• 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!)

So essentially expect something along the lines of a Steam Deck but with further capabilities for tech like RT, hopefully the battery life fares better than the Deck and ESPECIALLY the ROG Ally. (What even happened to that thing? Feels like it vanished after a week.)
better battery will be a given at least.

the ROG Ally is still selling. it just doesn't have the clout that being from valve brings, and it (like the Steam Deck) is a niche product
 
So essentially expect something along the lines of a Steam Deck but with further capabilities for tech like RT, hopefully the battery life fares better than the Deck and ESPECIALLY the ROG Ally. (What even happened to that thing? Feels like it vanished after a week.)
With everything kept in mind, that's a good expectation yeah. Just remember SD is a handheld PC and thus lacks specific optimization for games, so Switch 2 should come ahead in nearly every scenario.
 
better battery will be a given at least.

the ROG Ally is still selling. it just doesn't have the clout that being from valve brings, and it (like the Steam Deck) is a niche product
I mean, I'm still excited at the idea of a handheld console with RT.

To think in the space of two generations, Nintendo handhelds go from 240p to 4K. (Even if only when connected to an external display.)
 
I mean, I'm still excited at the idea of a handheld console with RT.

To think in the space of two generations, Nintendo handhelds go from 240p to 4K. (Even if only when connected to an external display.)
In the grand scheme of things... That's incredibly exciting yeah, but RT itself is far from being a big deal considering the designer of the hardware. Home consoles running AMD is what keeps it on this "impossible" light.
 
With everything kept in mind, that's a good expectation yeah. Just remember SD is a handheld PC and thus lacks specific optimization for games, so Switch 2 should come ahead in nearly every scenario.
I wouldn’t be surprised if this does end up being better than the first Steam Deck for gaming (Besides games not on the Switch.) on account of ports being specifically designed for the Switch 2 in mind, instead of relying on pure specs. Steam Deck 2 will likely vastly outperform but it won’t matter to the average consumer. Hopefully games like Elden Ring come to Switch 2, could see that one being the Skyrim moment for the reveal.
Cyberpunk 2077 won’t get ported cos that game struggles to run on many devices, including the Steam Deck lol.
 
I wouldn’t be surprised if this does end up being better than the first Steam Deck for gaming (Besides games not on the Switch.) on account of ports being specifically designed for the Switch 2 in mind, instead of relying on pure specs. Steam Deck 2 will likely vastly outperform but it won’t matter to the average consumer. Hopefully games like Elden Ring come to Switch 2, could see that one being the Skyrim moment for the reveal.
Cyberpunk 2077 won’t get ported cos that game struggles to run on many devices, including the Steam Deck lol.
The RT cores and some DLSS should help when porting that one, don't expect it with the Overdrive setting or anything crazy though, lol.
 
In the grand scheme of things... That's incredibly exciting yeah, but RT itself is far from being a big deal considering the designer of the hardware. Home consoles running AMD is what keeps it on this "impossible" light.
Side note, what is it with AMD that puts a mindset of RT being super difficult for consoles?
 
Side note, what is it with AMD that puts a mindset of RT being super difficult for consoles?
AMD seemingly didn't believe in RT when the call was put forward for DX12. they knew it was coming and that MS and Sony wanted it, but either they had no research in the best method or they just didn't want to spend silicon on acceleration
 
Doesn't really matter if a Steam Deck 2 or other future PC handheld outperforms it cause no one's coding native ports for those, no publisher is deciding between a SD2 port vs. a Switch 2 port, because they're all PCs and you can brute force any PC game you want on them, at whatever battery life you're willing to stomach. And we know they're not real competition.
 
Side note, what is it with AMD that puts a mindset of RT being super difficult for consoles?
AMD hardware doesn't actually have silicon on the die meant for raytracing, so it needs to brute force it through GPU cores. Oldpuck could put together a better summary, but that's pretty much it.
 
AMD hardware doesn't actually have silicon on the die meant for raytracing, so it needs to brute force it through GPU cores. Oldpuck could put together a better summary, but that's pretty much it.
Whats the RT core count for Switch 2 again? As well as what’s generally considered a good count for that
 
.
Doesn't really matter if a Steam Deck 2 or other future PC handheld outperforms it cause no one's coding native ports for those, no publisher is deciding between a SD2 port vs. a Switch 2 port, because they're all PCs and you can brute force any PC game you want on them, at whatever battery life you're willing to stomach. And we know they're not real competition.

I mean, the narrative a few months ago (and probably still circulating for many) was whether or not this device would match Steam Deck. I’d seen a lot of pessimism “because Nintendo.”

I genuinely didn’t think we’d see any “will it beat SD2?” discussions. I don’t care if it does or not. Giving a better experience than Deck (1) is going to make plenty of people happy already.
 
The issue with all these PC handhelds? The controls can be annoying and setting stuff up.. I love my Rog Ally

Switch is just so much easier and Stardew Valley plays great on Switch. Switch 2 day one.
 
Whats the RT core count for Switch 2 again? As well as what’s generally considered a good count for that
well cores aren't the limiting factor, it's your shader core performance. a big chunk of the render time is still shading pixels and building the BVH. if you need to rebuild/refit, then you have to use CPU power. and once that's done and you traverse the tree, you have to do all the programmed tasks for the pixel based on the data from the hit
 
0
I don’t care if it does or not. Giving a better experience than Deck (1) is going to make plenty of people happy already.
This is my point. It doesn't matter in a practical sense of whether or not this device can produce good looking games or receive plenty of current-gen third-party ports because we are expecting a sufficiently powerful hybrid device anyway. I understand some people want to participate in spec wars and I await the videos about the SD2 being a Switch 2 killer. But I can't think of a good reason to legitimately be concerned how the Switch 2 measures up to future handhelds. It's the usual 'console vs. PC' debate where convenience ends up being a major decisive factor, and the Switch's appeal is mostly its convenience.
 
.


I mean, the narrative a few months ago (and probably still circulating for many) was whether or not this device would match Steam Deck. I’d seen a lot of pessimism “because Nintendo.”

I genuinely didn’t think we’d see any “will it beat SD2?” discussions. I don’t care if it does or not. Giving a better experience than Deck (1) is going to make plenty of people happy already.
All these PC Handheld are just PC, Switch 2 is a console and devs will optimize special for it
 
I still think that if they do not stick with SD card, Nintendo will go with UFS. I believe there is already 2 lanes for UFS on the Orin board. It would be perfect if it was carried over the T239. One for internal storage and one for external. They probably using it for internal so it is simple to support it for external. It is designed to be energy efficient and thus suit the Switch perfectly unlike CFe and SDe which consume too much power at their rated speed. If they want backward compatibility, dual SD/UFS readers already exist though I imagined it will have to be updated to support the new standard.

Even if there haven't been recent development on the technology, 1.2gb/s is plenty. Especially since they will have to run CFe at a much lower speed to fit the power requirement so real world speed comparison will be much closer. And like CFe, it's essentially a plastic case around a much more actively developed standard. It would be simple to create a new card standard based on the current 4.0 standard for internal UFS.

No market share is a huge problem but Nintendo will have to kickstart the market either way whether they will be 100% or 90% of the market. It wouldn't be the first time either since I believed they were the biggest one to support SD card in the beginning. I don't know how many manufacturers there are for UFS but it's going by the size of the mobile phone market, there are plenty capable of adding a new product line.

Size is a big consideration for Nintendo. Putting in a CFe reader is basically another cartridge slot inside the Switch and there isn't any room for that. Viewing the internal, there's no comparison between the space for the microSD and the cartridge slot. Not to mentioned that CFe Type A is even bigger than a cartridge. Reports are hinting at a bigger Switch but not so much bigger that it will be an easy design. PSP did it by moving the reader away from the cramp center and into the control area. Joy Con are already cramp and detachable so that wouldn't work here. UFS on the other hand are already microSD sized as standard.

Also, Nintendo seems pretty averse to paying a license fee. They are a member of the SD card association so they won't have to pay a license for SDe but that standard have multiple problems. I don't know how much it cost to support CFe but I can imagine it being a deterrent for Nintendo. UFS is an open standard so there would be no license I believed.
 
TFLOPS don't compare well across architectures, unfortunately. In head to head matchups, Ampere fares worse than RDNA1/2 at similar TFLOPS. So we have to be careful about getting out over our skis on those comparisons.
Yes yes, I know that, and I agree that it is not very wise to compare two different architectures. Although I believe in a greater balance due to the PS5 not using a "full" RDNA 2 and Drake being more of a hybrid between Ampere and Ada.
 
This is my point. It doesn't matter in a practical sense of whether or not this device can produce good looking games or receive plenty of current-gen third-party ports because we are expecting a sufficiently powerful hybrid device anyway. I understand some people want to participate in spec wars and I await the videos about the SD2 being a Switch 2 killer. But I can't think of a good reason to legitimately be concerned how the Switch 2 measures up to future handhelds. It's the usual 'console vs. PC' debate where convenience ends up being a major decisive factor, and the Switch's appeal is mostly its convenience.
Even past all that and everything regarding hardware specs, Switch 2's power ballpark is everything Nintendo will ever need to produce stunning games in the foreseeable future and perhaps a little more. Until their HQ expansion takes place around 2028, Nintendo is unlikely to produce anything significantly more impressive than the very last PS4 games from a technical standpoint (if they even do). As long as this console is indeed everything the Drake chip claims to be, hardware power won't really be a limitation to their studios anymore,
 
Last edited:
0
Except in that case, the system would be docked, negating any power issues. That said, I do believe in some videos on YT, folks have tested power usage of gaming PCs with native resolution, and DLSS upscaling. I do believe the latter uses less power, but if I recall correctly, does not save a massive amount of power. Nothing is free of course as the tensor cores would be more utilized in this case. So using DLSS even in handheld mode I think would still result in power savings compared to native resolution.

Though, some of the other smarter Famis out there can go into more details on that.



You need to find yourself a good sipping tequila if you haven’t tried already. As a Bourbon drinker primarily, I quite like the profile of a good quality tequila that you can sit on the couch, and enjoy. My wife and I still have this bottle that looks like a flower vase, though can't remember the brand specifically. It is shockingly good imo.
Not gonna lie I completely forgot about docked mode smh, you're completely right.
As for the tequila I'll definitely have to try and find another brand I'll keep an eye out for a flower bottle!
 
Bit of both. A few key points that are new and will be of interest.
giphy.gif
 
Its actually really cool to see the gap between handhelds and consoles close in the field of specs.

In terms of raw power, the rog ally and upcoming legion go nearly have as much horse power as the PLAYSTATION 5!!! (an insane 8.6 tflops for both of them, yes i know tflops aren't everything)
even considering the switch 2, the gap between that and current gen consoles is noticeably smaller than the gap between switch and ps4/xbox one

I'm predicting around 2025/2026 handhelds and consoles will reach parity, we'll start seeing handheld PCs that are as if not more powerful than ps5/Series x. And yeah, the price of those portable pcs are nothing to laugh at, but it just goes to show how fastly we're technically advancing and its a fun to witness the rapid evolution of mobile tech.
 
Nate's podcast is going to be released tomorrow. Hoping there will be some extra tidbit to hear even though he did say his contact didn't seem too happy when Eurogamer broke the news (a little) too soon.
 
0
Its actually really cool to see the gap between handhelds and consoles close in the field of specs.

In terms of raw power, the rog ally and upcoming legion go nearly have as much horse power as the PLAYSTATION 5!!! (an insane 8.6 tflops for both of them, yes i know tflops aren't everything)
even considering the switch 2, the gap between that and current gen consoles is noticeably smaller than the gap between switch and ps4/xbox one

I'm predicting around 2025/2026 handhelds and consoles will reach parity, we'll start seeing handheld PCs that are as if not more powerful than ps5/Series x. And yeah, the price of those portable pcs are nothing to laugh at, but it just goes to show how fastly we're technically advancing and its a fun to witness the rapid evolution of mobile tech.
To clarify, the ROG Ally/Legion Go do not have the horsepower of the PS5.

Their 8+ TFLOP figures are incredibly irrepresentative due to how they are achieving it (so much so TechPowerUp retracted the TFLOP numbers back and redid them).

RDNA3 for most purposes have the same TFLOP calculations as RDNA2, this is doubled for some effective applications if a game can utilize Dual-Issue. The main benefit for Dual-Issue is RT Performance but calling it the same perf as PS5 is highly irrepresentative.

The ROG Ally/Legion Go/Any 7840U handheld is roughly in the 3-5TFLOP RDNA2 Range for most standard operations. And even then there's things like the Z1's Shaders getting power starved which heavily limit it in some scenarios versus the 7840U's GPU despite both having the same TFLOP value
 
ROG Ally's also hilariously bandwidth starved for what it is. Wild that Asus went with regular 6400 MT/s LPDDR5 when Phoenix officially supports up to 7500 MT/s LPDDR5X.

Edit: Right, speaking of memory bandwidth for mobile, LPDDR6 should be expected in about 2026? But that's based off of a roadmap Samsung presented at their Memory Tech Day last October. For this year, Samsung Memory Tech Day is on October 20.
 
So the Switch 2 will be the first time we see new Nvidia graphics tech (RT, Tensor, etc) on a dedicated, closed system and not a PC.

We know that developers can typically squeeze more performance out of a console relative to a PC of equivalent specs due to the fact that they can get closer to the hardware and know exactly what they're working with.

Does this apply to DLSS with Tensor Cores and RT performance on RT cores? For example, 12 RT cores doesn't sound like a lot - but since developers know exactly what they're working with, could they squeeze more performance out of those RT cores such that those 12 RT cores could match the performance of, say, a 3050 in terms of RT performance (which has 20 RT cores)? And could devs push the Tensor Cores to punch above their weight given that they know they have 48 to work with?
 
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