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

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

Welcome back! It really has been... 500+ pages since you hung out here. Woof! It'll be hard for me to remember what's transpired, exactly, though you're probably right to bow out because despite the sound and fury, not much has changed?

The chip is still Drake
Every time reliable hardware detail comes out, it just confirms prior knowledge/speculation about Drake/T239. Most recently, @LiC ran down a rabbit hole of shipping/customs documents, and found records of what are almost certainly devkits being shipped to/from an Nvidia division in India. Dates match the speculated tape-out timeline, explicitly mention T239, and confirm it's memory bus.

The storage is very fast
Multiple reports have come out about fast on-board storage and a new cartridge format. Here is Nate's podcast about loading times in the Gamescom demos. @Thraktor hunted down a candidate for the new cart technology

LCD screen
VGC reported that the device might have a LCD screen instead of an OLED...
* Hidden text: cannot be quoted. *


RAM
Nate has said he's gotten conflicting info on RAM but that "8GB is not one of the numbers I've heard." necrolipe, brazilian journalist and now thread semi-regular, has said he's heard 12GB for retail, with 16GB in devkits, matching prior speculation

The Gamescom Demos
VGC, Eurogamer, and Nate (previously linked) all reported on demos from Gamescom. At least one demo ran on "target spec hardware", not a devkit, and the demos were not hands on.

One demo was the Matrix Awakens, running with some kind of ray tracing. Reports have said things like "comparable to Series S" but without further details, it's unclear what that means. Here is a Digital Foundry video that I think dives into that pretty well but it matches "plausible but optimistic" descriptions of what T239 could be capable of that you likely remember

The other demo was a high frame-rate, high-res Breath of the Wild. According to Nate, the primary intent of that demo was to show that loading times have been "erased". Strong support for prior speculation on UFS based storage, and the powers of the File Decompression Engine.

Clock speeds
Not long after you left the thread, @Thraktor did some really smart analysis of power curves. The ultra-short version is that very low clocks on T239 don't make financial sense. This is also considered solid evidence of the process node.

That debate is as ongoing as ever, but the era of arguing that "Nintendo will clock this thing into the ground" is mostly over.

Battery Life
A loooooong while back I was reliably told that Nintendo was shooting for 3-6 hours of battery life. I have no idea if there is another source for that info, but I considered it stale enough for everyone involved to have plausible deniability, so I brought it up. Season that with salt to taste.

Gimmick? Camera?
Nothing about a new control scheme or feature has come out except for some likely-bullshit reports about a camera.

Launch Timing
@fwd-bwd has uncovered "for entertainment purposes only" some Taiwanese investment rags claiming that Foxconn expects manufacturing to happen later this year. They've doubled down on that report as recently as last week.

Nate also heard a context-less "March" rumor. Emphasis on context-less. Could be reveal, could be an internal deadline, could be just rumblings in as much as we have any info on it whatsoever.

I think that's the highlights?
Thanks for this! I don't check here often. Maybe twice a month. Good to have a summary.
 
Its going to be some sad days on this board if the next Investors meeting come and goes with no mentions of the next generation of hardware at all and that Nintendo Switch is still in the middle of its life cycle
 
this has to be a reference. I read this exact line somewhere else on this a thread
It is never my desire to pick on or antagonise new members or those who haven’t posted in here much. It was more an expression of frustration and exhaustion on my part, although I suspect I’m not alone… I have been pushing back against the policing of expectations on here since my second post. A lot of people felt “burned” by WUST and Switch speculation threads of the past, but the truth is that we’re all adults on here, and if something doesn’t happen, it doesn’t happen… that’s life. But it shouldn’t deter anybody from daring to expect and daring to expect again. So, the onus is on one to get a grip of one’s own emotions rather than WOO-HAH and demand everybody else keep their expectations in check. The cotton-wooling is very annoying, and all the more so in the face of what we know. 8nm might well happen, BUT if it does, it’s not because “Nintendo are cheap”, or some Internet villain that “hates certain very online fans” - Nvidia were selling multiple products on that process, after all. It would be because both have done enough due diligence to determine that it’s suitable for mass production of the next console. We don’t actually have a reason to believe that it spells doom, and the Orin tool which was touted much earlier in the thread was neither useful nor indicative of the successor’s performance. So, the reactionism towards this outcome hasn’t been appropriate.

By all means, feel free to believe in more pessimistic outcomes, and present a case - It would be a shame if others felt unwelcome for it, especially in a speculation thread, and it’s right that this place doesn’t become an echo chamber… BUT please don’t come in here telling everybody else to temper theirs, and don’t come here with a non-premise like “Because L-M-A-O, Nin-ten-D’OH!!” and expect to be taken seriously, or having not read the receipts in the threadmarks to arrive at whatever your point is. It’s disrespectful towards those who put a lot of work into giving clarity to members who aren’t entirely sure, and it’s frustrating as hell, especially when nobody in this thread has talked about matching the raw power of the latest AMD-powered home consoles, or “secret sauce” (DLSS isn’t that, it’s a product of smarter engineering, while neural units have existed in the portable space for quite a while). So, “Please, Stop this/Please don’t do this” is sparing everybody this rant and asking people kindly not to threadturd, derail, or bring “pissing in cornflakes” energy to the topic.

It's not like nvidia would be doing extra dlss work for nintendo, really. This is a product nvidia sells and they would develop dlss for the part because that's what they are selling... a mobile, low power chip that supports the feature.

I know nintendo fans are always looking for special sawce, and maybe that's what people are cheerleading for here, but nvidia already is supporting the chip they designed to be sold and used with the dlss feature lol
Which fans? RT and DLSS are lock-ins. We know this to be true. It’s a long-term partnership - Think about what that means. Nvidia WILL do the extra work by virtue of this being a CUSTOM chip (that’s the third time I’ve written that in CAPS, but nonetheless a point which ought not to be lost on people on here). It means if Nintendo wants a feature, they’ll have it. WHY will they do the extra work for a Nintendo product? Because they went to the trouble of courting them for years before the development of the Switch. Literally Because It’s Nintendo, and they’re the biggest design “win” they could get in this space. Because through them, they can get their tech into 140m+ pairs of hands. Because the Switch’s success is reason enough for them to want to keep their clients happy, to keep this endeavour alive. Also, Nintendo will have targeted specs, and asked them to deliver on that - there is a coalition of interests.
 
Last edited:
Welcome back! It really has been... 500+ pages since you hung out here. Woof! It'll be hard for me to remember what's transpired, exactly, though you're probably right to bow out because despite the sound and fury, not much has changed?

The chip is still Drake
Every time reliable hardware detail comes out, it just confirms prior knowledge/speculation about Drake/T239. Most recently, @LiC ran down a rabbit hole of shipping/customs documents, and found records of what are almost certainly devkits being shipped to/from an Nvidia division in India. Dates match the speculated tape-out timeline, explicitly mention T239, and confirm it's memory bus.

The storage is very fast
Multiple reports have come out about fast on-board storage and a new cartridge format. Here is Nate's podcast about loading times in the Gamescom demos. @Thraktor hunted down a candidate for the new cart technology

LCD screen
VGC reported that the device might have a LCD screen instead of an OLED...
* Hidden text: cannot be quoted. *


RAM
Nate has said he's gotten conflicting info on RAM but that "8GB is not one of the numbers I've heard." necrolipe, brazilian journalist and now thread semi-regular, has said he's heard 12GB for retail, with 16GB in devkits, matching prior speculation

The Gamescom Demos
VGC, Eurogamer, and Nate (previously linked) all reported on demos from Gamescom. At least one demo ran on "target spec hardware", not a devkit, and the demos were not hands on.

One demo was the Matrix Awakens, running with some kind of ray tracing. Reports have said things like "comparable to Series S" but without further details, it's unclear what that means. Here is a Digital Foundry video that I think dives into that pretty well but it matches "plausible but optimistic" descriptions of what T239 could be capable of that you likely remember

The other demo was a high frame-rate, high-res Breath of the Wild. According to Nate, the primary intent of that demo was to show that loading times have been "erased". Strong support for prior speculation on UFS based storage, and the powers of the File Decompression Engine.

Clock speeds
Not long after you left the thread, @Thraktor did some really smart analysis of power curves. The ultra-short version is that very low clocks on T239 don't make financial sense. This is also considered solid evidence of the process node.

That debate is as ongoing as ever, but the era of arguing that "Nintendo will clock this thing into the ground" is mostly over.

Battery Life
A loooooong while back I was reliably told that Nintendo was shooting for 3-6 hours of battery life. I have no idea if there is another source for that info, but I considered it stale enough for everyone involved to have plausible deniability, so I brought it up. Season that with salt to taste.

Gimmick? Camera?
Nothing about a new control scheme or feature has come out except for some likely-bullshit reports about a camera.

Launch Timing
@fwd-bwd has uncovered "for entertainment purposes only" some Taiwanese investment rags claiming that Foxconn expects manufacturing to happen later this year. They've doubled down on that report as recently as last week.

Nate also heard a context-less "March" rumor. Emphasis on context-less. Could be reveal, could be an internal deadline, could be just rumblings in as much as we have any info on it whatsoever.

I think that's the highlights?
Thank you again for the summary, this is a massive help and will hopefully let me keep pace with everything else being discussed 🙏

The storage being extremely fast sounds pretty promising, too. Just to clarify, is this in relation to the game cards, the internal flash memory, or both? Also, I read that post from Thraktor and about low-costs involved for the game cartridges. Is there any kind of speculative number attached to pricing yet, or all just guesswork? I guess in short I'm curious about the cost comparison between an 8GB, 16GB and 32GB Switch game cartridges, 16GB, 32GB and 64GB cartridge estimates for nuSwitch, and the 50GB/100GB Blu-Rays for PS5/XSX.

Like, is low-cost relative to what third party publishers currently pay for Switch, or could it actually be competitive enough with the cheapness of blu-ray discs that we might see fewer (if any) additional download games?

But also the notion of low clocks being removed from the equation is also a massive relief to hear. I'm quite content with 3-6 hours battery life as I rarely play handheld myself anyways, and usually not far enough away from a charging cable or power brick regardless.

A context-less March rumour is enticing to me, and while I'd certainly be happy if it were at least an announcement, it makes sense not to get hopes up incase this is an internal milestone. Either way, no sense holding my breath when we're still getting games now and most of the upcoming releases like SMWonder, SMRPG HD, Princess Peach and TTYD HD are unlikely to be super taxing that demand additional horsepower under the hood.
 
We are near the final stage of this thing. I'm still not quite sure how to fill the hole this thread will leave when the Switch 2 is announced and released.

Switch 2 pro speculation? 🤔
Will be fun, speculating about Switch 2 Pro, years go by and before you know it its the Switch 3 speculation thread.

Speaking from experience.
 
We are near the final stage of this thing. I'm still not quite sure how to fill the hole this thread will leave when the Switch 2 is announced and released.

Switch 2 pro speculation? 🤔

We probably won’t see any exclusives until June 2024 and won’t have the specs until near release which is still 10-12 months away.
 
Thank you again for the summary, this is a massive help and will hopefully let me keep pace with everything else being discussed 🙏

The storage being extremely fast sounds pretty promising, too. Just to clarify, is this in relation to the game cards, the internal flash memory, or both?
Both, from the sounds of it, but it’s unclear.

Also, I read that post from Thraktor and about low-costs involved for the game cartridges. Is there any kind of speculative number attached to pricing yet, or all just guesswork? I guess in short I'm curious about the cost comparison between an 8GB, 16GB and 32GB Switch game cartridges, 16GB, 32GB and 64GB cartridge estimates for nuSwitch, and the 50GB/100GB Blu-Rays for PS5/XSX.

Like, is low-cost relative to what third party publishers currently pay for Switch,
Relative to current costs for cartridges. I don’t think there is any real world where carts get down to disk prices. But something that makes larger carts financially viable
 
It doesn't have to be as complicated as needing a new lighter version of DLSS. If you're willing to settle for "4K but less good", you can...
*Instead DLSS to 1600p or 1800p or whatever you can afford and let the system's default scaler take it the rest of the way. End result: a 4K image that looks less good.
*If pre-DLSS your game is 1440p and can't afford the 4K upscale, you take it down to 1260p or something and see if you can afford the 4K DLSS now. End result: a 4K image that looks less good.

This is certainly possible, and the DF reviews for FSR 1.0 did show that it worked well with scaling 1440p up to 4K, but anything less than 1440p rendering resolution didn't have great results. Regardless if its FSR 1.0 are a more traditional scaler, it seems that 1440p scales up nicely to 4K and would result in good image quality, a significant upgrade over 1080p.

It does make me wonder though, if DLSS has a rather significant frame slice for going to 4K, does that leave the door open for less expensive reconstruction techniques? We have Nintendo's bespoke image reconstruction technology that they patented a couple years ago and implemented in Xenoblade Chronicles 3. This technology is taking 540p up to 1080p on the Tegra X1 and I would have to assume its relatively light weight, especially when running on a far more powerful chip like T239. There are absolutely various options available to get to the final 4k resolve that dont rely on DLSS doing any/all of the scaling.

Then I think about the Zelda BotW demo running at 4K via DLSS and at 60fps proving that a 4K output with DLSS is very viable even for 60fps games. However, this is a game that was originally designed for a 176 Gflop GPU in the Wii U, so what would the rendering time look like on T239 at 1080p 60fps or even 1440p 60fps? Even if DLSS has a rather high frame slice of 6ms, perhaps T239 has the power to render out BotW at 1080p or 1440p in just a few milliseconds?
 
Thank you again for the summary, this is a massive help and will hopefully let me keep pace with everything else being discussed 🙏

The storage being extremely fast sounds pretty promising, too. Just to clarify, is this in relation to the game cards, the internal flash memory, or both? Also, I read that post from Thraktor and about low-costs involved for the game cartridges. Is there any kind of speculative number attached to pricing yet, or all just guesswork? I guess in short I'm curious about the cost comparison between an 8GB, 16GB and 32GB Switch game cartridges, 16GB, 32GB and 64GB cartridge estimates for nuSwitch, and the 50GB/100GB Blu-Rays for PS5/XSX.
Supposedly if they're using Macronix's new 3D NAND storage cards were looking more at ~100-120GB Switch 2 cards for the price of current 8GB Switch cards.

 
Both, from the sounds of it, but it’s unclear.


Relative to current costs for cartridges. I don’t think there is any real world where carts get down to disk prices. But something that makes larger carts financially viable
Figured it was gonna be a pipe dream, but nevertheless worth asking just in case 😋
Supposedly if they're using Macronix's new 3D NAND storage cards were looking more at ~100-120GB Switch 2 cards for the price of current 8GB Switch cards.

That definitely puts into perspective how game carts that size would be financially viable 👀 Do we have an idea on what could be the floor for cartridge sizes? I can't imagine many games going forward except for the most niche indie titles that get a limited physical release would require something as low as 1GB or 2GB (and hell I'd be shocked if it even did that go low). That said, the lowest game cartridge size being, say, 8GB, and also being drastically cheaper compared to the current Switch cards of 8GB storage...idk where I'm going with this thought but I mean I feel like the ones who care most about their games going on the smallest viable cartridge would probably look at the Switch more favourably as a console to develop for, hopefully.

And very few if any games would need more than that.
Outside of CoD and GTA...yeah, I can't imagine much else, and surely some devs have to be at least a little cognisant of the growing need for improved file size compression
 
Figured it was gonna be a pipe dream, but nevertheless worth asking just in case 😋

That definitely puts into perspective how game carts that size would be financially viable 👀 Do we have an idea on what could be the floor for cartridge sizes? I can't imagine many games going forward except for the most niche indie titles that get a limited physical release would require something as low as 1GB or 2GB (and hell I'd be shocked if it even did that go low). That said, the lowest game cartridge size being, say, 8GB, and also being drastically cheaper compared to the current Switch cards of 8GB storage...idk where I'm going with this thought but I mean I feel like the ones who care most about their games going on the smallest viable cartridge would probably look at the Switch more favourably as a console to develop for, hopefully.


Outside of CoD and GTA...yeah, I can't imagine much else, and surely some devs have to be at least a little cognisant of the growing need for improved file size compression
I doubt they'd offer super small 1-2GB cards and then also 128GB cards, and if the cost of the latter is sufficiently small (like the same cost as 8GB now) I doubt it would really be economically viable to to offer anything as low as 8GB either. I'd guess something like 50GB/100GB/150GB being the main three tiers (assuming they use this tech).
 
Launch Timing
@fwd-bwd has uncovered "for entertainment purposes only" some Taiwanese investment rags claiming that Foxconn expects manufacturing to happen later this year. They've doubled down on that report as recently as last week.
Is that the right link? Making sure there's no tidbit I overlooked other than "Hongzhun expects growth in its three major businesses of light metals, cooling modules and game consoles this year." (and nothing else).
 
Quoted by: LiC
1
Rather than resolution I see Nintendo marketing the actual power increase more. Not directly, but through their games as they always do. Just seeing a Switch crossgen game at basically-4K (1440p DLSS) and an exclusive game that makes even their best looking titles archaic looking will be more than enough.
I don't necessarily think this approach would work. The last time Nintendo tried to appeal with graphical presentation "indirectly" was when Satoru Iwata was highlight a Mario & Sonic in the Olympics game and presenting it as a good showcase for the Wii U.

I think any appeal to a "power increase" will have to speak more with the library of Switch successor titles and 3rd party titles it will inherit from the other two rival consoles, as that speaks more to consumers who'd likely go "Oh, so it's powerful enough to run these titles at a great framerate and resolution? Sold!"

Selling the machine as simply "a Switch with 4K capability" is useless.
 
Is that the right link? Making sure there's no tidbit I overlooked other than "Hongzhun expects growth in its three major businesses of light metals, cooling modules and game consoles this year." (and nothing else).
The last few times they printed this expectation, they directly connected it to the release of a "new Japanese game console" in early 2024. They didn't reiterate it in this article, but there's not much else that could really be responsible for increased assembly of video game console in this year compared to previous years, if not for a new console. Something like PS5 Slim is just replacing existing models.

MoneyDJ is firmly in "entertainment purposes only" territory for now, though.
 
Supposedly if they're using Macronix's new 3D NAND storage cards were looking more at ~100-120GB Switch 2 cards for the price of current 8GB Switch cards.

I still have a hard time thinking prices could get that low. one of those "too good to be true" kinda things. maybe if it was 32GB cards for the price of 8GB cards, I'd more readily believe it
 
We are near the final stage of this thing. I'm still not quite sure how to fill the hole this thread will leave when the Switch 2 is announced and released.

Switch 2 pro speculation? 🤔
Hardware doesn’t start and end at the SoC. Also, there never will be a “Pro” because it would result in too many performance profiles and a split userbase. There is Labo, next Fit concept, new controller ideas, possibility of a new stylus, what’s happening in the mobile space or Nvidia’s tech, ARM CPUs, other tech advancements that could feature at some point. Other prospective concepts and form factors. So much more.
 
It really doesn't.
From his video, this part:
If it's a Tegra SOC 8nm design that's still unreleased, the chances are this could be for the Nintendo Switch 2. Because all of the other newer designs seem to not be on 8nm.
Then he says we don't know etc. But this is a strong one here. He then points out both this and his other findings are from the same exact location. So it feels like this might be leaning more towards 8nm 🤷🏻‍♂️
 
Quoted by: LiC
1
I don't necessarily think this approach would work. The last time Nintendo tried to appeal with graphical presentation "indirectly" was when Satoru Iwata was highlight a Mario & Sonic in the Olympics game and presenting it as a good showcase for the Wii U.

I think any appeal to a "power increase" will have to speak more with the library of Switch successor titles and 3rd party titles it will inherit from the other two rival consoles, as that speaks more to consumers who'd likely go "Oh, so it's powerful enough to run these titles at a great framerate and resolution? Sold!"

Selling the machine as simply "a Switch with 4K capability" is useless.
Isn't that literally what I said? Power increase =/= resolution, any exclusive game pushing polygon counts and raytraced lightning matching the likes of late PS4 and beyond is a huge af jump. Switch games will most likely run at 4K 95% of the time without the need for upscaling, but Nintendo isn't going to sell this based on backwards compatibility alone. It still needs its fair share of exclusives to blow away expectations in the graphical department, which encompasses a lot more than rendering resolution, and something people can perceive much more easily than image crispness.
 
Last edited:
0
Hardware doesn’t start and end at the SoC. Also, there never will be a “Pro” because it would result in too many performance profiles and a split userbase. There is Labo, next Fit concept, new controller ideas, possibility of a new stylus, what’s happening in the mobile space or Nvidia’s tech, ARM CPUs, other tech advancements that could feature at some point. Other prospective concepts and form factors. So much more.
We know Nintendo was considering a Switch Pro at some point. The reason it didn't happen seem to be timing and Mariko overclock being not great

We don't know the node and final specs but I wouldn't rule out a Switch 2 Pro. It's also possible we end up just like the Switch 2 where Pro discussion morphs Into Switch 3 talk if Nintendo ends up deciding to skip the Pro again.
 
0
From his video, this part:

Then he says we don't know etc. But this is a strong one here. He then points out both this and his other findings are from the same exact location. So it feels like this might be leaning more towards 8nm 🤷🏻‍♂️
The profile didn't say they were working on something or imply that it was unreleased. All it said was "currently working on," which cannot be accurate even if T239 is 8nm, because T234 and T239 were both over and done with last year, and any future Tegras are not 8nm.
 
Selling the machine as simply "a Switch with 4K capability" is useless.
I STRONGLY disagree, "4K Switch" is marketing GOLD. The fact they can use both is very encouraging for early generation adoption rates, provided they follow through on the marketing push.
 
I still have a hard time thinking prices could get that low. one of those "too good to be true" kinda things. maybe if it was 32GB cards for the price of 8GB cards, I'd more readily believe it
I'm paraphrasing from I believe Skittzo and z0m3le, but the whole point of the technology is that adding extra layers gives you big density/capacity increases, but those additional layers are cheap to create once you have the base.
 
I'm paraphrasing from I believe Skittzo and z0m3le, but the whole point of the technology is that adding extra layers gives you big density/capacity increases, but those additional layers are cheap to create once you have the base.
well, they give away 128GB flash drives at my microcenter on the regular. so if they can extend the lifespan by removing writing functions, then I guess it makes sense
 
The profile didn't say they were working on something or imply that it was unreleased. All it said was "currently working on," which cannot be accurate even if T239 is 8nm, because T234 and T239 were both over and done with last year, and any future Tegras are not 8nm.

Okay so

Most LinkedIn profiles are rarely updated so it could have been current when this text was written.

But this shows why trying to deduce timing from LinkedIn updates is very bad without additional info.
 
Hi everyone, this is my first post. I have followed this thread for quite a while and recently found a way to make a contribution.

I needed a refresher on one of the 3D modeling programs I use, so I decided to create a quick Switch 2 mock-up. My intent is outlined below the images:


z9ULWb6.jpg


5s6yXcX.jpg


91N0fvd.jpg


sOMIPxQ.jpg



1) Many design-oriented companies will revisit old concepts, whether re-imagining ideas that were implemented in past products or even looking for opportunities in the discard pile, and Nintendo is no exception. I set out to explore a fusion of Switch, Wii & 3D auto-stereoscopic display, while not veering too far from the initial Switch design

2a) I don’t care for the Joy-cons (in practice, great idea though). I feel them flexing when in handheld mode, and they are so diminutive that my hands cramp after a few hours. In tabletop mode or used as individual controllers the effect is even worse. Too small and not enough ergonomic design.

2b) Sliding the Joy-cons into place IS a satisfying experience; however, the way the controllers connect does not offer any lateral stability (which is why they flex). The thought here was to use magnets that connect all along the controller perimeter on two perpendicular surfaces. This would be akin to Apple’s MagSafe power bricks but with considerably more hold

3) Nintendo’s discarded concept of having a tablet screen with two adjacent Wii remotes has always been intriguing to me. The Wii remote always felt comfortable in hand, and the idea that you could have two identical controllers flanking the screen, interchangeable in either right or left position seemed like an apt way of advancing the controller concept

4) Face buttons, D-Pad, and Joystick would be modular; using similar magnetic tech mentioned in #2 to interchange and customize location in two possible areas on the controllers

5) I really loved the immersion created by the 3D effect on New 3DS and think now would be a great time to re-incorporate at a higher resolution and with the significant advancements in eye-tracking tech. the extent of what you see in the rendering above is just the inclusion of the dual front-facing cameras.

I honestly didn’t think about the dock design at all, but If I have more time in the future I may revisit that
 
Is there a reason this isn't talked about?
It has been talked about, and it would potentially disguise codes. However, Nintendo’s recommended graphics mode for their graphics library already does this same kind of trick for rendering.

I don't know if it's ever implemented on PC games, I think most non-Turing cards capable of DLSS would upscale so quickly it might not be worth bothering with it but Nvidia clearly states that Ampere is capable of this type of concurrency. Turing on the other hand is not, so using a 20xx series card as a benchmark for Switch 2 DLSS performance seems like it could be quite misleading.
What benchmark are you referring to? The only game-driven benchmark I am aware of that is on a 20 series card is Rich’s upcoming video, but he’s testing on a 2050M, which is Ampere, despite the name
 
Okay so

Most LinkedIn profiles are rarely updated so it could have been current when this text was written.

But this shows why trying to deduce timing from LinkedIn updates is very bad without additional info.
we can't deduce anything heavy but we can determine things that reinforce what we know. like the recent stuff on people in Nintendo working with AI/ML/Tensor math stuff.

it's is the first time a dedicated gaming device is using matrix math accelerators so putting effort into extracting the most out of it will be paramount for future research. even for Sony and MS (who leaked that they're looking in that direction for their next system)
 
I STRONGLY disagree, "4K Switch" is marketing GOLD. The fact they can use both is very encouraging for early generation adoption rates, provided they follow through on the marketing push.
I also strongly disagree with your position. Using power as a marketing angle isn't going to work well in the long term.
 
I think I will be pretty satisfied with what I have heard about the NG's GPU...

How are we feeling about the CPU? How would a game with the CPU utilization of Cyberpunk 2077 fair on NG?
 
It is never my desire to pick on or antagonise new members or those who haven’t posted in here much. It was more an expression of frustration and exhaustion on my part, although I suspect I’m not alone… I have been pushing back against the policing of expectations on here since my second post. A lot of people felt “burned” by WUST and Switch speculation threads of the past, but the truth is that we’re all adults on here, and if something doesn’t happen, it doesn’t happen… that’s life. But it shouldn’t deter anybody from daring to expect and daring to expect again. So, the onus is on one to get a grip of one’s own emotions rather than WOO-HAH and demand everybody else keep their expectations in check. The cotton-wooling is very annoying, and all the more so in the face of what we know. 8nm might well happen, BUT if it does, it’s not because “Nintendo are cheap”, or some Internet villain that “hates certain very online fans” - Nvidia were selling multiple products on that process, after all. It would be because both have done enough due diligence to determine that it’s suitable for mass production of the next console. We don’t actually have a reason to believe that it spells doom, and the Orin tool which was touted much earlier in the thread was neither useful nor indicative of the successor’s performance. So, the reactionism towards this outcome hasn’t been appropriate.

By all means, feel free to believe in more pessimistic outcomes, and present a case - It would be a shame if others felt unwelcome for it, especially in a speculation thread, and it’s right that this place doesn’t become an echo chamber… BUT please don’t come in here telling everybody else to temper theirs, and don’t come here with a non-premise like “Because L-M-A-O, Nin-ten-D’OH!!” and expect to be taken seriously, or having not read the receipts in the threadmarks to arrive at whatever your point is. It’s disrespectful towards those who put a lot of work into giving clarity to members who aren’t entirely sure, and it’s frustrating as hell, especially when nobody in this thread has talked about matching the raw power of the latest AMD-powered home consoles, or “secret sauce” (DLSS isn’t that, it’s a product of smarter engineering, while neural units have existed in the portable space for quite a while). So, “Please, Stop this/Please don’t do this” is sparing everybody this rant and asking people kindly not to threadturd, derail, or bring “pissing in cornflakes” energy to the topic.
Yeah I've spent quite some time lurking here and only made an account recently but I throughly agree with what is said. I hope my comment didn't come across as if I was challenging it. Rather I genuinely read a similar comment like a hundred pages ago and thought you were referencing that as an inside joke. I believe the topic was more related to someone saying this in opposition to an H1/March launch. On a slightly related note, my favourite inside joke on here is the August launch jokes that surround that one user on here who is convinced NG will launch in August. They know who they are and I look forward to reading their contributions on here even if I might quietly disagree with 90% of them. Personally, I believe that NG switch will not use 8nm. I came to this conclusion through reading various well-thought and mathematically supported arguments that suggest it to be to inefficient of a node with no financial gain to be had compared to 5nm. I'm saying this as someone who has followed this thread meticulously sifting through the many many many many offtopic and/or circular discussions that go on in this thread. I don't typically contribute because I like to be well educated before making any suggestions or ideas because I don't want to contribute the filler that goes on here when people ask questions or make suggestions based on outdated or incorrect information about what is being speculated on.

Story time incoming (skip if don't care):
When the first Switch was still being referred to as the NX, I was too young (like really) to know of these discussion boards and would fanboy over the hype-abusing youtubers and mald over the doomers too (lol). But past reddit I was unaware of what was and wasn't realistic I just wanted to see months of speculation turn into a tangible thing. 7 years later and I feel like a kid again reading all the speculation of the next Nintendo console and as I've gotten older I've grown to enjoy the more technically-inclined speculation which is why I closely follow Intel, AMD, Nvidia, and Apple forums to see read all about upcoming hardware. And it has actually informed many of my (my parent's cos I'm still young XD) purchases as I always wanted the best there was to offer out there. It even reflected in my degree of choice last year because I find it fascinating to talk about and work on these electronics and see innovative techniques such as DLSS make impossible experiences possible.

Now I'm a young adult, revisiting nintendo console discussions is like driving down an old trail with a new bike. It's nostalgic but different and I'm sure in 7 years time I'll say the same thing again as it really does feel like this kind of excitement towards new tech doesn't get old. I see many on here I assume are fully grown adults with their own busy lives who probably won't even have the chance to play as much as I might when this new system drops yet everyone here is actively bringing new ideas to the forum or waiting eagerly for something to discuss when everything has been exhausted. I'll probably continue lurking after this post but one day hopefully soon I plan to join in these discussions and help to bring something to the table. And who knows, depending on where my degree takes me I might get some insider information to share with everyone here 0-0 (jk, probably).
 
I'm paraphrasing from I believe Skittzo and z0m3le, but the whole point of the technology is that adding extra layers gives you big density/capacity increases, but those additional layers are cheap to create once you have the base.
That's more or less the idea, yeah. Cost scales mainly with wafer area and not much at all with overall volume or height. So the only real difference in cost between a typical 2D storage chip of a given area and a 3D NAND chip using this technology is in the initial R&D setup cost. Meaning the difference in cost in materials is negligible despite a massive gain in storage.

You can even get more storage for cheaper if you use a smaller wafer area for the chip too.

As far as I understand it anyway.
 
Nintendo has Mario, a stylistically consistent series, whose visual identity in 3D is now 27 years old.

Nintendo doesn’t need to talk power. They just show Mario. “Wow games can look like that??? In my hand???”

Done.

There is no value in talking about performance for any technology product unless you need that talking point to distinguish yourself from the competition or the previous gen. The numbers won’t tell that story any better than the images will. All it will do is invite comparisons to the bigger consoles.
 
Please read this staff post before posting.

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


Back
Top Bottom