• 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| (New Staff Post, Please read)

Reader mode
Bait02 said:
That's interesting.

Humble opinion.
I think 3DS has been, at least, as big as a blunder for Nintendo as the WiiU was.

Sure, after the terrible launch and the price cut (that hurt them a lot financially as you correctly say) they managed to steer the wheel to some degree and the end-of-life hardware / software sales look somewhat decent (as opposed the WiiU).
Then again, putting in perspective three things makes it look worse:
  • The growth trend of their handheld business since the original Gameboy, with every new console being more successful than its predecessor.
  • Their leadership position in the handheld space, that was untouched even by Sony at its peak with the PSP.
  • WiiU was competing with formidable direct competitors, expecially Sony with the PS4 got everything right. In comparison, the only direct competitor of 3DS was the PSVita (an even worse flop).
Another thing to consider for why it would be considered "a gut-punch", is that apparently Nintendo when designing the 3DS operated under ther wrong assumption that they could somehow make an incremental upgrade that could cater to both their more experienced audience of players, as well as the new audience brought in with the DS.
At the end, the 3DS design was sort of compromise bewtween the diverging expectations of these two groups and, as a result, it alienated a lot of players from both sides.
Although exactly the same reasoning could be applied to the WiiU, I think the 3DS launch was the first time they experienced this, and it could have been a sort of "wake up call".
Click to expand...

Excellent points. One thing I want to mention is the difference between disappointment, failed expectations, and outright customer faith loss. This is subjective but most people I talk to did not feel "betrayed" by the Wii-U. They were betrayed by the 3DS. In Japanese culture, Honor is a big deal. I know American media likes to make light of it and make it out to be dramatic but its important to them. The Wii-U was a total miss. The 3DS was a betrayal of customer faith due to the launch and pivot. Instead of Nintendo doubling down on 3D and correcting that specific point of the 3DS they gave up and shoehorned 3D in for the rest of the generation after Super Mario 3D land. I can understand how some would think that was the correct move but I'd wager in context of Japanese culture it would have been better to "go down with the ship" and do their damn best to justify 3D. They sort of did this with the Wii-U. They kept at it even though the tablet gaming concept truely sucked.

Again, Lots of subjective talk here. I think opinions on this are very wide ranging.

RealRaditude said:
doing my monthly mental health check in on the hardware thread users
Click to expand...

I know you think that was probably funny but you can miss us with that. People are here communicating for a variety of reasons and most are capable of understanding internal changes as things develop and not letting it affect our day to day. Its okay to disappointed when hearing about delays or some kind of hardware addition/subtraction. It doesn't mean or hint anything as to the mental health outside the internet.
 
I could think of some crazy stuff for the MIC when docked (paired with a camera), pretty much Kinect style. And I liked what I thought, but... that would have a RAM consumption that I definitely don't want lol

So, I imagine it's just for BC with handheld games that used a MIC, and I'm betting on some interesting uses in the next LABO.
 
Bait02 said:
Another thing to consider for why it would be considered "a gut-punch", is that apparently Nintendo when designing the 3DS operated under ther wrong assumption that they could somehow make an incremental upgrade that could cater to both their more experienced audience of players, as well as the new audience brought in with the DS.
At the end, the 3DS design was sort of compromise bewtween the diverging expectations of these two groups and, as a result, it alienated a lot of players from both sides.
Although exactly the same reasoning could be applied to the WiiU, I think the 3DS launch was the first time they experienced this, and it could have been a sort of "wake up call".
Click to expand...

Can you expand on this a bit more, the 'incremental upgrade' piece is new to me.
  • There were rumours 3D was bolted on relatively late, after SoC had already been selected, and it impacted battery life. My own assumption is it likely also impacted performance, as clearly there is a performance cost to 3D.
  • The original selection was apparently nvidia Tegra 2 but the chip was too power hungry so Nintendo went with another custom design, which had an underpowered CPU and a proprietary GPU that while powerful was sort of a turnkey design from 2005 and isn't not like there's a v2 of the DMP PICA, it seemed like a one and done thing so there isn't even a clean roadmap to a 3DS successor BC outside of just including the SoC in the device.
Nintendo certainly leanerd in heavily on the 3DS being a kind of a portable Wii/GameCube, i don't think they were shy about showing it off grafically. Their timing just sucked with Sony upstaging them with VITA and the general mobile space being very different.

My feeling was technologically Nintendo was in a lose lose position. mobile tech was moving so far between 2007-2015 literally went from single core ARM phones to quad then octo cores in less than a decade and the GPU power increased tremendously so anything they picked would be easily outdated and they had up to this point designed their own SoCs.

That's another place where Switch 2 had an advantage, it will come with a very strong SoC that will be very competitive and the technological landscape is completely different
 
Branduil said:
I think there's a few mistakes Nintendo made with the 3DS, some not applicable, some the context is different, some they can easily avoid:
  • Failed Gimmick: The 3D aspect of the 3DS, which contributed greatly to the price, turned out to not be the system seller Nintendo was banking on. And it never proved to be a game-changer in anywhere close to the way touch controls did; there was no real killer app for 3D. It could be an enjoyable feature, but it didn't sell people on the system itself.
    • Applicable to Switch 2? No. There's no indication of Nintendo banking on any particular gimmick with Switch 2, it appears to be a very iterative console.
  • High Price: Related to the above, the 3DS proved too pricey for the market. People did not value what 3D brought to the console like Nintendo thought they would. This is what forced the most dramatic price cut ever for a Nintendo console, just a few months after launch.
    • Applicable to Switch 2? Possibly. Nintendo needs to be careful with how they price the system, which is why I believe they won't go above $399.
  • Abandonment by Previous Market: 3DS launched during a time of turbulent changes in the gaming market. With the DS and Wii, they had gone all-in on "blue market" customers. Unfortunately for Nintendo, this market had little loyalty to them, and quickly gravitated towards smartphone and other casual games which did not require investment in a dedicated gaming console. The DS was rapidly declining by the time 3DS arrived.
    • Applicable to Switch 2? No. Switch has a much different customer base from the DS and Wii. They are much more reliable and interested in traditional console games.
  • Lack of Appealing Games: The 3DS launched without anything that could be called a killer app. It would take several months before it built up an appealing library, at which point the damage from the launch has been done.
    • Applicable to Switch 2? One would think not, given that the delay was supposedly done to improve the software situation for the Switch 2 at launch.
Overall, I think most of these can be avoided by Nintendo. The biggest question mark is the price.
Click to expand...

There's also the possibility of enhanced BC on the Switch 2, vs the 3DS BC that ranged from "zero improvements" to "actually worse" depending on which DS you were coming from.
 
Branduil said:
I think there's a few mistakes Nintendo made with the 3DS, some not applicable, some the context is different, some they can easily avoid:
  • Failed Gimmick: The 3D aspect of the 3DS, which contributed greatly to the price, turned out to not be the system seller Nintendo was banking on. And it never proved to be a game-changer in anywhere close to the way touch controls did; there was no real killer app for 3D. It could be an enjoyable feature, but it didn't sell people on the system itself.
    • Applicable to Switch 2? No. There's no indication of Nintendo banking on any particular gimmick with Switch 2, it appears to be a very iterative console.
  • High Price: Related to the above, the 3DS proved too pricey for the market. People did not value what 3D brought to the console like Nintendo thought they would. This is what forced the most dramatic price cut ever for a Nintendo console, just a few months after launch.
    • Applicable to Switch 2? Possibly. Nintendo needs to be careful with how they price the system, which is why I believe they won't go above $399.
  • Abandonment by Previous Market: 3DS launched during a time of turbulent changes in the gaming market. With the DS and Wii, they had gone all-in on "blue market" customers. Unfortunately for Nintendo, this market had little loyalty to them, and quickly gravitated towards smartphone and other casual games which did not require investment in a dedicated gaming console. The DS was rapidly declining by the time 3DS arrived.
    • Applicable to Switch 2? No. Switch has a much different customer base from the DS and Wii. They are much more reliable and interested in traditional console games.
  • Lack of Appealing Games: The 3DS launched without anything that could be called a killer app. It would take several months before it built up an appealing library, at which point the damage from the launch has been done.
    • Applicable to Switch 2? One would think not, given that the delay was supposedly done to improve the software situation for the Switch 2 at launch.
Overall, I think most of these can be avoided by Nintendo. The biggest question mark is the price.
Click to expand...
Gimmick is very much welcomed even the failed one.
Price, good marketing and games availability are what matters the most.
 
Branduil said:
Lack of Appealing Games: The 3DS launched without anything that could be called a killer app. It would take several months before it built up an appealing library, at which point the damage from the launch has been done.
Click to expand...
I don't disagree with this, but I want to point out that while the DS had a pretty solid launch line-up (and the original DS also came with a demo for Metroid Prime Hunters bundled) it didn't make the device explode in popularity. That came with the DS lite redesign in 2006 (which is still the most sold model of DS consoles) and the subsequent release of the Touch! Generations "casual games".

The main mistake with 3DS (and also Wii U) is that Nintendo tried to go after the "casual" market first to get this audience to jump to a new device rather than the more "hardcore" audience that is usually willing to embrace new technology early on.
 
LuigiBlood said:
Unless you RTX the microphone input into something more usable than a Wii Speak microphone
Click to expand...

This still seems like it would be better on the joycons itself unless the console comes with a Kinect with a long distance microphone.

How expensive are long distance mics that are given machine learning information from a camera to filter out noise these days?
 
0
Kise Ryota said:
I could think of some crazy stuff for the MIC when docked (paired with a camera), pretty much Kinect style. And I liked what I thought, but... that would have a RAM consumption that I definitely don't want lol

So, I imagine it's just for BC with handheld games that used a MIC, and I'm betting on some interesting uses in the next LABO.
Click to expand...
Given the placement of the MIC (console) and historic placement of the IR Motion Camera (controller)... I don't think you're actually that far off the mark?

See, putting a camera on the system itself limits where you can position it in a room if you want to use it. With a mic that's way less of a problem, you can place it in suboptimal locations but it would be difficult to place it somewhere truly wrong. Meanwhile, putting it on the controller means a wireless two way Bluetooth stream of audio, one into the controller for HD Rumble (which while low bitrate, isn't mono I don't believe) and one out for the mic. There's complications with that.

If the IR Motion Camera is substantially improved, it could, perhaps, allow you to place the Joy-Con down somewhere it can "see you" and act like a Kinect. If there's one on each controller this time, each person gets their own camera. While I doubt this is the implementation we see, it's a possibility that might be worth considering.

I'd like to see both controllers have at least one IR Motion Camera. Lay them on their sides for two Kinect-like sensors, or swing them around with the IR Motion Cameras providing additional spacial information for better motion controls.
 
How power hungry are the tensor cores relative to the cuda cores?

I was watching an old DF vid about DLSS on the Switch 2, way before the T239 leaks and they mentioned expecting around 20 TOPS out of the chip. Though I assume it’ll be much higher in the final gpu, could we expect Nintendo to drive more power to the tensor cores to improve DLSS performance?
 
rinku said:
How power hungry are the tensor cores relative to the cuda cores?

I was watching an old DF vid about DLSS on the Switch 2, way before the T239 leaks and they mentioned expecting around 20 TOPS out of the chip. Though I assume it’ll be much higher in the final gpu, could we expect Nintendo to drive more power to the tensor cores to improve DLSS performance?
Click to expand...
based on my limited understanding (others can correct me if I'm wrong) but I believe tensor cores to be more efficient compared to cuda cores, using lesser computation power because tensor cores does not have to be as "accurate" (precise calculations) as cuda cores (so that makes tensor cores more well suited for things like AI-assisted upscaling, ie: DLSS). Cuda cores perform better when precision is required (ie: physics engine)

But that would be ignoring a lot of factors like what tasks were tensor cores and cuda cores doing at the moment, and the amount (ratio to each other) of cores for each of those in the GPU
 
Last edited:
Concernt said:
Given the placement of the MIC (console) and historic placement of the IR Motion Camera (controller)... I don't think you're actually that far off the mark?

See, putting a camera on the system itself limits where you can position it in a room if you want to use it. With a mic that's way less of a problem, you can place it in suboptimal locations but it would be difficult to place it somewhere truly wrong. Meanwhile, putting it on the controller means a wireless two way Bluetooth stream of audio, one into the controller for HD Rumble (which while low bitrate, isn't mono I don't believe) and one out for the mic. There's complications with that.

If the IR Motion Camera is substantially improved, it could, perhaps, allow you to place the Joy-Con down somewhere it can "see you" and act like a Kinect. If there's one on each controller this time, each person gets their own camera. While I doubt this is the implementation we see, it's a possibility that might be worth considering.

I'd like to see both controllers have at least one IR Motion Camera. Lay them on their sides for two Kinect-like sensors, or swing them around with the IR Motion Cameras providing additional spacial information for better motion controls.
Click to expand...

Did any camera show up in the shipment data so far?
 
rinku said:
How power hungry are the tensor cores relative to the cuda cores?

I was watching an old DF vid about DLSS on the Switch 2, way before the T239 leaks and they mentioned expecting around 20 TOPS out of the chip. Though I assume it’ll be much higher in the final gpu, could we expect Nintendo to drive more power to the tensor cores to improve DLSS performance?
Click to expand...
Assuming they go with around 4 TFLOPS docked, that would mean around 32 TOPS for the Tensor cores, since Ampere has a roughly 1:2:8 ratio between its CUDA, RT, and Tensor cores. I assume the 20 TOPS figure is going off of Kopite's prediction of 8 SMs, since that's almost exactly what you get with a 1024-core GPU running at 1.2–1.3 GHz.
 
ExoticSpice said:
It’s obvious what Nintendo will do with the Switch 2. It have fun hardware features that the Switch lacked and the OS is likely to get the same treatment.
Click to expand...
There's not much to suggest that at present.

Hidden content is only available for registered users. Sharing it outside of Famiboards is subject to moderation.
 
Dekuman said:
Can you expand on this a bit more, the 'incremental upgrade' piece is new to me.
  • There were rumours 3D was bolted on relatively late, after SoC had already been selected, and it impacted battery life. My own assumption is it likely also impacted performance, as clearly there is a performance cost to 3D.
  • The original selection was apparently nvidia Tegra 2 but the chip was too power hungry so Nintendo went with another custom design, which had an underpowered CPU and a proprietary GPU that while powerful was sort of a turnkey design from 2005 and isn't not like there's a v2 of the DMP PICA, it seemed like a one and done thing so there isn't even a clean roadmap to a 3DS successor BC outside of just including the SoC in the device.
Nintendo certainly leanerd in heavily on the 3DS being a kind of a portable Wii/GameCube, i don't think they were shy about showing it off grafically. Their timing just sucked with Sony upstaging them with VITA and the general mobile space being very different.

My feeling was technologically Nintendo was in a lose lose position. mobile tech was moving so far between 2007-2015 literally went from single core ARM phones to quad then octo cores in less than a decade and the GPU power increased tremendously so anything they picked would be easily outdated and they had up to this point designed their own SoCs.

That's another place where Switch 2 had an advantage, it will come with a very strong SoC that will be very competitive and the technological landscape is completely different
Click to expand...
Ok, I'm sorry it took me a moment to figure out. I made a wording mistake (not a native speaker), when I wrote about the "incremental upgrade" I used the word "apparently" and it's incorrect. I meant it's what how I interpret their work. I've fixed that.

I consider the 3DS an incremental upgrade, since it retains the core DS design and adds incremental features to it (es. improved graphics, 3D, 16:9, analog, SteetPass, etc.).
And this is also why my argument is its design is kind of a compromise.
On one hand they didn't want to lose the "expanded" audience of the DS by offering a similar design, but in practice they didn't offer any novelty of real value for that audience (in terms of accessibility, price, ease of use, enabling of new game experiences). At the same time it still failed to meet the expectations of more experience players, offering 3D but lacking in other departments such as screen size and resolution, ergonomics, the lack of a second analogue stick, etc.
 
Last edited:
Concernt said:
Given the placement of the MIC (console) and historic placement of the IR Motion Camera (controller)... I don't think you're actually that far off the mark?

See, putting a camera on the system itself limits where you can position it in a room if you want to use it. With a mic that's way less of a problem, you can place it in suboptimal locations but it would be difficult to place it somewhere truly wrong. Meanwhile, putting it on the controller means a wireless two way Bluetooth stream of audio, one into the controller for HD Rumble (which while low bitrate, isn't mono I don't believe) and one out for the mic. There's complications with that.

If the IR Motion Camera is substantially improved, it could, perhaps, allow you to place the Joy-Con down somewhere it can "see you" and act like a Kinect. If there's one on each controller this time, each person gets their own camera. While I doubt this is the implementation we see, it's a possibility that might be worth considering.

I'd like to see both controllers have at least one IR Motion Camera. Lay them on their sides for two Kinect-like sensors, or swing them around with the IR Motion Cameras providing additional spacial information for better motion controls.
Click to expand...
Yeah. As long as it works fine on docked and it's not really expensive, there are multiple advantages to put the mic on the tablet even if its a bit more expensive:

  • For each tablet, they make 2 joycons and then there's extra controller sales. So, that's a lot more mics if it's in each joycon.
  • Space is at a bigger premium on the joycon.
  • You can have more people interacting with the console than the number of controllers
  • You can lay down the controllers for Kinect-like games as you said.
  • You can use OG controllers.

And there's probably more.
 
RennanNT said:
Yeah. As long as it works fine on docked and it's not really expensive, there are multiple advantages to put the mic on the tablet even if its a bit more expensive:

  • For each tablet, they make 2 joycons and then there's extra controller sales. So, that's a lot more mics if it's in each joycon.
  • Space is at a bigger premium on the joycon.
  • You can have more people interacting with the console than the number of controllers
  • You can lay down the controllers for Kinect-like games as you said.
  • You can use OG controllers.

And there's probably more.
Click to expand...

Why does it have to be in each joycon instead of just the left joycon.
 
redmutineer75 said:
Assuming they go with around 4 TFLOPS docked, that would mean around 32 TOPS for the Tensor cores, since Ampere has a roughly 1:2:8 ratio between its CUDA, RT, and Tensor cores. I assume the 20 TOPS figure is going off of Kopite's prediction of 8 SMs, since that's almost exactly what you get with a 1024-core GPU running at 1.2–1.3 GHz.
Click to expand...
32 TOPS would be fp32, which is typically not what is used for AI algorithm inference. The number cited in the DF video that was referred to was int-8, which I believe is 4x the TOPS vs. FP32. So it'd be more like 128 int-8 TOPS in that case.

Edit: Ampere GPUs typically have a 4-1 ratio between CUDA and tensor cores I believe, so 64 int-8 TOPS seems more likely (and more in line with what I remember calculating before).
 
RennanNT said:
It could, but then it would only receive input from 1 person (or 2 in a 4-player game like Mario Party).
Click to expand...
Yeah, a high quality stationary microphone with good noise rejection (Kinect, PlayStation Camera, Wii Speak) is generally better for multiplayer, even compared to a microphone you can pass around. Plus no people practically spitting onto your controller as they sing. They just do it vaguely towards the TV. The biggest barrier would be maybe having to open a cabinet when it's in use if you keep your dock in one.

Maybe I'm going a bit funny, but I can imagine having my new system in the dock, booting up Nintendogs+More and greeting a puppy from across the room, reaching out with a Joy-Con in hand to pet it, that sort of thing. It would also enable the possibly cancelled Seaman sequel planned for a Nintendo system at some point.

A high quality stationary microphone sounds like a minor, silly thing, and yeah, it kinda is, BUT it opens the door to lots of fun things, DS games, karaoke, all sorts!

I'm just imagining Spirit Tracks having me shout at the TV to blow air around. Heck, Phantom Hourglass having me remove and reinsert the console into the dock to copy a map.

The clearer the picture of this device becomes the more excited I am for the possibilities at hand, both old and new.
 
Zedark said:
32 TOPS would be fp32, which is typically not what is used for AI algorithm inference. The number cited in the DF video that was referred to was int-8, which I believe is 4x the TOPS vs. FP32. So it'd be more like 128 int-8 TOPS in that case.

Edit: Ampere GPUs typically have a 4-1 ratio between CUDA and tensor cores I believe, so 64 int-8 TOPS seems more likely (and more in line with what I remember calculating before).
Click to expand...
INT8 is used for SDR inputs. Assuming that Switch 2 is HDR (which is fairly reasonable), it would be using FP16. Thus, 32 TOPS.
 
Kise Ryota said:
I could think of some crazy stuff for the MIC when docked (paired with a camera), pretty much Kinect style. And I liked what I thought, but... that would have a RAM consumption that I definitely don't want lol

So, I imagine it's just for BC with handheld games that used a MIC, and I'm betting on some interesting uses in the next LABO.
Click to expand...
Best use of the docked mic is as a smart home style terminal. I don't think Nintendo will go for an AI assistant but they may enable voice activated functions like swapping games , turning on or off and doing other tasks and link to smartphones too if they want.
 
Dekuman said:
Best use of the docked mic is as a smart home style terminal. I don't think Nintendo will go for an AI assistant but they may enable voice activated functions like swapping games , turning on or off and doing other tasks and link to smartphones too if they want.
Click to expand...
I don't know that's the BEST use, but certainly a use we saw with Kinect and PlayStation camera.

The idea of using the mic and speakers of the new console to pair to a smartphone is pretty funny, one brrrrrings, the other beeps, and they're connected like an old school modem. Lol.

On a serious note, I hope they, like Xbox or the present NSO app, let us use our smartphones as an optional mic using the NSO app. A workaround for exceptional situations like the dock behind behind the TV, and nearly everyone has a phone with a mic and a WiFi connection nowadays
 
Concernt said:
I don't know that's the BEST use, but certainly a use we saw with Kinect and PlayStation camera.

The idea of using the mic and speakers of the new console to pair to a smartphone is pretty funny, one brrrrrings, the other beeps, and they're connected like an old school modem. Lol.

On a serious note, I hope they, like Xbox or the present NSO app, let us use our smartphones as an optional mic using the NSO app. A workaround for exceptional situations like the dock behind behind the TV, and nearly everyone has a phone with a mic and a WiFi connection nowadays
Click to expand...
I'm leaving the possibility open voice chat is still through their NSO app. 😬
 
Dekuman said:
I'm leaving the possibility open voice chat is still through their NSO app. 😬
Click to expand...
That makes some sense in a way, but personally I doubt it. I fully expect it to continue being an option however.

Being able to use it to directly push audio to the system for in-game chat like Fortnite or games with voice recognition would be a handy boon for the NSO app.
 
0
On the subject of the financially failed 3DS and Wii U, I'm extremely confident that these are are partially the reason the Switch 2 has taken so long. I don't want to say that Nintendo has leaned too heavily on the Switch's success out of fear, but they have a good thing going for them right now. Heck, they've got the best thing going for them now. The Switch era is the wealthiest in the company's history and with the Mario movie, upcoming Zelda movie, and themeparks? they're on top of the world all things considered. The last thing they want is a repeat of the Wii U that spoils all this momentum. Fortunately, the sheer amount of Switch and staying power has meant they can take as much time as they want to plan things out. Hence, the internal delay. They want to keep this good thing going for them and who can blame them for it.
 
After all the discussion and thinking, I'm starting to think that the additional face button on the controllers is the mic button, to mute it and unmute it, which can be lit up when active. However, it doesn't have to be limited to that. Consider a "Social" button. Press and hold to turn the mic on or off, or tap to open the "Social" menu - quickly view and share from your album, view your friend list and requests, manage your voice chat and microphone settings. Like an addendum to Quick Home but for "Social" stuff.
 
redmutineer75 said:
INT8 is used for SDR inputs. Assuming that Switch 2 is HDR (which is fairly reasonable), it would be using FP16. Thus, 32 TOPS.
Click to expand...
Right, that's true. We need to apply the same conversion to the DF video numbers, so the ratios would still be the same in that comparison
 
0
I'll be honest in that I'm surprised that there is a microphone integrated. I figured parents wouldn't want their kids talking to "online strangers", even if they're playing Mario Kart or something.

Yeah, I'm curious about if there's a camera because I recall that was one of the features that was talked about a long while ago alongside things like FF7Remake being ported over. Time will tell if that's real.
 
Pokemaniac said:
An AI "optimizing itself" is something that only happens in science fiction. A lot of that sort of talk originates from very handwavy (and frankly pretty cultish) "singularity" discourse that treats intelligence as some sort of fundamental quantity that's always going up which is just completely detached from the real world.

AI is not a young field of computer science, and it is one typified by boom and bust cycles where new developments get massively overhyped until the reality of what the techniques are actually capable of sets in and interest dries up. The next AI winter is most assuredly coming with the current "generative AI" bubble that's formed, and it may not even be all that far off.
Click to expand...
In a sense, neural networks are AI optimizing itself, but each layer has to be decided by a human.
Shareholder Chad said:
Another interesting one is apparently back when the Wii-u was winding down. Apparently there was a company wide initiative to allow for more younger developers to take an active role in idea creation and overall game direction. I vaguely remember this maybe being in an Iwata asks segment or something but they took it very seriously and that was viewed as a major success as the years went on. So much so the employees I spoke with were thrilled about it.
Click to expand...
If I recall correctly, splatoon was the first game to come out of this, and it was a roaring success. I imagine the attitude around the concept got really positive after that
 
gabeplays said:
I'll be honest in that I'm surprised that there is a microphone integrated. I figured parents wouldn't want their kids talking to "online strangers", even if they're playing Mario Kart or something.

Yeah, I'm curious about if there's a camera because I recall that was one of the features that was talked about a long while ago alongside things like FF7Remake being ported over. Time will tell if that's real.
Click to expand...
Most Switch users are adults. They will definitely be kid friendly but I think they realized lacking voice chat and communication features was a mistake that may have been too hard to patch in

The single A57 dedicated to the OS likely couldn't handle an integrated comms app with everything else it needed to do
 
rinku said:
How power hungry are the tensor cores relative to the cuda cores?

I was watching an old DF vid about DLSS on the Switch 2, way before the T239 leaks and they mentioned expecting around 20 TOPS out of the chip. Though I assume it’ll be much higher in the final gpu, could we expect Nintendo to drive more power to the tensor cores to improve DLSS performance?
Click to expand...
we don't know. there hasn't been any measure of tensor core power consumption. I would guess a benchmark has to be designed to do a task on compute only versus being done on tensor core only, and I don't think anyone has made such a bench.

the way tensor cores work is that they're in lockstep with the rest of the gpu unit (the TPC). increasing tensor performance, you have to increase clock speed of the whole TPC (the compute units, rt units, and tensor units performance goes up)
 
Does anyone think there will ever be fixed function CPU cores that only do directed graphs to massively improve gaming NPC performance?

I don't know if there's any fixed function CPU stuff out there at all...

Unless the decompression chips found in modern game consoles are fixed function CPU stuff? I don't think they are, but maybe I'm wrong...
 
Dekuman said:
Most Switch users are adults. They will definitely be kid friendly but I think they realized lacking voice chat and communication features was a mistake that may have been too hard to patch in

The single A57 dedicated to the OS likely couldn't handle an integrated comms app with everything else it needed to do
Click to expand...
DS could do it alongside a game. It probably wasn't a considerable technical hurdle, more one not worth jumping rather than an impossibility.
 
ItWasMeantToBe19 said:
Does anyone think there will ever be fixed function CPU cores that only do directed graphs to massively improve gaming NPC performance?

I don't know if there's any fixed function CPU stuff out there at all...

Unless the decompression chips found in modern game consoles are fixed function CPU stuff? I don't think they are, but maybe I'm wrong...
Click to expand...
Isn't the whole point of a CPU to do generalist tasks? That doesn't really fit with fixed functionality. And as far as I can tell, the decompression blocks in the PS5 and Xbox Series chips are on the I/O complex, not the CPU complex.
 
redmutineer75 said:
Isn't the whole point of a CPU to do generalist tasks? That doesn't really fit with fixed functionality. And as far as I can tell, the decompression blocks in the PS5 and Xbox Series chips are on the I/O complex, not the CPU complex.
Click to expand...
Yeah, it's not "in the CPU", but T239 does indeed have a suite of dedicated blocks. Decompression, tensor, OFA, etc.
 
Concernt said:
DS could do it alongside a game. It probably wasn't a considerable technical hurdle, more one not worth jumping rather than an impossibility.
Click to expand...
running the voice chat itself is possible as many Switch live service games integrate it into their apps, but the quality was not great. It just meant devs had to devote the game cores/memory to handling voice which is likely what DS , Wii and prior nintendo consoles did. Just have to budget some for the voice chat piece.

Switch having voice chat at the system level would offload that work from the game to the OS layer, plus there's the whole piece friendlist matching/connecting /communications dashboard to the OS core, which may not be up to the task, due to the very limited memory allocated to the OS.
 
Last edited:
Sin & Punishment said:
I don't think that's necessary.

Nintendo works with so many partners.

WayForward, Arc System Works, Platinum, Arika, Koei Tecmo, Ubisoft, MercurySteam, Iron Galaxy, Eighting, ArtePiazza, Atlus, Grezzo, the list goes on...

No need to acquire studios when you can just hire them for individual projects.
Click to expand...

I agree, only if one of those valuable partners are in "danger" of being acquired by some big company, Nintendo will buy them first, if not, there's no need to do it.
 
Dekuman said:
running the voice chat itself is possible as many Switch live service games integrate it into their apps, but the quality was not great. It just meant devs had to devote the game cores/memory to handling voice which is likely what DS , Wii and prior nintendo consoles did. Just have to budget some for the voice chat piece.

Switch having voice chat at the system level would offload that work plus there's the whole piece friendlist matching/connecting /communications dashboard to the OS core, which may not be up to the task, due to the very limited memory allocated to the OS.
Click to expand...
Your first point is true, absolutely, but the second point not so much. High quality voice chat might need a certain amount of memory it couldn't spare but a single core on Switch has many times the performance of some prior systems with voice chat, and just a few megabytes of memory would be needed. Nintendo Switch has more memory dedicated to the OS than Xbox 360 had memory period.

The problem is fitting that in, while possible, is still technical work, and the quality would be poor without further optimisations.

A mix of safety and juice not worth the squeeze.
 
Dekuman said:
When do you think realistically AMD will have a shot at getting Nintendo business again?
Click to expand...
Probably Never.

I can't see the future, but Nintendo probably value Nvidia partnership when they were at their lowest point and Nintendo using a cutsom made chip from Nvidia is quite telling.
Dakhil said:
In terms of a bespoke SoC existing, probably legit.

But as for AMD making a bespoke SoC for Nintendo, probably not legit.
Click to expand...

So MLID rumour of AMD trying to make a deal with Nintendo isn't true right?
 
Reader mode
You must log in or register to reply here.
Please read this new, consolidated staff post before posting.

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


Back
Top Bottom