• Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.

Nintendo's FY 2016 has officially begun - The Year of NX

Status
Not open for further replies.

10k

Banned
Good find. Should be an obvious thing but many people predicted that this wouldn't have been the case, so it's good to have another hint.


That ship sailed long ago, but it's about to face a storm from that date on.
And we're ready.
image.php



10k
Beware statements I make about the NX
(Today, 08:44 PM)
image.php
I love this tag haha. The avatar really completes it.
 

10k

Banned
I wonder...why didnt we hear anything from the WSJ for so long. It seemed they were pretty well informed and in the last 4 months we didn't hear anything.
Hmm. My speculation is a combination of Nintendo tightening up thier NDA's late last year.

There are journalists out there who sign NDA's all the time so perhaps a bunch of them were invited to see it at GDC and E3 and had to sign an NDA about it. Just a theory.
 
I'm actually getting pretty pumped for the NX release. But I've been lured into false hope before by Nintendo.

There's 33 titles in this list of NX games (which includes revealed, leaked and rumoured titles), but I think it does need more exclusive titles. Non-Nintendo exclusives, too. I wonder if Nintendo will splash the cash to make that happen.
 
I'm actually getting pretty pumped for the NX release. But I've been lured into false hope before by Nintendo.

There's 33 titles in this list of NX games (which includes revealed, leaked and rumoured titles), but I think it does need more exclusive titles. Non-Nintendo exclusives, too. I wonder if Nintendo will splash the cash to make that happen.

8.2% of that list will be released in the first year, 9.5% ever.
 
Didn't EA say they weren't doing mass effect remasters? So that's not true.

Found this but it is quite an old comment. Older than the leak about the games list.

To me it seems weirder they would release MEA and a Trilogy at the same time. Unless, the Trilogy isn't a remaster, but simply ports included with MEA - would not be a bad play as they can't assume a Nintendo fanbase will know the series.
 

maxcriden

Member

Hm, might be worthy of its own thread.

Confused by the part where they say the previous touch patent has images that evoke BotW, though:

One of the most well-known of these patents relates to a 'free-form screen' application, with images that now tellingly (in hindsight) draw comparisons to Breath of the Wild in terms of in-game items and so on. You can see the full details and those images in the article from that time.

I don't see that connection, but maybe I'm missing it.

-

So, this new controller/handheld is essentially a touchscreen-only system, but focused on vibration and sound, and then with a peripheral that puts controls on it?
 
Hm, might be worthy of its own thread.

Confused by the part where they say the previous touch patent has images that evoke BotW, though:



I don't see that connection, but maybe I'm missing it.

-

So, this new controller/handheld is essentially a touchscreen-only system, but focused on vibration and sound, and then with a peripheral that puts controls on it?

Nah this patent app really isn't suggesting any sort of touch screen or portable to me at first glance. Rather they're just using an easily described embodiment (single device with a screen) to sorta illustrate their primary invention which seems to be the interplay of 3D sound and vibration. Same concept is doable on a normal controller too.

I'll give it a closer read in a bit though.
 

maxcriden

Member
Nah this patent app really isn't suggesting any sort of touch screen or portable to me at first glance. Rather they're just using an easily described embodiment (single device with a screen) to sorta illustrate their primary invention which seems to be the interplay of 3D sound and vibration. Same concept is doable on a normal controller too.

I'll give it a closer read in a bit though.

Oh, I see. Thank you! I wonder how that button peripheral is related, then.
 
Oh, I see. Thank you! I wonder how that button peripheral is related, then.

Figure 8 is actually a separate embodiment from that of the previous one. Typically in patent applications the inventor will list many different embodiments (final forms the invention can take) and in this case the simple touch screen device is one example while the device with a screen and buttons is a different example, not a peripheral. The Nintendolife article is incorrect in this case.

I read over the application a bit more and I was right, it's essentially focused on coordinating sound out of the two speakers with the central vibration mechanism. The touch screen figures show examples of how this interplay can be used in gameplay- for example selecting a weapon can cause that weapon chime to play on one of the two speakers while vibrating the console in the center.

The claims are all about these functions, nothing about touch screens or portables with displays.
 

Eradicate

Member
I'm actually getting pretty pumped for the NX release. But I've been lured into false hope before by Nintendo.

There's 33 titles in this list of NX games (which includes revealed, leaked and rumoured titles), but I think it does need more exclusive titles. Non-Nintendo exclusives, too. I wonder if Nintendo will splash the cash to make that happen.

So...I went through the list of titles and did quick Googles of the middle group, the "confirmed by developer" titles. Not much found, but Lizardcube (making Wonder Boy: The Dragon's Trap) said on their Twitter them making the game for NX was a misunderstanding and they haven't announced platforms yet. There is an indie show taking place in a few days that they will be at in Kyoto, so maybe some information can come out from that?

Ready at Dawn (making Deformers) also is working on three titles currently: an AAA third person shooter, a VR game (teaming up with Oculus), and Deformers. No platforms yet for any of them from what I can tell!

Also, I think this list is a bit off. I could swear more developers said they were making games for the NX back in either of the "NX sources" threads, just a few more, such as some water themed puzzle game. We really ought to create our own list and keep checking on these guys! I "may" go back in and check through those threads and try to find them, but maybe someone else is keeping track?
 

Javier

Member
I'm actually getting pretty pumped for the NX release. But I've been lured into false hope before by Nintendo.

There's 33 titles in this list of NX games (which includes revealed, leaked and rumoured titles), but I think it does need more exclusive titles. Non-Nintendo exclusives, too. I wonder if Nintendo will splash the cash to make that happen.
Missing from that list is Pikmin 4, which IIRC was confirmed by Miyamoto himself. He didn't mention platforms but if the game was for Wii U or 3DS, they would have revealed it by now.
 
When you have patents like this people are going to spin it into different directions when it's not like what people think...it's so frustrating when people look at a patent picture and say "omg that's the NX?!?? No buttons!" It's so annoying because we've been over this over and over again.
 

maxcriden

Member
Figure 8 is actually a separate embodiment from that of the previous one. Typically in patent applications the inventor will list many different embodiments (final forms the invention can take) and in this case the simple touch screen device is one example while the device with a screen and buttons is a different example, not a peripheral. The Nintendolife article is incorrect in this case.

I read over the application a bit more and I was right, it's essentially focused on coordinating sound out of the two speakers with the central vibration mechanism. The touch screen figures show examples of how this interplay can be used in gameplay- for example selecting a weapon can cause that weapon chime to play on one of the two speakers while vibrating the console in the center.

The claims are all about these functions, nothing about touch screens or portables with displays.

Thanks for the info, it's much appreciated. I'm glad to have the clarification. I'm glad we have every indication that Nintendo plans to further innovate in this area.
 

10k

Banned
I'm actually getting pretty pumped for the NX release. But I've been lured into false hope before by Nintendo.

There's 33 titles in this list of NX games (which includes revealed, leaked and rumoured titles), but I think it does need more exclusive titles. Non-Nintendo exclusives, too. I wonder if Nintendo will splash the cash to make that happen.
Shinobi already debunked Mass Effect Andromeda for NX and Trilogy for any platform, didn't he?
 
Please. I could care less about past bad press, the main thing I cared about is he created one of my all time favorite games in Eternal Darkness, it was an extremely enjoyable discussion.

I was making a joke about how Dyack is a Gater but you seem well aware of that and welcome it, so that tells me quite a bit.
 

Shikamaru Ninja

任天堂 の 忍者
Please. I could care less about past bad press, the main thing I cared about is he created one of my all time favorite games in Eternal Darkness, it was an extremely enjoyable discussion.

I've always thought that Blood Omen: Legacy of Kain is his opus magnum. Eternal Darkness was slightly disappointing as a fan. I'm excited about Shadows of the Eternal, I just hope it actually releases.
 
Please. I could care less about past bad press, the main thing I cared about is he created one of my all time favorite games in Eternal Darkness, it was an extremely enjoyable discussion.

Really enjoyed this. I wish more people were able to separate personal opinion from professional opinion and give DD a chance to make another great game.
 
So I just rewatched the Wii U September 2012 conference, and Nintendo still hadn't settled on a launch line up, so I wonder whether we'll even find out launch games this year.
 
Please. I could care less about past bad press, the main thing I cared about is he created one of my all time favorite games in Eternal Darkness, it was an extremely enjoyable discussion.

It was a very good discussion and offered interesting insight into ED, silicon knights and their relationship with Nintendo.
 

Thraktor

Member
Interesting but it might be too soon for NX handheld is it's coming in March. If the handheld comes later UFS is definitely a storage medium Nintendo should consider, especially if they opt for solid state drives in the console.

I wouldn't really expect it for the handheld, just as I don't see a need for that kind of bandwidth in the handheld (unless the cost of a combo MicroSD/UFS card reader is relatively trivial), but I do think the embedded variety of UFS is an interesting proposition for the home console.

Nintendo has effectively six different options for storage in their next home console, assuming they want to design a system where they can alter the storage capacity by changing a single component. Roughly in increasing order of cost:

SATA HDD - 2.5"

Capacity - up to 2TB
Sequential reads - up to 120MB/s
Random reads - up to 100 IOPS

- The cheapest option per GB (although that advantage decreases each year)
- Physically much larger than most flash options
- Substantially increased failure rate over flash
- Nintendo have never used a mechanical hard drive before
- Could theoretically be user-replaceable

eMMC Flash

Capacity - up to 128GB
Sequential reads - up to 300MB/s
Random reads - up to 13,000 IOPS

- Likely the cheapest option for flash memory
- Used by Nintendo in both Wii and Wii U
- Small, single package BGA
- Could use the same storage option in the handheld
- Effectively EOL, as it's being replaced by eUFS, meaning sizes larger than 128GB are very unlikely, and even that may not last long

eUFS Flash

Capacity - up to 256GB
Sequential reads - up to 850MB/s
Random reads - up to 45,000 IOPS

- Small, single package BGA
- Successor to eMMC, so guaranteed to continue to be developed over the life of the console
- Could use the same storage option in the handheld

SATA Flash

Capacity - up to 2TB
Sequential reads - up to 550MB/s
Random reads - up to 100,000 IOPS

- Physically much larger than any other flash option
- Largest capacity of any flash option (for a price)
- Could theoretically be user-replaceable
- Potentially EOL during the life of the console, as NVMe replaces it

NVMe Flash - M.2

Capacity - up to 1TB
Sequential reads - up to 3,200MB/s
Random reads - up to 450,000 IOPS

- Smaller than SATA drives, but larger than BGA options
- Absurdly, over the top fast
- Seriously, that's one quarter the speed of Wii U's main RAM pool
- Could theoretically be user-replaceable
- Young standard, meaning prices will continue to come down and capacities will continue to increase for some time

NVMe Flash - BGA

Capacity - up to 512GB
Sequential reads - up to 1,500MB/s
Random reads - up to 190,000 IOPS

- Small, single package BGA
- Almost certainly more expensive than M.2 equivalent
- Only one manufacturer (Samsung), preventing Nintendo from shopping around for better price


Ordinarily I would have suspected Nintendo to stick with eMMC, as it's simple, cheap and capable of far higher speeds than any mechanical hard drive. With eUFS having already replaced it in all high-end phones, and likely to make in-roads into mid-range devices soon, eMMC is unlikely to find use in much more than low-capacity entry-level phones soon enough. That would place Nintendo in a difficult position, as it would prevent Nintendo from ever offering capacities over 128GB, and even the 128GB capacity is only being manufactured by one company (Samsung), meaning little scope for cost reduction over the life of the console.

Embedded UFS seems like a more sensible option, given the life-cycle of the console. It'll be available in capacities up to 256GB from launch, and is effectively guaranteed to see larger capacities developed, allowing Nintendo to easily release upgraded versions of the console over the course of the generation. It's also relatively cheap compared to other flash options, and it's very likely that Samsung's 256GB UFS2.0 module will actually be cheaper than Sony's 500GB HDD was in launch PS4s, while offering about 10x the sequential read speed and 500x the random read speed.

The question then comes to whether Nintendo stick with discs or move to game cards for NX. If discs, then mandatory installs are almost guaranteed, and in that scenario I can't see them getting away with any less than 256GB at a minimum, and preferably 500GB+. Embedded UFS remains an option, just about, but if they want any higher capacity beyond that, then they're basically forced onto mechanical hard-drives (unless they're willing to release an obscenely expensive console). If they use game cards, and hence avoid mandatory installs, then 128GB, or even 64GB in a basic model, would do the job, and they could stick with flash without breaking the bank.

The implications for game development also become pretty interesting for a console running off flash memory. Even eMMC is capable of vastly higher real-world performance than HDDs (which may look okay-ish in sequential reads, but for anything even remotely non-sequential they choke horribly). While obviously this would help bring down load times, it could significantly eliminate a bottleneck for a lot of developers, particularly devs creating open-world games, where efficiently streaming data from the HDD is the largest technical hurdle they have to jump over. If Nintendo could guarantee these developers hundreds of MB/s sequential reads and tens of thousands of IOPS then it would be very interesting to see what they could do with it.

Of course, mandatory installs onto a 850MB/s, 45K IOPS eUFS chip is one thing, but if Nintendo use game cards on the NX, and games are designed to run directly off the cards, then the game card is going to be the baseline case that developers will need to accommodate for. The question then becomes, how fast would the game cards be?

From my reading of it (and I'll admit a complete lack of expertise in this area), the answer seems to be "pretty much as Nintendo wants (or is willing to pay for)." Nintendo don't use flash memory in their game cards, but rather use ROM (or more specifically a custom variant of Macronix XtraROM). ROM behaves quite differently to NAND flash, because unlike NAND, where all data comes through a controller chip, which has to handle wear levelling, block management, caching, etc, when you read data off a ROM game card you're pulling data directly off the ROM chip itself. This makes ROM speed a pretty simple calculation based on the speed of the chip and the interface to the chip (much like RAM).

Macronix's 45nm XtraROM can currently achieve a read cycle time of 25ns (or 40MHz), and, going by past trends, we could expect their upcoming 32nm XtraROM to hit around 16ns (or 62.5MHz), which is what would almost certainly be used if Nintendo were to go the game card route with NX. On a 16-bit parallel interface, that would give Nintendo 125MB/s sequential reads, and although it's tricky to estimate IOPS values (strictly speaking they're only OPS on ROM), the RAM-like interface and low access latency means you should be able to get close to peak read speeds with random access patterns. For 4K random reads (a typical measure) you'd be looking at possibly 25,000 IOPS.

If they switched to a serial interface, the bandwidth could push up much higher. With 32-bit reads over an 8-bit, 250MHz serial interface they could get 250MB/s sequential reads and potentially 50,000 IOPS. Even 64-bit reads over a 500MHz interface, giving 500MB/s sequential and 100,000 IOPS would probably be technically possible if Nintendo wanted it. Moving to higher speed interfaces would certainly add to the cost, but only by a fairly small amount, and the matter will largely be a question of what Nintendo feels is enough. Given that a 16-bit parallel interface places the game cards firmly above PS4 and XBO's internal HDDs in terms of performance, that may well be enough for Nintendo.

Now, while a lot of people are probably reading this thinking "Nintendo would never use 800MB/s flash ... withered technology ... blah, blah, blah", there are a number of good reasons why it may end up being the case:


  1. Nintendo clearly dislikes moving parts in their consoles, and have never shipped a console with a built-in mechanical hard-drive before. Now that the cost difference between hard-drives and flash is at its lowest, I don't see them suddenly changing course on that one.
  2. As mentioned above, if they ever want to release an NX model with more than 128GB of storage, the cheapest way of doing so (and avoiding hard drives) would be eUFS. If that happens to give them 500+MB/s of bandwidth, then that's what they get.
  3. Nintendo's hardware and software divisions have always been closely linked, and their hardware decisions are very often made on the basis of their software engineers' requirements and issues with previous consoles. They've also been known to take these requirements to the extreme in some cases, such as when developers complained about the high latency of N64 RAM and they responded by using literally the lowest latency memory available in the Gamecube. With the Wii U, disc and storage read speeds are almost certainly the biggest issue Nintendo's internal software engineers have had. Xenoblade X quite obviously had pretty big problems with disc read speed, and I would guess it's been a big issue throughout Breath of the Wild's development. Outside of games, Nintendo has been criticised for the sluggishness of Wii U's OS, and they've gone as far as publicly apologising for it early on in Wii U's life (although the slow OS loads may be more to do with Starbuck struggling with crypto duties, Nintendo is almost certainly going to make a point of designing the NX to run a snappy OS).
  4. The speed of NAND flash is almost entirely constrained by the interface and flash controller, meaning that for a company in Nintendo's position (purchasing 10M+ units, no legacy interface restrictions) the cost difference between slow flash and blazingly fast flash is surprisingly low. Consider Samsung's new OEM NVMe M.2 drives, which are arriving only about a year and a half after their very first NVMe drive, yet are already almost saturating the PCI 3.0 NVMe interface (despite having over 6 times the bandwidth of the standard it replaced). What's more, the SM961 does this with the same MLC V-NAND as is used in the older SATA-based 850 Pro, the only change being the controller and interface. Similarly, Samsung's new "cheap" NVMe OEM drive, the PM961, gets 3,000MB/s and 360,000 IOPS out of the same TLC V-NAND used in Samsung's mass-market 850 Evo SATA SSDs. In low quantities, Samsung will likely charge about twice as much for the SM961 and PM961 compared to their SATA brethren, but that's almost all profit for them, as the added cost of the improved controller is fairly small in the context of the NAND cost. For the quantity Nintendo would be purchasing, though, Samsung would work on far lower margins, and the cost difference between the SATA and NVMe would be extremely small (and counting in the SATA drive's increased size, weight and mechanical complexity, the NVMe drive may actually end up being the cheaper option when everything's taken into account). In fact, I'd be pretty confident that Nintendo could get 256GB PM961s at a lower cost than Sony paid for the 500GB HDDs in launch PS4s. My expectation that they would use eUFS instead isn't even based on it being much cheaper (again, it's likely the same TLC V-NAND), but rather that the BGA package makes console assembly simpler.
  5. If Nintendo do go with game cards, then again the bandwidth they can achieve is pretty much purely a matter of the interface, although in this case there's no flash controller, and they get to design the interface themselves. Increased speeds would increase costs, potentially not by much, but on a per-game basis (rather than a per-console basis) Nintendo would want to squeeze every penny out of the costs. They would, however, be able to eliminate the flash memory used for saves on 3DS cards (on-console memory and cloud saves eliminate the need for it), which would reduce the number of components and the complexity of the card interface. They could also use a variable-speed serial interface, allowing basic games which don't require blazing fast data access to use cheaper, slower cards, while big open world games can ship on cards which cost a little bit more, but provide extremely high read speeds.
  6. Contrary to popular belief, Nintendo are quite happy to boast about the specs of their consoles, but only when they have something to boast about (for example they were happy to talk about Wii U having 2GB of RAM compared to the 512MB in XB360 and PS3). Going with a fast on-board SSD would be a very cheap way to get a spec win over their competitors, certainly compared to chasing after GPU Tflops or RAM bandwidth, and for a certain subset of games it would actually be the biggest bottleneck, so they could squeeze out some kind of positives even going up against Neo and Scorpio.
  7. Nintendo have had serious difficulty bringing down the cost of Wii U, due in large part to their using mature manufacturing processes and components, which have low scope for cost reduction over time. It would be natural to assume they're going to try to avoid this by choosing parts which are early-to-mid life cycle and have a number of suppliers, which increases their ability to reduce price over the life of the console. eUFS and NVMe components are likely to see much better cost reductions than eMMC and SATA SSDs over NX's lifespan.
There are, of course, a few issues with implementing such fast internal storage. One of them is the aforementioned crypto bottleneck in the Wii U, although fortunately new ARMv8 cores have massively improved crypto performance, with a single A53 at 1.3GHz managing 662MB/s AES throughput. A quad-core cluster running at ~2GHz could easily handle even the SM961 without breaking a sweat.

The other factor is that Nintendo would have to either eliminate the option of using USB hard-drives, or find some way to limit it to SSDs. If you design a game based on an assumption of even just 10,000 IOPS there's a very good chance that it simply won't run on a mechanical hard-drive, and I can see a lot of people plugging in cheap hard-drives and assuming the console's broken when their games don't work. Perhaps partnering with one of Samsung/Sandisk/etc to build a line of NX-branded external SSDs would help ensure there's a simple option for people looking to expand their console's storage.
 
R

Rösti

Unconfirmed Member
As previously mentioned, Nintendo rarely has a briefing for the first quarter earnings release. However, with the huge success of Pokémon GO positively affecting stock price and market cap, I cannot help but wonder if they might after all have a briefing for the FY 2016 First Quarter Earnings Release scheduled for July 27. Or maybe a news conference?
 

Plinko

Wildcard berths that can't beat teams without a winning record should have homefield advantage
Rösti;209767531 said:
As previously mentioned, Nintendo rarely has a briefing for the first quarter earnings release. However, with the huge success of Pokémon GO positively affecting stock price and market cap, I cannot help but wonder if they might after all have a briefing for the FY 2016 First Quarter Earnings Release scheduled for July 27. Or maybe a news conference?

Perhaps. It is the first real positive news in a while. Any update on that Nintendo patent for SCDs? Have they paid for renewal yet?
 
R

Rösti

Unconfirmed Member
Perhaps. It is the first real positive news in a while. Any update on that Nintendo patent for SCDs? Have they paid for renewal yet?
I'm monitoring that. Deadline for payment of issue fee is July 18. No sign of payment yet.
 
Status
Not open for further replies.
Top Bottom