Nintendo GOing Forward.

[Shifty Geezer]

Benefits for end-users would be huge, plenty of back catalog to play with if games dry up at any period, a lot of different kind of games... Huge benefit.

Why buy a new console to play old games? Why not buy a cheaper old console? This is tying in to the BC discussion, but for Nintendo I think it's even more prevalent. They have had BC as an important feature for generations. I'd love to see the stats on that. Is everyone buying a NES 7 really going to spending their time playing Wii and GC games? Also, the market for Nintendo's back catalogue is dwindling as are their current customers. Surely it'd make more sense to have a clean break and offer hardware that's much more appealing to a wider audience than cater to the ever-shrinking existing Nintendo fanbase?

Also, you can't provide BC without a Wiimote and a Wuublet. If the gamer is to provide these themselves, they'll already have the old systems. These old systems aren't big so it's not a huge thing to keep them over. And as discussed in the BC thread, there's always the option of a BC addon to utilise parts of the console to keep prices at a minimum for those who want BC without increasing the base price with an unwanted feature for those who don't. Or just a WiiU Slim download only unit that nicely fits under/on the NES 7 console.

 

[orangpelupa]

hopefully if they implement BC using emulator, it will make their older game looks much better. Like when i play on Dolphin, its muuuuuuuch better than when i play on Wii.

 

[Globalisateur]

Yes it would be better on almost all points if they use emulators to do BC.

I can't imagine another Nintendo Frankenstein hardware (think Wii U) with the 3DS successor if they again include, in a way or another, a whole 3DS hardware in its successor.

 

[function]

BC on the Wii3 would likely only require the Wuu CPU, and the die area it'd need would be tiny on 28nm. You could probably stick it into an APU in the same way on that MS put their Shape + Kinect stuff onto their APU.

Licensing and R&D costs permitting, of course. Still fairly expensive, but a lot better than including separate physical chips.

Given that the Wii U is building up a small but decent library of games now, a "full HD" Wii/U compatibility mode would help Nintendo's library early on. Because lets face it, no-one is going to play these (pretty cool) games on the Wii U ...

 

[Shifty Geezer]

As a coprocessor it makes some sense, like the background ARM chips in the current consoles. Probably overkill, but it could be used for something. Maybe also the old favourite audio processing? Enable it to clock up higher when not in BC mode. [shrug]

 

[function]

I suppose that's another way of looking at it. Those PPC cores are pretty efficient after all, and have been used in lots of embedded roles.

And I it might allow them to easily run their existing Wuu OS on the new device and so transition more easily, and also keep their entire back catalogue of virtual console stuff available without interruption. And it might allow Nintendo to avoid having to reserve ARM/x86 cores for the OS, like PS4Bone do...

 

[I.S.T.]

BC on the Wii3 would likely only require the Wuu CPU, and the die area it'd need would be tiny on 28nm. You could probably stick it into an APU in the same way on that MS put their Shape + Kinect stuff onto their APU.

Licensing and R&D costs permitting, of course. Still fairly expensive, but a lot better than including separate physical chips.

Given that the Wii U is building up a small but decent library of games now, a "full HD" Wii/U compatibility mode would help Nintendo's library early on. Because lets face it, no-one is going to play these (pretty cool) games on the Wii U ...

People spent years talking about a Wii HD that would play the games in 720p with 4X MSAA... I'd not get your hopes up.

 

[Fourth Storm]

Unless, of course they sell a home console that's a kind of hub and steroid boost for a 3DS successor, in which case you'd have everything required anyway...

I agree about that, your licence should carry on, you bought the title you should be able to benefit from it whichever machine you run it on, like a PC game...
Benefits for end-users would be huge, plenty of back catalog to play with if games dry up at any period, a lot of different kind of games... Huge benefit.

(I like to play Chrono Trigger for exemple, it's a great game and I find it better than many "modern" games...)

Chrono Trigger is indeed a joy. I still have to beat Lavos on the DS version. I must be underleveled, because he seems harder than he did when I was a kid.

I'd prefer they let their next handheld function as a new Gamepad rather than have the console act as a hub. I find the latter pretty unlikely given the comments coming out of Kyoto over the last year or so. I do kind of like Shifty's idea of a plug-in for BC support. With SATA or USB3.0, they should have enough speed to use some of the hardware of the main console. There is definite value in the Wii U's library. They need to carry it over either via the way we just discussed or via straight up hardware BC, because as Function says, not enough people are going to get to experience most of these quality games.

BC on the Wii3 would likely only require the Wuu CPU, and the die area it'd need would be tiny on 28nm. You could probably stick it into an APU in the same way on that MS put their Shape + Kinect stuff onto their APU.

Licensing and R&D costs permitting, of course. Still fairly expensive, but a lot better than including separate physical chips.

Given that the Wii U is building up a small but decent library of games now, a "full HD" Wii/U compatibility mode would help Nintendo's library early on. Because lets face it, no-one is going to play these (pretty cool) games on the Wii U ...

Agreed on that last bit, as I indicated above. Throwing Espresso on the SoC (shrunk down) is a possibility. One has to wonder how much integration that would take and whether it would stretch the bounds of "semi-custom," but MS has set a precedent with all their custom audio blocks and whatnot. It may be the only way, with IBM looking to sell off their foundries.

The SoC would likely be fabbed on 20nm tech as I'm anticipitating FinFET yields in 2016 will ward off Nintendo. The SRAM is going to take up alot of room. Considering the target size for the chip will probably be something like half that of the current Sony/MS machines, I'd say we're looking optimistically at specs comparable to Xbone.

 

[Goodtwin]

BC on the Wii3 would likely only require the Wuu CPU, and the die area it'd need would be tiny on 28nm. You could probably stick it into an APU in the same way on that MS put their Shape + Kinect stuff onto their APU.

Licensing and R&D costs permitting, of course. Still fairly expensive, but a lot better than including separate physical chips.

Given that the Wii U is building up a small but decent library of games now, a "full HD" Wii/U compatibility mode would help Nintendo's library early on. Because lets face it, no-one is going to play these (pretty cool) games on the Wii U ...

That's why I think they would just include the CPU, a dye shrink will make it very small, and probably pull 1-2 watts.

Lots of people are already playing some of those pretty cool games. Wii U already has multiple million sellers. Over two million people have bought Mario Kart in just one month, 4 million have bought NSMB:U, about 2 million Mario 3D World, Zelda WW HD about a million units. People need to stop acting like Wii U is a dead console, its a niche console, and a product that will likely have a larger userbase than the X1 until 2015 or perhaps even 2016. At least there is a solid reasons that's justify a Wii U purchase, exclusive software. Sony and Microsoft fight it out for those COD, Assassins Creed, and Batman gamers, all games that will be better on PS4, so other than Halo, there is very little reason to choose an X1 instead of a PS4, and very little inventive to own both an X1 and PS4. If your a PC gamer, then Wii U is probably the only console that justifies a purchase.

@Shifty, no one is buying a new console to play old games, but that doesn't mean people wont buy old games on their new console. There is a ton of revenue potential for virtual console games. A big virtual console catalog by itself wont sell new hardware, just like an internet browser or Netflix functionality wont sell new hardware, but its an attractive perk that can push someone over the edge if they were on the fence.

 

[DieH@rd]

If they really want HW BC, they should put 3DS GPU on its portable successor, and small WiiU CPU on its home console successor. Both chips are already very small, and could be made even smaller when made in 20nm, just small modues inside larger SoC/APUs.

 

[Rodéric]

If they really want HW BC, they should put 3DS GPU on its portable successor, and small WiiU CPU on its home console successor. Both chips are already very small, and could be made even smaller when made in 20nm, just small modues inside larger SoC/APUs.

Sony did that for the PS2, I think it was also the case for the Japanese PS3...

 

[Syferz]

I'll share my post from Neogaf yesterday in here, just some speculation but I hope it's interesting to everyone. (side note, I messed up on the gflops, I should of pushed the clock to match 176GFLOPs which would match Wii U's and not the 172 that I go on about in this post)

I've been thinking about this for a while now. The simple answer is that they will use an AMD APU. 2016 for their handheld and 2017 for their console. Here is how it would work and carry over BC as well as their current Wii U library of VC titles/eshop titles.

First some realities:
Nintendo's oldest partner in console making is the S3/ARTX/ATI/AMD team where they have hired over some of the staff themselves and have a very strong working relationship.

AMD's "Mullins" is a 28nm APU with 128 ALUs @ 500MHz (Wii U has 160 @ 550MHz = 172 GFLOPs) and a quad core 2GHz Puma core, these APUs run at 2.8 watts (which is lower than psp, but of course this is just the SoC and not the complete system)

AMD is moving to 20nm next year and 16/14nm in 2016 as well as new architecture for their cores (k12 arm/x86)

Wii U uses about 28watts without the disc drive, ~20% of that is likely PSU power efficiency loss, so you are closer to 22 watts, USB, Wifi, Wifi(gamepad signal), bluetooth, ram and other misc components will use up another ~8watts, this system is built on 45nm process for the CPU/GPU (GPU is possibly 40nm) these are drastically old parts that could be reduced as much as 3 times to reduce their power draw. Infact there is little in the Wii U that couldn't of literally been done in 2009 from a technical standpoint. This entire system can be created in a handheld size with 20nm process under the formfactor's power restrictions

The simple setup:

3DS uses 2 CPUs, one used in games and one used in the DS for BC and to power the OS, this was the key to where Nintendo is most likely going with their absorption of Wii U architecture, NNID and VC going forward.

They are likely to use an AMD APU for their handheld, possibly only 2 cores but I'd venture to guess a quad core is just as likely. For the GPU side you could see 3CUs running at 450MHz (192 ALUs @ 450MHz = 172 GFLOPs) with Espresso cores for OS/BC/VC compatibility with Wii U.

This is also the likely answer for Wii U's successor as well, Espresso as a co-processor makes a lot of sense.

Handheld: ~2016
[polygon performance is actually higher on GCN afaik, you get 2m poly per mhz, so a higher clock can push more polygons but more "shaders" or ALUs gives a better GFLOPs performance and is also more likely considering higher clocks effect battery more]

Pessimistic view:
CPU: Espresso co processor for VC/BC/OS; AMD ARM APU Dual core 1.6GHz on 20nm process
GPU: AMD APU 2CU (128 ALUs GCN2 @ 670MHz for 172GFLOPs) on 20nm process
RAM: 2GB
Screen: 2x ~4 inch 480p screens (Lower resolution gives a better performance so this might not be a huge negative)

Optimistic view:
CPU: Espresso co processor for VC/BC/OS; AMD ARM APU Quad core 2GHz on 20nm process
GPU: AMD APU 4CU (256 ALUs GCN2 @ 500MHz for 256GFLOPs) on 20nm process
RAM: 4GB
Screen: 2 ~5 inch 540p screens (qHD gives the best reduction of 1080p graphics which we would see a lot of in Nintendo's next console)

Console: ~2017

Pessimistic view:
CPU: Espresso co processor for VC/BC/OS; AMD ARM 8 core 2GHz on 20nm process.
GPU: AMD APU 14CUs (896 ALUs GCN2 @ 1GHz for 1792GFLOPs | ~PS4 spec) on 20nm process.
RAM: 8GB DDR4 ~150GB/s + 35MB edram

Optimistic view:
CPU: Espresso co processor for VC/BC/OS; AMD ARM 8 core 3GHz on 16/14nm process
GPU: AMD APU 20CUs (1280 ALUs GCN2 @ 1.2GHz for 3072GFLOPs)
RAM: 16GB DDR4 ~150GB/s + 35MB edram

[note: Sorry if the listed specs above is more of a range than some might be expecting, but the truth is I don't know the future and so speculating on specifications of a device without intimate knowledge is really just guess work, so rather than that, I looked at AMDs future products and selected likely candidates from 2 view points for both devices.]

This allows Nintendo to legacy their software through a service, connect both devices going forward into 1 dev cycle and allows the OS to become quite a bit more capable with background services (system wide voice chat, gameplay recording/uploading) all while keeping their current tools and apps alive.

I'd also like to point out that the OS they are using on Wii U could be carried over running right on espresso, this also means they never have to dump their previous catalog of games going forward again. Nintendo isn't Microsoft as we all know, so building up their current OS it probably a lot better for them than starting from scratch again considering how badly the Wii U ran 20 months ago.

 

[Wynix]

If the handheld SoC is 20nm it will be more powerful than the Wii-U, of this I have no doubt.

I just hope the memory configuration is good, a best case scenario would be 2-4GB of 2nd gen HBM(underclocked?) giving 80GB/s+ bandwidth to the APU.

 

[liolio]

I'm not sure I'm pleased by this talk about custom hardware and the various implication for the design from cost, to the choice wrt BC as well as other things.

I make sense for Nintendo to focus foremost on handheld, now I hope that they will properly evaluate how the handheld market has evolved through the year.

As those AMD representatives stated handheld still sold. Though there are changes, casual are gone and lost to mobile devices. They won't come back.
So Nintendo is left with more "core" gamers than it used to.
I also think the usage for the device changed: used to be used on the go by casual but quite possibly in a more sedentary manner by more core audience (from kids to feminine audience passing by testosterone train wrecks).
I always deemed the clamshell design superior to others approaches, but now as usage is I think evolving I'm no longer that sure. Other design have benefits, I find the new 2DS a lot more comfortable to handle than the 3DS or 3DS XL.
Clamshell design was great when people used to fit the device in their pockets, now most the people (if not all) I've seen using handhelds in the last years (from kids to workmates than used do play MP at lunch break) carry the device in a dedicated case along with the games.

So I think that opens the option for form factors. I actually wonder if both Sony and Nintendo got the last generation of handheld "wrong", aka they have not properly evaluated what that market could see as selling point for gamers that do not "jump in".
As I say I think that usage have slightly changed and that change makes room for reworked design: foremost I think both the PSV and the 3DS fails to deliver a screen of a decent size. By nowadays standard 5 inch is nothing special.

I think that battery life is another point that need to improve. They need to learn from Nvidia and its Shield.

As they should try to address more "serious" gamers they also badly need to ramp up their game wrt to hardware specifications. I think it is unreasonable to expect anything high end as seen in some phones sold +500$ but they can't release sucky hardware anymore.

I don't expect them to offer awesome bang for bucks either, but again they need good, well rounded hardware (SoC, screen, battery, etc.).
I think they should aim for 150$ max at release, adult core gamers may have agreed to sink more into a proper handheld but then it turns into an obstacle for parents.

For that kind of money the best you can buy are product akin to the Motorola Moto E for phones and those new Asus Memo Pad 7 (which use those nice BayTrail Intel is willing to let go at a discount), there are also nice product coming Xaomi among others.
Compared to though company Nintendo needs to present something a bit more rustic that parents would be comfortable handing to kids, they need to fit good and solid controls. They also need to provide good battery life while handling the single most demanding task a mobile (/self powered) device can run which is advanced games.
To keep its model going I think Nintendo needs to make money on hardware.

Now the bad part is that arbitrations need to be make to get there, that is my take obviously:
*they can't afford 2 screens whatever the impact on BC.
*they should pass on custom SoC and leave their option open as long as possible and go for the best possible deal, again with a possible impact on BC. The market is competitive there are lots of opportunities, as an example it seems that Intel is heavily discounting those baytrail to make a dent in the market, I suspect that Nvidia does/did the same at some point.
*BC should not be priority, Nintendo has done a good job on that front but like all others console manufacturers it did the wrong way. It should not be about backward compatibility through hardware but forward compatibility through software. Nintendo should no longer make a system that look back way too far in the past.

 

[Syferz]

Considering espresso is so tiny and costs them just a few dollars, it would be fairly stupid for them to drop it completely, it just shouldn't be the main cpu. Nintendo is the only console manufacturer that can do bc completely while going forward and they absolutely should.

 

[Fourth Storm]

Always enjoy some liolio speculation. :smile:

If Nintendo are doing what I think they are doing, I don't think you need to worry about the device being bogged down by BC. The 3DS SoC should be able to be shrunk down to a single small block on the new chip. If they are going with AMD, I can imagine they are getting a good deal. The part wouldn't be entirely custom, and AMD are looking to get a foot in on mobile/tablets, so pricing is probably a bit more flexible than going with the current major players, like Qualcomm.

On the home console side, they could go with either ARM or x86. If they are indeed developing an OS such as Android, there is little reason to go with ARM. Puma+ cores are also tiny and support multithreading and probably better SIMD (have to read up more on that) than the ARM A57. Again, they could add the Wii U CPU as a small module on the SoC, but that may end up being too complex an integration to be worth it imo.

I think you are correct on Nintendo's target audience. They can broaden it with a more attractive device. 3DS looks dated both in shape and technology. The 3D effect also places a concern on parents. At the last shareholders meeting, an investor brought up that parents have a concern that small screens can also cause eye strain in children, and I think he has a point. I believe for the initial version of the handheld, it should be a tablet, at least 6"-7" and 720p. One screen with modular controls as described in this Nintendo patent. Seems like a great gimmick that would appeal to kids and hardcore gamers alike and allow the standard unit to not ship with too many buttons. They could also potentially get 3DS BC to work in a vertical position with the right control modules.

I have doubts that consumers have issue with carrying more than one device. Kids and gamers especially usually have backpacks, laptop cases, or even car glove boxes. People use these devices alot in the home as well. The problem is the 3DS has not offered enough software outside the dedicated Nintendo/otaku fanbase. A better OS and a screen suitable for watching Netflix and Hulu is a requirement if they want people to bring this machine with them. Nintendo cannot and should not be looking to compete with smart phones, but they could make a compelling gaming tablet while pushing to get the most popular media apps and mobile games. Their best bet is returning to the quality over quantity philosophy they used in the NES days. Developing more educational software would also be to their benefit in adding value in the eyes of parents.

 

[liolio]

Some design bias I think would serve the system well:

Screen:
*6-7 inch (/whatever is readily available and fit their budget) screen with a 4/3 form factor.
*Used in a vertical manner (so 3/4 in fact), so it leaves a door open for BC and/or remastered version of old classics.
* capacitive screen.

Nintendo could keep the DS concept alive (2 different display) and have the game rendered on the upper area and reserve the lower part of the screen for UI and extra information as done on the DS line of consoles. Though I think it would be better to let developers (first or third party) choose what they want to do with the screen.

Shape:
I would stick to the 2DS design, so more like a tablet and with some meat to it. EDIT in fact a square , more compact gamepad.
Thanks to the 3/4 display the device and accounting for controls the device should be ~ scare, a more compact version of the Wii U gamepad while providing a bigger screen. Thinner bezels should help.
The "thick" design would allow Nintendo to use high capacity "cylindrical" battery as Nvidia did in the shield. It also allows for more comfortable triggers.

Input: similar to Wii U gamepad though I would invert the button and the analog stock on the right side (/Xbox lay-out).

Game: they should still be on ROM,
Increase in capacity should be reasonable/ low. I wish Nintendo would go with an hybrid model, akin to what happen with BF4 on the 360: the game card would include low quality assets, higher quality version of the asset would be downloaded from Nintendo's cloud.

Storage: tiny SD card, user replaceable. 2 slots

SoC: I think a mid-end SoC would do, they should go for the best deal they can get. The beefy (/thicker) casing (compare to tablet) should allows for a sane passive cooling solution that may prevent throttling. I would avoid X86 even if Intel is offering great discount as it would bind them to at best a couple of sellers.

Memory: I think that is an area where Nintendo should not fall for money pinching.
I think 4GB would be a great choice.

OS: grow from the 3DS they should try to keep the footprint tiny. They should not try to do everything but improve on what they are already doing.

Twistable: I think it could be a nice gimmick if the system ends up a perfect square. One could twist the system 90'

Portrait mode would be for gaming, with some (varying) amount of the screen dedicated to the UI and why not notification from a couple of relevant services, sort of following the step of the DS line. An untold benefit would be a significant reduction of the number of demanding pixels rendered on the screen.

Landscape mode would be dedicated for web/OS in that position thumbs would be way closer from the screen allowing one to type conveniently.
Some game could make use of the mode if they don't require std inputs.
Media/web would most likely be consumed that way too.

Overall philosophy:
*bring a bigger screen to gamers.
*Inheritage of the DS, the "odd" screen form factor that will be use more freely for games UI and/or services.
*Bog standard hardware with good amount of memory: try to get ports of games and support for the most relevant middleware providers.
*Good CPU performance. Sony stated that with the PS$ they want to reach users that passed on the ps360 and that there is a lot of content that could be reissued on their new console. If true it could also be a great new for Nintendo, with a sane SoC, games of the ps360 era could also be ported to their system, they won't look as good as on the ps4 or the One though the extra RAM will allow the system do a better showcase of the artists work (/assets) than last generation systems did and that with a portable system.
Good CPU performances along with the healthy amount of ram may also allow from some ports of the current gen at the costs of dumb down graphics (+lesser resolution).
*If possible bring them BC for the 3ds, but otherwise virtual console and the screen form factor should help with all the DS library (vanilla, and then 3/2 DS)
*Affordable price.
--------------------

Say they go with custom hardware, AMD more precisely. I hope they go with something akin to Durango and pass on what could be at first expensive tech (HMC) and may not come to mobile first (I've no idea).

I would see that kind of set-up "fit"
8 A53 with good amount of cache.
4 CU/SIMD GCN2
true audio2
16MB of eSRAM
128 bit bus to 4 gb of low power memory.
Runs on Mantle or a

 

[Fourth Storm]

Liolio: Your suggestion of a 4:3 screen got me to thinking. I think that's actually the way to go. In landscape, they could use scaling (as they do with DS games on 3DS) for BC. I think portrait would actually be somewhat worse for BC and make home console ports more difficult to implement, so following the idea in that patent, I'd put the slots for the control modules in landscape position. Nothing wrong with 4:3. They can get away with rendering less pixels and iPad still uses it. 1024x768 displays should be dirt cheap in a couple years (as they already are pretty much).

I agree with you on the thickness of the device for cooling purposes and grip, but I'd use the 2DS as a limit. I also wouldn't want to include the unsightly shelf or grips of the Wii U Gamepad. They can include 4 shoulder buttons, but arrange them all in a row as on the original Wii Classic Controller. I'd also like a much thinner bezel on the top and bottom of the screen.

I would stick with resistive touch, if only because they seem to have a good thing going on with art software as well as Miiverse sketches. Fujitsu offer dual touch resistive panels for less than the price of capacitive. That should also help keep costs in the $150-$200 range.

As far as the internals, I think a realistic spec would be as follows:

20nm Semi-Custom AMD SoC
CPU: Dual Core ARM A57 @ 800Mhz-1000Mhz
GPU: 128 GCN2 Cores @ 400Mhz-500Mhz
RAM: 2 GB LPDDR4 @ 3200 Mhz ~12.8 GB/s + 16 MB eSRAM (4GB LPDDR4 at double the bandwidth would sure be nice though)
3DS BC module on main SoC: Dual ARM11s, ARM9, PICA 200 GPU

 

[Wynix]

Again, they could add the Wii U CPU as a small module on the SoC, but that may end up being too complex an integration to be worth it imo.

This is such a far-fetched idea I am having a hard time figuring out if it's sarcasm or not.

 

[McHuj]

I don't think it's that far fetched. I believe IBM has stated that they are willing to license designs so I bet Nintendo could acquire the IP and manufacture on their own. From a die size, the WiiU CPU is aorund ~28mm on a 45nm process. That would be probably around 10mm on 28nm. And since it's a low clocked part, it should be portable to other manufacturers.

However, I'm not sure it's needed. The question I have would be possible to customize (or add some sort of hardware assist) to an existing ARM core to help with backwards compatibility. It's a simple RISC core without any fancy vector extensions/coprocessor so it may be feasible and cheaper.

I think the WiiU's GPU with it's embedded RAM is a bigger obstacle to backwards compatibility.

The way I see it, Nintendo's next handheld will launch on a very cheap and mature 28nm. I don't believe 20SOC will be cheap enough nor mature enough in the time frame that Nintendo want to launch (late next year, speculation).