• 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!)

I just don't get WHY THE HECK Nintendo doesn't use a peer-to-peer based game downloading service for the e-shop. If you're not interested in upgrading your servers, at least use an alternative that would improve download times.
Imagine if a peer manages to have a modified version of the files that can allow for exploits. There are some major security issues that have to be adressed with this worm of file sharing.

(Then again, they can always do hash checks for anything being downloaded with the eShop hosting the confirmed hash value.)
 
Steam Deck people saying that BG3 is mostly around 20-24 FPS in Act 3 (with drops to teens during major spells for all acts) after all major updates were completed.

Larian mostly focused on optimizing VRAM usage for the Series S, which I guess could be helpful for a Switch 2 port if they optimized more for slow RAM as well.

But the game looks like it could need a lot of CPU optimization for the Switch 2. Probably either turning off some stuff or moving stuff to the GPU.

But the confidence from some people that Act 3's CPU heavy nature would be cleaned up before they were done with the game on PC seems to have not come to fruition.

I imagine that this game and Elden Ring will have Nintendo and NVIDIA providing technical assistance for porting because Nintendo will desperately want both on the platform, but BG3 will probably be hard.

SD has a quad core CPU with SMT ( 4 cord, 8 threads) but that doesn’t make it equal to a Octo core CPU with 8 threads.

Switch 2 has an edge here with 8 cores but SD 4 cores will be clocked way higher
 
SD has a quad core CPU with SMT ( 4 cord, 8 threads) but that doesn’t make it equal to a Octo core CPU with 8 threads.

Switch 2 has an edge here with 8 cores but SD 4 cores will be clocked way higher

I'm pretty sure that the Deck CPU is usually clocked at around 2GHz (it has a max of 3.5 but it doesn't get anywhere near that while the GPU sits at its max of 1.6). That's around what we expect for Switch 2.
 
0
I expect Baldur's Gate to get more bespoke changes of they start porting the game to Drake. What those will be and what for they would take will be anyone's guess. But seeing as to how performance was improved for all systems when working on Series S, no doubt everything else would improve with work on Drake
 
I expect Baldur's Gate to get more bespoke changes of they start porting the game to Drake. What those will be and what for they would take will be anyone's guess. But seeing as to how performance was improved for all systems when working on Series S, no doubt everything else would improve with work on Drake

I’m playing it now on XSX, and split screen multiplayer aside I’m not sure why they couldn’t get a form of this game running on the next Switch. Probably about 12 hours in and visually yes it’s a step above D:OS2, and camera closeups for dialog are a new hotness, but it otherwise fundamentally feels like the same experience as D:OS2. Maybe I’m missing something.

Thus far I’m not sure why they said the game was “not possible on PS4.” CPU challenges? Studio resources? Because these visuals really should scale fine. They need to be better than D:OS2 on Switch but that was mostly an issue with resolution.
 
I’m playing it now on XSX, and split screen multiplayer aside I’m not sure why they couldn’t get a form of this game running on the next Switch. Probably about 12 hours in and visually yes it’s a step above D:OS2, and camera closeups for dialog are a new hotness, but it otherwise fundamentally feels like the same experience as D:OS2. Maybe I’m missing something.

Thus far I’m not sure why they said the game was “not possible on PS4.” CPU challenges? Studio resources? Because these visuals really should scale fine. They need to be better than D:OS2 on Switch but that was mostly an issue with resolution.
I haven't played BG3 but I've seen ppl mention Act 3 is where things get problematic for CPU, due to sheer amount of NPC actions or movements.

I think it's mostly a matter of if Switch 2 can handle Act 3, CPU-wise. We don't know everything about Switch 2 CPU yet.
 
I’m playing it now on XSX, and split screen multiplayer aside I’m not sure why they couldn’t get a form of this game running on the next Switch. Probably about 12 hours in and visually yes it’s a step above D:OS2, and camera closeups for dialog are a new hotness, but it otherwise fundamentally feels like the same experience as D:OS2. Maybe I’m missing something.

Thus far I’m not sure why they said the game was “not possible on PS4.” CPU challenges? Studio resources? Because these visuals really should scale fine. They need to be better than D:OS2 on Switch but that was mostly an issue with resolution.
act 3 is where a lot of people say the problems start. looking at videos on the steam deck and rog ally and the like definitely confirms that. folks are putting the game in ultra performance mode and are gaining no performance because of cpu bottlenecks in both the SD's 4 core and Ally's 8 core.

now that the Xbox versions are out, if the game was coming to Drake, that'll be the first thing they would tackle as it lifts all boats
 
Man, Retro Studios is still hiring artists:


Is it possible they have several projects on the way and not just Metroid Prime 4?

this job aplication, must be for a new game, Retro Studios is problaby working, when they are right now polishing Metroid Prime 4, for hopefully a 2024 release.
 
I don't believe that rumor for shit. they're most likely on their next project while wrapping up MP4

The job ads for cutscenes specifically said they were for Prime 4.

The issue is just the entire game is not done, not just the cutscenes.

Rumor people are trying to find a way to reconcile their "Prime 4 is basically done" prior rumors with Retro urgently hiring a bunch of people to work on Prime 4 cinematics with "well, everything is done but the cinematics!" which is just obviously false.
 
I’m playing it now on XSX, and split screen multiplayer aside I’m not sure why they couldn’t get a form of this game running on the next Switch. Probably about 12 hours in and visually yes it’s a step above D:OS2, and camera closeups for dialog are a new hotness, but it otherwise fundamentally feels like the same experience as D:OS2. Maybe I’m missing something.

Thus far I’m not sure why they said the game was “not possible on PS4.” CPU challenges? Studio resources? Because these visuals really should scale fine. They need to be better than D:OS2 on Switch but that was mostly an issue with resolution.

Graphically Switch 2 should be able to handle the game without issue, but as others have noted the CPU side is trickier, particularly in Act 3, where high NPC counts cause PS5 frame rates drop to the 20s, even in 60fps mode. It's not the kind of game where you can reduce NPC count, but I wonder if they can run AI asynchronously from rendering, or even just reduce how frequently they run AI logic for NPCs further away from the player. This affects all platforms, and is the biggest technical issue with the game, so hopefully they'll have some improvements in place by the time a Switch 2 version releases and we can get at least a steady 30fps.
 
Does anyone wanna bet we'll get a DK game by the end of 2024 to concide with the release of DKC in Nintendo World?
I used to wonder why they didn't just bring a single player version of GTA V to Switch 1. It would probably sell a few million units even without the online, why not do it?

Then I realised, its probably because the single player is the gateway drug to the online. They want anyone who has an itch to play GTA V to play it on a system that is getting the online updates as well. They don't want you having the option of playing it on a console that can't then lure you into playing the online mode right after.

So I think that finally explains why they didn't bring GTA V single player to Switch 1. But yeah, clearly R* is a fan of supporting Switch. We got RDR1 when PC still doesn't have that game lol and GTA trilogy remastered came to Switch and even sold better on Switch than any other platform except PS4 in Europe. LA Noire also sold decently on Switch.


if they really wanted to, I don't see why they couldn't add gta online.
I don't expect Nintendo to adopt CAMM2 due to CAMM2 probably taking a non-trivial amount of space, if CAMM is any indication. But JEDEC introduced CAMM2, which now includes LPDDR5/5X.
This is the first time I've heard of CAMM2. So it's a memory module that will replace the current modules in laptops (and mobiles) in due time. What are the advantages over the regular modules.. higher capacity RAM?



it's gonna be a lot more expensive than regular modules though. Will take a few years..
 
The job ads for cutscenes specifically said they were for Prime 4.

The issue is just the entire game is not done, not just the cutscenes.

Rumor people are trying to find a way to reconcile their "Prime 4 is basically done" prior rumors with Retro urgently hiring a bunch of people to work on Prime 4 cinematics with "well, everything is done but the cinematics!" which is just obviously false.
i really appreciate the fact that you stay true to your pessimism and being a contrarian, with regards to almost all things in this thread. even though you basically doodoo everywhere, it’s genuinely admirable lol.
 
i really appreciate the fact that you stay true to your pessimism and being a contrarian, with regards to almost all things in this thread. even though you basically doodoo everywhere, it’s genuinely admirable lol.

How is this even being pessimistic or contrarian.

"Video game is not done as shown by them hiring a ton of people (including contract workers who probably wouldn't stick around for the next game) recently to work on said video game"

"No, a game dev would not completely finish development before releasing that they need to shoot new cutscenes, that would be very incompetent."
 
Does anyone think Switch will outsell PS2 I hope so
I think that depends on their strategy with Switch 2. Given the state of the economy, I'm fairly sure that strategy will align with prolonged Switch 1 sales- the new device will almost certainly be far more expensive to produce and more expensive to buy, and this will be a barrier for lots of individuals and families. Keeping Nintendo Switch and Nintendo Switch Lite around as value offerings can really only help the brand as a whole- overwhelmingly the people most likely to pick up a Switch 2 are people who own a Switch.

What I wonder is how the pricing structure will change for the original Switch. We see this in effect already, with games given away by default with every variation. However, with a new generation, that means introducing a new model, and that's a good time to simplify your offerings. Personally I think, rather than dropping prices explicitly, the new pricing structure will be NG, OLED and Lite at 399, 299 and 199 respectively. This is sort of a price cut for OLED Model, but with the original design gone, the price of entry to own a home console hasn't gone down, which serves to preserve perceived brand value.

However, if Nintendo want to be aggressive, we could see drastic changes or a redesign. They may want to simplify their line down to just two models, a cost reduced V3 to replace all existing Switch models, and their premium next gen model. This would mirror what they did at Switch's launch, with a cost reduced New 2DS XL system and Switch being their two offerings for a time.

Personally I think keeping the Lite and OLED is the most likely - the OLED IS the die-shrunk complexity reduced TV mode capable model of Nintendo Switch already, although material choices and the screen drive up the price a bit, it already is production optimised.

At this point I think a TV only model is off the table, but I can't rid my mind of the idea that there may yet be one more model of original Nintendo Switch, this is the company that had six models of Nintendo 3DS after all. I think a cost reduced V3 model would be branded as just "Nintendo Switch", with no sub-brand, like the V2, but have visible changes. Lighter, a better kickstand, the dock with LAN Port, maybe a 7" LCD display, designed around 16nm allowing for a smaller motherboard, less internal complexity, share parts with the NG system, or reuse them from OLED Model. A real slap-together type system. Replacing the Lite, OLED and V2 and coming in at $249.99

A bit out there, maybe, but there's precedent and I think it's fun to think about.
 
The takes everyone gets mad at me about

1. The Switch 2 is going to be extremely powerful so I'm guessing it will be $499 based on how the Series S seems to still cost a ton of money to manufacture

2. 8nm can't be ruled out because all rumor and hack info lightly point there.

3. GTA6 will push the PS5 up to and probably past its limits and therefore would be incredibly difficult to port to Switch 2 without significant costs and Rockstar thus probably won't bother as Rockstar doesn't bother with things that cost resources that aren't GTA Online updates or the main games themselves.

4. MH Wilds and BG3 will be hard to port to the Switch 2 though both are possible (MH Wilds because Capcom has such a strong interest in bringing MH to the Switch and BG3 because it's not GPU heavy at all so I'm guessing there's some potential to move a lot of CPU functions to the GPU)

5. It's hard to tell if big budget UE5 games will run on Switch 2 as so far, UE5 games with complex assets have not run well

6. Ray reconstruction costs more than denoisers that are good enough in non-PT situations and thus is not viable on Switch 2. It could be viable if NVIDIA made a neural network that was much smaller and thus could run much faster.

7. There will be minimal usages of ray tracing on the Switch 2 as the system likely isn't powerful enough because 3 TFs is not that many for ray-tracing (especially once we get to 1.5 teraflops in handheld mode). The frequent major cuts in downports will probably be to ray-tracing and they should absolutely cut ray tracing before much more important (for marginal benefit on a 1.5 TF platform) effects such as per object motion blur.

8. Metroid Prime 4 entered the home stretch of development in Spring 2023 after they finished a huge hiring spree and games tend to take like two years to finish after this from what I can tell so I'm guessing Prime 4 is Spring 2025. If the game left pre-production September 2021, this timeline fits fairly normally.

9. The Switch 2 will have a promotional cycle like a normal console and will get a press release announcement 7-12 months before release. The system will launch in late 2024 or early 2025.

10. Due to the extremely weak yen, the Switch 2 will have a region lock to avoid exports limiting their Japanese audience. If it's literally possible to region lock only in one region, I would expect the region lock to only apply to systems sold in Japan.

IDK, these are not contrarian for the sake of it, these all seem fine.

There are Microsoft and Sega takes people are much more mad at me about, of course,
 
Samsung was the value for a foundry_init variable in a header file somewhere but it is possible for it to be a leftover value from Orin since T239's file is identical to T234's.
 
The takes everyone gets mad at me about

1. The Switch 2 is going to be extremely powerful so I'm guessing it will be $499 based on how the Series S seems to still cost a ton of money to manufacture

2. 8nm can't be ruled out because all rumor and hack info lightly point there.

3. GTA6 will push the PS5 up to and probably past its limits and therefore would be incredibly difficult to port to Switch 2 without significant costs and Rockstar thus probably won't bother as Rockstar doesn't bother with things that cost resources that aren't GTA Online updates or the main games themselves.

4. MH Wilds and BG3 will be hard to port to the Switch 2 though both are possible (MH Wilds because Capcom has such a strong interest in bringing MH to the Switch and BG3 because it's not GPU heavy at all so I'm guessing there's some potential to move a lot of CPU functions to the GPU)

5. It's hard to tell if big budget UE5 games will run on Switch 2 as so far, UE5 games with complex assets have not run well

6. Ray reconstruction costs more than denoisers that are good enough in non-PT situations and thus is not viable on Switch 2. It could be viable if NVIDIA made a neural network that was much smaller and thus could run much faster.

7. There will be minimal usages of ray tracing on the Switch 2 as the system likely isn't powerful enough because 3 TFs is not that many for ray-tracing (especially once we get to 1.5 teraflops in handheld mode). The frequent major cuts in downports will probably be to ray-tracing and they should absolutely cut ray tracing before much more important (for marginal benefit on a 1.5 TF platform) effects such as per object motion blur.

8. Metroid Prime 4 entered the home stretch of development in Spring 2023 after they finished a huge hiring spree and games tend to take like two years to finish after this from what I can tell so I'm guessing Prime 4 is Spring 2025. If the game left pre-production September 2021, this timeline fits fairly normally.

9. The Switch 2 will have a promotional cycle like a normal console and will get a press release announcement 7-12 months before release. The system will launch in late 2024 or early 2025.

10. Due to the extremely weak yen, the Switch 2 will have a region lock to avoid exports limiting their Japanese audience. If it's literally possible to region lock only in one region, I would expect the region lock to only apply to systems sold in Japan.

IDK, these are not contrarian for the sake of it, these all seem fine.

There are Microsoft and Sega takes people are much more mad at me about, of course,
problem is every opinion that hint sightly that the new system is not perfect and wont get every next gen without any major optimization get people annoyed, all of those are valid opinions tho so at least is good to have a different side on this thread to explore possibilities
 
0
Samsung was the value for a foundry_init variable in a header file somewhere but it is possible for it to be a leftover value from Orin since T239's file is identical to T234's.

Yes, it could definitely be an error, which is why I said lightly instead of strongly. I'm like 30% 8nm, 70% 4N.
 

If the Switch 2 can run this, even if it's at 1080P I will be quite happy


200w.gif
 
This is the first time I've heard of CAMM2. So it's a memory module that will replace the current modules in laptops (and mobiles) in due time. What are the advantages over the regular modules.. higher capacity RAM?


it's gonna be a lot more expensive than regular modules though. Will take a few years..
CAMM2 is smaller than SODIMM, and CAMM2 allows for modular LPDDR5/5X modules and higher I/O rates (e.g. 8533 MT/s for LPDDR5X vs 6400 MT/s for DDR5 SODIMM), due to CAMM2 having shorter connections between the RAM modules and the RAM controllers.
 
Damn well near everything
Yeah it’s a pretty odd denial on his part. I mean he works so hard at that I would have thought pessimism and contrarianism is something he would be openly proud of.

It have gotten so annoying despite him being on ignored list I’m actually starting to think about ignoring those who regularly engage with him
 
Last edited:
The takes everyone gets mad at me about

1. The Switch 2 is going to be extremely powerful so I'm guessing it will be $499 based on how the Series S seems to still cost a ton of money to manufacture

IDK, these are not contrarian for the sake of it, these all seem fine.

Thinking Nintendo are going to release a Switch 2 at $499 is a pretty bad start considering thats about as likely as the big Switch 2 launch title being "Killer Mario Haunted House Massacre - Bloodthirsty Edition"
 

If the Switch 2 can run this, even if it's at 1080P I will be quite happy


Didn’t know I was going to need to wear sunglasses while watching this. lol.

I think the lighting is a bit too over the top, almost reminding me of the original bloom lighting days of the mid to late aughts.

Otherwise, looks pretty cool.
 
Thinking Nintendo are going to release a Switch 2 at $499 is a pretty bad start considering thats about as likely as the big Switch 2 launch title being "Killer Mario Haunted House Massacre - Bloodthirsty Edition"
It's also not consistent with the realities of manufacturing. T239 may be powerful, yes, but on 4N it would be truly, utterly miniscule. The cost per SOC would be extremely small relative to the performance- because it's LITERALLY physically small relative to its performance! Smaller devices also take fewer materials, and lower power requirements mean smaller, often simpler cooling systems.

Since the most expensive part of a system that contributes to performance is indeed the SOC, we have to ask, would a "powerful" T239 cost extra? Well, it would be on a small node to achieve that kind of performance. But it gets cheaper, not more expensive, on that node. So the answer is a very clear no.

Performance and price are not directly correlated.

Pessimism isn't and never is in and of itself realistic. It's just pessimism. And it's not pleasant to read, in my view.
 
Thinking Nintendo are going to release a Switch 2 at $499 is a pretty bad start considering thats about as likely as the big Switch 2 launch title being "Killer Mario Haunted House Massacre - Bloodthirsty Edition"

It’s hard for me to reconcile

1. The Switch 2 will be extremely powerful, like 40-50% as powerful as the Series S (which probably still costs over $400 to make)

2. The Switch 2 will be a fraction of the Series S’s size and use a small fraction of its electricity.

3. The Switch 2 is either on a much worse node than the Series S or is on a node that is better, but significantly more expensive per cubic cm.

With a $400 Switch 2.
 
It's also not consistent with the realities of manufacturing. T239 may be powerful, yes, but on 4N it would be truly, utterly miniscule. The cost per SOC would be extremely small relative to the performance- because it's LITERALLY physically small relative to its performance! Smaller devices also take fewer materials, and lower power requirements mean smaller, often simpler cooling systems.

Since the most expensive part of a system that contributes to performance is indeed the SOC, we have to ask, would a "powerful" T239 cost extra? Well, it would be on a small node to achieve that kind of performance. But it gets cheaper, not more expensive, on that node. So the answer is a very clear no.

Performance and price are not directly correlated.

Pessimism isn't and never is in and of itself realistic. It's just pessimism. And it's not pleasant to read, in my view.
I don't think the costs of the display and the internal flash storage can be completely discounted, especially when Video Games Chronicle and NateDrake have heard of a ~8" LCD display being used to reduce costs, especially if Nintendo does plan to equip a significant amount of capacity for the internal flash storage.
 
you all know by now how I feel about them cheaping out on the display but over time I've come to accept that it's a sensible strategy. rather, at the very least it is congruent with a highly speculative guess of a sensible strategy:

  • making a more powerful switch is proving difficult for nintendo and nvidia due to power and size constraints
  • the solution is to make a physically larger system to afford more battery volume and cooling hardware
  • this inherently makes the device less appealing as a portable device to the mass market
  • it also inherently makes the device more expensive
  • accordingly costs need to be cut to bring the product down to a preferred price
  • since handheld mode is already less appealing, a cheaper, but bigger for design and marketing purposes, panel is used
  • the cherry on top: the OLED model already set a precedent that can be fulfilled when they manage to get the full package smaller and lighter

for now I'm going to solidify my guess that the new system is coming in calendar Q1 2025 (February or March) and will be followed in holiday 2027 by the concurrent launches of two models: an OLED model that is the same size as, or nominally smaller than, the 2025 model and a Lite that also has an OLED display. I think this is also when we get a new Animal Crossing, a perfect fit for a realistically portable device for smaller hands.
 
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