Why Nintendo's NX Console Needs To Be x86-Based (Op-Ed)

  • Thread starter Thread starter Guest
  • Start date Start date
Toggle sidebar Toggle sidebar
Status
Not open for further replies.
The machine ISA of console targets matters very little when development is now happening predominantly at a higher level.

Two generations ago it would have been a very valid claim to say consoles should homogenize on one architecture to minimize developer overhead in porting, but nobody is writing Xbone games in i386, and nobody is writing Wii U games (I don't think, at least) in PPC assembly. They use higher level programming langauges and the console manufacturers provide them compiler infrastructure to build for it. Xbone and PS4 already have different toolkits in this regard so any game targeting either must have two projects for the two different build systems plus whatever they are doing on PC (be it some visual studio project or something with makefiles or just a bundle for a game engine).

But we are already moving hugely away from in house engines anyway. Recent announcements and ongoing trends point to 2016-2018 seeing an extreme derth of in home games, where almost eveyone will be using Unity / UE4 / Frostbyte / Source 2 / etc. Because its the only thing that makes sense anymore - and in that case, the engine vendors handle the eccentric differences between the targets.

You can easily make a Wii U 2 using PPC to maintain backwards compatibility that is competitive. Nothing about the PPC architecture in recent years says you cannot optimize an out of order prefetching predicting chip competitive with what is in the Xbone / PS4 - look at how the Cell ten years ago was PPC based and decimated x86 in edge case SIMD operations. And for games, that is almost always what you are doing on your CPU cores.

The problem is just plainly that the Wii U uses such an underpowered SoC that it cannot even compete with high end smartphones today. They would have better horsepower throwing the next generation Qualcomm Snapdragon Chip in a plastic box and calling it their next console than what the Wii U has today. They just need a lot more horsepower, regardless of architecture, to be competitive.
 
if the next console is a living room console, all it needs to do is target 1080p 60fps from all first party games...

if its a handheld, while i love the idea of 3d, on that small a screen it doesn't add much and would be better off without 3d being considered... again... whatever they chose, so long as its an upgrade and targets 60fps.

as for needing x86... that has no benefit for me the consumer... in fact, in 10~ years time when the console is no longer new, and because of how hardware is on consoles sense ps2+ there is no guarantee you can even get working consoles after a certain length of time, and as the xbox emulation team said, x86 may not be emulateable so this generation may die with the consoles...

really, i don't want to see nintendo with ports... they haven't been the leader in hardware or visuals sense gamecube (its arguable xbox had the best hardware, but apparently gamecube still had life in it, in the form of an impressive tech demo, when it died, where as xbox had at its best ninja gaiden) and if i'm going to play a game thats ported across the consoles, i'm playing it on the pc...

nintendo may not make a crap ton of money (and at this point i would be ok if they just went full publisher instead of system manufacturer) but so long as they dont sell at a loss of commission billion dollar chips, even reduced (still large) sales would be just fine.
 


and here you see the games make the games, not the hardware.
cellphones could be supercomputers for all i care but their games made just for them suck and arent even worth getting them when they offer em up for free much less the dollar they normally cost.
 
Contrary to most opinions Nintendo actually knows what it's doing. Unlike the PS4 and Xbox One, Nintendo was able to make money on the console from day one. What hurt Sony was the fact that for years it was taking a loss on every PS3 sold, MS is in a similar position. Nintendo generally makes money every year, the same can't be said of Microsoft's and Sony's gaming divisions. Most companies release a huge set of tools that companies use to develop games for the consoles. Unless your a huge company (like EA or Ubisoft) you don't have the time to optimize down to a level where hardware differences matter. Even then most don't, Mantle is only available on PC, not Xbox One or PS4.

Hardware doesn't make the console. When the consoles were mostly PowerPC based Nintendo still had trouble attracting 3rd party devs, that doesn't change with hardware.

Nintendo has an image problem not a hardware problem.
 
IIRC, the cost of making the new gen x86 consoles (PS4,XBone) initially is much lower than when PS3 and X360 were first sold at a loss. Sony, I think, even went on record saying that they actually MADE money on each PS4 sold at launch.

But I think the main benefit of Nintendo going x86 has little to do with porting ease, and more to do with the massive amounts of R&D put into consumer x86 processors by AMD and Intel, versus IBM and its POWER line for anything but servers, etc.

Maybe an ARM-based Nintendo console would make sense, since the smartphone/tablet world certainly puts tons of development into ARM, although if i'm not mistaken, even the latest and most powerful ARM consumer-level chips cant compete with the mainstream x86 cpus.
 
Nintendo, bring back ROM cartridges! We have advanced enough now in solid state technology that this is inexpensive. No more installing console games or loading screens! Go back to the roots of what console gaming was all about. Add in to that an inexpensive x86 solution and gpu that can do 1080p and its a winner.
 
AMD is making the cpu for the next Nintendo console, and they just licensed an ARM ISA. Nintendo merged their mobile and console dev teams last year. It stands to reason that the next console might be based on an ARM core in order to be cross compatible with any new mobile device that they design.
 
i think you guys and the author are mistaken. i'm pretty freaking sure that this will be a sony vita tv/roku 3/ouya like competitor. it's going to be a cheaper option to play video games. it's not going to have the hardware to run newer games. i think the wii u will still be their flagship.

whatever jalapenoman is saying may just be a rumor, but it is confirming my suspicions. this is nintendo's way of broadening their IPs into other brands/platforms while still retaining control of them. I think the NX will be their "Google Nexus" platform, as a performance/developer standard basically.
 
Nintendo, bring back ROM cartridges! We have advanced enough now in solid state technology that this is inexpensive. No more installing console games or loading screens! Go back to the roots of what console gaming was all about. Add in to that an inexpensive x86 solution and gpu that can do 1080p and its a winner.
Click to expand...

I wanted to say this is actually a genius idea that I had not thought of before. Anymore the price of MMC or a small amount of NAND storage that could hold a game, although more expensive than an optical disk, would still be an affordable means of distributing games. In addition it would take up less space and would greatly help load times. Not to mention the physical console could be reduced in size and the disk drive is the largest part of any Nintendo game system since Gamecube.

I'm glad to see many of you reading my article with a variety of different views on the situation. I did want to point out that I am not attempting to say Nintendo is not profitable with this article, but the Wii U is not as profitable for Nintendo as the XB1 and PS4 are for Microsoft and Sony. The fact that a new console is in development already would point to Nintendo feeling the same.

I could be wrong, but I don't think the new console will be a handheld or OUYA like competitor. The New Nintendo 3DS devices will likely be Nintendo's competitor in the portable gaming sector for the next year at least, if not two. As for the OUYA like device, OUYA itself was not a commercial success and there is a large number of Android devices that can offer this type of experience. Nintendo has never been one to try and follow other companies, and tends to have fairly unique products. As a result it doesn't seem likely for Nintendo to pursue the home Android gaming market.

@zanny: You made some good points with your comment, but I think you are miss understanding me a little. I don't mean to imply that porting games from PPC to X86 is as hard as porting games and software was in the past. I know that these are running at a higher level with gaming engines and APIs that help to ease the porting of games. I am simply stating that it is more difficult to port a game to the Wii U's PPC CPU, than another x86 system. Even if there are numerous tools to aid in porting games, it still consumes more time and monetary resources. Likely with faster hardware this wouldn't be a major problem, but for two reasons it is a problem.

The first one is obvious, the WIi U has weaker hardware. Second, even in the x86 world between PCs and game consoles we have lately seen several games released that are poorly optimized and performance suffered as a result. The reason behind it is that game developers don't want to put additional time and money into optimizing games. When major performance problems are observed on the same architecture, you can only imagine what might happen going to an entirely different system.
 
I wanted to say this is actually a genius idea that I had not thought of before. Anymore the price of MMC or a small amount of NAND storage that could hold a game, although more expensive than an optical disk, would still be an affordable means of distributing games. In addition it would take up less space and would greatly help load times. Not to mention the physical console could be reduced in size and the disk drive is the largest part of any Nintendo game system since Gamecube.
Click to expand...

This is a foolish idea and I'm glad you're not at the helm. Modern games are often huge. It would not be uncommon to see a AAA title in the 20-50GB range. Sure some garbage flash is fairly cheap, but that still would add $5-10 to the cost. A reasonably fast external flash solution would add more, something like an extra $10-20 per game - I'd rather not have to pay that much extra for each title.

That doesn't even factor in downloads, which obviously still requires local mass storage, which puts you halfway back to where you started anyway. Add an optical drive and boom you've come full circle and landed back safely in sanity land.
 
i think you guys and the author are mistaken. i'm pretty freaking sure that this will be a sony vita tv/roku 3/ouya like competitor. it's going to be a cheaper option to play video games. it's not going to have the hardware to run newer games. i think the wii u will still be their flagship.
Click to expand...
This sounds plausible. If they really want to save the Wii U they need to reduce the price. It would sell like hotcakes at a more aggressive price point. Simultaneously, release a new model with a substantial increase in internal storage at the old price.
 
If the Nintendo NX running a x86 cpu can play games at 1080p or better they got my money no matter how pissed I am at the Wii U. At-least my daughter plays the hell out of it.

It would be great to see Microsoft be the lesser system as and Nintendo and Sony compete with 1080p gaming. Let the Xbox One 720p gamin compete for value.
 
While PPC is on the downslide and might be worth abandoning for Nintendo, x86 isn't necessarily the best option for them. While used on desktop PC's, x86 carries a lot of baggage from the early days of the design, where many current features weren't really anticipated and backwards compatibility needed to be maintained, even if it meant making a "bug" a "feature". ARM processors are a cheaper and viable alternative that are WIDELY developed for, especially as new designs push the performance envelope more aggressively than their typical power-efficient designs. Those power savings on CPU performance also provide a higher TDP for graphics power, which is where the focus should likely be, and with both Nvidia and AMD working on some ARM chips, it shouldn't be hard for Nintendo to get a customized SoC that works for them. I wouldn't expect Nintendo to necessarily try to compete frame for frame with the XBone or PS4, and honestly, that's okay. Particularly if the system sells for a fraction of those system's cost.

Virtual console also already answers a number of backwards compatability woes... the Wii U already has Wii Virtual Console games, for instance; Nintendo already has most of the emulators, and hey, they might even be able to steal some work from Dolphin for a new console.
 
Nintendo, bring back ROM cartridges! We have advanced enough now in solid state technology that this is inexpensive. No more installing console games or loading screens! Go back to the roots of what console gaming was all about. Add in to that an inexpensive x86 solution and gpu that can do 1080p and its a winner.
Click to expand...

For all the bitching devs did back in the 90s, ROM carts were an effective way to sell and produce games, especially for mobile consoles and have several benefits over disc based systems; including minimal load times, fast data transfer, and in the past, any extra space on the cart could be used as extra RAM, especially as larger carts became cheaper to produce. The latter of which was prevalent during the N64's end-of-life, when some roms reached 512Kb (64MB). While modern DDR3 RAM is still bounds faster then the fastest SSDs on the market, having reduced boot/load times and a smaller physical size is definitely something to consider as a selling point. But Nintendo has always been a step behind the competition in this regard, so I'm interested to see what they plan to do with NX.

However, I almost expect the disc-based system to vanish entirely on future home consoles, in favor of downloading them directly. Which means lower manufacturing and shipping costs. Also, with SSD's falling in price over the last year, along with recent advances in speed, I wouldn't be surprised to see the system becoming almost entirely solid state, with a cheaper variant using a 2.5" HDD.

X86 might not be a bad option to take, but Nintendo tries to see itself as "affordable," usually coming in ~$100 cheaper than the competition, and changes could mean higher costs. But factoring in the reduced cost of manufacturing by using a download model, it could pay off if more Devs open up to the idea.

Nintendo also mentioned in the past that if people want to continue to play their older games, they'll just keep their old system (I still own all my old Nintendo systems). But x86 is decent at emulating most other consoles, given enough power. Cartridge era and hand-helds especially, with a few exceptions. Nintendo disc based games are touchy and don't often play well with x86 emulators. But if Nintendo created compatible emulators that fix the issues present in current PC-based emulators, we could be looking at some serious potential.

But only time will tell what becomes of it.
 
As far as I read the cpu is the same from the wii but what was improved was the gpu. And that's the problem, x bone and ps4 developed games(paticularly 3rd party games are developed completely in high level language so the cpu has to process all data and gpu is being used less and less. That's why assassins creed had lower Fps then the Xbox one did. Nintendo is using engineers to optimize their games. Keeping their hardware cost efficent and keeping their first party games relevant is all that matters. It's business and they know what they are doing. Wii u started slow but all in all I would say it has the best line up of past and present games so far...
I'm very biased though as I can't stand the repetitive flood of shooters and sports games.
 
While PPC is on the downslide and might be worth abandoning for Nintendo, x86 isn't necessarily the best option for them. While used on desktop PC's, x86 carries a lot of baggage from the early days of the design, where many current features weren't really anticipated and backwards compatibility needed to be maintained, even if it meant making a "bug" a "feature". ARM processors are a cheaper and viable alternative that are WIDELY developed for, especially as new designs push the performance envelope more aggressively than their typical power-efficient designs. Those power savings on CPU performance also provide a higher TDP for graphics power.
Click to expand...

Depending on whether you want to believe Intel's propaganda or not (trying to sell those Atoms!), supposedly with die shrinks and new advanced power saving features, any inefficiency an x86 cpu has versus ARM is very small nowadays. Its just that we tend to compare the power hungry monsters in our desktops/notebooks with the super-efficient (but comparitively weak) ARM chips in our tablets/smartphones, when a more apt comparison would be with one of Intel's super low power consumption Atoms.
 
Contrary to most opinions Nintendo actually knows what it's doing. Unlike the PS4 and Xbox One, Nintendo was able to make money on the console from day one. What hurt Sony was the fact that for years it was taking a loss on every PS3 sold, MS is in a similar position. Nintendo generally makes money every year, the same can't be said of Microsoft's and Sony's gaming divisions. Most companies release a huge set of tools that companies use to develop games for the consoles. Unless your a huge company (like EA or Ubisoft) you don't have the time to optimize down to a level where hardware differences matter. Even then most don't, Mantle is only available on PC, not Xbox One or PS4.

Hardware doesn't make the console. When the consoles were mostly PowerPC based Nintendo still had trouble attracting 3rd party devs, that doesn't change with hardware.

Nintendo has an image problem not a hardware problem.
Click to expand...
Nintendo's problem is that they always choose lower performance hardware than their competitors. The graphics on the original Wii sucked when compared to the Xbox 360 or PS3, the only reason it was so successful is that everyone loved that new controller. Nintendo did the same thing with the Wii U, the only difference is that nobody wants a huge controller with a built-in screen.
 
Companies don't port for Wii U because of the small install base, they often have no problem porting to last gen PS360. So it is unlikely to be a problem with not having x86, it's not likely the code is written in assembly (though some of it might be). It is unlikely an x86-architecture would have made any difference for the Wii U, more power probably would have.

(the following assumes NX will launch in 2017, which seems most likely to me)
We still have no idea what NX will even be, only that it is a dedicated gaming machine. If it is a Wii U -replacement / successor, then it will come out right in the middle of the current gen, which positions it to compete badly against established platforms, and weakly against their successors. Didn't think I'd say it, but this time Nintendo needs gimmick. It's the console after NX (assuming it is a TV-console), that should release at a time where it can compete with next gen with power.
 


And IIRC Sony said they make a profit if the people buy both PS4 and a years worth of PSN sub. Either way, as you say, it is clearly a much cheaper console to produce than the PS3 (though, PS3 was able to compete with quite high end rigs at the time, and was the most affordable Blu -ray player even at its high launch price).



Many, if not most Nintendo games on Wii U already run at 1080p, so that is quite a poor measure. Devs could make 1080p/60fps games for PSone but they don't, because pretty screenshots are an easy sell.


And because Wii was such a failure, Nintendo started developing Wii U right after Wii. 😉

With a 5-year console cycle, Nintendo can't rest much when developing consoles, they always have the next one in development. NX was mentioned because people were freaking out about Nintendo going third party mobile. You can rest assured that even Sony and MS, with their longer console cycle, are thinking up and prototyping parts (hw and sw) for the next console already and will start prototyping the whole console in a couple of years.
 
Don't underestimate IBM, and a possible desire for them to stay on a console. IBM is a very good company for products that have a 5-year+ release cycle, which consoles have. They can't keep up with a race against x64, but they can occasionally leap ahead. IBM also might be able to help Nintendo on the software tools side. For example, IBM has a compiler, and IBM's developer documentation is about as good as Microsoft's.

That being said, I wouldn't be surprised if Nintendo goes to x64, but they'll have a few advantages staying with IBM
 
Nintendo, bring back ROM cartridges! We have advanced enough now in solid state technology that this is inexpensive. No more installing console games or loading screens! Go back to the roots of what console gaming was all about.
Click to expand...

Not only that, but there were custom IC's and such in those cartridges at times that were used to augment the console hardware. Starfox on the SNES is a great example. The SNES is absolutely incapable of those graphics by itself.

Bringing back this idea could really help keep the console current as far as hardware as long as they can keep costs down.

Cartridges also last a really long time. A lot of those old NES carts still work fine, whereas CD/DVD/blu-ray titles can easily get scratches and stop working.
 
Nintendo, bring back ROM cartridges! We have advanced enough now in solid state technology that this is inexpensive. No more installing console games or loading screens! Go back to the roots of what console gaming was all about. Add in to that an inexpensive x86 solution and gpu that can do 1080p and its a winner.
Click to expand...

There is simply NO reason to do this. A far more sensible and versatile approach is what Sony uses on the Vita. Generic flash memory that can be mass produced and not committed to a particular game until very late in the process.

Mask ROM was the only game in town for a long time that met the needs of the console business but had all kinds of hassles, especially the commitment required for production runs because changing masks was a huge hassle. One of the reasons optical media was so favorable wasn't just the huge capacity gain but also because a CD production line could stop on a dime and switch what content was being put on the disc. Flash memory cards have the same advantage.

But even those are a dying breed. My experience with the Vita has shown it is far more convenient to have a high capacity memory card to store a whole library. The Vita's memory cards are severely overpriced due to the problem of convincing retailers to sell something that will cut them out of future software sales: you have to give them a very good profit margin. But more standard flash memory formats are more affordable. 128 GB microSDXC cards are frequently down to $80 now and should get a lot lower towards the end of the year. Currently the only limit in the New 3DS for supporting such a card is Nintendo's refusal to license exFAT or implement an open source file system and give the 3DS the ability to reformat the cards. Right now to use greater than the 32 GB cards requires reformatting in FAT32 and taking some risk.

Increasingly, more and more software will only be sold as downloads. Consoles are making that transition, with handhelds farther along but still trailing phones and tablets, where it is the sole choice.
 
Nintendo has never cared much for ease of porting from competitors' platforms. Their primary concern has been bringing their existing developers forward. The SNES is an example, being one of the few major platform to use the 65816. This gave a 16 bit upgrade that was backward compatible to the 6502 but of the major users only Apple used it in the IIgs. Atari and Commodore chose to let their 8-nit systems fade away with no immediate compatibility options for the Amiga and Atari ST.

Nintendo's main concern was to reduce the learning curve for longtime NES developers. The machine was also intended to be backward compatible and have two cartridge slots but it was decided that this would only encourage the use of second-hand games. (Nintendo actually managed to severely restrict used game sales at retail in Japan for a few years and lobbied for it in the US but were told to go pound sand.)

With this in mind, if Nintendo drops PowerPC, the architecture that has the most to offer them is ARM. This unifies their work on tools for developers and allows for a platform unification going forward. The console can apply higher end ARM designs not suitable for a battery driven portable. Over time the platform should see engineering refinements that bring down to the price point and power envelope for a handheld. Games, especially smaller indie productions, that use only a fraction the console's power could be designed to run across both the console and handheld. More demanding games can hit the handheld market as the hardware migrates there, while a new performance level is introduced on the console side.
 
Status
Not open for further replies.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom