StarTopic Future Nintendo Hardware & Technology Speculation & Discussion |ST| (New Staff Post, Please read)

[ArchedThunder]

That's not true, at least not for Switch. Games can run fine with data spread across system storage and Game Card, that's how updates and DLC work with game cards at all.

Fair, but that’s not really the same as installing a portion of the game to speed up load times.

 

[Concernt]

That’s not really the same as installing a portion of the game to speed up load times though.

It really is, though. Kirby Star Allies literally installs a portion of the game to the system storage. Well, not the SAME, but definitely comparable.

All that said, I don't get the consternation over Game Cards. If Nintendo wants the storage to be faster, then they'll probably make the Game Cards faster too. It's just not something I'd worry about.

 

[ArchedThunder]

It really is, though. Kirby Star Allies literally installs a portion of the game to the system storage. Well, not the SAME, but definitely comparable.

All that said, I don't get the consternation over Game Cards. If Nintendo wants the storage to be faster, then they'll probably make the Game Cards faster too. It's just not something I'd worry about.

Wait, really? I don’t see anything from a quick google search and the box doesn’t mention anything about required storage space.

 

[Concernt]

Wait, really? I don’t see anything from a quick google search and the box doesn’t mention anything about required storage space.

Kirby Star Allies got nearly its own weight worth of post-launch updates including new story stuff, new levels, new characters, etc. They all integrate pretty seemlessly and mix stuff between the free DLC and main game regularly. Also Splatoon 2 where most of the stages are not on the game card.

 

[ArchedThunder]

Kirby Star Allies got nearly its own weight worth of post-launch updates including new story stuff, new levels, new characters, etc. They all integrate pretty seemlessly and mix stuff between the free DLC and main game regularly. Also Splatoon 2 where most of the stages are not on the game card.

That’s not installing a portion of the game, that’s new content added post launch.

 

[Concernt]

That’s not installing a portion of the game, that’s new content added post launch.

New content that forms a portion of the game.

The point is the same, Switch already supports games being split across the game card and system storage.

 

[Skittzo]

New content that forms a portion of the game.

The point is the same, Switch already supports games being split across the game card and system storage.

You're describing DLC in general here...

 

[ArchedThunder]

New content that forms a portion of the game.

The point is the same, Switch already supports games being split across the game card and system storage.

That isn’t the same as installing a portion of the data from the cart to speed up load times.

 

[Maxrunner]

It's largely that. CMA looks like they'll approve it and FTC won't have a valid reason to block on its own. The fact they are removing Nintendo from the discussions and trying to reshape the case to suit their agenda is becoming too transparent.

Whats their agenda? Liking Playstation ?

 

[Concernt]

You're describing DLC in general here...

Correct. If a system that can play games off removable media supports DLC, it can run a game across both its internal storage and something else. That's the point

Even Wii could do it.

 

[ArchedThunder]

Correct. If a system that can play games off removable media supports DLC, it can run a game across both its internal storage and something else. That's the point

Even Wii could do it.

Whether or not it can isn’t the subject, it’s whether it will have games install portions of the data from the cart to speed up load times.

 

[Concernt]

Whether or not it can isn’t the subject, it’s whether it will have games install portions of the data from the cart to speed up load times.

It could, maybe. It probably won't, though.

 

[ILikeFeet]

Dynamic (input) resolution support was added in DLSS 2.1. Normally, for a given target resolution and perf/quality mode, you're supposed to query the "optimal settings" for input resolution. So you say you want 4K via Performance mode, and the library will tell you it needs a 1080p input to do that. But when dynamic res is supported, it gives you minimum and maximum allowable input resolutions, in addition to the optimal value.

* Hidden text: cannot be quoted. *

DLSS scaling factors are arbitrary. You can do whatever you want, including 1/81 scale to 4k (240p)

 

[Concernt]

DLSS scaling factors are arbitrary. You can do whatever you want, including 1/81 scale to 4k (240p)

Call it sick curiosity, but I want to see what that would look like.

 

[Krikeno]

I hope that with the new model the games can occupy more than 15 GB. It is the worst thing of the Switch and the cartridges.

 

[Cybergatuno]

Didn't Just Dance 2020 release on the Wii?

How many years a platform gets releases ultimately depends on how active the install base remains. Third-parties will follow the money if it justifies the investment and indies would be foolish not to target the Switch for many years. Ambitious AAA games will skip the Switch, but they already do.

I do believe Nintendo wants a quick transition. OG Switch was hacked to the core and ROMs are dumped online weeks before release dates. But most games they publish are quick AA cash-grabs that don't need new hardware. Games like Mario Sports, Warioware, Kirbies, Brain Age, ports, etc could still make a lot of money on Switch 1 for years to come. Assuming there is backwards compatibility and each game offers a superior version on Switch 2, users are able to upgrade when they want without being completely locked-out. The Switch 2 may be supply-constrained for many years.

In the end, I don't think they currently have a strong opinion. They'll just wait for sales data and adapt accordingly.

 

[kueoe]

What? I actually like remakes. A lot of those games are games I missed because I'm too poor to have a console back when I was a kid.

In my opinion remakes are only acceptable when they are a reimagining, because then it becomes it own thing like FF7R or Pokémon Legends Arceus. Even then I still think it's just an excuse to just reuse ideas/concepts that are safe and proven rather than taking creative risks and venturing into the unknown, creating something new that disrupts the industry or introduces new ideas.

I'd rather they create an environment where games aren't left tied to an arbitrary piece of hardware that we all know will inevitably be replaced when the time comes. That way remakes are never needed. Everyone used to make the joke about remakes of remakes but it's already happening with Pokémon LGPE etc... This is why PC will always be king because in one shape or another everything comes to PC eventually.

There will indeed be many Switch games that will in 20 years time are only playable on the Switch line. It's just wasteful outside of instances where the experience simply cannot be replicated without the original hardware (imo Wii, DS and 3DS are good examples - even then you can use a Wiimote on Dolphin with 2 IR sensors and it works great)

 

[kueoe]

Didn't Just Dance 2020 release on the Wii?

How many years a platform gets releases ultimately depends on how active the install base remains. Third-parties will follow the money if it justifies the investment and indies would be foolish not to target the Switch for many years. Ambitious AAA games will skip the Switch, but they already do.

I do believe Nintendo wants a quick transition. OG Switch was hacked to the core and ROMs are dumped online weeks before release dates. But most games they publish are quick AA cash-grabs that don't need new hardware. Games like Mario Sports, Warioware, Kirbies, Brain Age, ports, etc could still make a lot of money on Switch 1 for years to come. Assuming there is backwards compatibility and each game offers a superior version on Switch 2, users are able to upgrade when they want without being completely locked-out. The Switch 2 may be supply-constrained for many years.

In the end, I don't think they currently have a strong opinion. They'll just wait for sales data and adapt accordingly.

Pretty accurate.

 

[Concernt]

Didn't Just Dance 2020 release on the Wii?

How many years a platform gets releases ultimately depends on how active the install base remains. Third-parties will follow the money if it justifies the investment and indies would be foolish not to target the Switch for many years. Ambitious AAA games will skip the Switch, but they already do.

I do believe Nintendo wants a quick transition. OG Switch was hacked to the core and ROMs are dumped online weeks before release dates. But most games they publish are quick AA cash-grabs that don't need new hardware. Games like Mario Sports, Warioware, Kirbies, Brain Age, ports, etc could still make a lot of money on Switch 1 for years to come. Assuming there is backwards compatibility and each game offers a superior version on Switch 2, users are able to upgrade when they want without being completely locked-out. The Switch 2 may be supply-constrained for many years.

In the end, I don't think they currently have a strong opinion. They'll just wait for sales data and adapt accordingly.

Did you just call Kirby "quick AA cashgrabs"?

 

[Cybergatuno]

Did you just call Kirby "quick AA cashgrabs"?

OK, Kirby and the Forgotten Land is pretty amazing and the next 2D port seems like it has a lot of love poured into. Definitely not a cash-grab but not exactly AAA system-seller either
It's the kind of family-friendly games that make sense to release on legacy platforms.

 

[LiC]

DLSS scaling factors are arbitrary. You can do whatever you want, including 1/81 scale to 4k (240p)

If you have a single input resolution, it can be anything. But if you're using dynamic scaling, then your input must be within 50% to 100% of the target resolution by axis.

So you can choose a custom input value lower than 50% (or use Ultra Performance where it's normally 33%) at the start, which will cause DRS to be turned off. Or, if you choose a valid value for DRS, but then later you pass input which is smaller when evaluating DLSS, it errors out.

 

[Dakhil]

---

 

[Maxrunner]

---

i've read something about N3 for samsung related. Some kind of improvement in the node.

 

[kalosmani]

There is huge evidence. Online stores have existed for 15 years, we have ample data. Besides, it's basic sampling data. Gamers who don't upgrade their hardware are either high cost conscious or less invested as gamers. People still on the older platform are almost by definition the people less likely to spend money on games.




There are plenty of reasons to stop supporting the base Switch. There are costs associated with continuing to maintain the old digital store front and printing the older cartridges. Maintaining the store front will also require continuing to support the base OS to keep it security patched and able to connect to modern servers.

Maintaining an old platform costs money, there is a reason Nintendo was so excited to nuke the Wii U shop for orbit.


The performance of the base system is somewhat irrelevant. Indie games are, for the most part, not ported by indie developers, who can't get access to devkits on their own. They're ported by publishers. Every platform added to the list of platforms supported increases the cost for the publisher, decreasing their profit margin.

Yes, games will continue to release on old platforms. The NES and the Gameboy continue to get games from indies to this day. No one would call them robust living platforms, however.


You're going to see it. It will stop the moment Sony wants it to stop, because they'll stop allowing it by shutting down submissions to the store. Sony has huge incentive to do so, because it will push games onto PS5, which increases sales of the hardware.





None of these games are "2D" games. They all run on 3D engines, Animal Crossing is no more 2D than Sackboy.

Of course, that doesn't matter. Ori and the Blind Forest is a "2D" game by this definition, and it's an absolute beast.


Nintendo is pushing their hardware harder than Sony is pushing theirs. That's why Nintendo games look so good despite the limited power of the console. The idea that Nintendo games are easy on the hardware just isn't factual.



The cross gen period is long this gen, for a lot of reasons. That doesn't mean it won't end.

Comparing to previous platforms is pointless. The transition from PS3 to PS4 won't be the same as PS4 to PS5. PS3 and 360 were very early systems that made digital games popular. However the PS4 store I feel has standardized the ecosystem. I don't think PS4 will ever be decommissioned like you think. It will be a platform like Steam, iOS, Android and the software will now uses the same architecture unlike previous generations. The boom of indie games also is something that got started with PS3 and has become even more mainstream with PS4. I don't think the PS4 store will ever be decomssuoned and as long games are able to run on PS4 and the sales of software haven't dropped considerably then games will continue to come out on PS4.

The same logic applies to Switch, indie games will continue to come out on Switch years from now because they can easily run on Switch and shares the same architecture with the successor. When I meant 2D games I meant games with less ambition and not 3D open world. Games like Animal Crossing and 2D Zelda use a birds eye view, these games can easily be scaled down to Switch 1. 2D Metroid, NSMB and etc as side scrollers and can also be easily scaled down to Switch 1. Then you have AA games like WarioWare, Kirby and etc that can easily be scaled down. I do think there will be a transition where eventually software sales will dry up but I am predicting this won't happen for many years, 4 years minimum. Same way I believe we will get indie games for another 2 years on PS4 hell maybe even until PS6 accouncement.

You said Sony will eventually stop submissions for PS4 games. The last PS3 game was Shakedown Hawai which released at the end of 2020. That's 14 years of games allowed to be submitted to PS3 which is a considerable amount of time. Which lines up with my prediction as the generation will end after the successor of the successor is announced.

 

[JoshuaJSlone]

Isn’t it the case more pixels that DLSS has to make (proportionally to the original res)the worse it gets, so even if they had a custom version that supported scaling from something like 480p to 4k wouldn’t the artifacts be much worse than 1440p?

I've seen a few YouTube videos that take things beyond Ultra Performance, but it's usually to some ridiculous amount where it's a lark. Like taking Control from 128x72 -> 2560x1440. Considering that's starting out with less than half the resolution of the original Game Boy, though, it actually comes out fairly decent until it goes completely bananas.


Here's another video on a simple animated model that goes beyond Ultra Performance to show off 4x4 and 6x6 scaling as well.

 

[ILikeFeet]

If you have a single input resolution, it can be anything. But if you're using dynamic scaling, then your input must be within 50% to 100% of the target resolution by axis.

So you can choose a custom input value lower than 50% (or use Ultra Performance where it's normally 33%) at the start, which will cause DRS to be turned off. Or, if you choose a valid value for DRS, but then later you pass input which is smaller when evaluating DLSS, it errors out.

as far as I know, from DLSS documentation, you can set your own input values. don't recall any hard turn-off with scaling values. I'll have to check when I get home

 

[Deleted member 887]

Comparing to previous platforms is pointless. The transition from PS3 to PS4 won't be the same as PS4 to PS5. PS3 and 360 were very early systems that made digital games popular. However the PS4 store I feel has standardized the ecosystem. I don't think PS4 will ever be decommissioned like you think. It will be a platform like Steam, iOS, Android and the software will now uses the same architecture unlike previous generations. The boom of indie games also is something that got started with PS3 and has become even more mainstream with PS4. I don't think the PS4 store will ever be decomssuoned and as long games are able to run on PS4 and the sales of software haven't dropped considerably then games will continue to come out on PS4.

The same logic applies to Switch, indie games will continue to come out on Switch years from now because they can easily run on Switch and shares the same architecture with the successor. When I meant 2D games I meant games with less ambition and not 3D open world. Games like Animal Crossing and 2D Zelda use a birds eye view, these games can easily be scaled down to Switch 1. 2D Metroid, NSMB and etc as side scrollers and can also be easily scaled down to Switch 1. Then you have AA games like WarioWare, Kirby and etc that can easily be scaled down. I do think there will be a transition where eventually software sales will dry up but I am predicting this won't happen for many years, 4 years minimum. Same way I believe we will get indie games for another 2 years on PS4 hell maybe even until PS6 accouncement.

You said Sony will eventually stop submissions for PS4 games. The last PS3 game was Shakedown Hawai which released at the end of 2020. That's 14 years of games allowed to be submitted to PS3 which is a considerable amount of time. Which lines up with my prediction as the generation will end after the successor of the successor is announced.

I’m not sure what your original assertion is, and I’m still not sure. I’m not following your train of thought, I was just responding to the specific factual errors in your previous post.

Generations lasting longer, with fuzzy lines between them is a reasonable extrapolation. Never ending platform support is not.

 

[LiC]

as far as I know, from DLSS documentation, you can set your own input values. don't recall any hard turn-off with scaling values. I'll have to check when I get home

The current 2.4 documentation does reference this limitation, although it doesn't specifically say there's a hard minimum or that it's 50% (some other documentation does say that "custom" scaling is within 50% to 100%).

 

[Pokemaniac]

Hence me saying there will probably be a cross gen period of around 2 years, which is pretty in line with Nintendo’s support of old consoles. Keep in mind this whole discussion started from someone suggesting Nintendo would continue to support the Switch with games until a 3rd generation Switch released, and then another said there wouldn’t be many big exclusives for Switch 2 until 3 or 4 years in.

It's been ages since Nintendo has had a system that's persistently successful for this long. Support for systems like the Famicom and Super Famicom (the latter of which continued to receive first party games in Japan throughout nearly the entirety of the N64's life) might be more relevant points of comparison. And this will be in an environment where any game that they continue releasing on Switch will be functionally cross gen, whether it gets explicitly ported or not.

 

[kalosmani]

I’m not sure what your original assertion is, and I’m still not sure. I’m not following your train of thought, I was just responding to the specific factual errors in your previous post.

Generations lasting longer, with fuzzy lines between them is a reasonable extrapolation. Never ending platform support is not.

Let me clarify. I believe Switch 1 will get software support from Nintendo AA software and Indies until Switch 3 is announced. Unlike previous Nintendo generations where their last system gets phased out after 1 year into the new generation.

 

[C_Ali]

Let me clarify. I believe Switch 1 will get software support from Nintendo AA software and Indies until Switch 3 is announced. Unlike previous Nintendo generations where their last system gets phased out after 1 year into the new generation.

Given they are only supporting the Switch family of systems, they are in a prime position to do so. Whether they will remains to be seen but I do believe they have the capability now.

 

[Deleted member 887]

Let me clarify. I believe Switch 1 will get software support from Nintendo AA software and Indies until Switch 3 is announced. Unlike previous Nintendo generations where their last system gets phased out after 1 year into the new generation.

Ah. Yeah, I’m not sure how to evaluate this statement since were not sure how Nintendo plans on positioning Drake, much less that there will be a switch 3, or how the cloud market might change in the next 5 years.

If I had to bet money, I’d bet against Nintendo developing games for the base Switch into 2029. But fortunately I’m not betting money.

 

[NintenDuvo]

I just had a crazy thought, what if Nate turns out to be the SuperMetalDave of this generation , for those not in the know SuperMetalDave was this "insider" during the Nintendo NX days when we knew nothing outside of a codename and used to always provide us tidbits here and there, the only reason we and especially me use to trust him because it turned out he did have one good source within Nintendo but the idiot completely threw that source under the bus exposing way too much that couldve got said source fired by revealing marketing materials and codenames. Anyways the Nintendo NX SuperMetalDave told us about would have teraflops that would match the Xbox One and that AMD was the front runner to provide the SOC. I used to hang on to this guy every word like i do Nate until the day of the Switch reveal and he did a live reaction where he had nothing but major egg on his face looking all kinds of befuddled,then had the nerve to diss it for just being a tablet. That was the last time i ever watched that clowns content

 

[D36]

I just had a crazy thought, what if Nate turns out to be the SuperMetalDave of this generation , for those not in the know SuperMetalDave was this "insider" during the Nintendo NX days when we knew nothing outside of a codename and used to always provide us tidbits here and there, the only reason we and especially me use to trust him because it turned out he did have one good source within Nintendo but the idiot completely threw that source under the bus exposing way too much that couldve got said source fired by revealing marketing materials and codenames. Anyways the Nintendo NX SuperMetalDave told us about would have teraflops that would match the Xbox One and that AMD was the front runner to provide the SOC. I used to hang on to this guy every word like i do Nate until the day of the Switch reveal and he did a live reaction where he had nothing but major egg on his face looking all kinds of befuddled,then had the nerve to diss it for just being a tablet. That was the last time i ever watched that clowns content

Deleted

 

[chocolate_supra]

.

 

[D36]

Yikes, Nintenduvo got threadbanned. I didn't think their post was THAT disruptive. Though, I guess it can come off as such, since it was referring to Nate.

 

[Deleted member 887]

Nate's podcast partner is a certified Nintendo Developer, and Known Technically Competent Person.

Leaking information that is potentially source burning would put MVG's career at risk. Meanwhile, MVG is clearly willing to disagree with Nate, and knows what he's talking about. The odds that Nate is source burning or wildly fabricating is pretty small. Not impossible, obviously, but when guys like MVG and John Linneman are on the podcast should be an indicator.

Of course it's possible that Nate's source is wrong, that Nate's interpretation is wrong, or some game of telephone is occurring with anything insidery he says, but SuperMetalDave he is not.

 

[Thugstas]

Nate's podcast partner is a certified Nintendo Developer, and Known Technically Competent Person.

Leaking information that is potentially source burning would put MVG's career at risk. Meanwhile, MVG is clearly willing to disagree with Nate, and knows what he's talking about. The odds that Nate is source burning or wildly fabricating is pretty small. Not impossible, obviously, but when guys like MVG and John Linneman are on the podcast should be an indicator.

Of course it's possible that Nate's source is wrong, that Nate's interpretation is wrong, or some game of telephone is occurring with anything insidery he says, but SuperMetalDave he is not.

I think i missed something

 

[Deleted member 887]

Yikes, Nintenduvo got threadbanned. I didn't think their post was THAT disruptive. Though, I guess it can come off as such, since it was referring to Nate.

I wonder if there was a pattern there? To be honest, my first read of the post I thought he was calling Nate a clown, but upon reread he clearly isn't.

I'm glad there is more cleanup in this thread, but I'm not sure it's 100% clear where the new lines are

 

[D36]

I wonder if there was a pattern there? To be honest, my first read of the post I thought he was calling Nate a clown, but upon reread he clearly isn't.

I'm glad there is more cleanup in this thread, but I'm not sure it's 100% clear where the new lines are

They did suggest that Nate could someone who is notorious exposing his source and being fraudulent, despite evidence of the contrary on Nate's part. I just think they were being careless.

 

[Thraktor]

I am fairly convinced that the "delay" actually is caused by a node shrink. It's just a node shrink that decided 14+ months ago, along with all the decisions that come with.

The wild "Drake is node shrunk Dane" theory neatly explains almost every weird bit of info we've heard.

Nintendo wants a more powerful Switch, they go to Nvidia, take a look at the Orin plans and see a DLSS demo. A 2x power increase is inline with the last three Nintendo generations, and 2-3x more power plus 4K hardware was the exact strategy for the midgen consoles from Sony and Microsoft.

Dane is exactly the device you would have expected if you had asked @Thraktor for predictions at the time. Its 4SMs, decent clock speed, it's got tensor cores for DLSS, it even has RT cores.

Devkits are out in 2021 with devs told to "make their games 4k ready". According to Nate, the upgrade is a 4k focused pro model, based on what he's heard from developers, and he expects a "real" next gen a few years later. 2022-2023 window seems consistent from multiple sources.

The problem is Dane sucks. 4 tensor cores at mobile clock speeds is barely enough power for DLSS 2. 4 Ampere SM is a solid 2.5x power upgrade over Switch, but you can't do anything with that extra power, because you spend it all trying to get frame time down to let DLSS 2 run. The RT is vestigial at that point. But pushing clock speeds or SMs any higher, and the battery life and heat get out of control.

Nvidia has its own challenges. The GPU market tanks, and it's clear that Lovelace is gonna be pricey. Ampere is going to stay around for a while. Meanwhile, the Tegra team that staffed up in 2019 has nothing to do - Atlan has been cancelled, and its replacement, Thor, has to wait until Blackwell's design is finalized before they can really get to work.

New Plan: Drake. The Tegra team will take Dane, triple it, and die shrink it. It'll be Nintendo's true next gen upgrade, and it will give Nvidia a second 5nm product they can use to buffer demand. Nvidia makes TSMC 5nm purchases in December of 2021, and the in-dev NVN2 is updated to use Drake's SOC information. Only a vestigial reference to Dane remains.

Which leaks in Spring of this year. Insiders and observant outsiders are all agog. How do they fit 12 SMs on Samsung 8nm? Where is the modest 4k Switch update? Surely something is up. Half the SMs in handheld mode? No CPU upgrade?

These ideas keep getting shot down as more and more info comes out. Meanwhile, some devs are talking. Reports of "PS4 plus DLSS" are coming out, matching Drake. Insiders are starting to hear about changes, but after the OLED debacle, no one is willing to talk about it until every possible avenue is covered.

In October partners are briefed and it is ... mixed. The device is now Switch 2, the next generation and will receive a huge marketing push. And a longer one. Nintendo doesn't want to repeat the short marketing cycle of the Wii U, they want to thoroughly sell the device, and prepare users to upgrade. They need 6-8 months to do that, plus they need to get their software library into shape.

DK, Mario, and any other unannounced game that wasn't finished are now all Drake exclusives. Everything else needs a gorgeous cross-gen upgrade. Every second party studio without anything else to do has been handed a 4k remake or remaster. This will all take time. Nintendo is planning to announce in early 2023, and launch by the holiday.

Nintendo will be fine, the Switch is still selling okay, and software sales are solid. But third parties are burned, and after a major rethink and delay, they're not trusting Nintendo's late 2023 timeline. Some start to talk, and those leaks hit us...

There's one big flaw in your reasoning here; the idea that Nintendo's original plans resembled my speculation at the time. If there's one thing Nintendo have committed to time and time again, it's that they'll never do anything I predict! Or maybe they saw my posts on the matter, and after realising I had predicted what hardware they were using they immediately changed their plans completely just to make sure I was wrong. That seems plausible.

Seriously, though, I'm a bit late to respond to this, but I've been thinking about it, and I'm coming around to the possibility of the cancelled device being based on an older version of the T239. Mainly because an overclocked Mariko would be a bit of a crap mid-gen upgrade, and there aren't really many other alternatives. It would require certain assumptions, though.

Firstly, it would require the Drake/T239 we see now (let's call it T239-12 for 12 SMs) to be, from Nvidia's perspective, the same chip as the older version planned for the mid-gen upgrade (let's say T239-6, assuming 6 SMs). That is, they've been calling it T239 the entire time, and even though the specifications, and possibly even the manufacturing process, have changed, they consider it just part of the development of a single chip. If this weren't the case, we would see references to T239-6 in the Nvidia hack, which by all accounts we don't. Personally I think this would be unusual, but it's not impossible.

It would simultaneously require Nintendo to consider the planned T239-6 (which I'll call Switch+) based Switch to be a different device from the upgraded T239-12 based Switch (ie Switch 2). Or, at least, communicate as much to developers. Hence "cancelled" rather than delayed.

I also would seem unlikely to me that any serious development was done for Switch+ before being cancelled. Part of my thinking here is that T239-6 obviously didn't tape out (otherwise it's definitely not the came chip as T239-12), which means Switch+ would have had to be cancelled at least a year before its intended launch, possibly more. Ports typically don't have multi-year development timelines, so it's possible any SDKs in the hands of third parties (ie software development kits designed to run on PCs with Ampere GPUs) were for evaluation and very early development, rather than full-scale productions. This would also explain a lack of public developer outrage resulting from the change of plans, as no significant work was really done.

Most importantly, it would require something to have changed regarding Nintendo's plans and expectations for T239-6 and Switch+ that couldn't have been predicted when they started the project. As I've said before, I don't think the clock speeds and thermal performance of Ampere on 8nm would have been any surprise to Nintendo or Nvidia, that would have been very predictable from the start of the project. One possibility here is that the hardware wasn't a surprise, but the software. Specifically, I wonder if Nvidia over-sold DLSS 2 a bit, and promised that they could get a version of DLSS 2 capable of generating a 4K output with just 6 SMs worth of Ampere tensor cores. Perhaps they couldn't achieve this without significant image quality loss (defeating the purpose of DLSS somewhat) and that forced a change of plan.

There is another scenario that I've though about which might explain things. Namely, that Nintendo was planning a new console to succeed the Switch, with a completely different form-factor or focus. Thinking about the first year or two of the Switch's life, while the console was pretty successful, there was no guarantee that it would go on to become one of the top selling consoles of all time. Interest could well have waned after a few years, and it wouldn't have been safe for Nintendo to assume they could release a simple Switch 2 and see it succeed.

So, let's say Nintendo was planning something different after the Switch. Rolling the dice again on a new form-factor, and not necessarily backwards-compatible with the Switch or using similar technology at all. It would have been in the relatively early stages of planning, and conscious of how risky it would have been, they decided to hedge their bets, "third pillar" style, with the Switch+. It would release a couple of years before this new form-factor console, with the capability to play Switch games at 4K, and more advanced capabilities like ray-tracing. If the new form-factor console is a failure, they could fall back on the Switch+, relying on its more advanced technology than the old TX1 to get them through a few years before something new can be brought out. If the new form-factor is a success, the Switch+ is treated more like the DSi or n3DS.

This would have been the plan in 2019 or early 2020, but after 2020 and into 2021 it would have become clear that the Switch was more than just a modest success. On the one hand they no longer saw the need for a mid-gen Switch+ to maintain interest, and on the other hand they decided to follow up with a much more straight-forward Switch 2 rather than the experimental form-factor they were previously considering. So, rather than starting from scratch on a Switch 2, they took what they already had been working on for the Switch+ and updated that to fit their requirements for Switch 2. This would push manufacturing of the chip back by a year or more, but would be a lot quicker (and cheaper) than starting from scratch, and would still allow them to release a follow-up to the Switch around the 6-7 year mark, before sales drop significantly.

As another possibility, what if the cancelled mid-gen upgrade that Digital Foundry are talking about never made it past the planning stages? That is, early on in Switch's life they considered the possibility of releasing a mid-gen upgrade and talked to some third parties about it. No chip was made for it, and no meaningful software development work was ever done for it, and they shelved the idea. Third party developers know about it, and know of its cancellation, because Nintendo consulted them, but it never existed outside of planning documents inside Nintendo.

 

[Hocot8]

They did suggest that Nate could someone who is notorious exposing his source and being fraudulent, despite evidence of the contrary on Nate's part. I just think they were being careless.

It was really just a random damn comment lmao. So unnecessary and needless.

 

[MP!]

My biggest worry is that insiders are going to continue to tease and titillate and we the regular thread goers will be the ones being moderated when the thread spirals out of control, and we try and save it.

That being said SMD clearly was mislead and completely wrong lol
Nate isn't like that at all
man what a time it was pre Switch launch

 

[kvetcha]

I wonder if there was a pattern there? To be honest, my first read of the post I thought he was calling Nate a clown, but upon reread he clearly isn't.

I'm glad there is more cleanup in this thread, but I'm not sure it's 100% clear where the new lines are

This will include posts such as personal attacks, attempts at “gotchas” for previous, incorrect predictions, and any post that is disruptive to this thread’s purpose of discussion of future Nintendo hardware.

I'm guessing it's that last part. Coming in here with 'haha, what if nate was a notorious fraud' isn't exactly conducive to the discussion at hand.

 

[Pokemaniac]

There's one big flaw in your reasoning here; the idea that Nintendo's original plans resembled my speculation at the time. If there's one thing Nintendo have committed to time and time again, it's that they'll never do anything I predict! Or maybe they saw my posts on the matter, and after realising I had predicted what hardware they were using they immediately changed their plans completely just to make sure I was wrong. That seems plausible.

Seriously, though, I'm a bit late to respond to this, but I've been thinking about it, and I'm coming around to the possibility of the cancelled device being based on an older version of the T239. Mainly because an overclocked Mariko would be a bit of a crap mid-gen upgrade, and there aren't really many other alternatives. It would require certain assumptions, though.

Firstly, it would require the Drake/T239 we see now (let's call it T239-12 for 12 SMs) to be, from Nvidia's perspective, the same chip as the older version planned for the mid-gen upgrade (let's say T239-6, assuming 6 SMs). That is, they've been calling it T239 the entire time, and even though the specifications, and possibly even the manufacturing process, have changed, they consider it just part of the development of a single chip. If this weren't the case, we would see references to T239-6 in the Nvidia hack, which by all accounts we don't. Personally I think this would be unusual, but it's not impossible.

It would simultaneously require Nintendo to consider the planned T239-6 (which I'll call Switch+) based Switch to be a different device from the upgraded T239-12 based Switch (ie Switch 2). Or, at least, communicate as much to developers. Hence "cancelled" rather than delayed.

I also would seem unlikely to me that any serious development was done for Switch+ before being cancelled. Part of my thinking here is that T239-6 obviously didn't tape out (otherwise it's definitely not the came chip as T239-12), which means Switch+ would have had to be cancelled at least a year before its intended launch, possibly more. Ports typically don't have multi-year development timelines, so it's possible any SDKs in the hands of third parties (ie software development kits designed to run on PCs with Ampere GPUs) were for evaluation and very early development, rather than full-scale productions. This would also explain a lack of public developer outrage resulting from the change of plans, as no significant work was really done.

Most importantly, it would require something to have changed regarding Nintendo's plans and expectations for T239-6 and Switch+ that couldn't have been predicted when they started the project. As I've said before, I don't think the clock speeds and thermal performance of Ampere on 8nm would have been any surprise to Nintendo or Nvidia, that would have been very predictable from the start of the project. One possibility here is that the hardware wasn't a surprise, but the software. Specifically, I wonder if Nvidia over-sold DLSS 2 a bit, and promised that they could get a version of DLSS 2 capable of generating a 4K output with just 6 SMs worth of Ampere tensor cores. Perhaps they couldn't achieve this without significant image quality loss (defeating the purpose of DLSS somewhat) and that forced a change of plan.

There is another scenario that I've though about which might explain things. Namely, that Nintendo was planning a new console to succeed the Switch, with a completely different form-factor or focus. Thinking about the first year or two of the Switch's life, while the console was pretty successful, there was no guarantee that it would go on to become one of the top selling consoles of all time. Interest could well have waned after a few years, and it wouldn't have been safe for Nintendo to assume they could release a simple Switch 2 and see it succeed.

So, let's say Nintendo was planning something different after the Switch. Rolling the dice again on a new form-factor, and not necessarily backwards-compatible with the Switch or using similar technology at all. It would have been in the relatively early stages of planning, and conscious of how risky it would have been, they decided to hedge their bets, "third pillar" style, with the Switch+. It would release a couple of years before this new form-factor console, with the capability to play Switch games at 4K, and more advanced capabilities like ray-tracing. If the new form-factor console is a failure, they could fall back on the Switch+, relying on its more advanced technology than the old TX1 to get them through a few years before something new can be brought out. If the new form-factor is a success, the Switch+ is treated more like the DSi or n3DS.

This would have been the plan in 2019 or early 2020, but after 2020 and into 2021 it would have become clear that the Switch was more than just a modest success. On the one hand they no longer saw the need for a mid-gen Switch+ to maintain interest, and on the other hand they decided to follow up with a much more straight-forward Switch 2 rather than the experimental form-factor they were previously considering. So, rather than starting from scratch on a Switch 2, they took what they already had been working on for the Switch+ and updated that to fit their requirements for Switch 2. This would push manufacturing of the chip back by a year or more, but would be a lot quicker (and cheaper) than starting from scratch, and would still allow them to release a follow-up to the Switch around the 6-7 year mark, before sales drop significantly.

As another possibility, what if the cancelled mid-gen upgrade that Digital Foundry are talking about never made it past the planning stages? That is, early on in Switch's life they considered the possibility of releasing a mid-gen upgrade and talked to some third parties about it. No chip was made for it, and no meaningful software development work was ever done for it, and they shelved the idea. Third party developers know about it, and know of its cancellation, because Nintendo consulted them, but it never existed outside of planning documents inside Nintendo.

A Mariko-based Switch Pro probably would have been a bit disappointing, but it's worth remembering we are talking about a cancelled system here. Would it really be that unreasonable that started work on such a system, to the point where some developers probably got briefed on it, but decided to cancel it fairly early specifically because the chip, while fine for the Switch Lite, didn't scale up how they wanted?

 

[Goodtwin]

Of course it's possible that Nate's source is wrong, that Nate's interpretation is wrong, or some game of telephone is occurring with anything insidery he says, but SuperMetalDave he is not.

From what I can tell there are two groups as it pertains to Nate. There is a group that 100% believe him to be very credible and want to believe what he has to say. Then there is the group that wants to believe what he has to say but have become very skeptical over time. Everybody is hungry for more information regarding Switch 2.

Also Nate's information from years ago pointed to a 4K capable system via DLSS,

This wont really convince anyone who is skeptical that he has insider information. The odds were always high that with the success of the Switch Nintendo would stick with Nvidia for its successor. 4K has become the standard so the next Switch is bound to support 4K and Nvidias DLSS technology makes that much more practical when dealing with a low power device. Pretty much anyone speculating a few years ago would assume 4k and DLSS were a lock for the Switch successor.

when guys like MVG and John Linneman are on the podcast should be an indicator.

For me this is the best reason to believe there is some real credibility. John and MVG have their own credibility to worry about so if they are doing a podcast with Nate, they must believe him to be on the up and up.

 

[ILikeFeet]

A Mariko-based Switch Pro probably would have been a bit disappointing, but it's worth remembering we are talking about a cancelled system here. Would it really be that unreasonable that started work on such a system, to the point where some developers probably got briefed on it, but decided to cancel it fairly early specifically because the chip, while fine for the Switch Lite, didn't scale up how they wanted?

A 2GHz cpu and 1.4GHz gpu and 6GB of ram wouldn't be disappointing as a pro system, imo.

Hitting that would have been the problem

 

[Gotdatmoneyy]

Yikes, Nintenduvo got threadbanned. I didn't think their post was THAT disruptive. Though, I guess it can come off as such, since it was referring to Nate.

No fuck that. Honestly, I'm tired of people with nothing of value to post just shit talking Nate or questioning him like he's on trial over and over.

For fucks sakes, enough is enough. Shit is of absolutely 0 value. Dude can go sit in the corner for a while. It's fine.

 

[pelusilla6]

Two questions dodged in the podcast regarding BC, we'll have to wait a bit more haha

 

[D36]

No fuck that. Honestly, I'm tired of people with nothing of value to post just shit talking Nate or questioning him like he's on trial over and over.

For fucks sakes, enough is enough. Shit is of absolutely 0 value. Dude can go sit in the corner for a while. It's fine.

Agreed, Nate has definitely proved his worth around here and while I was unsure of the ban, ultimately, I do believe it was necessary. This isn't really the place for that kind of talk. I should probably delete my initial response so as not to stoke the flames even further.