Nintendo GOing Forward.

[Grall]

http://kotaku.com/is-this-a-new-wii-u-gamepad-i-honestly-dont-know-1671135159

Just throwing this out there. It certainly looks different to the current pad; way bigger screen/smaller bezels. Hopefully, capacitive touch as well - or so I would hope myself anyway because the current touchscreen is just awfully bad. Heh.

 

[Deleted member 86764]

I hope the gamepad is dropped for the next machine.

 

[Grall]

I'm almost 100% convinced the pad will be dropped like it's radioactive.

It's been a spectacular failure in every respect so far.

 

[Rodéric]

I'm almost 100% convinced the pad will be dropped like it's radioactive.

It's been a spectacular failure in every respect so far.

I'd still like the handheld to be the pad for the next home console server/hub. (The purpose of which would be to display higher quality gfx on HDMI devices, and run game server software for the handhelds [ie Multiplayer server like CS/Quake...].)
Or something along those lines.

Play anywhere, either on your big TV screen with improve graphics or on your handheld, but never stop playing !
Use your handheld if someone is using the TV.


From a technical PoV if the hardware was identical (feature wise) but only changed in performance (so quite scalable), you'd run the same game code on both machines, picking the right configuration options depending on the hardware.
One could also imagine that the home console server would free cycles on the handheld allowing it to generate better gfx, or whatever MS intends to do with its XBone Cloud service, but locally... (so no lag, high bandwidth, much better .)

 

[Grall]

I'd still like the handheld to be the pad for the next home console server/hub.

I entertained that idea some time ago, but most other B3Ders who responded pooh-pooh'd it, for some reason. It would have potential, sure, but also many of the pitfalls of the current wuu architecture, namely that the pad competes with your attention from the TV, a handheld as your controller is obviously larger and heavier than a regular gamepad, and also: cost. This is surely one of the major reasons wuu sucks so hard performance-wise.

If Nintendo wants to sell the handheld separately at a 3DS-like price, and the stationary console at a wuu-like price, the stationary console would not be much more advanced or powerful than the handheld - if at all, really.

If I was to go completely insane, I would speculate that to get around the hardware power/pricing dilemma, the next Nintendo console would use the handheld as its main CPU, with what is essentially just a bigger GPU in the stationary box, with some relatively simple embedded ARM cores for system management and I/O.

However, that, of course, would be completely nuts!

 

[eastmen]

the pad is the best part of the system. I hope next gen they simply design their handheld to be the pad

 

[McHuj]

If I was to go completely insane, I would speculate that to get around the hardware power/pricing dilemma, the next Nintendo console would use the handheld as its main CPU, with what is essentially just a bigger GPU in the stationary box, with some relatively simple embedded ARM cores for system management and I/O.

It would be a cool idea, but I assume it would be rather difficult to pull off correctly. Video is a lot easier to stream than data, but I assume all you need to stream are the drawcalls to the GPU if you duplicate the assets on both machines (assume actually the home console contains higher res assets than the portable). Let's say the next portable has a 720x480 resolution and the next home console is optimized to render that at 1080p with some high quality AA.

 

[Shifty Geezer]

and also: cost. This is surely one of the major reasons wuu sucks so hard performance-wise.

If Nintendo wants to sell the handheld separately at a 3DS-like price, and the stationary console at a wuu-like price, the stationary console would not be much more advanced or powerful than the handheld - if at all, really.

If Nintendo want to sell the stationary console separately, games would have to work without a second screen. And with the second screen as an optional extra, would it ever gain traction as a gameplay feature? I doubt it. So I think the synergy would boil down to playing the same games on portable and home console, and with the right OS and APIs, that wouldn't nee dhardware compatibility, although that'd help.

Let's say the next portable has a 720x480 resolution...

Considering 720p 4" screens are a dime a dozen, why would this still not surprise me as a spec from Nintendo.

 

[eastmen]

http://www.sankeibiz.jp/business/news/141216/bsb1412161938002-n1.htm


Sharp is speeding up its production of a free-form display for 2016 instead of 17 for Nintendo. They are experimenting with a donut-shaped display.

"The 12.3-inch Free-Form Display prototype demonstrated by the company hints at potential automotive implementations, but the design canvas seems pretty much open to wearables, odd-shaped smartphones and tablets, and customized digital signage.

 

[McHuj]

I bet it's for their QoL thing. IMO, the next handheld has to be out sooner than 2017.

 

[Newguy]

I think AMD would be able to build a custom SoC with 4-8 Puma cores and 4-8 GCN 2.0 CUs using 16/20nm+HBM for a Nintendo handheld in 2016.

.

4 core Puma, 4 CU GCN 2.0 GPU (4/500MHz or would that be too much?) would be insane for a handheld, although maybe I'm not looking far enough into the future. I'm not sure how accurate "running circles" round the Wii U is but you'd definitely be getting close. 1/2GB, 720p max res (mainly for battery life), anything else? Just adding to your point eastmen:

Nintendo is expected to use the displays in its new portable game consoles to replace the Nintendo 3DS lines, or for a recently announced device to map sleep quality, which the company wants to market by March 2016, the sources said.

Nintendo is considering creating a hole at the center of the display, making it doughnut-shaped, one of the sources said.

http://www.japantimes.co.jp/news/20...campaign=sharp-supply-free-form-lcds-nintendo

 

[Deleted member 86764]

Nintendo is considering creating a hole at the center of the display, making it doughnut-shaped, one of the sources said.

Someone needs to have a quiet word with Nintendo.

 

[Grall]

I assume all you need to stream are the drawcalls to the GPU if you duplicate the assets on both machines

Yes, and it should be a much smaller amount of data too than a video stream. Assuming you have proper error correction implemented, data integrity shouldn't be an issue either. After all, wifi is well established tech today and we send all sorts of critical stuff over the airwaves these days, generally without any problems whatsoever...

If Nintendo want to sell the stationary console separately, games would have to work without a second screen. And with the second screen as an optional extra, would it ever gain traction as a gameplay feature? I doubt it.

Second screen gaming hasn't caught on even WITH a second screen as default. I'd almost expect the handheld console's screen to blank out when using as a controller for the stationary console, both to save power, and because you don't have two screens available when playing the same game on the portable in standalone mode. But, I'm just speculating rampantly here.

Who knows what they'll do. *shrug*

Nintendo is considering creating a hole at the center of the display, making it doughnut-shaped, one of the sources said.

Donut-shaped display - what the hell for? That's just insane. I'm almost entirely sure it's just another "anal"yst smoking crack again (I swear, it's like these guys don't know the first thing about the tech industry, they just make shit up to hopefully stir up waves on the stock market), but you never know.

 

[Deleted member 13524]

Let's say the next portable has a 720x480 resolution and the next home console is optimized to render that at 1080p with some high quality AA.

Why would Nintendo assume a 3:2 screen ratio if the initial render is done in 16:9?

4 core Puma, 4 CU GCN 2.0 GPU (4/500MHz or would that be too much?) would be insane for a handheld, although maybe I'm not looking far enough into the future.

I don't know about GCN 2.0, but assuming GCN 1.x proportions, 4 CUs at 500MHz would result in 256 Gflops.
In comparison, 5 months old Shield Tablet with Tegra K1 has its GK20A doing 384GFlops (or about 340GFlops at more "realistic" clocks), and the 2 month-old ipad air 2 with A8X is doing 240GFlops FP32 + 360GFlops FP16.

Given how we already have that kind of performance right now, a 4CU GCN GPU @500MHz in 2016 will be pretty mediocre. By then, I'm expecting Qualcomm's midranges (Snapdragon 4xx and 6xx) to be around that, to be honest.

If we were talking about a truly progressive hardware design, we'd need to look at the Vita's SoC in late 2011 compared to anything else. It bundled a SGX543MP4+ in December 2011, which was only surpassed by ipad 3's similar SGX543MP4 a few months later, and yet the Vita carried twice the Cortex A9 CPU cores and a TSV-ish memory configuration for increased memory bandwidth.
The Vita-equivalent of a powerful handheld in 2016 would have 6-8 GCN CUs at 800-1000MHz (high-end SoCs should be approaching 1TFlops by then), together with 6-8 CPU cores and HBM doing over 40GB/s.
And yes, this is biting on the toes of XBone's performance.



However, 4 GCN CUs at 500MHz is quite above of what I'd expect from Nintendo for a handheld SoC in 2016.
Heck, looking at the 3DS, they would probably pull it off with a 2 CU design at 250MHz and a dual-core Cortex A7 at 800MHz, using 28nm from GloFo.
People wouldn't stop at this from buying pokemon and mario, and that's as far as Nintendo can see.

 

[Shifty Geezer]

Second screen gaming hasn't caught on even WITH a second screen as default. I'd almost expect the handheld console's screen to blank out when using as a controller for the stationary console, both to save power, and because you don't have two screens available when playing the same game on the portable in standalone mode. But, I'm just speculating rampantly here.

My rampant advising would say go with a Ninty mobile, maybe, and have a dock for it in a fancy controller so that the controller for the home console turns the mobile into a usable handheld. People could buy mobile, mobile+controller to make handheld, and/or console. Games can add smart-controller functionality for those using the mobile-enhanced controller.

 

[Fourth Storm]

Why would Nintendo assume a 3:2 screen ratio if the initial render is done in 16:9?


I don't know about GCN 2.0, but assuming GCN 1.x proportions, 4 CUs at 500MHz would result in 256 Gflops.
In comparison, 5 months old Shield Tablet with Tegra K1 has its GK20A doing 384GFlops (or about 340GFlops at more "realistic" clocks), and the 2 month-old ipad air 2 with A8X is doing 240GFlops FP32 + 360GFlops FP16.

Given how we already have that kind of performance right now, a 4CU GCN GPU @500MHz in 2016 will be pretty mediocre. By then, I'm expecting Qualcomm's midranges (Snapdragon 4xx and 6xx) to be around that, to be honest.

If we were talking about a truly progressive hardware design, we'd need to look at the Vita's SoC in late 2011 compared to anything else. It bundled a SGX543MP4+ in December 2011, which was only surpassed by ipad 3's similar SGX543MP4 a few months later, and yet the Vita carried twice the Cortex A9 CPU cores and a TSV-ish memory configuration for increased memory bandwidth.
The Vita-equivalent of a powerful handheld in 2016 would have 6-8 GCN CUs at 800-1000MHz (high-end SoCs should be approaching 1TFlops by then), together with 6-8 CPU cores and HBM doing over 40GB/s.
And yes, this is biting on the toes of XBone's performance.



However, 4 GCN CUs at 500MHz is quite above of what I'd expect from Nintendo for a handheld SoC in 2016.
Heck, looking at the 3DS, they would probably pull it off with a 2 CU design at 250MHz and a dual-core Cortex A7 at 800MHz, using 28nm from GloFo.
People wouldn't stop at this from buying pokemon and mario, and that's as far as Nintendo can see.

Does anyone know the thermal budget in your typical handheld system? Mullins, in its current form, has a TDP of 3.95-4.5w and an SDP of 2.8 watts. Still seems a bit high for a handheld, although suitable for a tablet. And the clocks are 500 Mhz at the highest (likely not for long). Actually, I wonder if with this OS abstraction layer, Nintendo will opt for fluctuating clock speeds. I could see them going for 2 Cortex A57 cores and 2 Cortex A53 cores for maximum efficiency.

On the topic of dual screens, I think that Nintendo may want to keep a DS style form factor in order to stay different and continue that popular line. But I also see them wanting to release a tablet. Remember Iwata mentioned they are looking into additional form factors. It's possible that they might release a DS styled device , but have dual screen functionality baked into the OS, so that you get your map view or whatever if you want even on a single screen. The user could enable this on the console if they own the handheld as well. They could also do a type of PiP or similarly imitate multiple screens on a large-screened tablet.

 

[Newguy]

I don't know about GCN 2.0, but assuming GCN 1.x proportions, 4 CUs at 500MHz would result in 256 Gflops.
In comparison, 5 months old Shield Tablet with Tegra K1 has its GK20A doing 384GFlops (or about 340GFlops at more "realistic" clocks), and the 2 month-old ipad air 2 with A8X is doing 240GFlops FP32 + 360GFlops FP16.

Given how we already have that kind of performance right now, a 4CU GCN GPU @500MHz in 2016 will be pretty mediocre. By then, I'm expecting Qualcomm's midranges (Snapdragon 4xx and 6xx) to be around that, to be honest.

However, 4 GCN CUs at 500MHz is quite above of what I'd expect from Nintendo for a handheld SoC in 2016.
Heck, looking at the 3DS, they would probably pull it off with a 2 CU design at 250MHz and a dual-core Cortex A7 at 800MHz, using 28nm from GloFo.
People wouldn't stop at this from buying pokemon and mario, and that's as far as Nintendo can see.

You're absolutely right, completely forgot the K1 was about 900MHz and for some reason I was equalling CUs and SMX/SMMs. Maybe getting a 2 SMM Tegra M1 around 1000MHz next year on 28nm (pure speculation on my part) so 4 CUs @ 750MHz on 16FF should be fine but as you say that's probably the upper limit I'd go to. Although it seems far too dreamy, having more GPU performance in their handheld than their console.

 

[east of eastside]

Nintendo's design direction post-N64 has always been low cost, low power with a tendency to do its own thing hardware-wise, there is no reason it will stray from that now, none.

Personally, I expect Wii U software to be recycled (since it was good and nobody bought Wii U) and it will be the all-digital (at least no-optical) and console/handheld integration aspects of the platform that will be the selling points.

I expect the atrocious game pad to be abandoned and the GC controller to be gloriously resurrected and updated. Camera and voice accessory would not surprise me.

 

[Nesh]

Nintendo's design direction post-N64 has always been low cost, low power with a tendency to do its own thing hardware-wise, there is no reason it will stray from that now, none.

Personally, I expect Wii U software to be recycled (since it was good and nobody bought Wii U) and it will be the all-digital (at least no-optical) and console/handheld integration aspects of the platform that will be the selling points.

I expect the atrocious game pad to be abandoned and the GC controller to be gloriously resurrected and updated. Camera and voice accessory would not surprise me.

Not sure about the bold part. Nintendo wants their console to be family friendly and suitable for all ages especially youngsters. Buying digitally only isnt exactly kid friendly. Not to mention that they most likely would want kids to nag parents for tangible Nintendo "toys" when they get to stores,

 

[McHuj]

However, 4 GCN CUs at 500MHz is quite above of what I'd expect from Nintendo for a handheld SoC in 2016.
Heck, looking at the 3DS, they would probably pull it off with a 2 CU design at 250MHz and a dual-core Cortex A7 at 800MHz, using 28nm from GloFo.
People wouldn't stop at this from buying pokemon and mario, and that's as far as Nintendo can see.

Honestly, I think this is very inline with what to expect for the next handheld. Low power consumption and low cost. I don't think nintendo can release a handheld that costs more than $150.