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

[Mr.Gamerson]

Yeah that’s not Nintendo even a little bit.

The only thing that is allowing me to believe Nintendo may try to be more cutting edge than usual is the fact the Ceo said they were going be more aggressive from a technical stand point, and the fact that Nintendo just made a boatload of extra money because of Animal crossing and the pandemic. They have more money to invest in their new unified ecosystem and mobile hasn't taken off like they would have hoped it seems. If their Unified strategy starts receiving serious competition and starts declining, Nintendo will have some serious things to think about. Now would be the perfect time to start investing for the future of their platforms going forward. What if Apple, Google, Sony, Microsoft, Tencent, etc. decide to....

 

[fwd-bwd]

Have you noticed any substantial improvements to the Wi-Fi reception for the OLED model?

Everyone's wifi setup is different, but mine did improve from about 20Mbps to 30Mbps.

 

[ILikeFeet]

The only thing that is allowing me to believe Nintendo may try to be more cutting edge than usual is the fact the Ceo said they were going be more aggressive from a technical stand point, and the fact that Nintendo just made a boatload of extra money because of Animal crossing and the pandemic. They have more money to invest in their new unified ecosystem and mobile hasn't taken off like they would have hoped it seems. If their Unified strategy starts receiving serious competition and starts declining, Nintendo will have some serious things to think about. Now would be the perfect time to start investing for the future of their platforms going forward. What if Apple, Google, Sony, Microsoft, Tencent, etc. decide to....

wasn't that a mistranslation? not that they won't be investing in tech, but to be on the bleeding edge is a bit much

 

[Mori Medinayo]

The only thing that is allowing me to believe Nintendo may try to be more cutting edge than usual is the fact the Ceo said they were going be more aggressive from a technical stand point, and the fact that Nintendo just made a boatload of extra money because of Animal crossing and the pandemic. They have more money to invest in their new unified ecosystem and mobile hasn't taken off like they would have hoped it seems. If their Unified strategy starts receiving serious competition and starts declining, Nintendo will have some serious things to think about. Now would be the perfect time to start investing for the future of their platforms going forward. What if Apple, Google, Sony, Microsoft, Tencent, etc. decide to....

I think it hasn't a lot of sense to extrapolate that they have more money to invest due to Animal Crossing and the pandemic. That's pretty reductive. They have more money because Switch it's gonna be a top sales console so far. And because all the big franchises are selling like hotcakes, pandemic or not. What happened to Animal Crossing wasn't that much different from the rule. And I don't think it wouldn't have sold a lot less without the pandemic.

Anyway, I think you are right in your point with Nintendo investing more in R&D. But the reasons we know that is because we have official data for that.

Switch’s successor contributed to record annual R&D spend at Nintendo | VGC

Company reiterates next console will continue integrated hardware and software approach…

www.videogameschronicle.com

 

[BlackTangMaster]

I'm wondering if the tape out process hasn't been met yet because of possible tweaks to the architecture more-so and not just to the manufacturing process node of choice. RDNA2 definitely caught Nvidia by surprise just as Zen2 and Zen3 caught Intel resting on their laurels, so the advancements of Lovelace are either superficial in scope holding over until Hopper or Nvidia will try and tweak Ampere's shortcomings as much as possible.

Could we see a significant amount of architectural differences between desktop Lovelace and Nintendo’s Dane SoC is another question in the matter?

I believe that all these guys have been caught off guard by Apple having literally subsidised TSMC's especially their EUV toolings. They have most of TSMC's cutting edge 5 nm capacities and they will probably still buy a lot of 7 nm products reducing the remaining 7 nm capacities for other potential customers.

AMD/Intel/Nvidia will all have to fight against Apple with less performant nodes for chipsets within products generating less margins than Apple products. This is a vicious cycle made possible by ASLM being the only EUV toolings provider.

That said, Nintendo and Nvidia could circumvent the use of a cutting edge node by making the Switch 2 SoC as wide as Xavier was on 8 nm or targeting A12Z size on 7/6 nm. They will probably have to target tablets/ultrabook TDP (2 to 3W over the original 20 nm TX1 TDP instead of the Mariko mobile 5-7W TDP) which could be the case with Orin S being marketed as a 15W SoC derivative from the Orin architecture. A 10BTr chip made on 8 nm would probably be able to reach Xavier 15W mode performances consuming less than 10W for handheld mode. It would more or less be ~150-200mm2 considering the fact that the 17BTr GA104 is 399mm2 with HP libraries and has the same transistor count as the originally planned Orin chip with mobile libraries.

Having a 150-200mm2 chip will probably be necessary if Nintendo wants to be competitive with AMD offering with Renoir/VanGogh which are 152/170mm2 on TSMC's 7 nm and with mobile Android/Apple SoCs excluding M1. Especially if they are using 8 nm. Nvidia have already shown that they can make massive 350mm2 chips consume less than 10W on a less efficient node.

 

[ReddDreadtheLead]

The only thing that is allowing me to believe Nintendo may try to be more cutting edge than usual is the fact the Ceo said they were going be more aggressive from a technical stand point, and the fact that Nintendo just made a boatload of extra money because of Animal crossing and the pandemic. They have more money to invest in their new unified ecosystem and mobile hasn't taken off like they would have hoped it seems. If their Unified strategy starts receiving serious competition and starts declining, Nintendo will have some serious things to think about. Now would be the perfect time to start investing for the future of their platforms going forward. What if Apple, Google, Sony, Microsoft, Tencent, etc. decide to....

Sony isn’t, they tried, succeeded and then that flopped. Microsoft showed zero interest before and probably won’t now as they will leverage their massive scope and spread with features like xCloud for their gaming portability. Apple and Google are already the biggest players in the gaming space, they have no need to make a dedicated device really. Tencent is big but they would look to a PC portable like the Steam Deck rather than the switch. And as of current, their partner is Nintendo for China.

Amazon looks to be leveraging their widespread scope as well but for cloud gaming.

Nintendo is somewhat in a bubble on their own, but not a bubble that disallows an entry. But their only true competitor doesn’t want to try in that space anymore.


Plus, combined with redesigning for a higher node, and the R&D that went into Dane only to be scrapped, would actually be higher than half a billion in dough. If it’s just to have it on the better node that’s one thing, but combined the total would be north to nearing a billion I reckon. As nodes have became more complex the prices have increased dramatically as well.


While they can spend it, as they have that cash reserve, whether they spend it comes down to logistics of spending such an amount and cause the investors to ask serious questions. For one, where did and why did they spend money only to waste it and move onto something else.

And I say waste, as they would presumably not use it, but use the better node.

 

[Intoxicate]

I think it‘s interesting that the cooling in the OLED has been reduced, at least the heat pipe and the exhaust area on top of the Switch is smaller. That might show that they were testing the switch with higher performance settings in mind (especially docked). With the Swoled that seems to be completely off the table.

 

[Falcatops]

I think it hasn't a lot of sense to extrapolate that they have more money to invest due to Animal Crossing and the pandemic. That's pretty reductive. They have more money because Switch it's gonna be a top sales console so far. And because all the big franchises are selling like hotcakes, pandemic or not. What happened to Animal Crossing wasn't that much different from the rule. And I don't think it wouldn't have sold a lot less without the pandemic.

While I agree with the first part of you post (over half of Nintendo’s revenue comes from hardware sales), I disagree with the bolded part. Animal Crossing was the perfect storm and it’s success could be difficult to replicate: it released at the end of March ‘20 just when most countries were entering lockdowns and was the perfect getaway from the pandemic, it was THE social media darling game to buy at that time, and this pandemic boost was to such a large degree that it surpassed ACNL lifetime sales in just 6 weeks. ACNL was the previous best seller in the franchise with ~11M sales.

As you said a lot of Nintendo IP had big boosts in sales thanks to the “Switch factor” but the IPs with better “Switch game / previous best seller” ratios (like BotW selling 3 times more than the 2nd place) are those that released in key moments. Without the pandemic ACNH sales would be similar to that of the other high-end evergreens in the low 20Ms.

 

[Deleted User 1]

What happened to Animal Crossing wasn't that much different from the rule. And I don't think it wouldn't have sold a lot less without the pandemic.

100% agreed. The pandemic excuse for Animal Crossing sales reminds me of the "Easter" excuse used back in the day on GAF when the Wii took off.

 

[Mr.Gamerson]

100% agreed. The pandemic excuse for Animal Crossing sales reminds me of the "Easter" excuse used back in the day on GAF when the Wii took off.

The Switch also shipped like 40% more units, PS4's were selling online for more than msrp as was the Switch at times. It was likely the Pandemic that had a great influence in sales. There also weren't that many big releases in 2020 outside of Animal Crossing yet historic profits were had for Nintendo. Animal Crossing is a great game but not that great.

 

[ILikeFeet]

100% agreed. The pandemic excuse for Animal Crossing sales reminds me of the "Easter" excuse used back in the day on GAF when the Wii took off.

given that everything blew up in sales, it's very much a good reason

 

[Mr.Gamerson]

Sony isn’t, they tried, succeeded and then that flopped. Microsoft showed zero interest before and probably won’t now as they will leverage their massive scope and spread with features like xCloud for their gaming portability. Apple and Google are already the biggest players in the gaming space, they have no need to make a dedicated device really. Tencent is big but they would look to a PC portable like the Steam Deck rather than the switch. And as of current, their partner is Nintendo for China.

Amazon looks to be leveraging their widespread scope as well but for cloud gaming.

Nintendo is somewhat in a bubble on their own, but not a bubble that disallows an entry. But their only true competitor doesn’t want to try in that space anymore.


Plus, combined with redesigning for a higher node, and the R&D that went into Dane only to be scrapped, would actually be higher than half a billion in dough. If it’s just to have it on the better node that’s one thing, but combined the total would be north to nearing a billion I reckon. As nodes have became more complex the prices have increased dramatically as well.


While they can spend it, as they have that cash reserve, whether they spend it comes down to logistics of spending such an amount and cause the investors to ask serious questions. For one, where did and why did they spend money only to waste it and move onto something else.

And I say waste, as they would presumably not use it, but use the better node.

The PC space is growing and switch like devices are becoming more viable and common. Apple, Microsoft, Google, etc are rich enough to come in like Sony did and cause some trouble if Nintendo leaves gaping holes in advancements that can be filled. Microsoft is buying up everything right now because they know the gaming market is still young and they stand to gain a lot generations from now. Poor Sega, and Atari couldn't keep up. Nintendo is doing well enough to be Major player in the gaming space for at least this next gen based on how well they are doing with the Switch but afterwards they may be relegated to 3rd and fourth tier if they don't guard their spot well enough.

 

[Dakhil]

Having a 150-200mm2 chip will probably be necessary if Nintendo wants to be competitive with AMD offering with Renoir/VanGogh which are 152/170mm2 on TSMC's 7 nm and with mobile Android/Apple SoCs excluding M1. Especially if they are using 8 nm. Nvidia have already shown that they can make massive 350mm2 chips consume less than 10W on a less efficient node.

I highly doubt Nintendo cares about having Dane be competitive with AMD's APU offerings (e.g. Van Gogh, etc.). And I don't know if Nvidia necessarily cares about having Dane be competitive with AMD's APU offerings, considering that outside of the Tegra X1+ for the Nintendo Switch Lite, the Nintendo Switch, and the OLED model, Nvidia hasn't really offered consumer SoCs for a very long time.

Apple, Microsoft, Google, etc are rich enough to come in like Sony did and cause some trouble if Nintendo leaves gaping holes in advancements that can be filled.

Google's notoriously infamous for starting projects that Google abandons relatively quickly. In fact, there's a website dedicating to labelling all of the projects Google has started and then abandoned relatively quickly appropriately called Killed by Google. And Google Stadia seems to be no exception, considering Google's Stadia Games & Entertainment Studio was shut down at around 1 February 2021, several video game development projects as well as potential third party licencing deals have been cancelled due to Google's Stadia Games & Entertainment Studio being shut down, and Google Stadia's vice president and product head, John Justice, leaving Google around 3 May 2021. So I highly doubt Google is really a threat to Nintendo, considering Google hasn't really taken the video game industry seriously; and I'd argue Google currently isn't really taking the video game industry very seriously.

~

Anyway, Digital Foundry thinks that the RTD2172N chip in the OLED model's dock is a hint of Nintendo's future plans for a new console since a new console is in development. And Digital Foundry mentioned that the Tegra X1+ on the OLED model is only capable of up to 4K 30 Hz, so 4K 60 Hz support for the OLED model shouldn't be expected, although Digital Foundry is wrong, considering Nvidia explicitly stated that "Tegra X1 supports 4K 60 fps local and external displays with support for HDMI 2.0 interfaces and HDCP 2.2 copy protection."

 

[ReddDreadtheLead]

The PC space is growing and switch like devices are becoming more viable and common. Apple, Microsoft, Google, etc are rich enough to come in like Sony did and cause some trouble if Nintendo leaves gaping holes in advancements that can be filled. Microsoft is buying up everything right now because they know the gaming market is still young and they stand to gain a lot generations from now. Poor Sega, and Atari couldn't keep up. Nintendo is doing well enough to be Major player in the gaming space for at least this next gen based on how well they are doing with the Switch but afterwards they may be relegated to 3rd and fourth tier if they don't guard their spot well enough.

Rich enough to enter the space does not mean they have a need to enter the space. Google and Apple are already the biggest players in gaming without having to spend R&D on having a completely separate device that would have trouble being relevant to their other supported platforms. It would be like having the iPod when having an iPhone does everything the iPod does a thousand times better.

The game console in the switch format from them would need to convince developers a lot to get select titles that do not bring the same level of revenue as the general AppStore or Google PlayStore. There’s a difference between having the resources to do so and having a reason to expend resources into such an endeavor.

Microsoft already threw the towel for a dedicated device and their whole strategy isn’t about dedicated hardware, but hardware agnostic software and services. It’s their claim to fame. They leverage Azure for the cloud. They have no need for a dedicated hardware like a Switch. And this leads to my next point.

In the Pc space there is a rise of these portable devices like the SD, Aya Neo and GPD Win 3, however these are incredibly niche still and are for a very specific target audience. Microsoft already has those despite not making them, as their OS and their software stack that helps for games like DX.

We would have to wait a while before we see the effects of the SD on the PC market. By the next Switch would be out, not the Dane model, the one after.

The only thing Sony did to Nintendo is push them into the niche of new modes of play, or what people categorize as “gimmicks” to differently themselves from the competitors rather than chase the power and be the same.


Sega couldn’t compete due to their own issues that had nothing to do with the others. Almost self sabotage and not having the main driver of game consoles ready when they needed to be: Software.


Nintendo already had this issue at least twice in their last 30 or so years, only 1 was really done badly: Wii U

It lacked the right system pushing software and released when the previous system was already dead in the water.


I find it pretty difficult for a competitor to enter and somehow rock the boat for Nintendo especially when they are now. They considered Sony a really big threat in the portable scene as they had a huge gaming console mindshare fork the PS1 and PS2 was already doing pretty well by then.


And to round it off, Nintendo’s way of dealing with a competitor is not by releasing stronger hardware as a response, it’s by making their hardware unique as a result where it attracts consumers to them. They cannot in anyway possible compete with Apple and expect to win in hardware and stay afloat. What would they do? Add a gimmick that differentiates it from whatever Apple did. That’s the Nintendo way. When they have no direct competitor they act like a traditional Nintendo, aka, better version of the previous thing with no strings attached.

 

[ILikeFeet]

I finally watched 2kliksphilip's video on FSR vs DLSS, and he makes a glaring mistake. he claims that DLSS's scaling is unknown and thus uses the lowest presets for each FSR and DLSS. but that gives him two different input resolutions. Performance 720p is 360p in both methods, but he compares FSR's 360p with DLSS's 240p (and still thinks DLSS looks better)

 

[NineTailSage]

Could be half a billion

Yeah and although Nintendo undoubtedly has the money to bankroll a cutting edge SoC(that they could then use as the basis for the next-gen Switch) I expect them to not take such a drastic leap because it would dwarf the og Switch model in comparison.

The only thing that is allowing me to believe Nintendo may try to be more cutting edge than usual is the fact the Ceo said they were going be more aggressive from a technical stand point, and the fact that Nintendo just made a boatload of extra money because of Animal crossing and the pandemic. They have more money to invest in their new unified ecosystem and mobile hasn't taken off like they would have hoped it seems. If their Unified strategy starts receiving serious competition and starts declining, Nintendo will have some serious things to think about. Now would be the perfect time to start investing for the future of their platforms going forward. What if Apple, Google, Sony, Microsoft, Tencent, etc. decide to....

Even if that was exactly what Furukawa was hinting at, going with a Nvidia Lovelace based architecture while on Samsung's 8nm process would still be considered cutting edge.

I believe that all these guys have been caught off guard by Apple having literally subsidised TSMC's especially their EUV toolings. They have most of TSMC's cutting edge 5 nm capacities and they will probably still buy a lot of 7 nm products reducing the remaining 7 nm capacities for other potential customers.

AMD/Intel/Nvidia will all have to fight against Apple with less performant nodes for chipsets within products generating less margins than Apple products. This is a vicious cycle made possible by ASLM being the only EUV toolings provider.

That said, Nintendo and Nvidia could circumvent the use of a cutting edge node by making the Switch 2 SoC as wide as Xavier was on 8 nm or targeting A12Z size on 7/6 nm. They will probably have to target tablets/ultrabook TDP (2 to 3W over the original 20 nm TX1 TDP instead of the Mariko mobile 5-7W TDP) which could be the case with Orin S being marketed as a 15W SoC derivative from the Orin architecture. A 10BTr chip made on 8 nm would probably be able to reach Xavier 15W mode performances consuming less than 10W for handheld mode. It would more or less be ~150-200mm2 considering the fact that the 17BTr GA104 is 399mm2 with HP libraries and has the same transistor count as the originally planned Orin chip with mobile libraries.

Having a 150-200mm2 chip will probably be necessary if Nintendo wants to be competitive with AMD offering with Renoir/VanGogh which are 152/170mm2 on TSMC's 7 nm and with mobile Android/Apple SoCs excluding M1. Especially if they are using 8 nm. Nvidia have already shown that they can make massive 350mm2 chips consume less than 10W on a less efficient node.

So I'm leaning a couple of ways on a couple of things, from that great read on the Ampere architecture being extremely bandwidth heavy we are at this point almost certain to get 128bit wide bus.

You bringing up the possibility of a wider designed chip also makes me point back to that article of where they discussed how the 8nm process was not only more efficient but provided a higher transistor density when capped at lower clocks(which would benefit greatly anything under 1Ghz). This was looking at the Samsung Exynos 9820 design on the 8nm process

Nintendo were infamous of this using the GC architecture as the base for the Wii, so could we see a reverse of them doing a similar hat trick by designing an SoC that we don't fully see the entire scope of what it's capable of until it gets a die shrink to 5nm or smaller... A 150-200mm2 SoC moved to a much smaller process in the future could see Nintendo using essentially the same base design for 10 years (while providing a chip that can scale from under 2 Tflops on 8nm up to 10 Tflops in the future being on a more advanced process).

 

[uncleoptimus]

Now would be the perfect time to start investing for the future of their platforms going forward. What if Apple, Google, Sony, Microsoft, Tencent, etc. decide to....

Yea im with you on this. The real competition is for customer attention.

Nintendo produces first-rate game experiences but I'd like to see them do more to reduce friction when engaging w/ their platform. That includes investing in better online play experiences, player-to-player connection, even their eshop experience can and should be better.

On topic, I hope that means faster rate of hardware iteration, w/ a highlight on performance differentiation. As it is, since we just got OLED Model, I agree w/ others here that Spring 2023 looks like a cushy target date for the next machine. (So I guess BotW 2 couldn't be the new hardware's running mate. *Metroid Prime 4*!)

 

[Deleted member 645]

The only thing that is allowing me to believe Nintendo may try to be more cutting edge than usual is the fact the Ceo said they were going be more aggressive from a technical stand point, and the fact that Nintendo just made a boatload of extra money because of Animal crossing and the pandemic. They have more money to invest in their new unified ecosystem and mobile hasn't taken off like they would have hoped it seems. If their Unified strategy starts receiving serious competition and starts declining, Nintendo will have some serious things to think about. Now would be the perfect time to start investing for the future of their platforms going forward. What if Apple, Google, Sony, Microsoft, Tencent, etc. decide to....

Sorry we heard this before. More than likely for them that means investing in something like VR. I remember after the Wii everyone was so excited. How much money Nintendo made and was going to put into their next platform hardware and specs. We got the Wii U with a controller that cost too much and they spent way more than they should have making it BC. Wii U should have been a beast. Again Nintendo never again try and make a beast console. They will always favor small form factor, efficiency, and battery life. We have to deal with that.

 

[ReddDreadtheLead]

Yeah and although Nintendo undoubtedly has the money to bankroll a cutting edge SoC(that they could then use as the basis for the next-gen Switch) I expect them to not take such a drastic leap because it would dwarf the og Switch model in comparison.

No so much only that, but it’s the idea of just building it on a better node with a different technology which is what would be around half a billion. This doesn’t include the work done for the 8N Dane which for the convo sake is scrapped, which in total would be much higher than this combined.

I don’t think they are keen on spending over half a billion to near a billion to have two designs, one nearly done and then shelving it only to have another design that has its own R&D to have work on the node.

Things aren’t cheap.


However, we don’t know if 8nm is similar to 11nm from Samsung, where they have some backend stuff that allows porting from the 11 (really a 14nm) to the 10nm node.


8N might have back end from the 7nm which allows for porting to later. Or now.

 

[NineTailSage]

No so much only that, but it’s the idea of just building it on a better node with a different technology which is what would be around half a billion. This doesn’t include the work done for the 8N Dane which for the convo sake is scrapped, which in total would be much higher than this combined.

I don’t think they are keen on spending over half a billion to near a billion to have two designs, one nearly done and then shelving it only to have another design that has its own R&D to have work on the node.

Things aren’t cheap.


However, we don’t know if 8nm is similar to 11nm from Samsung, where they have some backend stuff that allows porting from the 11 (really a 14nm) to the 10nm node.


8N might have back end from the 7nm which allows for porting to later. Or now.

I don't think it's out of the realm of possibility for Nintendo to shell out close to a billion dollars for investing in future tech, they did so with IBM back during the GameCube's R&D days. Nintendo just as a whole seems to stray away from doing deals like that now, because it's failed them in the past. Since Switch is already established and cemented though, I think Nintendo could get more bang for their buck if they heavily invested now to establish something really cutting-edge on 5nm and could use this same design for the next 10 years.

I don't expect them to do anything remotely close to this anytime soon but it would be closer to what Apple did once upon a time and why their devices stay at the top performance wise(all things Nintendo doesn't put at the top of their priorities when designing hardware currently).

 

[Deleted member 645]

I don't think it's out of the realm of possibility for Nintendo to shell out close to a billion dollars for investing in future tech, they did so with IBM back during the GameCube's R&D days. Nintendo just as a whole seems to stray away from doing deals like that now, because it's failed them in the past. Since Switch is already established and cemented though, I think Nintendo could get more bang for their buck if they heavily invested now to establish something really cutting-edge on 5nm and could use this same design for the next 10 years.

I don't expect them to do anything remotely close to this anytime soon but it would be closer to what Apple did once upon a time and why their devices stay at the top performance wise(all things Nintendo doesn't put at the top of their priorities when designing hardware currently).

I also feel this way and really sad Nintendo probably won’t capitalize on this opportunity. Switch hybrid design is a home run. Gamers want to play high end AAA games on the go. Now is the perfect time for them to really go all out. Make the next switch cutting-edge in every was possible with a specific price point. I would say there next switch can’t be over $400.00. So make it as cutting edge as you can at the 400 price point. No gimmicks just hey this system is designed to play video games on the go and docked. Nvidia has the knowledge and tools. I shouldn’t talk about it because the possibility gets me excited and I don’t think Nintendo sees it the way I do.

 

[FrakEarth]

Sorry we heard this before. More than likely for them that means investing in something like VR. I remember after the Wii everyone was so excited. How much money Nintendo made and was going to put into their next platform hardware and specs. We got the Wii U with a controller that cost too much and they spent way more than they should have making it BC. Wii U should have been a beast. Again Nintendo never again try and make a beast console. They will always favor small form factor, efficiency, and battery life. We have to deal with that.

They tried to do something 'fresh' with the aging GCN/Wii architecture, and it probably cost them a pretty penny. I'd love to know what it actually cost them to retain that BC and ship a gamepad etc. comparative to building something like the Xbox One or PS4. I like to think they will be smarter about where they pump the money in future.

 

[Dakhil]

Now is the perfect time for them to really go all out. Make the next switch cutting-edge in every was possible with a specific price point. I would say there next switch can’t be over $400.00. So make it as cutting edge as you can at the 400 price point. No gimmicks just hey this system is designed to play video games on the go and docked. Nvidia has the knowledge and tools. I shouldn’t talk about it because the possibility gets me excited and I don’t think Nintendo sees it the way I do.

Outside of the process node rumoured to being used, which is Samsung's 8N process node, Dane, which is rumoured to be a custom variant of Orin, and features a GPU based on the Lovelace architecture, the successor to the Ampere architecture, is rumoured to be leading edge, as far as Nvidia's Arm based SoCs are concerned. In fact, I think Dane's the best Arm based SoC from Nvidia that Nintendo can use, considering Nvidia's next Arm based SoC, Atlan, won't be available until around 2025.

 

[Deleted member 1324]

Glad to find you all again (I was Euphoria on era) Love the hardware speculation!

I got an OLED to tide me over. Incredible premium feeling device and an amazing battery life upgrade over my launch Switch and Lite.

To tie into the topic do we expect this screen, dock and battery life to be the basis for the DLSS model?

Cheers!

 

[Dakhil]

Glad to find you all again (I was Euphoria on era) Love the hardware speculation!

I got an OLED to tide me over. Incredible premium feeling device and an amazing battery life upgrade over my launch Switch and Lite.

To tie into the topic do we expect this screen, dock and battery life to be the basis for the DLSS model?

Cheers!

Welcome.

I think the DLSS model* will use an OLED display like the OLED model. But I don't know if Nintendo will continue to use a 720p OLED display, like the OLED model, or if Nintendo will use a higher resolution OLED display (≥ 1080p), for the DLSS model*.

I also think the DLSS model*'s dock will be very similar to the OLED model's dock in terms of the design; and with similar, if not the same, chips being used.

As for battery life, I think it depends on how conservative or aggressive Nintendo is with Dane, in terms of CPU and GPU configurations, how high Nintendo decides to set the CPU and GPU frequencies (and perhaps RAM frequency as well if Nintendo decides to add RAM channels in Dane, like with the Apple M1).

 

[ReddDreadtheLead]

I don't think it's out of the realm of possibility for Nintendo to shell out close to a billion dollars for investing in future tech, they did so with IBM back during the GameCube's R&D days. Nintendo just as a whole seems to stray away from doing deals like that now, because it's failed them in the past. Since Switch is already established and cemented though, I think Nintendo could get more bang for their buck if they heavily invested now to establish something really cutting-edge on 5nm and could use this same design for the next 10 years.

I don't expect them to do anything remotely close to this anytime soon but it would be closer to what Apple did once upon a time and why their devices stay at the top performance wise(all things Nintendo doesn't put at the top of their priorities when designing hardware currently).

Shelling out a billion for a thing they ended up using which they stretched out for several years past it’s waypoint isn’t really the same as shelling out close to a billion for two and only using one in the end. Having an 8nm design, shelving it, and then paying upfront for the R&D costs for a design that’s similar on a much better node that is more expensive long-term than the 8nm version isn’t really… realistic even if it wasn’t Nintendo.

And it’s unlikely that they would be using the same chip for 10 years in the same way that the GCN was essentially stretched and then modified later on in the Wii U scenario but still being the original device at its heart.

They’ll presumably have a new chip by 2026 either way, so I don’t see any logical reason in doing such a thing really.

Capitalizing on the switch concept should not be taken as being more generous with their cash for anything because they can, it’s irresponsible. There’s always a smart way for this.


Besides, as mentioned, we don’t know if Samsung 8nm is similar to Samsung 11nm, and judging by the 9820 to 9825, it probably is similar and they can move it to the 7nm range which offers a die shrink and energy savings or it offers more perf if Nintendo chooses to use that for the model. We’ll have to wait and see.

Plus, it isn’t like this thing is dated outside the node which is possible considering the thing is supposedly a modified and unique version only for nVidia that became available last year for the mass market, it’s still fairly cutting edge. Based on the consumer GPUs, it’s around 11% less dense than the TSMC 7nm that they used in a MCM design.

We’ll have to see how 8N actually pans out for a mobile device like this. Could end up better than we expect

 

[ILikeFeet]

I really hope Dane can run lumen very well. it has a lot of potential

This scene is lit with 100% realtime lighting taking full advantage of Unreal Engine 5 latest advancements of Lumen & Global Illumination and Virtual Shadow Maps. There are zero Light Bakes in this scene.

click for video

https://www.artstation.com/artwork/Vg5ZeN

 

[Hermii]

Outside of the process node rumoured to being used, which is Samsung's 8N process node, Dane, which is rumoured to be a custom variant of Orin, and features a GPU based on the Lovelace architecture, the successor to the Ampere architecture, is rumoured to be leading edge, as far as Nvidia's Arm based SoCs are concerned. In fact, I think Dane's the best Arm based SoC from Nvidia that Nintendo can use, considering Nvidia's next Arm based SoC, Atlan, won't be available until around 2025.

I agree. While 8nm will be dated by the time it releases, it isn’t as bad as 20nm was for the og model.

I think that was when Nintendo really had a chance to go big. 2017 smartphone tech had cought up to current consoles in a lot of ways.

 

[NineTailSage]

Shelling out a billion for a thing they ended up using which they stretched out for several years past it’s waypoint isn’t really the same as shelling out close to a billion for two and only using one in the end. Having an 8nm design, shelving it, and then paying upfront for the R&D costs for a design that’s similar on a much better node that is more expensive long-term than the 8nm version isn’t really… realistic even if it wasn’t Nintendo.

And it’s unlikely that they would be using the same chip for 10 years in the same way that the GCN was essentially stretched and then modified later on in the Wii U scenario but still being the original device at its heart.

They’ll presumably have a new chip by 2026 either way, so I don’t see any logical reason in doing such a thing really.

Capitalizing on the switch concept should not be taken as being more generous with their cash for anything because they can, it’s irresponsible. There’s always a smart way for this.


Besides, as mentioned, we don’t know if Samsung 8nm is similar to Samsung 11nm, and judging by the 9820 to 9825, it probably is similar and they can move it to the 7nm range which offers a die shrink and energy savings or it offers more perf if Nintendo chooses to use that for the model. We’ll have to wait and see.

Plus, it isn’t like this thing is dated outside the node which is possible considering the thing is supposedly a modified and unique version only for nVidia that became available last year for the mass market, it’s still fairly cutting edge. Based on the consumer GPUs, it’s around 11% less dense than the TSMC 7nm that they used in a MCM design.

We’ll have to see how 8N actually pans out for a mobile device like this. Could end up better than we expect

Don't get me wrong I fully understand that Dane is currently in the works on 8nm and all of the wishlist stuff is just hypotheticals of what the device could be if they went a different route. The device we end up getting will undoubtedly make sense in comparison to the current Switch that is on the market, because it will look like a gradual increase in performance and not something that is a 10x generational leap...

 

[Deleted member 1108]

Given that in 5 month’s time Switch will be 5 years old, I wonder when are we going to start hearing about the successor.

 

[SiG]

Given that in 5 month’s time Switch will be 5 years old, I wonder when are we going to start hearing about the successor.

Most likely towards the latter half of 2022. Maybe to coincide with the 2022 releases, perhaps? It's very hard to say.

 

[ReddDreadtheLead]

Don't get me wrong I fully understand that Dane is currently in the works on 8nm and all of the wishlist stuff is just hypotheticals of what the device could be if they went a different route. The device we end up getting will undoubtedly make sense in comparison to the current Switch that is on the market, because it will look like a gradual increase in performance and not something that is a 10x generational leap...

Well, the device will still be a lot more powerful than the current switch, especially the CPU which can give it 7-12 times the boost range, A57 has a lot of architectural inefficiencies vs A78, even on an equivalent node and its also a lot less performant per clock.

GPU is still roughly 4 generational differences between the Maxwell in the Switch(the dev told us the documentation for the Switch is Maxwell, not Pascal, despite similarities) and the Ampere/Lovelace in the other model so there has to be improvements between the two uArch, coupled with DLSS.

And it should be twice as efficient with the memory bandwidth vs the current switch on top of having faster and more RAM.

I believe LPDDR5 is also more efficient at the same workloads than LPDDR4/4X on top of being faster due to how it channels the data.

regardless of how it is sliced it still ends up as a generational upgrade that is a huge bump over the current switch.

 

[Deleted member 1324]

Welcome.

I think the DLSS model* will use an OLED display like the OLED model. But I don't know if Nintendo will continue to use a 720p OLED display, like the OLED model, or if Nintendo will use a higher resolution OLED display (≥ 1080p), for the DLSS model*.

I also think the DLSS model*'s dock will be very similar to the OLED model's dock in terms of the design; and with similar, if not the same, chips being used.

As for battery life, I think it depends on how conservative or aggressive Nintendo is with Dane, in terms of CPU and GPU configurations, how high Nintendo decides to set the CPU and GPU frequencies (and perhaps RAM frequency as well if Nintendo decides to add RAM channels in Dane, like with the Apple M1).

Makes sense. Thanks for the reply.

 

[Deleted member 1324]

Don't get me wrong I fully understand that Dane is currently in the works on 8nm and all of the wishlist stuff is just hypotheticals of what the device could be if they went a different route. The device we end up getting will undoubtedly make sense in comparison to the current Switch that is on the market, because it will look like a gradual increase in performance and not something that is a 10x generational leap...

Won’t Dane be a 10x generational leap even with specs being kept on the conservative side?

Switch is what around 200gflops in its highest portable clock so you’d need a theoretical 2tflop GPU. But DLSS changes everything due to it sort of cheating the GPU fill rate needed to get to higher resolutions if you know what I mean so a 1.2-1.5tflop GPU is hitting the theoretical performance of a 2tflop GPU.

Not sure on the CPU side but I read many times on the era thread people expecting a 6-8x leap in compute even on the cautious side.

RAM is obviously just going up to 16gb in a best case scenario (probably 8gb imo) but Series X and PS5 also only had a 2x leap in memory.

All of this is of course for nothing if you can’t show a full generational leap on screen but that has far more to do with how long devs have to support the original Switch and how much Nintendo are prepared to spend on software development (mainly outsourcing). It would be really interesting to know how much something like Ratchet & Clank Rift Apart cost to develop (only using this as it’s a similar art style to AAA Nintendo games like Super Mario Odyssey).

*sorry for double post.

 

[Deleted member 645]

They tried to do something 'fresh' with the aging GCN/Wii architecture, and it probably cost them a pretty penny. I'd love to know what it actually cost them to retain that BC and ship a gamepad etc. comparative to building something like the Xbox One or PS4. I like to think they will be smarter about where they pump the money in future.

Well back on the old old site there was a figure given. I don’t remember the actual number but the amount they spent to customize that they could have easily had a SOC that performed with ps4 and XB1. It also didn’t help they spent so much on the controller. It was a bad design setup from day 1. Not a bad concept just a bad design.

Outside of the process node rumoured to being used, which is Samsung's 8N process node, Dane, which is rumoured to be a custom variant of Orin, and features a GPU based on the Lovelace architecture, the successor to the Ampere architecture, is rumoured to be leading edge, as far as Nvidia's Arm based SoCs are concerned. In fact, I think Dane's the best Arm based SoC from Nvidia that Nintendo can use, considering Nvidia's next Arm based SoC, Atlan, won't be available until around 2025.

We shall see. I don’t have my eggs in any basket really. I’m just hoping for the best. I do feel though that this is the perfect time with the momentum they have to really go all our with tech.

 

[ReddDreadtheLead]

Something to keep in mind is that Arm CPUs are based on reference designs that Arm sells to third parties, Nintendo can customize the A78 however they please if its the reference design, we just would not know what exactly they would change in this scenario. The help of nV +Nintendo could make it different than the standard used in flagships today.

 

[BlueManifest]

Given that in 5 month’s time Switch will be 5 years old, I wonder when are we going to start hearing about the successor.

If your talking about when is Nintendo going to talk about it I think they talk about it in January 2023 after holidays are over

 

[ILikeFeet]

If your talking about when is Nintendo going to talk about it I think they talk about it in January 2023 after holidays are over

I'm betting on summer. with more details about the dev kits late this year or winter 2022

 

[NineTailSage]

Well, the device will still be a lot more powerful than the current switch, especially the CPU which can give it 7-12 times the boost range, A57 has a lot of architectural inefficiencies vs A78, even on an equivalent node and its also a lot less performant per clock.

GPU is still roughly 4 generational differences between the Maxwell in the Switch(the dev told us the documentation for the Switch is Maxwell, not Pascal, despite similarities) and the Ampere/Lovelace in the other model so there has to be improvements between the two uArch, coupled with DLSS.

And it should be twice as efficient with the memory bandwidth vs the current switch on top of having faster and more RAM.

I believe LPDDR5 is also more efficient at the same workloads than LPDDR4/4X on top of being faster due to how it channels the data.

regardless of how it is sliced it still ends up as a generational upgrade that is a huge bump over the current switch.

The device will no doubt be more capable but I'm not expecting anything crazy... The biggest jump will be in CPU performance for sure, but it all depends on what they prioritize in such a small format factor and TDP budget. Ampere and most likely Lovelace need the bulk of the memory bandwidth so again what will they prioritize in order to get DLSS(or similar solution) to working in tandem with CPU tasks and audio(which I'm sure will get an upgrade as well).

Won’t Dane be a 10x generational leap even with specs being kept on the conservative side?

Switch is what around 200gflops in its highest portable clock so you’d need a theoretical 2tflop GPU. But DLSS changes everything due to it sort of cheating the GPU fill rate needed to get to higher resolutions if you know what I mean so a 1.2-1.5tflop GPU is hitting the theoretical performance of a 2tflop GPU.

Not sure on the CPU side but I read many times on the era thread people expecting a 6-8x leap in compute even on the cautious side.

RAM is obviously just going up to 16gb in a best case scenario (probably 8gb imo) but Series X and PS5 also only had a 2x leap in memory.

All of this is of course for nothing if you can’t show a full generational leap on screen but that has far more to do with how long devs have to support the original Switch and how much Nintendo are prepared to spend on software development (mainly outsourcing). It would be really interesting to know how much something like Ratchet & Clank Rift Apart cost to develop (only using this as it’s a similar art style to AAA Nintendo games like Super Mario Odyssey).

*sorry for double post.

I don't think that's a fair comparison, I'm looking at what will the differences be from handheld mode to handheld vs (docked to docked)...
General consensus is that it will roughly be around XboxOne S/PS4 without DLSS which still doesn't equate to 10x generational leap in raw performance for graphics. Now on the other hand because of modern GPU libraries and capabilities, I fully expect the next Switch to trounce the current model in visual fidelity but on paper you probably won't expect what the end result shows on screen.

 

[Z0m3le]

Don't get me wrong I fully understand that Dane is currently in the works on 8nm and all of the wishlist stuff is just hypotheticals of what the device could be if they went a different route. The device we end up getting will undoubtedly make sense in comparison to the current Switch that is on the market, because it will look like a gradual increase in performance and not something that is a 10x generational leap...

CPU wise is going to be around 10x the performance.

GPU is going to be 3x to 5x before DLSS which basically doubles performance, so 6x to 10x is realistic here too.

RAM should be twice the capacity at 4x the bandwidth... What isn't a generational leap here?

Storage should also be much bigger, now that the OLED model has 64GB, I think it is fair to assume we will see 128GB to 256GB, maybe even multiple SKUs based on storage size, while speed is likely to be UFS this time as they have continued to work with Samsung on RAM and Samsung has a RAM+UFS storage option which as a combo is cheaper than buying things individually.

Dane model is designed as a next generation device, Tegra X1 is a 2015 SoC made on a process node from that time, things have drastically improved since then. There is other benefits here too, flash capacity has continued to grow, meaning game storage on the game cards should launch with 64GB available at a reasonable price (likely similar to Switch's 16GB size we see with botw at launch) and 128GB should also be available, which matches Blu-ray UHD capacity while being capable of far faster read speeds, though we will have to wait and see.

 

[mariodk18]

CPU wise is going to be around 10x the performance.

GPU is going to be 3x to 5x before DLSS which basically doubles performance, so 6x to 10x is realistic here too.

RAM should be twice the capacity at 4x the bandwidth... What isn't a generational leap here?

Storage should also be much bigger, now that the OLED model has 64GB, I think it is fair to assume we will see 128GB to 256GB, maybe even multiple SKUs based on storage size, while speed is likely to be UFS this time as they have continued to work with Samsung on RAM and Samsung has a RAM+UFS storage option which as a combo is cheaper than buying things individually.

Dane model is designed as a next generation device, Tegra X1 is a 2015 SoC made on a process node from that time, things have drastically improved since then. There is other benefits here too, flash capacity has continued to grow, meaning game storage on the game cards should launch with 64GB available at a reasonable price (likely similar to Switch's 16GB size we see with botw at launch) and 128GB should also be available, which matches Blu-ray UHD capacity while being capable of far faster read speeds, though we will have to wait and see.

If they do go with UFS, what are the chances of Swane also having a micro SD slot?

 

[Dakhil]

If they do go with UFS, what are the chances of Swane also having a micro SD slot?

I assume there will still be a microSD card slot on the DLSS model* at the very least for the purpose of allowing Nintendo Switch games to be stored on the microSD card as well as load Nintendo Switch games from the microSD card.

 

[mariodk18]

I assume there will still be a microSD card slot on the DLSS model* at the very least for the purpose of allowing Nintendo Switch games to be stored on the microSD card as well as load Nintendo Switch games from the microSD card.

I would hope so. It would be similar to HDDs with current consoles I assume. Also, wasn't there talk about some alternative fast storage solution that also allowed for micro SD within the same slot for BC?

 

[ILikeFeet]

MicroSD card will still be used to load Dane games, there's no other reasonable option IMO. mainly because I don't expect SD Express to really take off. there's still a lot of room for storage speeds to increase without jumping off of EMMC/SD Cards

 

[fwd-bwd]

UFS Card reader is compatible with microSD Cards. AFAIK, Orin supports UFS.

 

[Z0m3le]

The UFS storage I'm talking about would be the internal storage. Yes I fully expect microsd card slot on the device, they will likely use internal storage for any games that require high speed storage, microsd cards can be used for anything else.

 

[Deleted member 1108]

Also please Nintendo add achievements

 

[Dakhil]

Also, wasn't there talk about some alternative fast storage solution that also allowed for micro SD within the same slot for BC?

Yes, that would be SD Express 7.0. Unfortunately, the current iteration of SD Express 7.0 cards right now can reach surface temperatures as high as 96°C! So far, the current iteration of SD Express 7.0 cards are not viable for the DLSS model*.

There's also UFS Card 3.0. However, I believe Samsung's the only company currently designing and manufacturing UFS cards. So unless Nintendo decides to aggressively adopt UFS cards as an external storage option for DLSS model* exclusive games and contact other companies (e.g. Kioxia, etc.) to manufacture UFS cards, I don't expect UFS cards to really drop in price, at least as quickly as SD/microSD cards.

~

Anyway, since I've seen people taking about wanting Nintendo and Nvidia to use leading edge process nodes to fabricate Dane, I thought I'd talk about the potential caveats of using leading edge process nodes outside of higher prices (Dr Ian Cutress has a very informative video about how much TSMC's 7 nm** process node actually costs based on various factors, and there's an issue brief from CSET about a rough estimation of the price of TSMC's process nodes from 90 nm to 5 nm** on pp. 44-45).

One potential caveat is that not only is TSMC's IPs are not compatible with Samsung's IP, process nodes begin to use EUV lithography starting from TSMC's N6 process node and Samsung's 7LPP process node, which is not compatible with DUV lithography, which Samsung's 8N process node, which is rumoured to be used to fabricate Dane, is using. So if Nintendo and Nvidia want to use a leading edge process node to fabricate Dane, then Nintendo and Nvidia would have to redesign Dane with EUV lithography in mind, which means more money and time are wasted, which I don't think would make Nintendo's and Nvidia's investors and shareholders happy.

Another potential caveat is that there's uncertainty about if Nintendo and Nvidia can secure enough capacity for leading edge process nodes needed to produce enough SoCs to minimise the impact of the global chip shortage. I don't believe secure enough capacity for leading edge process nodes is a task that can be done relatively quickly, especially with very long lead times for equipment needed for leading process nodes, as well as electronic components. Not even Apple is immune to the shortages of electronic components.
And there are also many big companies besides Nintendo and Nvidia that are trying to secure enough capacity for leading edge process nodes as well (e.g. Apple, AMD, Mediatek, etc.). So I imagine Nintendo and Nvidia would need to spend a huge amount of money to simply have a chance of securing enough capacity for leading edge process nodes before the other big companies do, which isn't even guaranteed, especially when there are big companies that have priority over other big companies.
Although I imagine Dane will have at least considerably better yields compared to the APUs for the Xbox Series X|S and the PlayStation 5 due to Dane probably having at least a considerable smaller die size compared to the APUs for the Xbox Series X|S and the PlayStation 5, and therefore the DLSS model* won't be as impacted by the global chip shortage compared to the Xbox Series X|S and the PlayStation 5, especially in the scenario when using a leading process node to fabricate Dane, I don't think the DLSS model* is immune to the effects of the global chip shortage, for the reasons I've stated above.


I'm sure there are people who disagree with me when it comes to leading edge process nodes. But I want to discuss why I think using leading edge process nodes may not be a viable choice outside of higher prices.

** purely a marketing nomenclature from all the foundry companies
~

Anyway, the next Nate the Hate episode is coming tomorrow.

 

[mariodk18]

Yes, that would be SD Express 7.0. Unfortunately, the current iteration of SD Express 7.0 cards right now can reach surface temperatures as high as 96°C! So far, the current iteration of SD Express 7.0 cards are not viable for the DLSS model*.

There's also UFS Card 3.0. However, I believe Samsung's the only company currently designing and manufacturing UFS cards. So unless Nintendo decides to aggressively adopt UFS cards as an external storage option for DLSS model* exclusive games and contact other companies (e.g. Kioxia, etc.) to manufacture UFS cards, I don't expect UFS cards to really drop in price, at least as quickly as SD/microSD cards.

~

Anyway, since I've seen people taking about wanting Nintendo and Nvidia to use leading edge process nodes to fabricate Dane, I thought I'd talk about the potential caveats of using leading edge process nodes outside of higher prices (Dr Ian Cutress has a very informative video about how much TSMC's 7 nm** process node actually costs based on various factors, and there's an issue brief from CSET about a rough estimation of the price of TSMC's process nodes from 90 nm to 5 nm** on pp. 44-45).

One potential caveat is that not only is TSMC's IPs are not compatible with Samsung's IP, process nodes begin to use EUV lithography starting from TSMC's N6 process node and Samsung's 7LPP process node, which is not compatible with DUV lithography, which Samsung's 8N process node, which is rumoured to be used to fabricate Dane, is using. So if Nintendo and Nvidia want to use a leading edge process node to fabricate Dane, then Nintendo and Nvidia would have to redesign Dane with EUV lithography in mind, which means more money and time are wasted, which I don't think would make Nintendo's and Nvidia's investors and shareholders happy.

Another potential caveat is that there's uncertainty about if Nintendo and Nvidia can secure enough capacity for leading edge process nodes needed to produce enough SoCs to minimise the impact of the global chip shortage. I don't believe secure enough capacity for leading edge process nodes is a task that can be done relatively quickly, especially with very long lead times for equipment needed for leading process nodes, as well as electronic components. Not even Apple is immune to the shortages of electronic components.
And there are also many big companies besides Nintendo and Nvidia that are trying to secure enough capacity for leading edge process nodes as well (e.g. Apple, AMD, Mediatek, etc.). So I imagine Nintendo and Nvidia would need to spend a huge amount of money to simply have a chance of securing enough capacity for leading edge process nodes before the other big companies do, which isn't even guaranteed, especially when there are big companies that have priority over other big companies.
Although I imagine Dane will have at least considerably better yields compared to the APUs for the Xbox Series X|S and the PlayStation 5 due to Dane probably having at least a considerable smaller die size compared to the APUs for the Xbox Series X|S and the PlayStation 5, and therefore the DLSS model* won't be as impacted by the global chip shortage compared to the Xbox Series X|S and the PlayStation 5, especially in the scenario when using a leading process node to fabricate Dane, I don't think the DLSS model* is immune to the effects of the global chip shortage, for the reasons I've stated above.


I'm sure there are people who disagree with me when it comes to leading edge process nodes. But I want to discuss why I think using leading edge process nodes may not be a viable choice outside of higher prices.

** purely a marketing nomenclature from all the foundry companies
~

Anyway, the next Nate the Hate episode is coming tomorrow.

UFS 3.0? Weren’t we discussing 2.0 recently? How much better and mite expensive is it?

I think it’s likely IF Nintendo does go with UFS or something similar they require installation. And to get a $399 price point they may go for 128 GB with 256 at $449? Also can game card speeds be as fast as UFS?

 

[ReddDreadtheLead]

UFS Card reader is compatible with microSD Cards. AFAIK, Orin supports UFS.

really? where did you find this?

it reduces a lot of the worry over storage options.