Kevin
Chain Chomp
Seems unlikely.
because it's happening in 2022
Ha ha I like your optimism!
Seems unlikely.
because it's happening in 2022
The eShop not being a native app is most likely a very intentional decision. There are some very real maintenance and future proofing benefits of doing it this way500MB is the absolute minimum, but itās more like 750MB-1GB for the OS and 1 core.
Switch eShop is so slow because itās not a native app, maybe it all be a native app on the succ/2/pro with enough RAM.
Because itās not a native app wouldnāt stronger hardware still mean better performance?500MB is the absolute minimum, but itās more like 750MB-1GB for the OS and 1 core.
Switch eShop is so slow because itās not a native app, maybe it all be a native app on the succ/2/pro with enough RAM.
It is an intentional design, it would have caused more bloat and slowdown which they wanted to avoid. They did want lean and minimalism after allā¦.. eShop being one of the āpaths of least resistanceā where it doesnāt matter it being a native app for the console or not probably made it easier since they didnāt really lose anything by making it a browser.The eShop not being a native app is most likely a very intentional decision. There are some very real maintenance and future proofing benefits of doing it this way
For the eShop? No. This is more of an issue off the switch than on the switch.Because itās not a native app wouldnāt stronger hardware still mean better performance?
You've actually got it backwards. The web eShop and NSO app would be way faster and less bloated as native apps. They were made web apps for maximum flexibility in terms of updates.It is an intentional design, it would have caused more bloat and slowdown which they wanted to avoid. They did want lean and minimalism after allā¦.. eShop being one of the āpaths of least resistanceā where it doesnāt matter it being a native app for the console or not probably made it easier since they didnāt really lose anything by making it a browser.
Though, the native app thing is still my dream.
Same for the NSO āappā
That doesnāt make sense to me, the reason the web features are so slow compared to other platforms should be in large part due to the tiny amount of resources allowed to be used for them.For the eShop? No. This is more of an issue off the switch than on the switch.
I mean they can market and choose to use it however they want, but that statement is going to be a really hard sell.
You don't need 1,534 cuda cores and a vastly superior set of cpu cores like A78's, to run games designed for 256 Cuda cores and a quad-core A57, but in 4k with better performance. It's such vast overkill for that job, that saying it was designed to do that job sounds goofy.
If you wanted a switch designed to do ā4Kā switch games, here Iāll design a system for you:
512 Maxwell based CUDA cores, clocked to 460MHz portable and 921MHz docked
6GB LPDDR4X memory with 51GB/s, 128-bit
4 A72 cores running at 1.5-1.8GHz
This could be enough to offer what the PS4 Pro did to the PS4 games. And you know what res most of those games were.
And what technique it used at times for thatā¦CBR anyone?
8nm node.
471GFLOPs portable, an uplift of over twice for most switch games. And 943GFLOPs docked, would have traded blows with XBox One GPU wise.
Reminder that the PS4 Pro was a 2.27x increase over the PS4. And some games managed to be 4K on that. Most didnāt, but for the market Nintendo operates in doesnāt matter since you need a 4K Tv anyway.
Games would have run notice smoother on this vs the base switch. Dynamic would be closer to pretty much locked. Faster loading, nicer images on both modes, etc.
I updated the CPU cores too, because A72 exists on that node , A57 doesnāt I think.
You donāt need Drake for that with the subject is switch games.
This is exactly what i'm thinking. The idea that drake is to play Switch games in 4K no more and no less is really really baffling to me.
Yeah, I've said it before, but Drake really doesn't look like hardware designed to play Switch games, but better. There are much simpler ways to achieve that that would have required a lot less work from Nintendo/Nvidia. If that was all they wanted to do, switching to a new, incompatible GPU arch seems actively counterproductive. Drake looks like hardware intended to enable games not possible on the current Switch, and eventually become the new baseline.
Itās more so than that, you can have the Fidelity for a 720p display but the output resolution is 1080p. Imagine 360 games but at sayā¦ 4k. At itās core itās still a 360 game and looks very much of its era.
Nintendo couldāve done a ā4Kā switch that takes the fidelity of the current switch games, but raise the resolution. This makes them look nicer with more memory to work with and a better GPU for the target display.
Drake on the other hand is going above and beyond just a simple resolution focus, itās also doing a whole generational leap in terms of fidelity.
That is nothing to scoff at. Calling it a āproā system does a disservice to what the device is or seems to be.
Which is why still entertaining the idea of a "pro" console more than 5-6 years into the life of a machine which demonstrably didn't need that at all to become one, if not the most successful console of all time is kind of weird, in my opinion.
Whatever comes next is a successor in the switch line, and I don't even expect Nintendo to hide it.
I don't think Nintendo needs to sell drake at a loss to make a reasonable (AKA $400) price.
They're making a huge profit margin now and outside of the GPU, CPU and RAM they don't really need to change that much. I'm pretty sure they can find a way to make that work at 400.
I honestly don't think "Look, this game you can buy on Switch 1 now you can play it better on Switch 2 too" is the kind of ad to build hype with a new console. They need some new games exclusive for Switch 2, even if there is a long crossgen period.
IMHO Nintendo's counting on a lot of people doing this. I feel BOTW2 will be a cross gen Switch 2/Pro/4K launch window title. They don't have to do much, just advertise the improved performance and IQ. And it serves to address their existing market while appealing to power users. It's the best ace they have.
Assuming hardware is coming this year, this is my sentiment as well. Although I'd have thought they'd need to have something in July, not August. They're running out of time to market any additional major titles launching before October.
Sorry for the quote blastsā¦Iām not able to post very often
I donāt think itās overkill for DLSS functioning, though.
What we seem to be getting spec wise, is basically the minimum to effectively run decent DLSS upscaling and some RT functions at an extremely low power draw. No?
They didn't advertise that for Botw 1 tho. Just said it has better environmental sound. In my opinion Nintendo will never make a show of a console's power.IMHO Nintendo's counting on a lot of people doing this. I feel BOTW2 will be a cross gen Switch 2/Pro/4K launch window title. They don't have to do much, just advertise the improved performance and IQ. And it serves to address their existing market while appealing to power users. It's the best ace they have.
I feel like I ask this every so often; so what's the current consensus? Switch 2 or Switch Pro in the same vein as the PS4 Pro and the XBOX One X?
effectively switch 2. what they actually call it is up for discussion
The idea that itād only ābarely be able to do DLSSā is basically a left over from Digital Foundryās speculative video that came out months before we found out the system would have way more SMs than anyone expected. With what we know now it should be able to very comfortably do DLSS.There are no longer practical Nvidia products in existence for Nintendo to use that do not have tensor cores or rx cores. The last time Nvidia created a product without tensor cores was...... Irrc pascal. Creating a custom solution to remove these things would be much more expensive. It's not about Nvidia convincing Nintendo of dlss. It's that a practical product that does not contain the hardware dlss runs on no longer exists and making one would be paying much much much more for an inferior product.
The vast majority of what is needed for DLSS is performed by the tensor cores. Not on the shader cores, which render a scene.
The tensor cores come packaged with a certain amount of alu's per sm, Tensor cores, Ray tracing cores, and cuda cores. Iirc it's 128 cuda cores, 1RTX core and.... 4 Tensor cores (this is for the a100 series irrc, Drake may be different in specific number per sm, but the practice is the same). This is decided by Nvidia on a architectural scale. All their architectures of a lineage will have fixed numbers. A customer demanding anything deviating from this will be paying for a very very very expensive custom job.
In order to get enough tensor cores to make dlss practical, whatever that may have been, each they would need to add SM's with a certain number of Tensor cores, like for example 4, until they got the desired number of Tensor cores for practical dlss performance that meets their target.
Drake has 12 SM's , if they soley designed a system around the number they wanted for their performance target of dlss, it took 12 SM's to reach.
Each sm, also comes with 128 Cuda Core shaders. They needed 12 SM's, to get the tensor cores needed for DLSS. That's 1,536 Cuda Core shaders.
Once again I am pointing out, it is the tensor cores that perform the DLSS workload. Not the Cuda Cores. They still handle crunching polygons, and graphical effects, bells, and whistles, along with other various tasks. They are not impacted by having to do dlss, that is offloaded to the tensor cores.
The original switch had 2 SM's, with 256 shader cores for producing graphics/shader tasks.
Drake has 12 SM with 1,536 shader cores for producing graphics/shader tasks. And it's a seperate group of processors in the GPU that do dlss.
This is a massive and inescapable increase in raw power. In order to get the tensor cores, to 'barely be able to do dlss' 6x the number of shader cores had to come along for the ride.
It makes the most sense. Call it a switch 4K and be done. It will get third party games current switch canāt get. It will be backwards compatible. 1st party games will be available on both. You can keep the switch name alive and well.Effectively a Switch mid-gen upgrade.
I think people misunderstand how Nintendo is going to position it and utilize it because of its modern architecture and power differential.
I'm not seeing how going from 256, 12nm Maxwell cores to 1536, 7/5nm Ampere cores is a "mid gen upgrade". that's a bigger paper jump than the PS4 to PS5Effectively a Switch mid-gen upgrade.
I think people misunderstand how Nintendo is going to position it and utilize it because of its modern architecture and power differential.
I don't think they were talking about the specs upgrade, rather how it will be positioned.I'm not seeing how going from 256, 12nm Maxwell cores to 1536, 7/5nm Ampere cores is a "mid gen upgrade". that's a bigger paper jump than the PS4 to PS5
now I'm curious as to what a proper next gen system even means in your eyes, because I doubt it physically exists right now
The idea that itād only ābarely be able to do DLSSā is basically a left over from Digital Foundryās speculative video that came out months before we found out the system would have way more SMs than anyone expected.
This is ABSOLUTELY the right time to push a iterative mid gen forward looking āproā type model.
They expect the current Switch lifespan to last another 4-5 years. This model is absolutely essential to keep Switch gaming engagement high during that time for the core gamers who might lose interest for what they feel is increasingly outdated looking visuals/performance.
Thatās what a mid gen system is designed to do. All about software engagement.
Even the OLED model, I am positive, made Switch gamers who bought it, buy and play more Switch games over the last year than they would have otherwise on their launch model. It breaths new life in the games, even just the slightly bigger screen and OLED.
The Drake will breathe even more life into Switch gaming engagement for much longer. Thatās all it needs to do.
Nintendo aren't the kind of company to advertise relative graphical power. It's not like the PS5 or XSX where they'll talk about how many teraflops the GPU is capable of. So while the performance leap will seemingly be very large, I don't think it's unreasonable to speculate that this might not factor hugely into how Nintendo markets the productI'm not seeing how going from 256, 12nm Maxwell cores to 1536, 7/5nm Ampere cores is a "mid gen upgrade". that's a bigger paper jump than the PS4 to PS5
now I'm curious as to what a proper next gen system even means in your eyes, because I doubt it physically exists right now
I don't think they were talking about the specs upgrade, rather how it will be positioned.
Yeah just realized that he was replying to feet saying the opposite.I am very distinctly not getting this impression.
Particularly since both feet and I prefaced everything with 'Nintendo will market it however they want, but the specs are'
Why you and others cant say: āIn my opinionā¦ā But you always act as if you are the CEO of the company themselfā¦
I'm pulling for $400. Any higher and it better have the power to go with it.
My bets 400 also, but that was before inflation so now wouldn't be surprised. But also imagine oled gets price dropped to 300 shortly after, and Redbox discontinued
I can only see $400 if they drop the OLED price end of this year. I wasnāt expecting thatās something theyād do only a year after launch but I guess it couldnāt happen?
Iām guessing itāll be $450. OLED stays the same, and 2017 sees an official price cut.
The reason I think we need a new system soonish is due to third parties, now they are getting to grips with PS5/XsX the gap between them and Switch is huge so chances are less and less devs will be bothered with Switch ports/versions unless they get a boosted system where they donāt need to downgrade things so much that itās more hassle than itās worth.
What do you mean? They couldn't have downloaded more ram through an update.yeah, i mentioned that, and thats the reason why they could have improved it if they wanted, ba replacing it with a System Update with an Native App.
As it stands, it was clearly not a priotiry
I recall they mentioned the resolution difference and faster loading times alongside the improved sounds. The context is different, too. They wanted to highlight the Switch's portability when it was first revealed, which is a more striking feature than the minor performance improvements.They didn't advertise that for Botw 1 tho. Just said it has better environmental sound. In my opinion Nintendo will never make a show of a console's power.
Unrelated to that:
unpopular opinion but I 100% expect original Switch games to run in "Switch Mode" on the next console. Like Wii on Wii U, DS on 3DS and even normal 3ds games on new 3ds. It seems to me that Nintendo cares much more about accuracy than improvements haha.
Wii U was Nintendo chasing the tailwind of PS3 and XBox 360 at a too little, too late time and grossly underestimating the resources to needed to stream to the GamePad.I just find it strange that Nintendo is suddenly going back to the Game Cube philosophy of "an equal but more powerful console" for a successor. I need to see something else, new and different, to truly believe it.
It's hard to me to believe that they are going to change their philosophy after Wii, Wii U, DS, 3DS and Switch.
I mean, not to the same extent as MS and Sony, but certainly more than nothing.Nintendo is not releasing this new model for 3rd party multiplats. They donāt even factor that in on any design decisions they do.
And I donāt think we will see any 3rd party multiplat support (ports) over the next few years different than we have seen on the current models the last 3 years.
Having to deal with the same joycons, the same interface and the same online service until 2028 is a nightmare for me, tbh. I want something new and exciting.
The Switch's launch price was the same as the PS4 Slim's at the time. I agree that Nintendo would rather avoid a high MSRP, but not because of potential comparisons with PS5/XSX.I could see 450 now with inflation but I feel they would want to avoid 499 to not be compared as much to the PS5
Two things, neither of which I'm an expert on:
1) Monster Hunter Rise on PC now supports DLSS. Do with that what you will.
There are no longer practical Nvidia products in existence for Nintendo to use that do not have tensor cores or rx cores. The last time Nvidia created a product without tensor cores was...... Irrc pascal. Creating a custom solution to remove these things would be much more expensive. It's not about Nvidia convincing Nintendo of dlss. It's that a practical product that does not contain the hardware dlss runs on no longer exists and making one would be paying much much much more for an inferior product.
The vast majority of what is needed for DLSS is performed by the tensor cores. Not on the shader cores, which render a scene.
The tensor cores come packaged with a certain amount of alu's per sm, Tensor cores, Ray tracing cores, and cuda cores. Iirc it's 128 cuda cores, 1RTX core and.... 4 Tensor cores (this is for the a100 series irrc, Drake may be different in specific number per sm, but the practice is the same). This is decided by Nvidia on a architectural scale. All their architectures of a lineage will have fixed numbers. A customer demanding anything deviating from this will be paying for a very very very expensive custom job.
In order to get enough tensor cores to make dlss practical, whatever that may have been, each they would need to add SM's with a certain number of Tensor cores, like for example 4, until they got the desired number of Tensor cores for practical dlss performance that meets their target.
Drake has 12 SM's , if they soley designed a system around the number they wanted for their performance target of dlss, it took 12 SM's to reach.
Each sm, also comes with 128 Cuda Core shaders. They needed 12 SM's, to get the tensor cores needed for DLSS. That's 1,536 Cuda Core shaders.
Once again I am pointing out, it is the tensor cores that perform the DLSS workload. Not the Cuda Cores. They still handle crunching polygons, and graphical effects, bells, and whistles, along with other various tasks. They are not impacted by having to do dlss, that is offloaded to the tensor cores.
The original switch had 2 SM's, with 256 shader cores for producing graphics/shader tasks.
Drake has 12 SM with 1,536 shader cores for producing graphics/shader tasks. And it's a seperate group of processors in the GPU that do dlss.
This is a massive and inescapable increase in raw power. In order to get the tensor cores, to 'barely be able to do dlss' 6x the number of shader cores had to come along for the ride.
The idea that itād only ābarely be able to do DLSSā is basically a left over from Digital Foundryās speculative video that came out months before we found out the system would have way more SMs than anyone expected. With what we know now it should be able to very comfortably do DLSS.
reduce the overehead by having a natively running app instead of using a browser (or applet or how you wanna call it) and running the eshop as an webapp in that.What do you mean? They couldn't have downloaded more ram through an update.
There is about 303 pages of comments in here of people speculating on timing.Since we haven't got any kind of announcement, I was wondering when the presentation of the new console could be. August? September? Next year?
I'm not seeing how going from 256, 12nm Maxwell cores to 1536, 7/5nm Ampere cores is a "mid gen upgrade". that's a bigger paper jump than the PS4 to PS5
now I'm curious as to what a proper next gen system even means in your eyes, because I doubt it physically exists right now
Just curious, how many Capcom games currently support DLSS? I genuinely do not know.
Nintendo Switch first look was in October for a March release, I think we will see a similar 6 month ish period between the event and release date, so 6 months before release is my bet. The console existence will get announced earlier tho, so if it's early 2023 we should get news before SeptSince we haven't got any kind of announcement, I was wondering when the presentation of the new console could be. August? September? Next year?
The past two revisions were revealed in July. If we don't see anything by the end of July it's likely we won't see anything until next year.Since we haven't got any kind of announcement, I was wondering when the presentation of the new console could be. August? September? Next year?
Im mixed. I dont want to pick a price point, but while in the short therm they will sell out (heck, i think even the Wii U sold out at the beginning), if they want a widespread adoption like with the switch, in a time when people start to struggle financially because of inflation...It's all about perceived value because the early adopters have the means. The steamdeck and PS5 / Series S all set a pretty comfortable price range for the Switch 2 to sell at Given its speculated performance
A portable device with AI upscaling and some RTX features will be kind of a big deal.
Anyone blindly picking a price like $450 and declaring the device is DOA if it launches at that price can expect to get egg on their face. I personally think $399 makes the most sense for Nintendo. But not ruling out $449
I needed a night sleep before going back to this and I think you are right with your classification. There might always be room for DLSS in the frame budget but that would imply that the developers would knowingly and deliberately sacrifice shader and post-effect quality to 'shoe horn' it in.I've learned a lot from hanging around here, including from your posts. I have a technical background, but in a totally different field. It's awesome that we can have an ongoing conversation.
It definitely doesn't. I realize I was a little unclear here. What I was getting at was the tradeoff between tensor cores and shader cores. On games where DLSS is NOT useful, how much better would those games look/run if Nintendo had replaced the tensor cores with just more raw shader cores?
It's a really hard question to answer, but probably not much? The GTX 1650 attempted to do just that, relative to the 1660, and doesn't seem to have managed to squeeze any extra performance out relative to its die size/power draw. Someone here might have a better answer. But think of it this way. There are roughly three classes of games
Group 3 needs a really weird performance profile. If a game runs like a powerpoint at the target res, then they'd need huge amounts of extra shader perf, and sacrificing tensor cores probably won't get you there. If the game runs, say, at a stable 30fps at 1080p, then you probably can get to 720p60fps with enough room in the frame budget for DLSS to get you back up to 1080p60fps.
- Games that aren't going to have a problem reaching "max" resolution using just shader cores. Anything pure pixel art, for example.
- Games that benefit from DLSS. Imagine a game running a comfortable 1080p60fps on the core hardware, that can get down to 720p90fps, leaving them ample time in their framebudget to DLSS up to 1440p60fps, and still look good.
- Games that push the hardware, but can't create the room in their framebudget for DLSS without looking bad. These are the games that, in theory, would benefit from dropping Tensor Cores and replacing that with more shader cores or higher clocks.
The place where you really want to toss Tensor Cores for Shader Cores is something like a game running 55fps at your target res. You're close enough that a little bit of extra GPU perf is going to push you over the finish line, but DLSS might be a noticable IQ drop for a tiny boost in performance, and would rather just hand tune themselves back up to 60fps. I think the number of games in that bucket is going to be small, and regardless, the sacrifices those games will have to make will also not be dramatic.
You can innovate with new hardware without ditching the hybrid idea, they are not contradictoryThey can easily offer ānew and excitingā gameplay and ways to game different as a peripheral to the Switch ecosystem.
They donāt need to solely devote to a single console for it.
But I agree with the other poster who feels the Switch hybrid type ecosystem will persist as the mainline Nintendo console for another decade or two, minimum.
This is where I think Tensor cores comes in. Yes dlss is nice, but not every game needs it. If any dev will get creative with dedicated ml hardware, I believe itās Nintendo.You can innovate with new hardware without ditching the hybrid idea, they are not contradictory
If it works I'm scared that from now on Nintendo will create a situation were there is a constant hype cycle for the next Nintendo event instead of a clear schedule of 3 directs + extra smaller ones per yearGetting hyped for another potential Direct/unveiling while the one planned for today wasn't even aired yet?
I have the sensation that some actual Nintendo workers are involved in this forum. The transition from 'get hyped for the direct!' to 'get hyped for the NEXT THING!' was a little too smooth. It felt engineered.
The past two revisions were revealed in July. If we don't see anything by the end of July it's likely we won't see anything until next year.
Getting hyped for another potential Direct/unveiling while the one planned for today wasn't even aired yet?
I have the sensation that some actual Nintendo workers are involved in this forum. The transition from 'get hyped for the direct!' to 'get hyped for the NEXT THING!' was a little too smooth. It felt engineered.
we don't mind you, Nintendo workers.
Personally I don't think they'd sabotage their holiday sales by announcing new hardware before the holidays that you can't buy during the holidays.Iām all in on it releasing in March, so I think a reveal around October, like with the original Switch, makes a lot of sense.
It aired hours ago.