (which has additional optimizations.) 

An RX 6400 is about 90% of a Series S in theoretical performance. The Steam Deck is about 45% of an RX 6400 in terms of theoretical performance.

Ostensibly you could make that difference by scaling internal resolution to something like 45% of 720p, but we don't see that. Even at internally 260p the game isn't running at 30fps. Scaling isn't linear on the same architecture here.

So something in the render pipeline is bottlenecking performance on the Steam Deck. RT seems to be the natural answer of what that might be. The Steam Deck's CPU certainly isn't the issue because CPU utilization remains pretty low when playing the game and it has a decent CPU.

Maybe...but I guess we'll probably never know, since I doubt Ubi will decide to throw money down the drain and make custom Deck port.

It would be interesting to see what's the bottleneck, since, as I said, Ubi games tend to like AMD GPUs, apart from the fact that Deck, compared to Switch 2 in handheld mode, is not weak per se.



I know I'm likely beating a dead horse here (and I'm sure this will get a thread eventually) but do we think a hypothetical Switch 2 Pro could functionally match a PS5/Series X in docked mode even if not matching them in raw power? Switch 2 is closer to the eye in performance to Series S and Xbox One X than some expected. And it outpaces PS4 Pro in a lot of ways.
Provided it makes economic sense with tariffs and stuff among other things, I wonder if reaching close to PS5 performance with a Pro in 2029-2030 is feasible.



Lifetime Sales Predictions 

Switch: 161 million (was 73 million, then 96 million, then 113 million, then 125 million, then 144 million, then 151 million, then 156 million)

PS5: 122 million (was 105 million, then 115 million) Xbox Series X/S: 38 million (was 60 million, then 67 million, then 57 million. then 48 million. then 40 million)

Switch 2: 120 million (was 116 million)

PS4: 120 mil (was 100 then 130 million, then 122 million) Xbox One: 51 mil (was 50 then 55 mil)

3DS: 75.5 mil (was 73, then 77 million)

"Let go your earthly tether, enter the void, empty and become wind." - Guru Laghima

Wman1996 said:

I know I'm likely beating a dead horse here (and I'm sure this will get a thread eventually) but do we think a hypothetical Switch 2 Pro could functionally match a PS5/Series X in docked mode even if not matching them in raw power? Switch 2 is closer to the eye in performance to Series S and Xbox One X than some expected. And it outpaces PS4 Pro in a lot of ways.
Provided it makes economic sense with tariffs and stuff among other things, I wonder if reaching close to PS5 performance with a Pro in 2029-2030 is feasible.

Nah even in 2029-2030 hitting PS5 performance levels in a handheld form factor will be tricky, and Nintendo won't push for high end tech given how expensive it is nowadays.

In addition to the thermal/spatial/wattage limitations of mobile tech, Pro will also be "held back" by the base model.