• Hey everyone, staff have documented a list of banned content and subject matter that we feel are not consistent with site values, and don't make sense to host discussion of on Famiboards. This list (and the relevant reasoning per item) is viewable here.

StarTopic Future Nintendo Hardware & Technology Speculation & Discussion |ST| (Read the staff posts before commenting!)

Ehh. N64, Wii U, 3DS fall in that category I suppose. But SNES, GBA, DS make three as well, so seems pretty even.
Yeah, that's the thing, right? GBA and DS were both follow-ups to extremely successful generations, and both did extremely well. I don't think the curse is real, realistically.
 
In almost every scenario where Nintendo has had to follow up a major successful hardware generation, They screw it up.

Maybe this excruciating wait may be the key to go back to back. Let them cook


main-qimg-90a289d0075204939bcd4508bc370ba9
So my gut tells me that Furukawa, who comes from the marketing department, is more rational and cautious than Iwata at this point, and even though he doesn't know anything about gaming, he should have a good understanding of what it takes to get the attention of gamers when a console is released.
 
In almost every scenario where Nintendo has had to follow up a major successful hardware generation, They screw it up.

Maybe this excruciating wait may be the key to go back to back. Let them cook


main-qimg-90a289d0075204939bcd4508bc370ba9

Nintendo has been on top, and they’ve hit rock bottom, so they have a lot of experience on what not to do.
 
So my gut tells me that Furukawa, who comes from the marketing department, is more rational and cautious than Iwata at this point, and even though he doesn't know anything about gaming, he should have a good understanding of what it takes to get the attention of gamers when a console is released.
I think the most interesting aspect of Furukawa is probably that he worked at Nintendo of Europe as a accountant and rose the ranks.

Plus I think he and the staff at Nintendo has done an incredible job at maintaining engagement for the Switch.

But his main hardship will probably be the transitional period, but his mention of a smooth transition gives me hope that he took noticed from his competitor, like Sony with PS4-PS5

Edit: he also likes golf story, which is a fun tidbit.
 
Last edited:
Neither, because df also mentioned that Nintendo custom hardware is not something that can be simply applied to x86 hardware measurements, and predictably looking at just paper floating point data switch2 will only be at the level of ps4 and ps4+, however that conclusion ignores too many things.
Like the switch, the device's performance is that of a 7th gen console (x360), yet all of its first party game resolutions and rendering techniques actually surpass the x360 and are on the same standard as 8th gen in terms of rendering techniques.

These are not entirely different camps, there is a significant ( ~ 2X) performance gulf between handheld and TV mode. Half of "nearly a Series S" is "basically a PS4" in terms of FLOP performance. Both are very likely true; one of these describes TV Mode, the other Handheld and Tabletop.

DLSS is its own can of worms where you can say silly things like, it has an "effective" performance above PS5 because in scenes with PS5 rendering 4K natively, NG Switch can use "only" 720p and upscale to 4K, using a third the performance to render one ninth the pixels, making it appear even stronger. That's nonsense in a whole bunch of ways, but like, you can say silly things like that based on real things, so we have an idea of a system "cheating" its way into trading blows.

Sorry I should have mentioned I was looking for "Docked Performance" expectations only. I was just trying to gauge if you guys were thinking 4tf in docked was viable or a lower base number closer too Steam Deck.
 
Yeah, that's the thing, right? GBA and DS were both follow-ups to extremely successful generations, and both did extremely well. I don't think the curse is real, realistically.
Yeah, doomer bias is real. Handhelds generally succeed after a success compared to home consoles so
 
Sorry I should have mentioned I was looking for "Docked Performance" expectations only. I was just trying to gauge if you guys were thinking 4tf in docked was viable or a lower base number closer too Steam Deck.
I think oldpuck's analysis of steam deck vs switch2 is worth looking at, steam deck has a 30% or so raster performance advantage (note that the discussion is based on portable mode), but switch2 has near-magical (But there is a cost.)dlss with great RT capabilities that basically make up for that 30% rasterperformance gap.
 
Yeah, the modes will likely be similar to a PS4 and PS4 Pro, but with much faster storage, a much better CPU, more advanced architecture, more RAM...
"similar to PS4 Pro" performance with faster storage, a better CPU, more advanced architecture and more RAM descibes the Xbox Series S. A Gen 9 home console.

Being, let's call it optimistic, and say DLSS gives us a doubling to a four times performance uplift all else being equal, and that puts the effective performance of NG Switch on par with the other home consoles, in any situation where they're rendering natively and NG Switch is upscaling, it's going blow for blow. Really impressive in a handheld body!
 
I think oldpuck's analysis of steam deck vs switch2 is worth looking at, steam deck has a 30% or so raster performance advantage (note that the discussion is based on portable mode), but switch2 has near-magical (but costly) dlss with great RT capabilities that basically make up for that 30% rasterperformance gap.
I believe he specified however he was looking at handheld performance. We can, again, expect TV mode to nearly double that.
 
Yeah, doomer bias is real. Handhelds generally succeed after a success compared to home consoles so
Mostly because
1. They don’t have competition in that field

2. Most of the competition Nintendo has received in the handheld market (Sony) gave up, mostly because of the output of games and the difficulty of handling a console and handheld division.

Also the reason why Nintendo cut away the GBA support is mostly because of the PSP. Which in the end helped them in the long run.
 
I think oldpuck's analysis of steam deck vs switch2 is worth looking at, steam deck has a 30% or so raster performance advantage (note that the discussion is based on portable mode), but switch2 has near-magical (but costly) dlss with great RT capabilities that basically make up for that 30% rasterperformance gap.
Speaking of rasterization, isn't that advantage based on AMD's desktop GPUs? Because from what I understand, what partially accomplishes that is from Infinity Cache, which none of their APUs contain, including the Steam Deck.
 
"similar to PS4 Pro" performance with faster storage, a better CPU, more advanced architecture and more RAM descibes the Xbox Series S. A Gen 9 home console.

Being, let's call it optimistic, and say DLSS gives us a doubling to a four times performance uplift all else being equal, and that puts the effective performance of NG Switch on par with the other home consoles, in any situation where they're rendering natively and NG Switch is upscaling, it's going blow for blow. Really impressive in a handheld body!
While I'm already convinced that dlss is magic, we can't ignore the fact that there is a cost to be paid for this, requiring you to lower the native resolution to reduce the frame time cost, however, since the 360p native resolution scaled to 1080p is pretty mind blowing, I'd be completely confident in this, such as the switch2 portable mode session 540p to 720p dynamic resolution as the nativeResolution then scaled to 1080p.And the visible future is that Nintendo first parties will be heavily customizing their game engines and rendering technology around dlss & RT, which is certainly a good thing for Nintendo.
 
Sorry I should have mentioned I was looking for "Docked Performance" expectations only. I was just trying to gauge if you guys were thinking 4tf in docked was viable or a lower base number closer too Steam Deck.
Well as I've said, being pessimistic is not the same as being realistic.

Think about it for a moment, if handheld mode is about half the performance of TV mode, and TV mode is PS4 levels of performance, that puts NG Switch's handheld performance in the realm of 800-900GFLOPS; maybe lower. That's simply not realistic with T239, it's not even clear it would WORK at speeds low enough for that. What's the point of T239, when that's the performance category of Tegra X1+?

So when people, especially technical folk, talk about PS4 or the Deck, they do mean handheld mode, because that's what T239's design implies to be the case, it would seem. Remember again that this would be HALF of TV mode's performance, and twice the PS4 is close to the Series S.

In terms of it being an actual 4TF beast? I think that's extremely unlikely, for thermal and power consumption reasons. I'm certain T239 can achieve it, but I don't think it's likely at all. As others, I believe, have mused, 4TF may have been the target peak for Nvidia and they may have achieved that, but real world sustained performance I would expect to be in the 3-3.5TF range in TV mode. Not a Series S, not exactly, but really rather close. And all that before upscaling.
 
Last edited:
Speaking of rasterization, isn't that advantage based on AMD's desktop GPUs? Because from what I understand, what partially accomplishes that is from Infinity Cache, which none of their APUs contain, including the Steam Deck.
According to oldpuck's analysis, the steam deck doesn't affect AMD's raster performance advantage too much despite the lack of Infinity Cache, but you're also right that the drake is bigger and has a higher number of shaders, so I'm not sure about the final comparison.
 
0
Mostly because
1. They don’t have competition in that field

2. Most of the competition Nintendo has received in the handheld market (Sony) gave up, mostly because of the output of games and the difficulty of handling a console and handheld division.

Also the reason why Nintendo cut away the GBA support is mostly because of the PSP. Which in the end helped them in the long run.
Furukawa's main role during his tenure has been to hold the fort, and I don't think anything will go wrong with Nintendo for at least the next six to seven years by following the set of directions and rules that Iwata ended up laying out for the next generation.
 
0
I believe they would make statements like “considering innovative idea for next Generation, exploring way to smoothly transition current gen to the next”
Very sure they will cut switch production this FY. With the lack of new software lineup this year, they need at least a hint to sooth the shareholders
 
While I'm already convinced that dlss is magic, we can't ignore the fact that there is a cost to be paid for this, requiring you to lower the native resolution to reduce the frame time cost, however, since the 360p native resolution scaled to 1080p is pretty mind blowing, I'd be completely confident in this, such as the switch2 portable mode session 540p to 720p dynamic resolution as the nativeResolution then scaled to 1080p.And the visible future is that Nintendo first parties will be heavily customizing their game engines and rendering technology around dlss & RT, which is certainly a good thing for Nintendo.
This is me going full silly mode, BUT... With DLSS Ultra Performance Mode targeting 4K with 3.3TF of performance, you're rendering at 720p and sacrificing about a third of your performance to DLSS, leaving you with 2.2TF to render 720p. That pixel count nonuples (9X) through DLSS to give you 4K, giving you an effective performance of 19.8TF. That is to say for a machine to achieve the same scene complexity as NG Switch and render it natively at 4K, it would need nearly 20 teraflops of compute, and all that would get you is better image quality. Even after the sacrifice to DLSS, the performance uplift compared to native rendering is INSANE! I think it'll really turn heads come release.

But again, that's being silly, it's not realistic to call NG Switch a "20TF class device", it isn't. 😛
 
Last edited:
Ehh. N64, Wii U, 3DS fall in that category I suppose. But SNES, GBA, DS make three as well, so seems pretty even.
SNES sold around 12-13 million less than NES' 62 mil. It was ok. It's definitely nowhere near DS.

But Switch is still a portable so I wouldn't put it in the home console bracket. Nintendo's portables do very well. It's their home consoles that were selling less with every new generation bar the Wii.
 
SNES sold around 12-13 million less than NES' 62 mil. It was ok. It's definitely nowhere near DS.

But Switch is still a portable so I wouldn't put it in the home console bracket. Nintendo's portables do very well. It's their home consoles that were selling less with every new generation bar the Wii.
No, it's not, the biggest reason SFC sales are lower than FC is because its actual lifespan is only 6 years, they released the N64 in 1996, and the introduction of the ps1 out of nowhere in 1994 hurt SFC's sales results somewhat, but even so, SFC was definitely a very successful product.As others have said, we basically can't expect to replicate the success of the switch on switch2.If we can draw an analogy to SFC, then switch2 is a huge win.
 
Last edited:
While I'm already convinced that dlss is magic, we can't ignore the fact that there is a cost to be paid for this, requiring you to lower the native resolution to reduce the frame time cost, however, since the 360p native resolution scaled to 1080p is pretty mind blowing, I'd be completely confident in this, such as the switch2 portable mode session 540p to 720p dynamic resolution as the nativeResolution then scaled to 1080p.And the visible future is that Nintendo first parties will be heavily customizing their game engines and rendering technology around dlss & RT, which is certainly a good thing for Nintendo.
DLSS certainly does have a cost in terms of frame time, but, it was also demonstrated how that can be done concurrently.

In fact, from what I understand, AMD's FSR (and RT?) cannot be done concurrently, so it ultimately does require going from start to finish for a given frame.
This is me going full silly mode, BUT... With DLSS Ultra Performance Mode targeting 4K with 3.3TF of performance, you're rendering at 720p and sacrificing about a third hour performance to DLSS, leaving you with 2.2TF to render 720p. That pixel count nonuples (9X) through DLSS to give you 4K, giving you an effective performance of 19.8TF. That is to say for a machine to achieve the same scene complexity as NG Switch and render it natively at 4K, ir would need nearly 20 teraflops of compute, and all that would get you is better image quality. Even after the sacrifice to DLSS, the performance uplift compared to native rendering is INSANE! I think it'll really turn heads come release.

But again, that's being silly, it's not realistic to call NG Switch a "20TF class device", it isn't. 😛
I'm not quite understanding this statement. DLSS is cutting into overall performance, even though it is using separate, dedicated hardware?
 
I'm not quite understanding this statement. DLSS is cutting into overall performance, even though it is using separate, dedicated hardware?
DLSS uses the tensor cores, but it doesn't JUST use the tensor cores, there's a performance cost to be paid for every part of it, because it needs the GPU and CPU to dedicate some of their time and resources that could have otherwise gone to raster performance or physics.

The tradeoff is almost always worth it, and certainly more than worth it with Ultra Performance Mode, but it does exist - DLSS is not "free".

(Also I had a typo, it should be "of your", not "hour".)

These numbers also aren't being totally accurate, I'm being silly and making approximations.
 
In fact, from what I understand, AMD's FSR (and RT?) cannot be done concurrently, so it ultimately does require going from start to finish for a given frame.

I'm not quite understanding this statement. DLSS is cutting into overall performance, even though it is using separate, dedicated hardware?
FSR is just a shader, and it is possible to run shaders concurrently. But it's the same set of resources as other shaders, so concurrency buys you very little. It eats into the render budget no matter what.

In RTX cards, tensor cores, shader cores, and RT cores are bundled together in groups, the SM. In RTX 20 cards, only one kind of core in an SM can be active. RTX 30 and 40 allow each core to be active without blocking the other.

DLSS is a black box, but it most likely uses tensor cores to accelerate the slowest, most quality sensitive part of the operation, and still uses shader cores for the setup and the final drawing phases of the operation. So when you run concurrently, unlike FSR, a decent chunk of the process is "free" as it doesn't compete with the shader cores. But there are still some chunks that run on the shader cores and have to compete with other shaders for time.
 
I think everyone does? They’re in the process of leaving hardware, lol.
They've explicitly stated they're making more hardware and that the current Xbox ecosystem is profitable. It's neither failed nor are they "leaving hardware". They are changing their strategy; but Sony is too, stating their intention to bring their first party games to platforms outside of PS, PC and mobile. Whether that means what it sounds like, and PS exclusives end up on Xbox eventually, we can't know right now. The modern high performance console market is a tough one for all involved.
 
Are shader cores the same things as CUDA cores?

Camp 2: System isn't even on par with a Steam Deck or PS4 base model but does feature DLSS. Digital Foundry definitely deflated the hype with their speculation video but of course they could only use off the shelf components and a lot of guessing. No on the market handheld seem capable of the Xbox Series S visuals (at least as far as I'm aware). Especially at a $400 price point.

Hmm? It's been a while since I saw the video, but I don't recall getting the impression that they thought switch 2 would only be around base PS4 power. In fact there are many points that DF and myself were impressed, like the fact that it ran A Plague Tale ( A PS5 exclusive) at 1080p and consistent 30 FPS. I think they were trying to say that it wasn't going to reach PS5 or Xbox series X level, which, to be fair I don't think anybody believed either at the time or now.
 
Last edited:
This is me going full silly mode, BUT... With DLSS Ultra Performance Mode targeting 4K with 3.3TF of performance, you're rendering at 720p and sacrificing about a third of your performance to DLSS, leaving you with 2.2TF to render 720p. That pixel count nonuples (9X) through DLSS to give you 4K, giving you an effective performance of 19.8TF. That is to say for a machine to achieve the same scene complexity as NG Switch and render it natively at 4K, ir would need nearly 20 teraflops of compute, and all that would get you is better image quality. Even after the sacrifice to DLSS, the performance uplift compared to native rendering is INSANE! I think it'll really turn heads come release.

But again, that's being silly, it's not realistic to call NG Switch a "20TF class device", it isn't. 😛
DLSS uses the tensor cores, but it doesn't JUST use the tensor cores, there's a performance cost to be paid for every part of it, because it needs the GPU and CPU to dedicate some of their time and resources that could have otherwise gone to raster performance or physics.

The tradeoff is almost always worth it, and certainly more than worth it with Ultra Performance Mode, but it does exist - DLSS is not "free".

(Also I had a typo, it should be "of your", not "hour".)

These numbers also aren't being totally accurate, I'm being silly and making approximations.
Comparing to engines because that's a language I speak a little better...

Engines displace a certain amount of air based on how large their combustion chambers are. That's where we hear about 2 litre engines, 2.5 litre, 3 litre, etc.

If you take an engine that is, say, 3 litres, and it makes around 200hp, then you add turbochargers which compress the incoming air, that 3 litre engine will effectively be displacing 6 litres if the turbo is compressing it to double the atmospheric pressure. People don't then call it a 6 litre engine but it will be moving 6 litres of air (technically a little less because of inefficiency but for sake of conversation). Then that same 200hp engine will suddenly be making 400hp, which is roughly what would be expected from a natural 6 litre engine. And it does continue to scale (not perfectly, but still) if you compress the incoming air even further.

SO... a smaller unit, but with a power adder, and the power adder does take some energy to run but is designed to make up for its energy cost by drastically scaling up the effective size/power of the main unit (without physically changing its actual size), making the unit just as powerful and fast as a unit literally double its size...

Yeah I don't think you're actually being as silly as you think you are.

and yes I'm saying the 2witch is gonna be the gaming equivalent of a 2JZ
 
Last edited:
SNES sold around 12-13 million less than NES' 62 mil. It was ok. It's definitely nowhere near DS.
It sold less, but I don't think selling 20% less in a market with stronger competition is enough to land in the screwed-it-up pile.
No, it's not, the biggest reason SFC sales are lower than FC is because its actual lifespan is only 6 years, they released the N64 in 1996, and the introduction of the ps1 out of nowhere in 1994 hurt SFC's sales results somewhat, but even so, SFC was definitely a very successful product.As others have said, we basically can't expect to replicate the success of the switch on switch2.If we can draw an analogy to SFC, then switch2 is a huge win.
You mention PS1, but on the whole I think Genesis was a bigger deal. Prevented Nintendo from just walking away with 90+% of the market again. PS1 hit Japan late '94, but it wasn't pushing big numbers for a few more years. Either way, there's little sign of any new Genesis or PS1 appearing to meddle with Switch 2.
 
No, it's not, the biggest reason SFC sales are lower than FC is because its actual lifespan is only 6 years, they released the N64 in 1996, and the introduction of the ps1 out of nowhere in 1994 hurt SFC's sales results somewhat, but even so, SFC was definitely a very successful product.As others have said, we basically can't expect to replicate the success of the switch on switch2.If we can draw an analogy to SFC, then switch2 is a huge win.
6 years isn't really an "only". The PS1 had 5.5 years and sold around 102 mil. The PS2 had 6.5 years and sold between 155-160 mil.

Also, when the next console is out, the previous console still has life in the market. Doesn't mean it completely stopped selling then and there. So you can add a year or two to its lifetime easily. Those aren't really valid reasons.
 
0
I remember getting the Switch in May 2017 and being super disappointed at how little of a boost it was over the Wii U as a home console, which is how I would be using it. My only consolation was that as mobile-powered tech, it would surely be on a faster upgrade cycle. If you had told me back then that we wouldn't get an upgrade/sequel console until 2025, I simply could never have believed you. That was an impossible nightmare-world scenario, "Nintendo gonna Nintendo" taken to an absurdity. I would have bet my entire life savings against it. So for me, the whole wait is viewed in that context.
I'm not sure where this idea even came from in the first place: handhelds aren't going to suddenly be immune to all the complexities of developing a console. If anything, faster upgrade cycles are less difficult for stationary consoles than handheld ones: with the former, people don't care as much if the system is bulkier and runs at a higher TDP, but with a handheld there's a lot more reason to scrutinize these things. Though I should probably disclose that this is all coming from the perspective of someone who straight up didn't want a Switch Pro, I've previously gone over the reasons why.
I don't know of he counts three. 2 at most, Lovelace and Blackwell. The thing is that, we're still getting the node and efficiency upgrade, which is by far the most significant feature of Lovelace. Frame gen woudnt have been particularly useful in a handheld either way. A 4 nm ampere chip is definitely not night and day worse than a Lovelace chip.
Yeah, I don't really see how they're reading this as "missing three generations of improvements"
 
It sold less, but I don't think selling 20% less in a market with stronger competition is enough to land in the screwed-it-up pile.
Yea I definitely didn't say that. But I wouldn't put it in the GBA, DS, Wii pile either. I'd put it in the middle, with the GBC and 3DS.

I'm obviously separating GBC sales from the GB variants here

So I see the Switch 2 at least falling in the middle there. I don't see it doing any worse.
 
0
Since the Switch 2 has been delayed by a whole year, is there any chance it may receive any upgrades?
I think the rumor is more that it's been delayed 3-5 months, from Q4 24 to Q1 25.

I'm not a tech expert at all, but I think the most common opinion is that the CPU/GPU cannot be notably changed with that small of a delay.
 
"similar to PS4 Pro" performance with faster storage, a better CPU, more advanced architecture and more RAM descibes the Xbox Series S. A Gen 9 home console.

Being, let's call it optimistic, and say DLSS gives us a doubling to a four times performance uplift all else being equal, and that puts the effective performance of NG Switch on par with the other home consoles, in any situation where they're rendering natively and NG Switch is upscaling, it's going blow for blow. Really impressive in a handheld body!

Well, the Series S is beyond both in terms of RAM bandwidth (the PS4 Pro also has more bandwidth than Switch 2 but I don't think it was able to take nearly as much advantage of tiled rendering?) and I think the raster is better due to RDN2, but yeah, you're essentially right. I do hope we see that level of grunt. DLSS may or may not be a major performance booster since the current consoles can use FSR2, but it'll sure as hell close the gap in image quality. It kind of depends on the angle you look at it from.
 
Since the Switch 2 has been delayed by a whole year, is there any chance it may receive any upgrades?
seems like it's the same as appeared in the Nvidia leak & was finalized in 2022, for better or worse (hopefully better if designed for a smaller node than SEC 8nm). there's a chance the delay could have increased the RAM quantity if it ends up being 16gb.
 
Well, the Series S is beyond both in terms of RAM bandwidth (the PS4 Pro also has more bandwidth than Switch 2 but I don't think it was able to take nearly as much advantage of tiled rendering?) and I think the raster is better due to RDN2, but yeah, you're essentially right. I do hope we see that level of grunt. DLSS may or may not be a major performance booster since the current consoles can use FSR2, but it'll sure as hell close the gap in image quality. It kind of depends on the angle you look at it from.
For a related question about ps4pro bandwidth you can look up the discussion from a few days ago, GDDR5 has a higher bandwidth requirement than LPDDR5, so the bandwidth advantage of the ps4 series is not significant, and for the switch2, a portable gaming console, the bandwidth of LPDDR5 (102.4GB/s) is able to perform at most of the upper limits of performance for the current state of the art.
 
0
I think in regards to the hardware internals of the Switch successor, I'm having a hard time with setting expectations. There are what I feel to be two camps.

Camp 1: This system is a monster. Series S capabilities + DLSS and better raytracing support. Those GDC rumors from last year of the Matrix tech demo running on Switch 2 specs and looking comparable to the PS5 version (also 4K/60 BOTW with zero load times) would certainly suggest it would need to be very beefy to run this well. There were even reports that the lighting looked even better then PS5 versions.

Camp 2: System isn't even on par with a Steam Deck or PS4 base model but does feature DLSS. Digital Foundry definitely deflated the hype with their speculation video but of course they could only use off the shelf components and a lot of guessing. No on the market handheld seem capable of the Xbox Series S visuals (at least as far as I'm aware). Especially at a $400 price point.

Like I'm leaning towards this thing basically being a base PS4 but with DLSS capabilities but at the same time there is no way that Matrix demo would run on such weak hardware. Digital Foundry said they couldn't even get it to boot on their test system.

So yeah. I have no idea where I should place my expectations. Nintendo hasn't targeted higher end specs in decades though and so I'm kind of leaning toward the system being a lot lower powered then we are all hoping for. Maybe a base 2013 PS4 (in terms of raw power) but with DLSS and a few other modern features sets.

So what do all of you guys think? Will Switch 2 be a monster with Xbox Series S levels of power plus DLSS and raytracing or will the Switch 2 be more closer to a base PS4 or Steam Deck with DLSS?

Are you Camp 1 or Camp 2? Seriously I have no idea what to expect anymore lol. [Looking for docked performance estimates only]
camp 1 = TSMC 4N/5nm
camp 2= SEC 8nm

something inbetween is possible but feels like that's the biggest determining factor here.
 
camp 1 = TSMC 4N/5nm
camp 2= SEC 8nm

something inbetween is possible but feels like that's the biggest determining factor here.
This is an oversimplified and crude understanding, or even just looking at paper floating point data, and is misleading.
 
0
I think in regards to the hardware internals of the Switch successor, I'm having a hard time with setting expectations. There are what I feel to be two camps.

Camp 1: This system is a monster. Series S capabilities + DLSS and better raytracing support. Those GDC rumors from last year of the Matrix tech demo running on Switch 2 specs and looking comparable to the PS5 version (also 4K/60 BOTW with zero load times) would certainly suggest it would need to be very beefy to run this well. There were even reports that the lighting looked even better then PS5 versions.

Camp 2: System isn't even on par with a Steam Deck or PS4 base model but does feature DLSS. Digital Foundry definitely deflated the hype with their speculation video but of course they could only use off the shelf components and a lot of guessing. No on the market handheld seem capable of the Xbox Series S visuals (at least as far as I'm aware). Especially at a $400 price point.

Like I'm leaning towards this thing basically being a base PS4 but with DLSS capabilities but at the same time there is no way that Matrix demo would run on such weak hardware. Digital Foundry said they couldn't even get it to boot on their test system.

So yeah. I have no idea where I should place my expectations. Nintendo hasn't targeted higher end specs in decades though and so I'm kind of leaning toward the system being a lot lower powered then we are all hoping for. Maybe a base 2013 PS4 (in terms of raw power) but with DLSS and a few other modern features sets.

So what do all of you guys think? Will Switch 2 be a monster with Xbox Series S levels of power plus DLSS and raytracing or will the Switch 2 be more closer to a base PS4 or Steam Deck with DLSS?

Are you Camp 1 or Camp 2? Seriously I have no idea what to expect anymore lol. [Looking for docked performance estimates only]


I know others have already said this but it needs to reiterated that whenever anyone (even DF) are comparing Switch 2 to PS4 or Steamdeck, we are talking about handheld first. Most have gone on record to say of course we expect docked to be more powerful than PS4, but how close that reaches to Series S is something else entirely...

I also want to point out, that when we are comparing memory bandwidth with PS4 and Steamdeck.
To the best of our knowledge (based on Gamescom information) where Switch 2 will overcome these other devices is by having a newer generation I/O that supports on the fly asset streaming.

The PS4 probably couldn't capitalize on that 179GB/s memory bandwidth because it was still at the mercy of a horrible CPU and an even slower internal HDD. Steamdeck has to adhere to some of this philosophy because the base SD still uses slow internal memory, so they can't design games to fully take advantage of the full system like Switch 2 will be capable of...

I really stand on the best features this generation that came from PS5 and Series X are them doubling down on fast I/O, because it changes how games can be made and developers are no longer at the mercy of having to load a full scene into memory.
 
Last edited:
So here's a question I've also been kind of curious about. We all know about ray tracing and we all know about dlss, but what other fancy new graphical features will switch 2 be capable of that switch wasn't?
 
So here's a question I've also been kind of curious about. We all know about ray tracing and we all know about dlss, but what other fancy new graphical features will switch 2 be capable of that switch wasn't?
Off the top of my head there should be HDR and proper Anti-Aliasing
 
I believe he specified however he was looking at handheld performance. We can, again, expect TV mode to nearly double that.
Yea but if we're comparing raw raster only, rdna 2 has 30% more performance per teraflop (or does ampere have 30% less? Not sure).

Either way it's far from a 1:1 comparison.
 
Are shader cores the same things as CUDA cores?
Yes. Just fancy Nvidia naming

just saw this over on Resetera. Don't know what any of this means but figured people here might be interested.
It's more than likely this is just an "uncapped framerate" mode that bezel engine already does (and that most engines do). This isn't a hint to VRR support
 
I just listened to the Nate podcast where he still stand by the late 22/23 info as a cancelled pro model- Respectfully disagree, there's no way there ever was a pro supposed to release in that timeframe.

But it roughly matches up to the timeframe Drake was finished. So in my headcanon the info was really about Drake powered devkits.
 
Please read this staff post before posting.

Furthermore, according to this follow-up post, all off-topic chat will be moderated.
Last edited:


Back
Top Bottom