Switch 2 Speculation

Additional gpu in the dock is still not out of question right?

Anyway, nowadays APUs are very scalable. So the more cooling solution you can provide (e.g. Via a dock), the faster it'll be, and in substantial enough amount.
My understanding is that additional gpu power in the dock is doable, but you'll have to sacrifice the Switch core feature, which is to be able to dock/undock in a split second.
Mostly because there is no advantage to run an external GPU with no dedicated VRAM, and moving all the data required to have the external gpu ready to render the next frame is way too long for the quick dock/undock thing to be possible.

Anyway, like you said, APU are very scalable and capable, so if the "mobile" device thermal solution can deal with say 60W when docked, then you have a 60W desktop console.
Problem is your heat dissipation system has to fit into a super small enclosure. The key might be to design an additional active cooler in the dock. But let's say existing solutions like 'fans pads' are not super efficient. Nintendo has to figure out smth else... or live within the constraints of a boosted portable device.
I'll add that the definition of a portable device vary of course. In this regards, I'm pretty pleased with the form factor of the Switch (vs the Steam Deck for example), and I may buy a better Switch with the same form factor, but I won't buy a Switch 2 with the Steam Deck form factor. The switch is the limit for me I'd say.
 
Additional gpu in the dock is still not out of question right?

Anyway, nowadays APUs are very scalable. So the more cooling solution you can provide (e.g. Via a dock), the faster it'll be, and in substantial enough amount.
Well if you still want the switch to support “hot unplug” and meanwhile keep the cost down, then I’d say it is out of the question
 
Well if you still want the switch to support “hot unplug” and meanwhile keep the cost down, then I’d say it is out of the question
Boosted clocks and higher power limit with extra cooling in the dock would probably be the preferred option if at all.

The current 16nm shrink Mariko can clock the CPU at 2397MHz and GPU at 1267MHz (Max handheld 405MHz and 768MHz docked) if overclocked.
 
Boosted clocks and higher power limit with extra cooling in the dock would probably be the preferred option if at all.

The current 16nm shrink Mariko can clock the CPU at 2397MHz and GPU at 1267MHz (Max handheld 405MHz and 768MHz docked) if overclocked.
That is definitely possible, though I doubt the how much an improvement it can be.
i'm not a hardware guy so don't take my words for granted: i don't think there's an easy way to add additional cooling system when the entire console is already wrapped and enclosed for handheld purposes -- like the place that requires the most cooling is the SoC and i can't think of a way to make the dock's cooling device go in.
In addition, the thermal design of the console also needs to consider the handheld positions, so places that would be ideal for air exchange/heat disspation might not work as it will hurt the handheld feeling.
 
As someone who used to play on a windows tablet and a thin and light laptop, the sustained tdp can be almost doubled just by letting it getting more natural air.

more than double sustained tdp can be achieved by using additional USB fan that is stuck on the exhaust. Like this one

It would make switch 2 dock larger and more expensive to manufacture tho
 
As someone who used to play on a windows tablet and a thin and light laptop, the sustained tdp can be almost doubled just by letting it getting more natural air.

more than double sustained tdp can be achieved by using additional USB fan that is stuck on the exhaust. Like this one

It would make switch 2 dock larger and more expensive to manufacture tho
That's the thing. We don't need "double TDP" in this case to unlock modern SoC full potential.
We're basically stuck with ~5-6W in handheld, mostly because of battery life concerns. So I'd say more like x5-6TDP (~30W) for the added cost and complexity to pay off visually.

But maybe Nintendo can decide to go the extra mile and put in an oversized heatsink.
With oversized vents at the bottom and top. The vents in the bottom would basically be left unused in handheld mode, but could serve as the main cool air intake while docked, thx to a more powerful fan in the dock...
Maybe it's doable. But it will be more expensive, add some weight and thickness to the console, and won't be as fool-proof as the OG concept as the console would need to be perfectly set in the dock for optimal use.

I don'k know. Part of me wants to get the best of the device potential, with super low TDP in handheld and max TDP in desktop. And at the same time, I'm perfectly fine with the Switch as it is today...
 
Proper alignment is already required for the USB-C connection.
I'd say "yes and no". the male USB-C connector has some degrees of freedom. You can have the Switch properly connected and not have it sit perfectly "flat" inside the dock. I the dock blows air from below, you wan't to force it through the bottom vent, not have most of it just go elsewhere, and this require some tighter fit. Anyway, I'm sure clever engineers could get smth working.

Real question is still : would it be worth it? (perceived image quality increased vs additional complexitiy, cost, weight and thickness)...
I think "yes, of course"!
At the same time I know my two boys would say "But dad, games are already super beautiful!"

That's the thing. The Switch is not a home console. We have 3 at home. If it was more expensive, it's likely we would have "only" 2. And the boys mostly play in handheld mode, like many people, so all this added cost/complexity will only be relevant to some customers. I guess only Nintendo knows what its true customer base demographics is, and what are their true expectations.
Maybe the next Switch will have the super dock discussed earlier (with dedicated GPU and all) because Nintendo knows no one really takes advantage of the Switch 'switch' feature and customers can deal with a loading screen when they switch :D
Or maybe they will stick to the original formula, maybe simplify it even further (like a larger Switch Lite... who really need detachable joy-cons? :D) because they know a lower cost is key to their success.
Exciting times ahead :)
 
That's the thing. We don't need "double TDP" in this case to unlock modern SoC full potential.
We're basically stuck with ~5-6W in handheld, mostly because of battery life concerns. So I'd say more like x5-6TDP (~30W) for the added cost and complexity to pay off visually.

But maybe Nintendo can decide to go the extra mile and put in an oversized heatsink.
With oversized vents at the bottom and top. The vents in the bottom would basically be left unused in handheld mode, but could serve as the main cool air intake while docked, thx to a more powerful fan in the dock...
Maybe it's doable. But it will be more expensive, add some weight and thickness to the console, and won't be as fool-proof as the OG concept as the console would need to be perfectly set in the dock for optimal use.

I don'k know. Part of me wants to get the best of the device potential, with super low TDP in handheld and max TDP in desktop. And at the same time, I'm perfectly fine with the Switch as it is today...
My point was that the extra oomph gonna be facilitated by the dock. With beefy fan.

The handheld device will be the same.
 
There have been rumors of Switch Pro before but nothing, apart from the slight OLED evolution, has emerged. However, clues pointing to a console capable of providing 4K gaming have circulated for a time.

And it already starts with the statements of Robert Wu, CEO of Sharp, who would have indicated to his investors a collaboration with Nintendo on the LCD screen of the future console.
Asked about the subject, he reportedly stressed: "I cannot comment on details concerning particular customers. But about a new console, we were involved in the R&D phase," Bloomberg said.

Robert Wu does not name Nintendo or specify that it is the Switch 2, but the timing of these details and the close ties between Sharp and Nintendo point to this scenario.

In a sign of the sensitivity of the subject, the mention of Sharp's role as a supplier for a new console has been erased from the presentation, says Bloomberg, which specifies that Sharp will set up a pilot production of LCD screens for the new console this year.
 
it will surely be a 720-1080p screen that does 4k dlss maybe even dlss3. Something in the 2-4tflop region based on a love lace likely and on 5nm in late 2024 or early 2025.

I think with the release of the new zelda looking like a potato even them getting between xbox one /ps4 and xbox series s in terms of graphics would be a huge shock to their development teams / budgets / time lines
 
I think with the release of the new zelda looking like a potato even them getting between xbox one /ps4 and xbox series s in terms of graphics would be a huge shock to their development teams / budgets / time lines
It's almost like they were incentized to keep with lower quality assets to fit in the smaller 16 GB cartridge limits.
 
it will surely be a 720-1080p screen that does 4k dlss maybe even dlss3. Something in the 2-4tflop region based on a love lace likely and on 5nm in late 2024 or early 2025.

I think with the release of the new zelda looking like a potato even them getting between xbox one /ps4 and xbox series s in terms of graphics would be a huge shock to their development teams / budgets / time lines
Series is twice as powerful as series S 😂 I think the GPU will be close to PS4 but the CPU will definitely be stronger
 
it will surely be a 720-1080p screen that does 4k dlss maybe even dlss3. Something in the 2-4tflop region based on a love lace likely and on 5nm in late 2024 or early 2025.

I think with the release of the new zelda looking like a potato even them getting between xbox one /ps4 and xbox series s in terms of graphics would be a huge shock to their development teams / budgets / time lines
I'm definitely against the "zelda looks like a potato" hyperbole (because it obviously does not....), and I don't see what such a stance adds to the discussion.

But I 100% agree that a jump to a more modern and more powerful hardware can be a shock to their dev teams.
That's why I don't get the complains about Zelda technical (well, visual really) limitations: should the game be on a more powerfull hardware, it would have cost more and it would have taken more time to make. Meaning we would still have to wait for it a few months/years. Obviously, if we could have it all, all the best, but we live in a world with constraints, and we can't expect to have everything.

Other studios may choose to put more work and money towards bleeding edge graphics tech. That's cool. But those games cost more to make and often lack the polish some of us gamers can expect. The only studio I can think of that pushes everything (cost, polish, story, gameplay and all) is Rockstar... and well, they make a game every 5/6 years now. We just can't have it all guys, and it's ok this way. It's ok to live with a little frustration when it's just about video games. This is what turn those forums into a lively thing after all :)

---

Back on topic: we still have no new rumors, and no solid clue as to when the new console will arrive. But the consensus that devkits were sent to devs, then called back have me wondering... why?
Why would Nintendo call back devkits, given the new console will arrive no matter what. Having the dev to work on games "early" is always better than "lately". And obviously, the "new" devkits will be more powerful than the "old" ones... So what could justify the call back?
- T239 proven not efficient enough: no need to recall as a move to a newer node would have no impact on software
- T239 proven not powerfull enough: need to tweak it (new RAM controller, wider bus, wider GPU, new node)... well the new chip would still be a "super T239", and thus the software would still run on the updated platform
- found a pretty nasty security flaw in T239: avoid having flawed devkits in the wild could make sense
- Nintendo pissed off by the NVidia leak, new Switch to be based on smth else: is this even possible as it would mean no backward compatibility moving forward? We know a pissed off Nintendo can do weird things (I remember some rumors about a cancelled Zelda TV show because of a leak)... but ditching NVidia?

I just don't get how recalling devkits makes sense "just" because of a delay of the actual console...
 
I'm definitely against the "zelda looks like a potato" hyperbole (because it obviously does not....), and I don't see what such a stance adds to the discussion.

But I 100% agree that a jump to a more modern and more powerful hardware can be a shock to their dev teams.
That's why I don't get the complains about Zelda technical (well, visual really) limitations: should the game be on a more powerfull hardware, it would have cost more and it would have taken more time to make. Meaning we would still have to wait for it a few months/years. Obviously, if we could have it all, all the best, but we live in a world with constraints, and we can't expect to have everything.

Other studios may choose to put more work and money towards bleeding edge graphics tech. That's cool. But those games cost more to make and often lack the polish some of us gamers can expect. The only studio I can think of that pushes everything (cost, polish, story, gameplay and all) is Rockstar... and well, they make a game every 5/6 years now. We just can't have it all guys, and it's ok this way. It's ok to live with a little frustration when it's just about video games. This is what turn those forums into a lively thing after all :)

---

Back on topic: we still have no new rumors, and no solid clue as to when the new console will arrive. But the consensus that devkits were sent to devs, then called back have me wondering... why?
Why would Nintendo call back devkits, given the new console will arrive no matter what. Having the dev to work on games "early" is always better than "lately". And obviously, the "new" devkits will be more powerful than the "old" ones... So what could justify the call back?
- T239 proven not efficient enough: no need to recall as a move to a newer node would have no impact on software
- T239 proven not powerfull enough: need to tweak it (new RAM controller, wider bus, wider GPU, new node)... well the new chip would still be a "super T239", and thus the software would still run on the updated platform
- found a pretty nasty security flaw in T239: avoid having flawed devkits in the wild could make sense
- Nintendo pissed off by the NVidia leak, new Switch to be based on smth else: is this even possible as it would mean no backward compatibility moving forward? We know a pissed off Nintendo can do weird things (I remember some rumors about a cancelled Zelda TV show because of a leak)... but ditching NVidia?

I just don't get how recalling devkits makes sense "just" because of a delay of the actual console...

You can pick any Nintendo game on the switch you want , they all look like wii/ wii u games (the hardware was not that different) . It's like Nintendo is stuck in the mid 2000s. Nintendo has a tech debt of at least a decade if not closer to two decades and the next console is going to end up having to pay for that or nintendo will have released a new high end system and still release games looking like zelda botw which is a wii u game
 
You can pick any Nintendo game on the switch you want , they all look like wii/ wii u games (the hardware was not that different) . It's like Nintendo is stuck in the mid 2000s. Nintendo has a tech debt of at least a decade if not closer to two decades and the next console is going to end up having to pay for that or nintendo will have released a new high end system and still release games looking like zelda botw which is a wii u game
Tears has tons of modern techniques. The hardware hasn't moved all that much, but the team isn't sitting in place. It's low resolution and some of the polygon counts aren't huge, but it's a very impressive game technically. The team would have no trouble making something impressive on ps5 or xsx too.
 
Tears has tons of modern techniques. The hardware hasn't moved all that much, but the team isn't sitting in place. It's low resolution and some of the polygon counts aren't huge, but it's a very impressive game technically. The team would have no trouble making something impressive on ps5 or xsx too.
It looks amazing to me, especially comparing it to BOTW, which still looked great. They have definitely increased the visual fidelity and art direction for TOTK, which is impressive considering the Nintendo Switch hardware is pretty much garbage at this point.

I prefer gameplay and clever art direction over pure fidelity.
 
You can pick any Nintendo game on the switch you want , they all look like wii/ wii u games (the hardware was not that different) . It's like Nintendo is stuck in the mid 2000s. Nintendo has a tech debt of at least a doecade if not closer to two decades and the next consle is going to end up having to pay for that or nintendo will have released a new high end system and still release games looking like zelda botw which is a wii u game
Hm, ok, I guess. All of this is overly exagerated to the point there is no truth left in there. I can only agree on the fact that moving to a really more powerful hardware will require a lot of work from Nintendo dev teams. They will be fine. And maybe they'll choose to stop at the PS4 level for this new gen.
I'm not sure this industry can live with all vendors and studios following the path of multi-million dollars/5 years+ AAA games.
If a gamer can only enjoy a game if it's 4K/120FPS, then there are machines for him. If he can live with sub par graphics and still enjoy the hell out of some of the best game ever made, then, there is also a machine for him (plus this one gamer will spend less on his hardware, and less on his electricity bill). I see no looser here.

Tears has tons of modern techniques. The hardware hasn't moved all that much, but the team isn't sitting in place. It's low resolution and some of the polygon counts aren't huge, but it's a very impressive game technically. The team would have no trouble making something impressive on ps5 or xsx too.
This. I think the last DF video on TotK makes mentions a few of such modern techniques.
 
Back on topic: we still have no new rumors, and no solid clue as to when the new console will arrive. But the consensus that devkits were sent to devs, then called back have me wondering... why?
Why would Nintendo call back devkits, given the new console will arrive no matter what. Having the dev to work on games "early" is always better than "lately". And obviously, the "new" devkits will be more powerful than the "old" ones... So what could justify the call back?
- T239 proven not efficient enough: no need to recall as a move to a newer node would have no impact on software
- T239 proven not powerfull enough: need to tweak it (new RAM controller, wider bus, wider GPU, new node)... well the new chip would still be a "super T239", and thus the software would still run on the updated platform
- found a pretty nasty security flaw in T239: avoid having flawed devkits in the wild could make sense
- Nintendo pissed off by the NVidia leak, new Switch to be based on smth else: is this even possible as it would mean no backward compatibility moving forward? We know a pissed off Nintendo can do weird things (I remember some rumors about a cancelled Zelda TV show because of a leak)... but ditching NVidia?

I just don't get how recalling devkits makes sense "just" because of a delay of the actual console...
Some devkits are just prototypes for evaluation purposes and it's far from certain that console manufacturer's current supplier will offer a successor with a trivial backwards compatible implementation. A console's firmware isn't some common application that can run on just any hardware configuration. The firmware in the case of a console often depends on running with very specific sets of hardware that's far from guaranteed to be provided even with the same vendor ...

As an example both Microsoft and Nintendo still supplied GPUs from AMD on their previous platforms but their execution of backwards compatibility was far from trivial. Nintendo made a request to AMD to include a 2nd GPU for their semicustom SoC which is basically dead silicon for purposes other than backwards compatibility and that just bloats the hardware's design and costs as a negative. Microsoft still had to do GPU side software emulation for backwards compatibility when they still partnered with AMD and were transitioning from the Xbox 360. Don't look to previous consoles like the PS4/X1 and their successors PS5/XSX when sticking with the same hardware vendor as examples for backwards compatibility because they are the *exceptions* rather than the rule. Even if Nintendo does decide to stick with the same partner, expect Nvidia to do anything else regarding binary compatibility but follow AMD's most recent example ...

In knowing how hard it'll be for Nintendo to make the jump without losing backwards compatibility they now have other options on the table to consider ...

Nintendo could opt in for Qualcomm as their next partner for their integrated wireless connectivity technology to push for more social experiences on their next platform ...
If one of Nintendo's main design goals is to have tighter security measures to make piracy harder for their new system then they may partner with AMD once again since they currently have the best anti-emulation solution since their modern GPUs have never been successfully emulated and it'll make life harder on developers/users of Nintendo's hardcore reverse engineering scene often don't have AMD graphics hardware ...
Nintendo could also choose to go with other mobile GPU vendors for more efficient designs and better deals ...

They're taking their time specifically for their new platform because they know that they don't have any obvious successors to do a smooth transition like their competitors so they're optimizing their software release cycle in case backwards compatibility doesn't work out ...
 
Some devkits are just prototypes for evaluation purposes and it's far from certain that console manufacturer's current supplier will offer a successor with a trivial backwards compatible implementation. A console's firmware isn't some common application that can run on just any hardware configuration. The firmware in the case of a console often depends on running with very specific sets of hardware that's far from guaranteed to be provided even with the same vendor ...

As an example both Microsoft and Nintendo still supplied GPUs from AMD on their previous platforms but their execution of backwards compatibility was far from trivial. Nintendo made a request to AMD to include a 2nd GPU for their semicustom SoC which is basically dead silicon for purposes other than backwards compatibility and that just bloats the hardware's design and costs as a negative. Microsoft still had to do GPU side software emulation for backwards compatibility when they still partnered with AMD and were transitioning from the Xbox 360. Don't look to previous consoles like the PS4/X1 and their successors PS5/XSX when sticking with the same hardware vendor as examples for backwards compatibility because they are the *exceptions* rather than the rule. Even if Nintendo does decide to stick with the same partner, expect Nvidia to do anything else regarding binary compatibility but follow AMD's most recent example ...

In knowing how hard it'll be for Nintendo to make the jump without losing backwards compatibility they now have other options on the table to consider ...

Nintendo could opt in for Qualcomm as their next partner for their integrated wireless connectivity technology to push for more social experiences on their next platform ...
If one of Nintendo's main design goals is to have tighter security measures to make piracy harder for their new system then they may partner with AMD once again since they currently have the best anti-emulation solution since their modern GPUs have never been successfully emulated and it'll make life harder on developers/users of Nintendo's hardcore reverse engineering scene often don't have AMD graphics hardware ...
Nintendo could also choose to go with other mobile GPU vendors for more efficient designs and better deals ...

They're taking their time specifically for their new platform because they know that they don't have any obvious successors to do a smooth transition like their competitors so they're optimizing their software release cycle in case backwards compatibility doesn't work out ...
Indeed, I should have say moving to AMD "or any other vendor" "could" mean no backward compatibility.
As you say, BC is not a simple thing, but granted there is no legal stuff preventing Nintindo to emulate its previous NVidia based hardware, a technical solution may ultimately be found.

Now, I really think Nintendo is sticking to NVidia with Switch 2. It just makes sense, the rumored T239 chip sounds just about perfect for the new console.
I'm just really curious as to what was the reason to recall devkits as this action does not make any sense to me. I just think that letting the devs work on deprecated devkits is better than recalling them and ditch whatever work they were doing. And I guess the only reason this could make sense (to me) is the move to another vendor indeed. But moving to another vendor does not make sense either so...
I'm looking forward the next batch of rumors regarding this situation. Fun times ahead :)
 
Back
Top