• 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!)

Might as well delete all the evidence/info in the OP. The T239 doesn't exist anymore either, it was probaby destroyed in a disaster.

VvBeRGql.jpeg
T239 died on the way back to his home planet.
 
this sound like they are hiring for a port of Dragon Quest 12 for Nintendo next hardware
I would expect SE (via the primary dev) to do that themselves, like with 11S

I'm not expecting it to be constantly at its max for minutes on end. I doubt PS5 or Series do that outside of loading screens. But they could definitely being using it to its max across a handful of frames, enough to where you don't even realize something was being loaded in. For a game running at 60fps, PS5's 5GB/s rate is about 83.33MB/frame (which is roughly around how long it takes for PS4 to load sequential data from its HDD in one second), and Series' 2.4GB/s is about 40MB/s. Assuming Drake's storage ran on 100MB/s, that's about 1.66MB/frame.
well the first problem is that you're using theoretical maximums, which rarely happen outside of rare instances. those numbers are sequential, which is less useful for games anyway. Digital Foundry shown that one of the prominent games used to highlight the move to an SSD (Ratchet and Clank) doesn't even break 1.x GB/s during data streaming. maybe something more modern like Burning Shores push it harder, but someone will have to actually test that
 
So, what kind of tech are you guys thinking Nintendo can use on the cartridges? Is there anything with fast read and surpass the 50 GB without cost too much?
 
this sound like they are hiring for a port of Dragon Quest 12 for Nintendo next hardware
Dragon Quest XII certainly isn't an unannounced AAA game based on Japanese IP, at least going by Area 35's Visual Effects Artist job listing, and the Senior Environment Artist job listing that no longer exists, but is in that video.

And speaking of, Orca's been hiring for a bunch of artist positions for Dragon Quest XII. So I don't know where Area 35's going to fit in.

So, what kind of tech are you guys thinking Nintendo can use on the cartridges? Is there anything with fast read and surpass the 50 GB without cost too much?
There's a rumour of Macronix sending samples of Macronix's 48-layer 3D NAND flash memory to Nintendo. But I don't know if that went anywhere.
 
Last edited:
I think that it’s obvious that storage speeds for REDACTED will lag behind current consoles, just like CPU and GPU perf. I am dubious about that being the critical bottleneck.

It’s obvious that games natively targeting REDACTED’s performance envelop are well served by 100MB/s. Last gen games, indies and Nintendo first party games will all be in excellent shape.

Yes, other current-gen-only multiplats will have this as yet another pain point, but there is a reason that PS4 ports to the Switch were called miracle ports. I don’t expect REDACTED to get more support than it’s predecessor in that regard.

And if it does there are a number of mitigating strategies available. We don’t know how much RAM Nintendo intends to provide, but 12GB would provide a nice buffer over the Series S for preloading assets. A lower resolution target - inevitable with the much less capable GPU - will reduce asset sizes.

This discussion has ranged a bit from your initial assertion, which was that Nintendo will mandate installs from cartridges. I think it’s obvious that the majority of games won’t need faster-than-HDD speeds, and as such, Nintendo is unlikely to mandate installs by default.

I think there are economic disincentives here as well, and that Nintendo simply doesn’t have a lot of cost effective options for massive, ultra fast storage which isn’t a big power drain. Whether or not that will prove problematic for Nintendo remains to be seen, but even if it is, thems the breaks when you’re dealing with a mid budget mobile device
I'm just having a hard time accepting that Nintendo would limit themselves this way when we so far have an idea that they are pushing for more in a small package than they have in the past, using a custom SoC based on Orin, and adding features like hardware decompression. We didn't end up with the original plan of the Switch that would have been weaker than the Wii U because Nvidia convinced them to go further, and 3rd-parties convinced Nintendo to not limit the Switch's RAM capacity.

My assertion with regard to mandatory installs was because of the cost associated with carts, which right now is pretty expensive for something that is upwards of 32GB and operates at 15MB/s sequential read speed by my own tests. Even just skipping a capacity increase for Drake and pushing for 100MB/s on the carts is going to be even more costly that devs would have to deal with. Hence if it came down to mandatory installs, then why not use an internal storage solution that would also have the benefit of being faster? Switch already uses eMMC that handle raw reads of around 300MB/s, but games can't utilize it at nearly that rate because the CPU is too busy decompressing data. Hell, they could just swap it with UFS 1.0 (if that's still being made) for the same 300MB/s, which has better latency and is more energy efficient than eMMC, but there is also UFS 2.0 that is 600MB/s per lane if they needed it. Anything beyond that is overkill imo.
 
I'm just having a hard time accepting that Nintendo would limit themselves this way when we so far have an idea that they are pushing for more in a small package than they have in the past, using a custom SoC based on Orin, and adding features like hardware decompression. We didn't end up with the original plan of the Switch that would have been weaker than the Wii U because Nvidia convinced them to go further, and 3rd-parties convinced Nintendo to not limit the Switch's RAM capacity.

My assertion with regard to mandatory installs was because of the cost associated with carts, which right now is pretty expensive for something that is upwards of 32GB and operates at 15MB/s sequential read speed by my own tests. Even just skipping a capacity increase for Drake and pushing for 100MB/s on the carts is going to be even more costly that devs would have to deal with. Hence if it came down to mandatory installs, then why not use an internal storage solution that would also have the benefit of being faster? Switch already uses eMMC that handle raw reads of around 300MB/s, but games can't utilize it at nearly that rate because the CPU is too busy decompressing data. Hell, they could just swap it with UFS 1.0 (if that's still being made) for the same 300MB/s, which has better latency and is more energy efficient than eMMC, but there is also UFS 2.0 that is 600MB/s per lane if they needed it. Anything beyond that is overkill imo.
Why are assuming the limits to carts are costly. If the bottleneck is due to the pinout, that's not a large expense to overcome. If the bottleneck is due to the cpu, that's solved with drake's cpu and separate decompressor. So where is this hypothetical cost coming from?

Even in your own example, you show that it doesn't matter how fast your storage is because the bottlenecks are elsewhere. So why the hullabaloo about speed? Just fixing the listed issues gives you magnitudes is speed increases as is.
 
It could be something different instead of Fire Emblem, like Dragon Quest 12.

that would be weird since Engage is IS's first Unity game. maybe they're in the evaluation phase and wanted to use both to their fullest before committing?

this sound like they are hiring for a port of Dragon Quest 12 for Nintendo next hardware
I'm not sure if it's Fire Emblem either, but the job posting does seem to target something related to Switch or Switch 2

It's also a support studio job posting, so can;t rule out something a 3rd party is doing that is being contracted out.
 
Why are assuming the limits to carts are costly. If the bottleneck is due to the pinout, that's not a large expense to overcome. If the bottleneck is due to the cpu, that's solved with drake's cpu and separate decompressor. So where is this hypothetical cost coming from?

Even in your own example, you show that it doesn't matter how fast your storage is because the bottlenecks are elsewhere. So why the hullabaloo about speed? Just fixing the listed issues gives you magnitudes is speed increases as is.
But that's assuming the limitations are elsewhere besides the memory chips in the carts themselves. They use XtraROM, which is already slower than traditional flash memory because they were designed for security and data retention time. That is part of the cost. And carts have always been a more expensive medium, especially as the size increases, which for Switch is why various devs opt for using smaller carts and require mandatory downloading, or skip physical carts completely and throwing a download code into the game case. Now imagine even bigger cart sizes, and being faster in the process for Drake.

As for folks that disagree with my numbers for the Switch cart speeds, I can post captures of them shown on the Switch itself, and these folks can provide their own evidence of their games loading multiple times faster on their own hardware.
 
Dragon Quest XII certainly isn't an unannounced AAA game based on Japanese IP, at least going by Area 35's Visual Effects Artist job listing, and the Senior Environment Artist job listing that no longer exists, but is in that video.

And speaking of, Orca's been hiring for a bunch of artist positions for Dragon Quest XII. So I don't know where Area 35's going to fit in.


There's a rumour of Macronix sending samples of Macronix's 48-layer 3D NAND flash memory to Nintendo. But I don't know if that went anywhere.

I'm not saying it's DQXII, but it's often normal for many studios to be subcontracted in a large project.
Therefore, knowing one of DQXII's subcontractors does not invalidate the fact that there are other subcontracted companies.
 

I think it's worth mentioning Area 35's connection to Yusuke Kozaki isn't really strong evidence that this new project is Fire Emblem. Kozaki is a freelance character designer that works on a lot of different projects. He did the character designs for Awakening and Fates, but he had nothing to do with the three most recent games (Echoes, Three Houses, and Engage).
 
But that's assuming the limitations are elsewhere besides the memory chips in the carts themselves. They use XtraROM, which is already slower than traditional flash memory because they were designed for security and data retention time. That is part of the cost. And carts have always been a more expensive medium, especially as the size increases, which for Switch is why various devs opt for using smaller carts and require mandatory downloading, or skip physical carts completely and throwing a download code into the game case. Now imagine even bigger cart sizes, and being faster in the process for Drake.

As for folks that disagree with my numbers for the Switch cart speeds, I can post captures of them shown on the Switch itself, and these folks can provide their own evidence of their games loading multiple times faster on their own hardware.
There's an entire custom chip that sits between the cart slot and the rest of the system. That's probably a significant factor in how fast the data can be loaded in.

As for the numbers, they don't really seem to gel with what has been observed in practice, such as in this Digital Foundry video from around when the system launched:


Cartridges and the SD card generally seem to come in pretty close to one another here, with the internal storage being notably faster. If there was an order of magnitude between the cart and SD card, one would think the cart would be the outlier. I have to wonder if there might be some flaw in methodology in the test causing the cart to bottleneck in a way that it shouldn't.
 
The illegal Nvidia leaks, and Nvidia's commits on Linux and GitHub, made no mention of which type of internal flash storage is being used.

So nobody knows which type of internal flash storage Nintendo's going to ultimately use for Nintendo's new hardware.
"Slow" is not just sequential reads, latency for random access will not be a problem even with a slow sequential read. File Decompression Engine would increase the sequential read speed even when using the same internal memory from the Original Switch.
 
0
There's an entire custom chip that sits between the cart slot and the rest of the system. That's probably a significant factor in how fast the data can be loaded in.

As for the numbers, they don't really seem to gel with what has been observed in practice, such as in this Digital Foundry video from around when the system launched:


Cartridges and the SD card generally seem to come in pretty close to one another here, with the internal storage being notably faster. If there was an order of magnitude between the cart and SD card, one would think the cart would be the outlier. I have to wonder if there might be some flaw in methodology in the test causing the cart to bottleneck in a way that it shouldn't.

Again, an important factor with games is that you're not going to hit advertised sequential read speeds because games are made up of countless different assets spattered across the medium, relying on how quick the medium can handle random access. It's like comparing the reading of a 1GB file to 1000 1MB files or even 1 million 1KB files. It's going to be quite a bit faster for the single 1GB file than either of the others because it's not having to jump all over the place. Honestly, if sequential read speeds was all that mattered and random access had absolutely no delay, then loading of games from internal storage would fill RAM in mere seconds because it can handle sequential reading @ 300MB/s. But it doesn't because of the random access delays and that assets are compressed. Faster than microSD as shown by the BotW video, but not by the magnitude when comparing sequential speeds of internal vs microSD. That is because internal storage handles random access somewhat better, but not massively better.

Here are the pics I mentioned of my benchmarking and tests dumping.

NeP4cJo.jpg

vdNx0bG.jpg


And just for good measure, here is the benchmark of that same microSD card on my computer using a USB reader.

iE2KALG.png


Here are images screenshot from in the Switch itself when dumping games from cart and internal storage.

8tfV1my.jpg

yJoYPdc.jpg
 
Well, as soon after the next direct is out, we will finally get news about the Switch 2 in a few weeks.
Would love to and I also expect it, as otherwise Nintendo got even more tight lipped to releases as before. Some hype should be real, like know roughly all releases within 1 year.
But for sure, next Direct will be interesting, likely already their IR this year, as they will be for sure questioned about their future, even if they are asked just for switch releases.
 
0
…..next direct isn’t even until June.

The financial briefing is in May.

We are only in the second half of April…
 
I'm not sure if it's Fire Emblem either, but the job posting does seem to target something related to Switch or Switch 2

It's also a support studio job posting, so can;t rule out something a 3rd party is doing that is being contracted out.
true, and they never specified the consoles Dragon Quest 12 is gonna be released
 
Any speculation on what the Switch 2 new gimmick could be? It’s bound to have one right, I’d still like to see a mic and camera’s return so we can get Face Raiders 2!
 

Wonder if he deleted the tweet because it was getting unnecessary attention and felt bothered and/or because he thought it might affect his “industry sources”.

I’m tired of the pro, give me the anti


Nintendo Switch Anti: now makes your games literally worse. That’ll be $599 and two jobs, thanks.

No BC included.
lmao sounds like that FFVII NES Demake
 
Uhhh... don't wanna sound rude but what's the validity of this "store uncle" claims again?
Nothing. The reddit post is just an update to show that it was nothing this whole time too. The retail uncle is someone who works in the Japanese Gaming industry but I wouldn't put much weight into anything they say at all.
 
Dragon Quest XII certainly isn't an unannounced AAA game based on Japanese IP, at least going by Area 35's Visual Effects Artist job listing, and the Senior Environment Artist job listing that no longer exists, but is in that video.

And speaking of, Orca's been hiring for a bunch of artist positions for Dragon Quest XII. So I don't know where Area 35's going to fit in.


There's a rumour of Macronix sending samples of Macronix's 48-layer 3D NAND flash memory to Nintendo. But I don't know if that went anywhere.
if this AAA game based on a japanese IP is not Dragon Quest 12?what it could be? it cant be Fire Emblem, Fire Emblem Engage was just released this january.
 
FE Engage wrapped up development in 2021.
Chinese capture leaks indicated that in mid-2022 it was still under testing, what was finished was the base development, polishing took another year.


Anyway, about the game listed in these job offers, without knowing the development status, it can be any game announced or to be announced.
 
Chinese capture leaks indicated that in mid-2022 it was still under testing, what was finished was the base development, polishing took another year.


Anyway, about the game listed in these job offers, without knowing the development status, it can be any game announced or to be announced.
Would polishing take a big chunk of staff, or just some part it?

With already existing assets, I don’t think it’d take them that much to develop another game. But taking your info into consideration, I don’t think the game will release this year. Maybe the next one
 
Please read this staff post before posting.

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


Back
Top Bottom