Darknut85
#TeamRemake
He knows nothing. And they‘ll doom anyway.
-
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!)
- Thread starter Dakhil
- Start date
Vantell
Does several things in the gaming industry
Not really. But I wouldn't worry too much either. Don't compare the Switch succesor with a PS5/Xbox Series/High-End PC. It should be more compared with other handhelds on the market like Steamdeck, ROG Ally etc. Even then, Nintendo knows very well how to push their hardware so the leap from NS1 to NS2 will be big
. I am not really worried it being outdated at all.Not really. The Nintendo Switch got already hinted at as NX because of the situation Nintendo was at. Current Nintendo is not in a situation where they need to say anything. They can just focus on for example the opening of their other parks, or other ways they will be working on their IP's. (Movies, series etc). A console can also be revealed in September or October and then be released in March. This doesn't mean they can't say anything of course, but it not being there next week doesn't mean anything at all.
Nintendo's official denial tweet from 2021 has proven accurate, 2.5 years later. No one has given any evidence of any development kits beyond ones for the retail models released, not pro or succ, and there's lots of reasons to believe Nintendo's own software pipeline was still very focused on Switch games right through to this year.
There's also reason to suspect that Bloomberg just got their wires crossed, maybe asked their contacts the wrong questions ('have you seen any new hardware?" - to which the answer was yes because OLED), got caught up in all the circular reporting from the analysts and bloggers, put 2 + 2 together to make 4k. The thing that has been most apparent is that people are much more happy to trust rumors than official lines, even when the official lines turn out to be completely accurate in hindsight. So that keeps this whole thing going. Even the 'delay' is taken as fact around the internet, despite it being entirely the creation of bloggers who heard one thing, then another. No one knows or has proven that the first thing was ever accurate.
People sometimes say 'but Nintendo denied DS Lite the day before they until announced' but in fact that was Nintendo UK. They just didn't know about DS Lite either. So it's actually more reason to think very, very few people will know about what Nintendo is releasing next, until it's announced.
Aufhebung RPG
History is tragedy becoming comedy
I've been saying for a while now that these bloggers have been giving Nintendo their own "original release date" and then learning about Nintendo's "delays" from sources they don't know if they're credible or not, and what's even more amazing is that the internet chooses to believe these leaks unconditionally!
How does a machine that has never been officially announced have an "original release date" and at the same time be "postponed"?
Last edited:
Stinky Horse
Bob-omb
Dr. Toto says
But avoiding repeating any part of the Wii U schedule might be a good idea.
But they did just that in 2011 with (the then unnamed) Wii U, saying that they'd show it at E3.
But avoiding repeating any part of the Wii U schedule might be a good idea.
Danny
Like Like
- Pronouns
- He/Him
I'm more interested in speculating if Switch 2 has been delayed multiple times. The SoC was finished around 2022? Meaning it will be released 3-4 years later after it was finished. Was that their original plan or did Covid and continued good Switch sales lead to them just sitting on a finished SoC for years?
In the previous investor Q&A this was the topic of the first question:
'It takes a long time and thorough planning to get ready for new hardware, and those plans are not impacted by whatever the latest business conditions might be.' (https://www.nintendo.co.jp/ir/pdf/2024/240207e.pdf)
Danny
Like Like
- Pronouns
- He/Him
So the SoC was finished 2022 but then they started to develop games for new hardware after the SoC was finished. So none of the games are ready for release meaning the Switch 2 is not ready either.
P4bl0
Project Oz H2 2024
No ? You don't need a finalized SoC to develop games.
Yea, I do think this is a strong possibility. I mean Iwata suggested more of a phone like upgrade model for the NX, where hardware updates dropped more frequently and it didn't mean a clean break from the previous gen. Nintendo have obviously abandoned that mindset now, but we don't know when they did.
eye0h
"aggressively persue a multiplatform strategy"
- Pronouns
- He/Him
Except we will because the console will release eventually, there's the silver lining
It is common for game console to have preliminary devkits before final hardware is ready. Early 360 devkits were famously Macs. Even though there isn't evidence, I don't think Orin NVN2 devkits ever existing is that unlikely.
BloodMoon
Boo
- Pronouns
- He/Him
2017 - Launch
2019 - Lite and V2
2021 - OLED
2023 - Nothing
So presumably somewhere after 2021, something didn’t quite workout or plans simply changed. I think most people expected the successor in 2023 for this exact reason. It just made the most sense at the time.
Kise Ryota
My body's ready for SEC 8nm
You know, I never saw what Iwata said in 2014 as a hint to a phone model of releasing more powerful hardwares every couple of years of something like that. The only thing I believe he hinted was the possibility to release a number of devices with different form factors, but even then he said that they could end up needing just one device, as consumers would ultimately decide what would be the best approach.
The smoother transition (for a new gen) would happen because of the strategy of having a common platform/architecture/way of programming.
KarlRemarx
Cappy
I don't think they've abandoned that mindset at all when by all accounts the next hardware generation has backwards compatibility, and they did indeed release 2 different hardware revisions this generation, one of which being handheld-only.
They've literally done what Iwata originally espoused:
To cite a specific case, Apple is able to release smart devices with various form factors one after another because there is one way of programming adopted by all platforms. Apple has a common platform called iOS. Another example is Android. Though there are various models, Android does not face software shortages because there is one common way of programming on the Android platform that works with various models. The point is, Nintendo platforms should be like those two examples. Whether we will ultimately need just one device will be determined by what consumers demand in the future, and that is not something we know at the moment.
No sane platform holder would look at the Switch's performance and think that there is some gap in demand that a mid-gen higher spec model was going to address. You'd simply roll with the gigantic userbase you have and develop for that baseline accordingly, because framerate/resolution drops in some titles simply weren't affecting sales performance. Couple this with COVID and working with the success they had has been a no-brainer, especially compared with how the rest of the industry reacted to COVID.
And, I just want to stress - again - that people are dooming about hardware specs (i.e. being "only" a 2022 SoC or whatever), based on what they think performance will be relative to PS5/XBS, which is a completely incorrect basis for Nintendo to use. Nintendo have just had an unprecedented success that is completely detached from the PS/Xbox paradigm, because both their software and hardware offering is aiming for a different, wider population than these systems. Based on all reported information this system can run current generation game engines, which already puts it in a better spot than Switch 1. This is enough to garner major 3rd party support but at the end of the day, until Sony or Microsoft actually decide they want to compete in the part of the market where Pokemon/Animal Crossing/Mario/Mario Kart etc resides, Nintendo will continue to be extremely successful, regardless of delusions that e.g. a 6-month internal delay suddenly makes T239 obsolete relative to PS5.
The presence of major indies like Silksong and Hades 2 will be far more relevant to Switch 2's fortunes than whether 90% of e.g. EA's output runs well on it. EA have already trained their userbase to be on PS or Xbox (often even to the detriment of their PC customers).
This is just projecting that Nintendo were going to release new hardware every 2 years based on a total of... 2 datapoints. And the bigger leap, that this 3rd point would also have been a hardware upgrade and not simply a different form factor.
Individualised
Bob-omb
Unless you're talking about something else, I believe what you're referring to was Iwata talking about the original plans for the 3DS and Wii U successors pre-NX - where Nintendo would have still released two distinct portable and home consoles, but with a shared architecture. The intention was to make software developers' lives easier by making both systems act more as a single hardware target. This was the original reasoning for merging the software and hardware divisions. There was no need for this when NX happened though, as that system was powerful enough to be both a 3DS and Wii U successor. NX was late 2014, so this is what Iwata was probably referring to when he said he didn't know if Nintendo needed more than one device, as they would have probably already started talks with Nvidia and had doubts about the viability of INDY and the old plan.
TheGreatMightyPoo
Like Like
lol
SoldierDelta
Designated Xenoblade Loremaster
- Pronouns
- he/him
Welcome to the Time Void, we have snacks (vanilla ice cream monster munch packets), drinks (Bloody Mary cocktails with extra tomato juice) and entertainment (Season 4 of BBC Sherlock as recorded on your Grandma's VHS tape). A lifetime of eternal waiting and pain awaits.
Steve
Koopa
- Pronouns
- he/him
If the Switch 2 isn't acknowledged in their investor meeting at 7th of May, it must mean we'll be hearing from their next investor meeting in the holiday.
But thankfully we aren't arguing over linking accounts... Hopefully not in the future
Aufhebung RPG
History is tragedy becoming comedy
I was probably able to get through Mcluhan's book before switch2 was officially announced, and it largely deepened my understanding of the medium's subject matter.
YugamiSekai
T239 vs. Picross 4K
- Pronouns
- He/Him
Just got my first personal 4K TV yesterday, so...
...Ready whenever you are, Nintendo.
They'll never be ready, won't they
...Ready whenever you are, Nintendo.
They'll never be ready, won't they
KeijiKG
Moblin
I think it'll just be the same "we'll talk when we're ready" or "if we do it, it's more than a hardware refresh" response, focusing only on the current product and how business is going now.If the Switch 2 isn't acknowledged in their investor meeting at 7th of May, it must mean we'll be hearing from their next investor meeting in the holiday.
Doesn't mean a reveal or an event teasing the reveal won't happen later in the year, just that putting stock of it happening in the investor's meeting might be a little overexpectant.
Switcharoo
Piranha Plant
They may not need to say anything, but having it acknowledged for the next FY does mean they have preparations regarding new generation hardware, which takes a lot of investment. NX wasn't the only thing acknowledged in advance. Other Nintendo hardware have been acknowledged as well, such as Revolution and even the Wii u. They could do things different this time, but it's also about marketing the software. Also, for many investors having movies and other parks, is a very predictable thing and not that surprising. I believe even @ShareholderChad had said something along the lines of this in the past.
Not saying they can't do a short announcement to release cycle, they definitely can, which is also what I am leaning towards
Last edited:
The times where they done it before, their current platform did not still have sales momentum. That's the difference.
Switcharoo
Piranha Plant
Gameboy I believe did have a pretty good momentum before DS got announced, if I recall correctly. Also, momentum for Switch has been on decline YoY. But then again, Information now travels faster. This isn't the days of gaming magazines and G4 news. So there is absolutely an argument for short marketing cycle.
Last edited:
Shoulder
Like Like
Correct. You could, however, alter the quantity of Ram by increasing it further, provided it still maintains the same bit bus, and type of Ram used. The common thread here is 12GB of LPDDR5, possibly LPDDR5x given the efficiency improvements over the former. But Nintendo could bump that up to 16GB provided the price is right.
You could also improve other areas of the system with its internal storage for games, maybe a better HD rumble, or an improved gyro sensor for better aiming. Heck, even the screen could be updated.
But the System-On-Chip which houses the CPU and GPU, that is set in stone.
eye0h
"aggressively persue a multiplatform strategy"
- Pronouns
- He/Him
You can make the image smaller before posting btw.
I think there are rumours that 8 and 9 were developed at the same time so I'd imagine they aren't doing anything too crazy that the switch 2 couldn't handle but that's assuming they didn't vastly increase the scope of the project.
Steve
Koopa
- Pronouns
- he/him
Makes sense, because it started development in 2017 and is stated to be open world and the most ambitious RE game to date.
But I’m curious if Capcom would like to develop for the Switch 2, since I’m guessing they already have the Dev-Kit.
But Monster Hunter is pretty much a get go to be early for the Switch 2, since Nintendo has a pretty sizeable install base of MH fans.
eye0h
"aggressively persue a multiplatform strategy"
- Pronouns
- He/Him
They're pretty much all in on multiplatform and doing quite well. I think they'd attempt stuff for Monster Hunter and depending on how successful there might be a push to do more, especially because they are trying to develop a lot of new IP like pragmata and that new game 'The Descendants' that just got leaked. I don't think every game will reach the system though. I don't see something like dragons dogma 2 for example but I could be wrong!
Not many people expect Series S much less Series S with DLSS and better raytracing. Definitely not enough to be a camp.
Then again, few here that believe the leak is thinking Steam Deck or PS4 level either with 12SM at current Switch clock easily surpassing them.
Most here are in between I feel. So it's mostly speculation where it lands since there is a big gap between Series S and PS4.
I personally would be happy with PS4 Pro level with much faster CPU, DLSS, more RAM, more updated GPU, and faster loading.
Concernt
Optimism is non-negotiable
- Pronouns
- She/Her
Sorry, but isn't this a little contradictory?
To repeat myself a little; PS4 Pro level raw performance with a faster CPU and a modern architecture IS a Series S. Series S and PS4 Pro are in the same raw performance bracket.
That puts you firmly in the hoping for Series S camp.
Now as I've said, these camps don't really "work" to describe a system with multiple performance profiles out the box, but my own camp is very much "won't be as fast as a Series S for TDP reasons", but with, of course, better upscaling and RT, for TV mode anyway.
Maybe I'm mistaken but I remember reading that there's a bigger gap between PS4 Pro and Series S. Googling gives Pro TFLOP above Series S but I remembered reading that the way TFLOP number are calculated are not comparable. Also, camp 1 is stated to be Series S before the DLSS. But yes, I overestimate the difference between Pro and S I completely agree with your expectation.
comparing GCN 2.0 with RDNA 1.75 means that they're not comparable, yea. FLOPS are FLOPS, but that's just goes to show why FLOPS aren't end-all-be-all measures of performance. like comparing a tractor to a sports car, they have the same horsepower, but their performances are very different
Like has been explained in detail by better people than me, these comparisons to these other platforms are imperfect because Drake is RTX, which has inherently different tradeoffs than competing RDNA 2 platforms. It can be simultaneously more powerful and less powerful, depending on the workload. It has more focus on RT and scaling tech, while having less raw raster grunt.
I really like Oldpucks example, where Drake can acheive better image quality and effects compared to SD, but turn the settings way down on both and RDNA2s raster advantage will give it a huge frame rate win. So we can end up with a scenario where Drake has a better fidelity mode compared to other machines, but can't hold a candle in performance mode.
oldpuck
Like Like
- Pronouns
- he/they
Thraktor
"[✄]. [✄]. [✄]. [✄]." -Microsoft
- Pronouns
- He/Him
The last tweet is the most important one. Nintendo have zero obligations to say anything to investors. They have to provide their accounts for the last year, and financial forecasts for the next year (as far as I know even these forecasts aren't a legal requirement in many places, but are generally expected), but they have no obligations to talk about specific products or unit sales expectations. Nintendo is actually very unusual in both reporting and forecasting unit sales, and it's not something that any of their competitors do.
Serkan Toto also glosses over a pretty important point: Nintendo didn't start forecasting Switch sales until after it released. Here's Nintendo's financial report from May 2016, with forecasts for the financial year including Switch's launch. You'll find a couple of mentions of "NX", as it had already been announced, but they didn't forecast unit sales for it.
Nintendo can (and I expect will) do the same for Switch 2, simply not provide forecasts for it. They would have to include revenues/costs in their overall financial forecast for the financial year, but with it likely appearing right at the end of the FY it wouldn't have a huge impact on those numbers.
for more on mesh shaders, ironically, look at UE5's Nanite. it pretty much does the same thing, but in compute. until the triangles are large enough that Nanite uses mesh shaders instead. but, effectively, they're the same
then comes the question of "why do one when I can do the other". as I mentioned, Nanite already does both! it also highlights SW and HW strengths. when triangles and clusters get super small (think pixel or subpixel in size), hardware rasterizers get overburdened and compute rasterizers become much faster. but when you have large chunky polygons, hardware rasterizers are faster. there was a research paper that I unfortunately can't find anymore that had some pretty charts showing this
mesh shaders do seem to be picking up steam as we're seeing adoption by some big studios like Ubisoft Massive (Avatar and Star Wars Outlaws), Remedy (Alan Wake 2), and Virtuos (a lot of games)
here's Virtuos's presentation on it from VGC
gdcvault.com
here's AMD's blog post showing delicious sexy mesh shader grass
gpuopen.com
then comes the question of "why do one when I can do the other". as I mentioned, Nanite already does both! it also highlights SW and HW strengths. when triangles and clusters get super small (think pixel or subpixel in size), hardware rasterizers get overburdened and compute rasterizers become much faster. but when you have large chunky polygons, hardware rasterizers are faster. there was a research paper that I unfortunately can't find anymore that had some pretty charts showing this
mesh shaders do seem to be picking up steam as we're seeing adoption by some big studios like Ubisoft Massive (Avatar and Star Wars Outlaws), Remedy (Alan Wake 2), and Virtuos (a lot of games)
here's Virtuos's presentation on it from VGC
Maximizing Graphics Performance with Flexible Virtualized Geometry (Presented by VIRTUOS)
A technical walkthrough on the development of a virtualized geometry technology that can be integrated with any game engine, and an overview of the challenges faced. nThe talk explores how Virtuos achieved optimal performance with ultra...
gdcvault.com
here's AMD's blog post showing delicious sexy mesh shader grass
Procedural grass rendering - Mesh shaders on AMD RDNA™ graphics cards
The fourth post in our mesh shaders series takes a look at the specific example of rendering detailed vegetation.
gpuopen.com
quelseydilla
zoom?
- Pronouns
- she/they
Targeted dates being internal don't mean they don't exist. Nintendo undoubtedly has internal and external teams/partners who were given an anticipated release date, or at least anticipated time frames. Shifting those to later dates can still be a delay/postponement, even if that isn't consumer-facing.
- Pronouns
- he / him
Even if it's being run fast enough to be very comparable in direct GPU performance, I'd still expect games (that aren't appearing on PS4) to run into real-world bottlenecks due to things like CPU and memory bandwidth on Switch 2 first, enough for me to not think "they're the same picture".
Concernt
Optimism is non-negotiable
- Pronouns
- She/Her
I agree in large part. Though the volume of Drake at that point was very low, I would suspect. Parties outside of Nintendo and Nvidia maybe could have gotten some samples, but I suspect the 4K development environment for Nintendo platforms, to be as broad as possible, was and is still largely software, with dev kits in as few hands as they can reasonably allow without hurting the release schedule. That's pure speculation on my part mind.
More saliently though, Nvidia had a major breach we learned a lot from. The SOCs they planned and produced suitable for a hybrid start and end with the Tegra X1(+) and T239. If a "Pro" existed, I stand by my view that it was one possible outcome of developing OLED Model, and with the amount of '4K' hardware and software implemented in it, it may have gotten very far like this. In the end they chose against that approach. It was the first hybrid designed around the Tegra X1+ processor, it absolutely could have targeted a higher performance profile than the original Nintendo Switch, but didn't.
But that means the "Pro" that could have been, using Tegra X1+, was on track for 2021, not 2022 or 2023. By the end of 2022, T239 was very likely already being produced in some capacity.
that's why I don't think there was ever a "delay". expected release periods are something that goes out to all partners as they all need to be ready at the same time. probably one of the safest things to leak. but we never got a leak of the initial release timing like we did this delay
Concernt
Optimism is non-negotiable
- Pronouns
- She/Her
I think that's true to some degree, but I have to question how many current generation games are really stressing out the Series S' CPU.
My idea of "what will NG Switch games look like?" has remained pretty consistent, I expect lower scene complexity in general next to the home consoles, but after upscaling, probably higher resolutions than Series S, and I think that's fair. There's also the question of to what degree CPU tasks on Series S could be reorganised on NG Switch, and then it comes down to "how much are developers willing to optimise".
lattjeful
Like Like
- Pronouns
- he/him
If it means more cool educational posts from oldpuck, Thraktor, and LiC, then that won’t be a bad thing.
Did some calculating regarding the general power level of the Switch 2 a few days back and wrote a pretty long post, but because everyone was still busy posting about the recent Pyoro tweets and the following ESRB rating leak it ended up sitting in my draft for days. So now that the situation has calmed down i feel this is the right time for it. But i'm rewriting it all due to how unsatisfied i was with the first draft. (And because i hate myself.)
It's just some food for thought, i'm not that tech savvy tbh. but always interested into this kind of stuff.
So please don't think too much into what i'm about to write here, however wrong it may be.
I think a lot of it comes down to THE 1 big factor: The damn process node.
TSMC 4N would allow it to come within swinging distance of Series S.
Samsung 8nm would essentially turn it into a dockable plug&play Steam Deck that uses DLSS to match or exceed it.
I will focus on the worst case, which would be Samsung 8nm, because i believe it's important to set realistic expectations.
Remember what happened with the Switch 1 and people believing Tegra X2 would be used? Or believing that the X1 on Switch would closely match Nvidia Shield TV clock speeds but instead it ended up being heavily downclocked? That's why.
Now this might be a hot take, but i believe that even at its absolute worst possible scenario with Samsung 8nm being the node of choice and Nintendo clocking this device to Ampere's absolute lowest possible clock speed of just 420MHz, it will still qualify as a huge generational leap in my book and may be Nintendos biggest one since the Gamecube. Allow me to explain:
First let's take a look at portable mode:
Going from Switch 1's GPU performance of 307.2MHz in portable mode to Switch 2's 420MHz might sound like a very minimal increase, but since the GPU is so much bigger this time, it is still an 8.2x increase from a measly 157.28GF to 1.29TF. It's already matching Xbox One's raw GPU performance of 1.31TF in portable mode alone.
And this is ignoring the fact that Switch 2 has the much more modern and efficient Ampere architecture, which also provides access to the best in class upscaler and RT hardware, and additionally having its own file decompression engine and much faster storage transfer speed.
Considering all of this, it actually comes much closer to a Steam Deck than any last gen home console. Switch 2 can just render at 540p and let DLSS upscale to 1080p, whereas the Deck needs to either render its full res (720-800p) or destroy its image quality via FSR to get acceptable framerates or battery life.
Even if the Steam Deck may be theoretically more powerful, the Switch 2 will punch above its weight thanks to its better feature-set and having native ports.
And since Nintendo is so keen on keeping the battery life high and offering a true plug&play style (which the Deck fails at), it makes it the more attractive choice if we ignore exclusive games and such.
And this is ignoring the fact that Switch 2 has the much more modern and efficient Ampere architecture, which also provides access to the best in class upscaler and RT hardware, and additionally having its own file decompression engine and much faster storage transfer speed.
Considering all of this, it actually comes much closer to a Steam Deck than any last gen home console. Switch 2 can just render at 540p and let DLSS upscale to 1080p, whereas the Deck needs to either render its full res (720-800p) or destroy its image quality via FSR to get acceptable framerates or battery life.
Even if the Steam Deck may be theoretically more powerful, the Switch 2 will punch above its weight thanks to its better feature-set and having native ports.
And since Nintendo is so keen on keeping the battery life high and offering a true plug&play style (which the Deck fails at), it makes it the more attractive choice if we ignore exclusive games and such.
Now let me do some more calculating for possible docked GPU clock speeds:
Doubling the GPU clock speed from 420MHz to 840MHz when docked we jump up to 2.58TF and we're already surpassing PS4 & Steam Deck GPU perf by a huge margin BEFORE applying DLSS. With DLSS enabled we theoretically are at PS4 Pro levels of GPU power.
I believe docked will end up a bit higher than just doubling from portable mode, since the Switch 1's portable GPU clock speed of 307.2MHz is actually just 40% of the full docked 768MHz GPU clock.
So calculating from there i think 1050MHz docked is a real possibility since 420MHz is just 40% of that.
It would put the Switch 2 GPU at 3.22TF when docked. Coming quite close to Series S already, so imagine what we'd get if calculating with TSMC 4N.
So in general i wouldn't worry too much about docked mode this time. The portable mode is already such a big upgrade from Switch 1 that even a 540p game DLSS upscaling to 1080p should make pretty good looking docked experience by simply using the higher clocks to DLSS upscale from higher internal resolutions like 720p to 1440p or even 900p to 1800p if possible.
I'm not expecting to see many games do 4K either natively or via DLSS upscaling, since the cost in render time is simply too high for that.
Maybe with Switch 3.
I believe docked will end up a bit higher than just doubling from portable mode, since the Switch 1's portable GPU clock speed of 307.2MHz is actually just 40% of the full docked 768MHz GPU clock.
So calculating from there i think 1050MHz docked is a real possibility since 420MHz is just 40% of that.
It would put the Switch 2 GPU at 3.22TF when docked. Coming quite close to Series S already, so imagine what we'd get if calculating with TSMC 4N.
So in general i wouldn't worry too much about docked mode this time. The portable mode is already such a big upgrade from Switch 1 that even a 540p game DLSS upscaling to 1080p should make pretty good looking docked experience by simply using the higher clocks to DLSS upscale from higher internal resolutions like 720p to 1440p or even 900p to 1800p if possible.
I'm not expecting to see many games do 4K either natively or via DLSS upscaling, since the cost in render time is simply too high for that.
Maybe with Switch 3.
What about the CPU?
This is one of the biggest question marks yet. We have a good idea what CPU they went with as there's really only one possible choice with an ARM based 8-core cluster, that being the Cortex A78C, but what about the clock speed? Truth be told, i really don't know how much the process node can affect CPU clock speed, but the A78C cores will already be a huge upgrade even without increasing the clock speeds by much since they're again much more modern & efficient. And let's not forget the Switch 2 will have access to 4 more cores this time too.
Since the AMD Jaguar Cores on PS4/XBO were so terrible the Switch 2 can just blast through any last gen game no problem.
The Steam Deck might have a higher CPU clock, but it only has a 4C/8T CPU, while the Switch 2 will have an 8C/8T CPU.
Given how CPU intensive some games already get on PS5/XSX, i think the Switch 2's CPU setup is more future proof than the Deck.
Switch 2 will likely be a close match for the position the Switch 1 was CPU-wise compared to its competition. Ports of current gen games that have no last gen version will require largely the same amount of optimization to get them working at acceptable levels on Switch 2.
In theory the Deck may have better CPU performance, however it has to rely on a translation layer to play standard PC games, whereas the Switch 2 will get native ports that utilize the hardware to its fullest potential. I'd say they're pretty evenly matched here.
But Switch 2 will never match Xbox Series S in raw CPU performance though, that's for sure.
I can only see the Switch 2 surpassing the Series S in things like resolution thanks to DLSS and raytracing, but as for the overall asset quality i think we will see slight downgrades across the board and frequent halfed framerates for games that are 60fps or offer it via an option on competing hardware.
Since the AMD Jaguar Cores on PS4/XBO were so terrible the Switch 2 can just blast through any last gen game no problem.
The Steam Deck might have a higher CPU clock, but it only has a 4C/8T CPU, while the Switch 2 will have an 8C/8T CPU.
Given how CPU intensive some games already get on PS5/XSX, i think the Switch 2's CPU setup is more future proof than the Deck.
Switch 2 will likely be a close match for the position the Switch 1 was CPU-wise compared to its competition. Ports of current gen games that have no last gen version will require largely the same amount of optimization to get them working at acceptable levels on Switch 2.
In theory the Deck may have better CPU performance, however it has to rely on a translation layer to play standard PC games, whereas the Switch 2 will get native ports that utilize the hardware to its fullest potential. I'd say they're pretty evenly matched here.
But Switch 2 will never match Xbox Series S in raw CPU performance though, that's for sure.
I can only see the Switch 2 surpassing the Series S in things like resolution thanks to DLSS and raytracing, but as for the overall asset quality i think we will see slight downgrades across the board and frequent halfed framerates for games that are 60fps or offer it via an option on competing hardware.
And the RAM situation?
That's a bit tricky. Nintendo COULD've seen 8GB as enough initially, i recall reading somewhere they wanted to settle for just 2GB for the Switch 1 but were pushed for at least 4GB, so if this was true and history repeats itself, we could see 16GB as a possibility.
But the pessimist in me says it will be 12GB for the consumer model and 16GB for the dev-kit.
IIRC the Switch OLED dev-kit has a bit more RAM, totalling at 6GB to be exact, so yeah, make of that what you will.
I might have said a year or so earlier that 8GB could be enough because HorizonOS doesn't eat as much RAM as Playstation's or Xbox's OS do, but given how the industry has evolved, i certainly can no longer say the same with confidence anymore. It would be a major blow if 8GB would actually end up happening.
But the pessimist in me says it will be 12GB for the consumer model and 16GB for the dev-kit.
IIRC the Switch OLED dev-kit has a bit more RAM, totalling at 6GB to be exact, so yeah, make of that what you will.
I might have said a year or so earlier that 8GB could be enough because HorizonOS doesn't eat as much RAM as Playstation's or Xbox's OS do, but given how the industry has evolved, i certainly can no longer say the same with confidence anymore. It would be a major blow if 8GB would actually end up happening.
All in all Switch 2 is in a much better position than Switch 1 because it isn't using an off the shelf product such as the Tegra X1. The T239 is a custom designed and manufactured SoC. Nintendo will have settled with Nvidia for the best possible price/performance/efficiency ratio they can possible achieve with this SoC. Whether it will be TSMC 4N or Samsung 8nm we can't know until someone actually tears it apart. (Or a funcle leaks it)
We as a consumer can only vote with our wallets whether the device will be good enough for us or not, and that's a very important thing to remember. The only reason the Switch 1 got so successful with third parties despite being massively underpowered compared to the competition is because so many people were buying it. If you have a huge install base the 3rd parties will want a piece of that cake too, resulting in the hardware receiving as much support as possible, as seen with Switch 1.
Feel free to let me know about potential errors so i can correct them if necessary.
tl;dr: I'm between camp 1 and camp 2.
Okay i'm getting sick from re-reading this so i'm gonna post it NOW!
Last edited:
Danny
Like Like
- Pronouns
- He/Him
But if production on Switch 2 is starting this fiscal year we should be able to see that from the increased R&D spending from their report right? If Nintendo forecast a huge jump in their R&D spending then we can use that to gauge Switch 2 release timings as well.
The big reason for the high RAM bandwidth is because it still relied on traditional rendering at the time, where rendering was done in the main RAM itself. This differs from the Switch which used tile-based rendering, holding tiles being rendered in cache until they were finished before flushing them to main RAM. So while things like textures would be cached for each system, the PS4 would still have to interact with the buffers in main RAM for each pixel with reads and writes, but the Switch did not until the very end. The Xbox One did something similar to Switch through the use of its eSRAM (and Xbox 360 with eDRAM), but the overall bandwidth wasn't enough to match what PS4 had. Combined with the weaker GPU, it rendered the same games as PS4, but at a lower resolution.
The CPU only held back the frame rate, pushing many games into the 30fps range, but planning with 30fps in mind meant double the frame time vs 60fps, and devs would plan accordingly, and could push the GPU that much more per frame because of the available time. With the slow HDD, most of the time it was during loading screens or planned script events, so it had no real affect on bandwidth.
Stinky Horse
Bob-omb
Last edited:
