Nintendo GOing Forward.

I don't think we will see 64 gigs in the next gen. Maybe 8 gigs of fast stacked ram and another 16 gigs of slower ram
 
See my last post. Flashy visuals sell.
So PS4's sales dominance has nothing to do with the fact that it's a powerful console and has the best looking games ?..
Graphics are part of the puzzle. If you're selling the same games to people as your rival, or you're selling a new generation of hardware, having flashier visuals helps. If you are selling your machine based on a gimmick like Waggle, flashy visuals aren't necessary. Or at least, Nintendo doesn't think so, hence them putting the worst ever console hardware in Wii and still making a bundle. Wii could have been a far better machine, not as powerful as PS360 to remain small and cheap but still better than Nintendo released, yet Ninty went with a 'graphics don't matter' attitude. They've done similar with WiiU, putting in limited power from the off.

So what's the incentive to put in more powerful current-gen graphics in their next machine? If the games are first party only, Nintendo knows from experience that their games are well received even using outdated tech (see Mario Galaxies rim-lighting shader for example, or Captain Toad: Treasure Tracker which isn't doing anything particularly fancy using baked lighting, but people love the look). Something of the power of PS4 will be plenty for Nintendo's first-party titles. I'd dare say it'd even be more power than Nintendo could get a handle on; they don't like big, expensive AAA games by and large. Mario Kart will look just as good on a machine with half PS4's power as a machine with 2x PS4's power, so where's the incentive to push the hardware envelope? Especially, as others say, if you're after handheld synergy. A massive differential between home and handheld power won't do Nintendo any favours. If they can create something that looks good on handheld, and then scales up to HD with minimal effort, that's more in keeping with their vibe, IMO. A game with two render pipelines, one simple for handheld and the other a complex, voxelised GI lighting model, could just as readily be two games that share a save file meaning no need for a shared hardware architecture.

If sharing architecture, a powerful home console seems very much off the cards.
 
Graphics are part of the puzzle. If you're selling the same games to people as your rival, or you're selling a new generation of hardware, having flashier visuals helps. If you are selling your machine based on a gimmick like Waggle, flashy visuals aren't necessary. Or at least, Nintendo doesn't think so, hence them putting the worst ever console hardware in Wii and still making a bundle. Wii could have been a far better machine, not as powerful as PS360 to remain small and cheap but still better than Nintendo released, yet Ninty went with a 'graphics don't matter' attitude. They've done similar with WiiU, putting in limited power from the off.

So what's the incentive to put in more powerful current-gen graphics in their next machine? If the games are first party only, Nintendo knows from experience that their games are well received even using outdated tech (see Mario Galaxies rim-lighting shader for example, or Captain Toad: Treasure Tracker which isn't doing anything particularly fancy using baked lighting, but people love the look). Something of the power of PS4 will be plenty for Nintendo's first-party titles. I'd dare say it'd even be more power than Nintendo could get a handle on; they don't like big, expensive AAA games by and large. Mario Kart will look just as good on a machine with half PS4's power as a machine with 2x PS4's power, so where's the incentive to push the hardware envelope? Especially, as others say, if you're after handheld synergy. A massive differential between home and handheld power won't do Nintendo any favours. If they can create something that looks good on handheld, and then scales up to HD with minimal effort, that's more in keeping with their vibe, IMO. A game with two render pipelines, one simple for handheld and the other a complex, voxelised GI lighting model, could just as readily be two games that share a save file meaning no need for a shared hardware architecture.

If sharing architecture, a powerful home console seems very much off the cards.

Most of this view point is based on the success of the Wii though. They went with a similar ideal for WiiU (low powered hardware with a 'unique' controller) and it has failed momentously.

Remember WiiU's final specs were a surprise to a LOT of people (just ask the guys from the WiiU speculation thread on gaf). It was expected to have a decent 3 core CPU, 3GB's of decent RAM and around a 1TFLOP GPU.

Do people really think that Nintendo were serious about third parties by going for an overclocked Gamecube CPU, 1GB of extremely slow RAM for games and a 176GFLOP GPU with WiiU ?. I personally believe the WiiU's specs were significantly downgraded between E3 2011 and E3 2012. It's the reason most third parties dropped out and the reason we got Mario 3D World instead of Mario Galaxy 3 and the reason we are getting a Zelda game that looks to be a cross between Wind Waker and Skyward Sword visually instead of the highly stylized realism of the Zelda WiiU tech demo.

I do take your point that if they go too powerful with their next console it wouldn't be beneficial to making games across handheld and console but they do have to show a large leap on screen from WiiU to entice people to upgrade and that's harder to do with Nintendo's art direction.

If all Nintendo's next console ends up being is PS4 levels of performance then I will personally be over the moon, I really think Nintendo would blow other developers away with a GPU that is 10x as powerful as WiiU's esp if it's combined with even 8GB's of RAM. I just think that AMD will offer them far, far more performance per doller in 2016 than PS4 if they are going for a modern SoC architecture.
 
Nintendo wont make a console 50%-100% more 'powerful' than the PS4. Nintendos consumer base is largely kids at this point. It will be difficult to pull over alot of PS4 & Xbox gamers, who don't know if their favorite developers will support this upcoming Nintendo platform IN THE LONG TERM and who's exclusive titles outside of bayonetta and a couple others are not mature themed. Their core demography are kids, and the parents generally spend less money on kids videogame consoles, compared to adults and older teenagers who have different tastes and whos who can also put in their own money because they have a job.

The console needs to hit a lower price point than Xbox1 and PS4.
 
Last edited:
I just think that AMD will offer them far, far more performance per doller in 2016 than PS4 if they are going for a modern SoC architecture.

Only problem is if it really is that much more powerful, why not save a bit and have around PS4 performance. It's still a large jump from their previous effort, third parties will know what they're doing (they know what to aim for, similar targets since they're used to PS4, little time required to devote) and they can release at a good price point. Don't get me wrong I would like them to try it but I don't think they'd risk it.
 
Most of this view point is based on the success of the Wii though. They went with a similar ideal for WiiU (low powered hardware with a 'unique' controller) and it has failed momentously.

Remember WiiU's final specs were a surprise to a LOT of people (just ask the guys from the WiiU speculation thread on gaf). It was expected to have a decent 3 core CPU, 3GB's of decent RAM and around a 1TFLOP GPU.

Do people really think that Nintendo were serious about third parties by going for an overclocked Gamecube CPU, 1GB of extremely slow RAM for games and a 176GFLOP GPU with WiiU ?. I personally believe the WiiU's specs were significantly downgraded between E3 2011 and E3 2012. It's the reason most third parties dropped out and the reason we got Mario 3D World instead of Mario Galaxy 3 and the reason we are getting a Zelda game that looks to be a cross between Wind Waker and Skyward Sword visually instead of the highly stylized realism of the Zelda WiiU tech demo.

I do take your point that if they go too powerful with their next console it wouldn't be beneficial to making games across handheld and console but they do have to show a large leap on screen from WiiU to entice people to upgrade and that's harder to do with Nintendo's art direction.

If all Nintendo's next console ends up being is PS4 levels of performance then I will personally be over the moon, I really think Nintendo would blow other developers away with a GPU that is 10x as powerful as WiiU's esp if it's combined with even 8GB's of RAM. I just think that AMD will offer them far, far more performance per doller in 2016 than PS4 if they are going for a modern SoC architecture.
If you expected the wii u to be 1TF then there was no hope for you. It was expected that the max performance they could hit would be 400GFLOP from the power consumption of the device. The people on Neogaf just have no reference for hardware specs if they thought the wii-u was anywhere near that powerful. The reason third party dropped out had little to do with the performance of the wii-u. They dropped because it wasn't economical to port their game to a system with so little interest in buying their games and how bad nintendo's dev support is.
 
Most of this view point is based on the success of the Wii though. They went with a similar ideal for WiiU (low powered hardware with a 'unique' controller) and it has failed momentously.

Remember WiiU's final specs were a surprise to a LOT of people (just ask the guys from the WiiU speculation thread on gaf). It was expected to have a decent 3 core CPU, 3GB's of decent RAM and around a 1TFLOP GPU.

Do people really think that Nintendo were serious about third parties by going for an overclocked Gamecube CPU, 1GB of extremely slow RAM for games and a 176GFLOP GPU with WiiU ?. I personally believe the WiiU's specs were significantly downgraded between E3 2011 and E3 2012. It's the reason most third parties dropped out and the reason we got Mario 3D World instead of Mario Galaxy 3 and the reason we are getting a Zelda game that looks to be a cross between Wind Waker and Skyward Sword visually instead of the highly stylized realism of the Zelda WiiU tech demo.

I do take your point that if they go too powerful with their next console it wouldn't be beneficial to making games across handheld and console but do have to show a large leap on screen from WiiU to entice people to upgrade and that's harder to do with Nintendo's art direction.

If all Nintendo's next console ends up being is PS4 levels of performance then I will personally be over the moon, I really think Nintendo would blow other developers away with a GPU that is 10x as powerful as WiiU's esp if it's combined with even 8GB's of RAM. I just think that AMD will offer them far, far more performance per doller in 2016 than PS4 if they are going for a modern SoC architecture.

We've had numerous devs come forward and say that the early dev kits were actually less capable. The GPU was clocked at 400 Mhz initially and the CPU ~1 Ghz, I believe. Third parties knew it was going to be a low power console right from the very first meeting.

As to the style of their current offerings, I hardly believe that this is due to technical limitations. Mario Galaxy 3 would have looked spectacular on Wii U hardware, as does 3D World. Similarly, the Zelda tech demo's style could have been achieved had they wanted to. That was just TP assets at higher resolution with some fancy lighting. Games like Skyrim pull off realism on last-gen hardware just fine. It doesn't matter if the textures are created from photographs or drawn by a cartoon artist...

AMD will create an SoC under the guidelines of what Iwata, Takeda, and company ask of them. If that recent job listing is any indicator, having a low power console and low cost hardware is still very much part of the "Nintendo DNA." Miyamoto and Takeda are currently training people to carry on their hardware ideals after they retire, so unless this next generation completely revoke those lessons in a few years, Nintendo's philosophy does not look to change any time soon.
 
Here is a question then, do you guys think Nintendo can show a graphical generational leap over WiiU if the next console is another cheap, tiny, low powered offering ?.
 
Here is a question then, do you guys think Nintendo can show a graphical generational leap over WiiU if the next console is another cheap, tiny, low powered offering ?.

Depends on how cheap, tiny, and low powered Nintendo go. They could shrink the existing Xbox One console CPU/GPU and get a graphical generational leap over the WiiU.
 
I think they can get somewhere in between Xbone and PS4 if they choose to stick with a TDP of ~30 watts and TSMC's 16nm or GF's 14nm are available. Let's look at some of AMD's GPUs out there. According to Wikipedia, the Radeon HD 8770 measures 160 square mm and has a max TDP of 85 watts (and that's the whole card w/ GDDR5, mind you). TSMC promises their finFET process reduces power consumption by 70% over 28nm. So that's 25 watts for 1792 gflops. Nintendo might choose to clock their SoC a bit more conservatively, but if they get on the right process node, then a low power system in the ballpark of current gen Sony/MS consoles is entirely feasible.
 
@Fourth Storm

In early 2012 'Bgassassin' who was a vetted neogaf 'insider' said several times that WiiU would aim for a GPU around the 1TF range, he later lowered that estimate to 600GFLOPs around late 2012 just before launch.

WiiU ended up being 176GFLOPs according to you and many others best guesses based on the chip plans.
 
IGN heard the 1 TF figure as well. Rumors happen. Bgassassin was going on the best info he had, but I think that the info was corrupted somehow by the time it got to him. In any case, we had a couple devs come onto GAF and verify that early dev kits were, in fact, weaker. 400 Mhz GPU. There was also an article which backed up this assertion. I can't recall who published it atm, unfortunately.
 
I do take your point that if they go too powerful with their next console it wouldn't be beneficial to making games across handheld and console but they do have to show a large leap on screen from WiiU to entice people to upgrade and that's harder to do with Nintendo's art direction.
Upgrade from what? From PS4/XB1, ain't gonna happen. Even a substantially faster console isn't going to show enough difference to offset the advantages of the older systems. It'd need to be a generational advance (~10x more powerful) to attract the enthusiast console gamer with graphics. And if you meant upgrade from Wii U, PS4 power, less even, will be enough of an improvement to justify it to Ninty fans. Especially if it has added synergy with the handheld.

If all Nintendo's next console ends up being is PS4 levels of performance then I will personally be over the moon, I really think Nintendo would blow other developers away with a GPU that is 10x as powerful as WiiU's esp if it's combined with even 8GB's of RAM. I just think that AMD will offer them far, far more performance per doller in 2016 than PS4 if they are going for a modern SoC architecture.
As mentioned previously, the performance advance is debatable at this point. A 10x advance would be ~40 billion transistors (less whatever advances they can achieve in the same number of transistors) at the same clock, or clocked much higher, neither of which looks particularly promising in anything but monster chips in the next few years, and even if possible it'd be in a monster machine with a crazy pricetag. In 2016, definitely not. The CPU can be much improved but the GPU is still going to be in the same ballpark regards what's on screen. A bit more AA, a stabler framerate, but that's about it I reckon. Given more power == higher price point with little to gain below a certain 'next gen' threshold, there seems zero incentive to push the envelope. Unless Nintendo want to try and win back third parties, keeping it at PS4 level would offer enough of a next-gen upgrade for Nintendo fans while remaining cheap. And also allow for handheld hardware synergy.
 
IGN heard the 1 TF figure as well. Rumors happen. Bgassassin was going on the best info he had, but I think that the info was corrupted somehow by the time it got to him. In any case, we had a couple devs come onto GAF and verify that early dev kits were, in fact, weaker. 400 Mhz GPU. There was also an article which backed up this assertion. I can't recall who published it atm, unfortunately.

Yeah I really like BG, not trying to bad mouth him, I was just responding to someone who said that I was mad for expecting that kind of performance out of WiiU. At that time 'Durango' was rumoured to have a 3-4TF GPU aswell so 1TF for WiiU wasn't really that far fetched esp as we didn't know it's power consumption limitations at the time.

Good work on the chip diagrams by the way, it was much appreciated by everyone who read who didn't have an account to say thanks.
 
Upgrade from what? From PS4/XB1, ain't gonna happen. Even a substantially faster console isn't going to show enough difference to offset the advantages of the older systems. It'd need to be a generational advance (~10x more powerful) to attract the enthusiast console gamer with graphics. And if you meant upgrade from Wii U, PS4 power, less even, will be enough of an improvement to justify it to Ninty fans. Especially if it has added synergy with the handheld.

Upgrade over WiiU. They will need to show a graphical leap of some kind to tempt people into upgrading, esp if they are essentially going to abandon WiiU after four years instead of the usual six years.

I don't agree at all that it's only PC gamers who value good graphics.

Nintendo will be able to get close to CGI with PS4 level specs though imo so that should be more than fine. A cheap price point and good branding / advertising are more important than specs though for their next console.
 
Yeah I really like BG, not trying to bad mouth him, I was just responding to someone who said that I was mad for expecting that kind of performance out of WiiU. At that time 'Durango' was rumoured to have a 3-4TF GPU aswell so 1TF for WiiU wasn't really that far fetched esp as we didn't know it's power consumption limitations at the time.

Good work on the chip diagrams by the way, it was much appreciated by everyone who read who didn't have an account to say thanks.

I'm just glad we could finally get to the bottom of it all with Latte. Wouldn't have been able to do it without some of the people on this board--it was truly a team effort.

Bg's estimates were based on the presumption that the Wii U case could handle a higher TDP (before Iwata told us it averaged about 40 watts). He may be right as the thing hardly feels hot even after a decent amount of play time. Nintendo were likely very conservative on this issue, however, and didn't want to risk a RROD scenario.
 
Gobs of RAM won't really make a sense for the next-gen platforms. Those are still primary gaming and media consumption devices, not workstations. I would rather see the same amount of memory, but much more tightly integrated with the host SoC (interposer, stacking, etc.) that will result in lower access latency, higher throughput and power efficiency. A much more urgent need is implementing better storage sub-system, using solid-state medium of course, preferably directly attached to the peripheral interface (SATA Express or M.2). By the 2016-2017 time frame, V-NAND supply should press the prices down enough to make TB-sized storage OEM-friendly.
 
Gobs of RAM won't really make a sense for the next-gen platforms. Those are still primary gaming and media consumption devices, not workstations. I would rather see the same amount of memory, but much more tightly integrated with the host SoC (interposer, stacking, etc.) that will result in lower access latency, higher throughput and power efficiency. A much more urgent need is implementing better storage sub-system, using solid-state medium of course, preferably directly attached to the peripheral interface (SATA Express or M.2). By the 2016-2017 time frame, V-NAND supply should press the prices down enough to make TB-sized storage OEM-friendly.
Ram is cheaper than a SSD, SSD still has a long way to go to get to HDD prices and so long as that is true, consoles will use HDD. The more ram the console has the bigger the cache to the harddrive is and the more efficient the streaming into ram becomes. Ram is always needed with larger and larger games. A lot of developers were saying that the ps4 and xbone could use more ram. Ram is something you can always use in a game as long as your whole game isn't in memory, you can always use more ram.

An SSD solves next to nothing when you are doing linear reads 90% of the time when gaming.
 
I would rather see the same amount of memory, but much more tightly integrated with the host SoC (interposer, stacking, etc.)
Can't do interposer on a high-power ASIC, you'll cook your chip without it having direct contact to the heatsink. An alternative would be through-via stacking and sticking the RAM under the ASIC instead, but the need for thousands of vias to support the main ASIC would make the RAM die larger and less cost efficient than otherwise needed.

A much more urgent need is implementing better storage sub-system, using solid-state medium of course, preferably directly attached to the peripheral interface (SATA Express or M.2).
Why is this so urgent? It's basically the least urgent thing of all, pretty much. PS4 even has its HDD connected via USB3 for chrissakes... Consoles aren't I/O intensive, at all.

An SSD solves next to nothing when you are doing linear reads 90% of the time when gaming.
Well, a good 4x PCIe SSD today can do upwards of 1TB/s linear reads, meaning level load times would be pretty much instantaneous. :) SSD will likely be too costly still to be viable for next-gen consoles though, but I wouldn't expect (console) harddrives to hang on longer than that. They will probably get phased out sometime during the next gen, I'd think...
 
Back
Top