• Hey everyone, staff have documented a list of banned content and subject matter that we feel are not consistent with site values, and don't make sense to host discussion of on Famiboards. This list (and the relevant reasoning per item) is viewable here.

StarTopic Future Nintendo Hardware & Technology Speculation & Discussion |ST| (Read the staff posts before commenting!)

Playing devil's advocate for a moment. How do we know Nate's figures are correct? Did he ever cite a source, or is it based on what his own "sources" tell him? I'm not disputing his numbers or anything, just asking for clarification, especially since this thread is all about providing sources, and link's to sites.

14 dollars does seem about right, though I would imagine by this year, that figure would have dropped a bit to maybe 10-12 dollars per 16GB cart. That said, I have doubts a 32GB cart would be literally double the price at about 25 dollars. Though, it would help explain why TOTK was sold for 70 dollars rather than the typical 60 for US markets.

EDIT: Looking a couple posts down, a Tweet said 32GB carts are 60% more expensive than a typical 50GB Blu-Ray. Does that mean those discs are more than 4 dollars to print? 8 dollars? 10 dollars? Assuming 10 dollars for a 50GB Blu-Ray, that's only 16 dollars for a 32GB Cart, which isn't THAT bad. And that Tweet was back in 2017. So what is the case nowadays?

It just seems like something doesn't add up.

EDIT2: Again, asking for clarification.
Pretty sure the cited price was referring to the total cost per unit including licensing and distribution fees and all that stuff. Not just the physical BoM. Stuff like licensing and distribution/shipping wouldn't necessarily decrease in price over time. In fact due to COVID shipping prices probably went up for at least a bit.
 
Pretty sure the cited price was referring to the total cost per unit including licensing and distribution fees and all that stuff. Not just the physical BoM. Stuff like licensing and distribution/shipping wouldn't necessarily decrease in price over time. In fact due to COVID shipping prices probably went up for at least a bit.

Zero chance he's including licensing, Nintendo's cut is 30% from third-parties which is $18 for games that launch at $60 (which is basically all games that use a 16 GB cart)
 
I take your point on the general arc of RAM usage, and it's certainly true that RAM is at a premium in the current gen machines.

I'm not a professional game developer, so there is a certain element of me talking out my ass, caveat emptor, but - there doesn't seem to be a feeling that the last gen consoles were particularly RAM constrained. There isn't a change in rendering technology this gen like the switch to physically based rendering in the last gen.

If you're doing a last-gen port, 8GB + Nintendo's smaller OS is a very comfortable place to be on a 2 TFLOPS machine.

If you're doing a current-gen multi plat, I don't think the 8GB is your primary worry. It's less of a squeeze than the GPU or the CPU relative to, say, the Series S. More RAM is nice, but so is more everything. Cutting resolution and complexity will likely be required just to hit your GPU targets, and that will naturally reduce RAM usage.

12GB, plus Nintendo's OS, means as much RAM available to games as Series S. At 16GB you can keep asset quality the same as PS5, and still have room left over. And then you'll start cutting the RAM usage of those assets anyway, because you need to get GPU load down to PS4 levels. That's why it feels like overkill to me.

Switch's 3GB of available RAM was a huge leap over the 360, despite similarly powerful GPUs, but the PS4/XB1 era saw physically based rendering, and that RAM leap was necessary to employ modern rendering techniques. The closest thing to that this gen is RT, which the other consoles are barely tapping, an 4 extra GB of RAM is enough to keep all of Manhattan in Spider Man: Mile's Morales BVH tree. I think that the CPU and the RT clocks will be the primary limiter at that point.

Not to say that exclusives can't take advantage of 16GB of RAM, but that the amount of advantage over 12GB is vanishingly small.

I'm not a professional game developer either, so I'm definitely talking out of my ass, but I don't think RAM usage in consoles is that closely related to either GPU performance levels or output resolution. Buffer objects themselves take up a relatively small proportion of a console's memory, with even a full 4K FP16 RGBA buffer clocking in at only around 66MB by my count. Even with additional Z buffers and G buffers and other intermediate buffers, I'd be surprised if PS5 or XBSX games allocate more than low hundreds of MBs to buffers. Hence why the reduction in RAM is one of the big issues brought up with developing for the Series S vs the Series X; dropping the resolution alone likely only saves 100MB or 200MB directly, and developers then have to find an extra 5GB+ of savings elsewhere.

To my knowledge, most memory in games consoles is used by game assets, particularly textures, and to a lesser extent geometry. There is, in theory, a relationship here between output resolution and how much RAM is required. If the textures you have in memory are high enough resolution to provide the highest mip level required by their size on screen and the resolution, then you could argue you have enough RAM. However, in a world where almost every game features asset streaming, having textures which are significantly higher resolution than required by that frame in memory means you have a safety net when it comes to streaming in assets. There's more room to allow players to behave unpredictably or move very quickly before the streaming system has to catch up. Even then, I don't think we're really approaching that point. The PS4, which as you say wasn't short on RAM, still clearly wasn't reaching the limits of texture fidelity at 1080p, outside perhaps some very linear and constrained environments like in Naughty Dog games.

There is one technology which I do think will have an impact on memory usage relative to performance, though, and that's DLSS. DLSS lowers the GPU performance required to hit a given output resolution by rendering at a lower res and temporally upscaling, but importantly it still expects textures to be sampled for the output resolution. Put another way, DLSS allows games to use higher quality assets without increasing GPU performance proportionally, but will still need a proportional RAM increase to actually hold these assets.

RT will also have an impact, although I'm not familiar with what kind of BVH sizes are typical at the moment. Still, more RAM to hold more detailed BVHs and get more accurate shadows/reflections/etc. would I'm sure be appreciated.

I should also say that I'm more concerned about exclusives than ports. PS4 era ports should be fine in any case, and anything from later in the generation will have been built to leverage the 12GB Xbox One X, so I can see how 16GB wouldn't be useful there, but being overkill for old games isn't something that really bothers me. In terms of PS5/XBSX ports, I agree that other factors like CPU performance will be a much bigger restriction, but having 16GB of RAM would give them one less thing to have to deal with. Even if they're not using PS5 quality assets, the system also isn't going to be able to stream data in as quickly as a PS5, so having more RAM can make up for slower storage.

It's really Nintendo's exclusives I'm most interested in, though, and I'd be very surprised if they couldn't leverage 16GB of RAM if it was available to them. Personally I'd be fine with 8GB, particularly if it's combined with a fast storage solution and the ability to quickly decompress that data (and we know it has the latter), and I'd be happy with 12GB, and even happier with 16GB. But that's more based on the realities of the DRAM market right now than any kind of inherent limit on how much they can utilise. I'd honestly say that even 32GB would provide a noticeable benefit for first party exclusives, even if I don't think it would be a sensible use of a limited system budget.

Incidentally, it occurred to me that 36GB isn't actually the limit for a 128-bit LPDDR5 bus. There are 32-bit 16GB LPDDR5 modules, and with four of those you could get to 64GB. Or, if they're using Grace's LPDDR5X controllers, then they can also use a pair of the 60GB 64-bit LPDDR5X modules which Grace is using, which would give them 120GB of RAM (and ECC to boot). I'm quite happy to accept that would be overkill, though!
 
Zero chance he's including licensing, Nintendo's cut is 30% from third-parties which is $18 for games that launch at $60 (which is basically all games that use a 16 GB cart)

Hmm, some people say that Nintendo/Sony/Microsoft only have a 15% cut for physical sales which would make it literally possible if a 16 GB cart and the box only costs $5.

But probably need to ask for clarification.

(Also, have to imagine that Sony, Microsoft, and Nintendo want physical games gone ASAP if they're losing $9 to $10.50 a sale on every game sold on their platform at a games store)
 
0
Nate says that Nintendo charged $14 per cart for 16 GB carts to third-parties last year.


I'm going to put that in the field of speculation on his part unless he provided sources or other credible sources such as the DF guys corroborated that info. Now if that pricing were to include the packaging and shipping cost, then sure, that would be believable, but $14 for just the cartridge itself? Nah, I'm calling BS on that.
 
EDIT: Looking a couple posts down, a Tweet said 32GB carts are 60% more expensive than a typical 50GB Blu-Ray.
Printing costs being the primary driver of per unit costs, 60% sounds huge. For a $50 dollar game that means an extra 20% of sales to make the equivalent profit for the other platforms.
 
0
(It's more likely that Monster Hunter Next is multiple years away as they probably would have announced it already if it was coming soon by 2024)
didn't it get leaked in that big capcom hack for Q2 of 2023 (FY)? I think it could've just been pushed back or something and it'll get shown at tgs
 
didn't it get leaked in that big capcom hack for Q2 of 2023 (FY)? I think it could've just been pushed back or something and it'll get shown at tgs

Basically nothing in the Capcom hack came close to hitting its release date, lol.

Dragon's Dogma 2 is 2025 at the earliest and was scheduled for release in 2022.

RE: Outrage and RE: Apocalypse were also cancelled (as probably were Final Fight and Power Stone and Omnimusha) while SF6 missed its release date as well by many months.
 
$14 was the cost to partners -- including licensing and such. Cost to Nintendo would be roughly 30% lower.

This implies a $4 cut to Nintendo from physical game sales of third-party titles which is obviously wrong as Nintendo gets an $18 cut from $60 digital game sales and would be encouraging no third-party to release any physical games if the gap was so outrageous.

But manufacturing+licensing being $14 implies that 16 GB carts are only $5 which is fine then.
 
This implies a $4 cut to Nintendo from physical game sales of third-party titles which is obviously wrong as Nintendo gets an $18 cut from $60 digital game sales and would be encouraging no third-party to release any physical games if the gap was so outrageous.

But manufacturing+licensing being $14 implies that 16 GB carts are only $5 which is fine then.

Does anyone know what the retailers cut is for physical games? I seem retailers doing sales on $60 games down to about $40 quote often, so I would assume retailers are paying roughly that for $60 games. I cant imagine they are taking a loss every time they do a sale on games. Unless of course publishers have a rebate program in place when retailers do a sale on games. Still, if we assume the a new $60 game cost WalMart roughly $45, all of a sudden the $14 cart plus licensing fee seems about right and is a modest increase over other platforms.
 
It's really Nintendo's exclusives I'm most interested in, though, and I'd be very surprised if they couldn't leverage 16GB of RAM if it was available to them. Personally I'd be fine with 8GB, particularly if it's combined with a fast storage solution and the ability to quickly decompress that data (and we know it has the latter), and I'd be happy with 12GB, and even happier with 16GB.
I realize we're a couple steps down the road now, but yes, you're totally right on all counts. I'm just not worried about Nintendo hardware giving Nintendo the performance that Nintendo games need. When I hear someone saying there would be developer backlash to 8GB, I'm assuming they're talking about third parties. While devs might not be super happy with 8GB of RAM, it is in keeping with the across the the board constraints of modern systems - and that includes the majority of PC GPUs.

But I take your point on overkill and I rescind it as exaggeration :)
 
Does anyone know what the retailers cut is for physical games? I seem retailers doing sales on $60 games down to about $40 quote often, so I would assume retailers are paying roughly that for $60 games. I cant imagine they are taking a loss every time they do a sale on games. Unless of course publishers have a rebate program in place when retailers do a sale on games. Still, if we assume the a new $60 game cost WalMart roughly $45, all of a sudden the $14 cart plus licensing fee seems about right and is a modest increase over other platforms.

Retailers get 30% so it's $18 for new AAA Switch games.

So a third-party would get $28 for each physical sale of a $60 Switch game on a 16 GB cart.
 
I'm beginning to wonder if even Nintendo had no clue how successful the Switch would be, so with Macronix (that's who manufactures the Game Carts, correct?), they could not negotiate a great deal back then, but would be looking to reverse that for the NG Switch to keep prices competitive.

I only say this because if we want Game Carts to be much much faster than current ones for Switch (assuming UFS internal storage and such, plus faster microSD), the price would be going up correct due to new tech being used, or at least a larger bus, more pins, etc?

But at the same time, the Switch has over 1 BILLION in Software sales total, so even if only 1/3rd of that is Physical sales, that's still 330-340 million Game Carts, and that I consider to be a conservative figure. I'd be curious what the overall all-time Physical sales are, but I believe Nintendo does not disclose that figure except that Physical vs. Digital for general sales in individual games are about 50/50 split. Does that mean there are about a half billion Game Carts been manufactured?

Again, something just doesn't add up with all this. How can the carts be that fricking expensive, especially when there's been hundreds of millions of these damn things manufactured already, and looking to be another few hundred million or so being manufactured for NG Switch with the new Game Cart standard?

Then again, Nintendo could simply opt for the exact same Game Carts as before, but given their history with carts seems unlikely.
 
if nintendo starts designing their games for 4K output, then 32GB might not be enough. if even a couple of gigabytes of Zelda is devoted to textures, that's a 4x increase in file size
most of Nintendo franchises has/have artstyle that dont need a huge amount of textures, to work, they are king on otimization/compression, even if the next 3D Mario, Legend of Zelda, Metroid and so on, has more detailed assets/textures i dont expect a subtancial increase in file sile, even Tear of the Kingdom that a huge game is just 16.4GB in file size,


 
Basically nothing in the Capcom hack came close to hitting its release date, lol.

Dragon's Dogma 2 is 2025 at the earliest and was scheduled for release in 2022.

RE: Outrage and RE: Apocalypse were also cancelled (as probably were Final Fight and Power Stone and Omnimusha) while SF6 missed its release date as well by many months.
the pandemic problaby played a part on this.
 
0
most of Nintendo franchises has/have artstyle that dont need a huge amount of textures, to work, they are king on otimization/compression, even if the next 3D Mario, Legend of Zelda, Metroid and so on, has more detailed assets/textures i dont expect a subtancial increase in file sile, even Tear of the Kingdom that a huge game is just 16.4GB in file size,




I mean, this is what people said about Bethesda Games Studio for a long time and now Starfield is 125 GBs, lol.

Obviously Nintendo has much more reason to make their games as small file size as possible.
 
I mean, this is what people said about Bethesda Games Studio for a long time and now Starfield is 125 GBs, lol.

Obviously Nintendo has much more reason to make their games as small file size as possible.
Bethesda is pretty terrible at otimazing they games, they need to learn at Nintendo how to proprely otimize a game, in fact most of the game industry need to learn with Nintendo, how to proprely otimize games, when the current gen begin, everyone trough, nice SSD will make everything fast and the file sizes of the games small.
 
Ty for the detailed explanation. Could anyone go into more detail on the changes that indicate the next gen system stuff?
The specifics are getting kind of deep in the weeds, but basically we've been seeing several new features added that either don't apply to the current hardware (such as support for device trees) or appear to be present but disabled on the current hardware.
This was largely a hard rewrite. Early on they may have ported certain critical userspace processes from Horizon CTR as a starting point (NintendoSDK 0.7.0 had a SRV0 process that at a glance looks suspiciously similar to 3ds srv), but that was all long gone well before any game devs had hardware.
Rewrite and fork are not mutually exclusive. Rewrites often aren't done from scratch, and frequently use the previous version of the thing as a starting point.
 
I'm beginning to wonder if even Nintendo had no clue how successful the Switch would be, so with Macronix (that's who manufactures the Game Carts, correct?), they could not negotiate a great deal back then, but would be looking to reverse that for the NG Switch to keep prices competitive.

I only say this because if we want Game Carts to be much much faster than current ones for Switch (assuming UFS internal storage and such, plus faster microSD), the price would be going up correct due to new tech being used, or at least a larger bus, more pins, etc?

But at the same time, the Switch has over 1 BILLION in Software sales total, so even if only 1/3rd of that is Physical sales, that's still 330-340 million Game Carts, and that I consider to be a conservative figure. I'd be curious what the overall all-time Physical sales are, but I believe Nintendo does not disclose that figure except that Physical vs. Digital for general sales in individual games are about 50/50 split. Does that mean there are about a half billion Game Carts been manufactured?

Again, something just doesn't add up with all this. How can the carts be that fricking expensive, especially when there's been hundreds of millions of these damn things manufactured already, and looking to be another few hundred million or so being manufactured for NG Switch with the new Game Cart standard?

Then again, Nintendo could simply opt for the exact same Game Carts as before, but given their history with carts seems unlikely.
nah, considering Nintendo is Macronix's biggest customer, and the only customer of XtraROM, Macronix doesn't exactly have any legs to stand on

most of Nintendo franchises has/have artstyle that dont need a huge amount of textures, to work, they are king on otimization/compression, even if the next 3D Mario, Legend of Zelda, Metroid and so on, has more detailed assets/textures i dont expect a subtancial increase in file sile, even Tear of the Kingdom that a huge game is just 16.4GB in file size,



part of the problem with using this method of texturing is that the higher in fidelity you go, the more complicated your shader stack has to be. it's good for repetitive faces and objects, but when you start dressing the world in more objects, that complicated shader stack starts putting a load on the gpu. so they have to ask if they want to save on memory or gpu resources
 
Bethesda is pretty terrible at otimazing they games, they need to learn at Nintendo how to proprely otimize a game, in fact most of the game industry need to learn with Nintendo, how to proprely otimize games, when the current gen begin, everyone trough, nice SSD will make everything fast and the file sizes of the games small.

? Skyrim was 3.8 GBs, lol. Less than Skyward Sword on Wii.
 
OKAY SO

I bothered my wife about this and feel terrible for wasting her time (she says she thought it was fun but...)

Anyway, she says the MoneyDJ article is just talking about the financial report of the company (saying it did well) and randomly states "the supply chain suggests a new Japanese game console will be released in Q1 2024 which should also help their financials as they are partners with this Japanese game console company."

No links to anything are given and no suggestion is given by the author that this is taken from their financial reports.

The summaries in the meeting minutes for PixArt do not suggest anything.

Basically, the MoneyDJ writer seems to just be giving her own speculation about PixArt's upcoming financial performance, using her own prior rumor about a Q1 2024 release date.

WccfTech bizarrely misread a summary of the article and thought that the financial report and the rumor were connected, that the company had reported that they were expecting higher revenues due to the release of a new Japanese game console in Q1 2024. This does not seem to be the case based on everything I've seen.
Yo shout out to the wife that's really cool
 
0
I'm not a professional game developer either, so I'm definitely talking out of my ass, but I don't think RAM usage in consoles is that closely related to either GPU performance levels or output resolution. Buffer objects themselves take up a relatively small proportion of a console's memory, with even a full 4K FP16 RGBA buffer clocking in at only around 66MB by my count. Even with additional Z buffers and G buffers and other intermediate buffers, I'd be surprised if PS5 or XBSX games allocate more than low hundreds of MBs to buffers. Hence why the reduction in RAM is one of the big issues brought up with developing for the Series S vs the Series X; dropping the resolution alone likely only saves 100MB or 200MB directly, and developers then have to find an extra 5GB+ of savings elsewhere.

To my knowledge, most memory in games consoles is used by game assets, particularly textures, and to a lesser extent geometry. There is, in theory, a relationship here between output resolution and how much RAM is required. If the textures you have in memory are high enough resolution to provide the highest mip level required by their size on screen and the resolution, then you could argue you have enough RAM. However, in a world where almost every game features asset streaming, having textures which are significantly higher resolution than required by that frame in memory means you have a safety net when it comes to streaming in assets. There's more room to allow players to behave unpredictably or move very quickly before the streaming system has to catch up. Even then, I don't think we're really approaching that point. The PS4, which as you say wasn't short on RAM, still clearly wasn't reaching the limits of texture fidelity at 1080p, outside perhaps some very linear and constrained environments like in Naughty Dog games.

There is one technology which I do think will have an impact on memory usage relative to performance, though, and that's DLSS. DLSS lowers the GPU performance required to hit a given output resolution by rendering at a lower res and temporally upscaling, but importantly it still expects textures to be sampled for the output resolution. Put another way, DLSS allows games to use higher quality assets without increasing GPU performance proportionally, but will still need a proportional RAM increase to actually hold these assets.

RT will also have an impact, although I'm not familiar with what kind of BVH sizes are typical at the moment. Still, more RAM to hold more detailed BVHs and get more accurate shadows/reflections/etc. would I'm sure be appreciated.

I should also say that I'm more concerned about exclusives than ports. PS4 era ports should be fine in any case, and anything from later in the generation will have been built to leverage the 12GB Xbox One X, so I can see how 16GB wouldn't be useful there, but being overkill for old games isn't something that really bothers me. In terms of PS5/XBSX ports, I agree that other factors like CPU performance will be a much bigger restriction, but having 16GB of RAM would give them one less thing to have to deal with. Even if they're not using PS5 quality assets, the system also isn't going to be able to stream data in as quickly as a PS5, so having more RAM can make up for slower storage.

It's really Nintendo's exclusives I'm most interested in, though, and I'd be very surprised if they couldn't leverage 16GB of RAM if it was available to them. Personally I'd be fine with 8GB, particularly if it's combined with a fast storage solution and the ability to quickly decompress that data (and we know it has the latter), and I'd be happy with 12GB, and even happier with 16GB. But that's more based on the realities of the DRAM market right now than any kind of inherent limit on how much they can utilise. I'd honestly say that even 32GB would provide a noticeable benefit for first party exclusives, even if I don't think it would be a sensible use of a limited system budget.

Incidentally, it occurred to me that 36GB isn't actually the limit for a 128-bit LPDDR5 bus. There are 32-bit 16GB LPDDR5 modules, and with four of those you could get to 64GB. Or, if they're using Grace's LPDDR5X controllers, then they can also use a pair of the 60GB 64-bit LPDDR5X modules which Grace is using, which would give them 120GB of RAM (and ECC to boot). I'm quite happy to accept that would be overkill, though!
Nintendo exceeding the RAM capacity of PS5/XS would be very funny. Probably not happening, but still. That would certainly be one way to counter any concerns of slow storage.
 
It feels like I live in the Groundhog Day. When it comes to non-English rumors, always check the primary source instead of trusting these content farms.

Edit: Just to be clear, I'm complaining about the content farms, not people sharing stuff on this thread, which is good!
I said this on a previous page, but I feel like yours is a valid but minor point, which people are now taking too far in the opposite direction as if the reporting is debunked. The interpretation you raised is just that the author isn't claiming to have supply chain information about Nintendo from PixArt, but from other sources (Foxconn) which, if true, would probably have implications for PixArt's business since they're a Nintendo supplier. Correct?

I don't even want to give Wccftech a click on their crappy article, but I did see the headline, and it's certainly inaccurate since it calls PixArt a Chinese SoC manufacturer, when they're actually a Taiwanese image sensor supplier. And I assume from the headline they're claiming the inside info is from PixArt. Maybe it is, maybe it isn't, but either way, what those MoneyDJ articles are claiming isn't too ambiguous and I think the focus should be on the legitimacy of that claim.
 
I am a developer but not a game developer so I don't know what I'm talking about. It just seems that many of the recent optimization tricks are ram heavy. I think it would be worthwhile to go to 16gb instead of "matching" similar ratio to other consoles at 12gb. It's the one thing they can do better than the other consoles at a relatively low cost. Like instead of a few extra hundreds MHz on the CPU or GPU that still pale to the competitions. A relatively overspec RAM amount would be a unique hardware identifier that could really attract devs and make for unique features. A secret sauce if you will.
 
Last edited:
With RDR1 remaster (whatever it ends like being) probably heading to Switch I would say that:

1) Its extremely likely we will see RDR2 on Switch NG (or even an impossible port in Switch 1).

2) Its not really an entire dream to think of something like GTAVI releasing in Switch NG. Probably a year later than PS5/XBSeries versions (that I guess will be late 2024 titles).

We only need Ubisoft leaks/rumors that they are planning to add Switch NG in their current pipeline plans.
 
Basically nothing in the Capcom hack came close to hitting its release date, lol.

Dragon's Dogma 2 is 2025 at the earliest and was scheduled for release in 2022.

RE: Outrage and RE: Apocalypse were also cancelled (as probably were Final Fight and Power Stone and Omnimusha) while SF6 missed its release date as well by many months.
I think MH6 could still be the game they're talking about though. Matches the internal delay pattern I suppose. Tsujimoto also said to look forward to TGS. Gives them like a good half a year to market the game.
 
Last edited:
I think MH6 could still be the game they're talking about though. Matches the internal delay pattern I suppose. Tsujimoto also said to look forward to TGS. Gives them like a good half a year to market the game.
MH6 would cannibalize Dragon's Dogma 2, so I can't see that happen
 
nah, considering Nintendo is Macronix's biggest customer, and the only customer of XtraROM, Macronix doesn't exactly have any legs to stand on
I wouldn't be surprised if they were the only customer now, but XtraROM was heavily used in arcade machines in Japan, where piracy protection and long lifetimes matter/mattered.
 
0
Have no clue if this has been linked before, but with all the talk concerning the Game Cards for the Switch, XtraRom, Macronix, I just did a quick interwebs search, and found some interesting stuff, including die shots of the rom chips themselves, and such, plus other info:

https://forums.atariage.com/topic/270216-nintendo-switch-cartridge-eprom-or-maskrom/

https://hackaday.com/2020/12/03/game-cartridges-and-the-technology-to-make-data-last-forever/

I'm sure there's a rabbit hole to dig into concerning the XtraRom chips, but I'm also sure someone in this thread has already gone through all of this in page 100 or something for all I know.
 
most of Nintendo franchises has/have artstyle that dont need a huge amount of textures, to work, they are king on otimization/compression, even if the next 3D Mario, Legend of Zelda, Metroid and so on, has more detailed assets/textures i dont expect a subtancial increase in file sile, even Tear of the Kingdom that a huge game is just 16.4GB in file size,



It's always crazy how Sunshine continues to remain a good-looking game to me.
 
Sometime next year split between summer & holidays for when the launch happens.
Does anyone really think summer is more likely than early/spring? Holiday is the "normal" time to launch a console, and if they're launching outside that, it would make more sense to predict a window they have launched at least some systems in the past.
 
Does anyone really think summer is more likely than early/spring? Holiday is the "normal" time to launch a console, and if they're launching outside that, it would make more sense to predict a window they have launched at least some systems in the past.
I could have gotten the date wrong for which we were discussing for the non-holiday period.
 
0
Rewrite and fork are not mutually exclusive. Rewrites often aren't done from scratch, and frequently use the previous version of the thing as a starting point.
Right. Though here it's more complicated. They didn't start with a fork of their old repo, they started with a whole new repo. Any porting was done on a per-component basis and was generally a temporary solution to help with their development timeline, since having things like a working ServiceManager implementation is required to work on most other things.
I don't think calling the whole thing a fork when they started with a new repo and 90% of their code was written from scratch (EDIT: the other <10% being externals) is fair.
 
I'm beginning to wonder if even Nintendo had no clue how successful the Switch would be, so with Macronix (that's who manufactures the Game Carts, correct?), they could not negotiate a great deal back then, but would be looking to reverse that for the NG Switch to keep prices competitive.

I only say this because if we want Game Carts to be much much faster than current ones for Switch (assuming UFS internal storage and such, plus faster microSD), the price would be going up correct due to new tech being used, or at least a larger bus, more pins, etc?

But at the same time, the Switch has over 1 BILLION in Software sales total, so even if only 1/3rd of that is Physical sales, that's still 330-340 million Game Carts, and that I consider to be a conservative figure. I'd be curious what the overall all-time Physical sales are, but I believe Nintendo does not disclose that figure except that Physical vs. Digital for general sales in individual games are about 50/50 split. Does that mean there are about a half billion Game Carts been manufactured?

Again, something just doesn't add up with all this. How can the carts be that fricking expensive, especially when there's been hundreds of millions of these damn things manufactured already, and looking to be another few hundred million or so being manufactured for NG Switch with the new Game Cart standard?

Then again, Nintendo could simply opt for the exact same Game Carts as before, but given their history with carts seems unlikely.

As others have mentioned, Nintendo accounts for a large part of Macronix's business, possibly all of their XtraROM business, so they're definitely in a better negotiating position.

The big issue with game card capacities is that manufacturing advancements for XtraROM have stalled. Macronix have a "Milestones" section in their Annual Report, which among other things lists advancements in their various technologies. Here are the references to XtraROM:

2006 - Mass production of 100nm XtraROM®
2007 - Mass production of 75nm XtraROM®
2008 - Mass production of 65nm XtraROM®
2012 - Mass production of 45 nm XtraROM®
2014 - Mass production of 32 nm XtraROM® products

They're still on the 32nm process. Usually you'd expect newer manufacturing processes to come along to increase capacities and bring down prices, but this hasn't happened with XtraROM. There are obviously clear incentives for Macronix to continue advancing XtraROM, as without doing so they risk losing their biggest client, but there's been no movement in almost 10 years, after pretty regular advancements up to that point (and they've continued to improve the processes for their other product lines like NAND flash).

Have no clue if this has been linked before, but with all the talk concerning the Game Cards for the Switch, XtraRom, Macronix, I just did a quick interwebs search, and found some interesting stuff, including die shots of the rom chips themselves, and such, plus other info:

https://forums.atariage.com/topic/270216-nintendo-switch-cartridge-eprom-or-maskrom/

https://hackaday.com/2020/12/03/game-cartridges-and-the-technology-to-make-data-last-forever/

I'm sure there's a rabbit hole to dig into concerning the XtraRom chips, but I'm also sure someone in this thread has already gone through all of this in page 100 or something for all I know.

It's NROM. See here and here.
 
Last edited:
I realize we're a couple steps down the road now, but yes, you're totally right on all counts. I'm just not worried about Nintendo hardware giving Nintendo the performance that Nintendo games need. When I hear someone saying there would be developer backlash to 8GB, I'm assuming they're talking about third parties. While devs might not be super happy with 8GB of RAM, it is in keeping with the across the the board constraints of modern systems - and that includes the majority of PC GPUs.

But I take your point on overkill and I rescind it as exaggeration :)

I'm actually really curious how developers would work with a console with an obscenely over-specced amount of RAM. I think in one of your posts a while back you used a Wii with 16GB of RAM as an extreme case of overkill, but I'd be really interested in finding an alternate universe where that actually happened. I suspect the limit in extreme cases like that is how quickly you can load data into RAM and the audience's patience for loading screens. The Wii supported 8.5GB discs, and with a plausible 2:1 compression ratio that could mean 16GB of game data, so you could actually fill the entire RAM with assets, so long as players are willing to put up with probably 20 or 30 mins of loading as it pulls all the data off the disc and decompresses it on the Wii's CPU. The Wii really would have benefitted from a modern standby mode in that case!

Nintendo exceeding the RAM capacity of PS5/XS would be very funny. Probably not happening, but still. That would certainly be one way to counter any concerns of slow storage.

The strange thing is that it's entirely possible that Nintendo's new hybrid system launches at around the same time as the PS5 Pro, and has more usable RAM for games. The PS4 Pro had the same RAM capacity as the PS4, so I wouldn't be surprised to see the PS5 Pro stick to 16GB as well. I don't think 16GB is particularly likely for Switch NG, but the fact that RAM prices have plummeted so much over the past year makes me think it's at least somewhat plausible, and if they go that route then it's a safe bet that they'll reserve less RAM for the OS than Sony or MS do.
 
Capcom has given no date, officially.
Then I don't see why MH6 coming out early next year is unbelievable. The monster hunter guy like the guy himself alluded that there would be news at tgs and if the report is right they were probably talking about monster hunter as I don't really see any other game doing millions in that short amount of time. Paired with the data leak, they already projected the game would be out within a few months (granted that was a while ago and the leak was pretty inconsistent with release schedules), shift around the dates and things check out no?

sorry this has NOTHING to do with hardware I'm just rambling
 
Right. Though here it's more complicated. They didn't start with a fork of their old repo, they started with a whole new repo. Any porting was done on a per-component basis and was generally a temporary solution to help with their development timeline, since having things like a working ServiceManager implementation is required to work on most other things.
I don't think calling the whole thing a fork when they started with a new repo and 90% of their code was written from scratch (EDIT: the other <10% being externals) is fair.
If it's known that they started from scratch, that'd be one thing, but I'd question how that's known. Being a new repo and there being little to nothing remaining of the previous thing aren't super strong indicators by themselves. You'd have to have some pretty deep insight to know if the starting point was an empty repo or a copy of the 3DS one.

That said, calling it a rewrite is probably fair, and a more useful description of what happened. I'm just not aware of any evidence that really supports it being a true "burn it to the ground" rewrite, especially since it seemingly kept the name and there's some evidence of 3DS component use.
 
I read today on twitter, that some chinese manufacteur informed about manufacturing part for a new console to be released in Early 2024.

Can somebody tell me what fami thinks about this, since there are way too many pages right now to read through xD

Thanks for any clarification!
 
Here are the references to XtraROM:

2006 - Mass production of 100nm XtraROM®
2007 - Mass production of 75nm XtraROM®
2008 - Mass production of 65nm XtraROM®
2012 - Mass production of 45 nm XtraROM®
2014 - Mass production of 32 nm XtraROM® products

They're still on the 32nm process.
Didn't know it was this bad. From a layman perspective, my first thought is that Nintendo and Macronix decided to move away from XtraROM, but had to stick with it this gen.

Maybe they thought prices for 32nm wafers would go down enough for 32GB be at acceptable prices and wasn't worth the R&D on a tech about to die.
 
I'm actually really curious how developers would work with a console with an obscenely over-specced amount of RAM. I think in one of your posts a while back you used a Wii with 16GB of RAM as an extreme case of overkill, but I'd be really interested in finding an alternate universe where that actually happened. I suspect the limit in extreme cases like that is how quickly you can load data into RAM and the audience's patience for loading screens. The Wii supported 8.5GB discs, and with a plausible 2:1 compression ratio that could mean 16GB of game data, so you could actually fill the entire RAM with assets, so long as players are willing to put up with probably 20 or 30 mins of loading as it pulls all the data off the disc and decompresses it on the Wii's CPU. The Wii really would have benefitted from a modern standby mode in that case!



The strange thing is that it's entirely possible that Nintendo's new hybrid system launches at around the same time as the PS5 Pro, and has more usable RAM for games. The PS4 Pro had the same RAM capacity as the PS4, so I wouldn't be surprised to see the PS5 Pro stick to 16GB as well. I don't think 16GB is particularly likely for Switch NG, but the fact that RAM prices have plummeted so much over the past year makes me think it's at least somewhat plausible, and if they go that route then it's a safe bet that they'll reserve less RAM for the OS than Sony or MS do.
As a general rule, Nintendo's handhelds (except the DS until the DSi) have tended to have significantly more RAM than the home console that they're most comparable to, despite pushing lower resolutions, but not by orders of magnitude. You could even kinda include Switch and Wii U in that. Unfortunately, there are a lot of confounding factors there, though.
 
Capcom has stated they have an unannounced game that'll be released by March 31st 2024. How likely do you think it could be delayed and become a Switch 2 launch title(I personally doubt it) and how likely could it be that it's MegaMan X9?

Source
 
I said this on a previous page, but I feel like yours is a valid but minor point, which people are now taking too far in the opposite direction as if the reporting is debunked. The interpretation you raised is just that the author isn’t claiming to have supply chain information about Nintendo from PixArt, but from other sources (Foxconn) which, if true, would probably have implications for PixArt’s business since they’re a Nintendo supplier. Correct?
I don’t think the distinction that I raised was minor, otherwise I wouldn’t have bothered. Yes, the meat of these MoneyDJ articles is that they received inside info (may or may not be correct) of the Switch NG launch window, and the source was probably from within Foxconn as I suggested here. However, any connections to PixArt or Weltrend are suppositional on the part of the reporter—this is of note because:
  1. According to factory uncle #2, the Joy-Con has been redesigned, therefore PixArt’s involvement isn’t a given. When people take MoneyDJ’s conjecture as a fact, speculations of the new Joy-Con inheriting the current IR camera or adding any new image sensors start taking a life of its own without any evidence to substantiate it.
  2. When translating foreign rumors, accuracy matters more than usual, because most readers would not check the source before parsing the words and building new narratives. For instance, Mochizuki baselessly claimed that Nintendo’s FY03/24 projection does not include any new hardware, and to this day many still cite that misinformation. And in this particular case, the speculative connection made by the reporter somehow was taken as a disclosure made by PixArt itself.
It’s never my intention to change any minds or harsh anyone’s buzz, but to encourage more discerning examinations of foreign rumors mediated by content farms.
 
Please read this new, consolidated staff post before posting.

Furthermore, according to this follow-up post, all off-topic chat will be moderated.
Last edited:


Back
Top Bottom