No, you're right as far as custom chips in the past. The use of the off the shelf X1 was the historical outlier. I guess my point was and my hunch is that the Orin NX is "close enough" to what could serve as a suitable successor and, coupled with Nintendo's conservative approach to its tech and use if the X1, I wouldn't expect a custom chip. I'd be pleased to be wrong, though.
You are using an out of date browser. It may not display this or other websites correctly.
You should upgrade or use an alternative browser.
You should upgrade or use an alternative browser.
Switch 2 Speculation
- Thread starter Goodtwin
- Start date
I'm confused why you guys aren't discussing T239 codename "Drake" that leaked on March 1st this year in the big Nvidia data breach, Drake is what Nvidia's/Nintendo's NVN2 API calls for and is Ampere based GA10F and comes with 12SM or 1536 cuda cores, has Tensor cores and RT cores.
I'm confused why you guys aren't discussing T239 codename "Drake" that leaked on March 1st this year in the big Nvidia data breach, Drake is what Nvidia's/Nintendo's NVN2 API calls for and is Ampere based GA10F and comes with 12SM or 1536 cuda cores, has Tensor cores and RT cores.
Interesting, thanks for the links! It looks like T239 is a "customized" T234 Orin, which itself is the Orin chip underlying the Jetson AGX Developer Kit with up to 2048 CUDA Cores (16 SM). So, will "Drake" be a cut-down version or, as suggested, a "customized," unique piece of silicon?
Perhaps more interesting in my mind is whether Nintendo will try to scale the docked experience higher than the current Switch does. If Nintendo sticks with a 720p handheld display, "Drake" presumably could clock very low given to hit desirable frame rates at that resolution. The Orin silicon that NV is selling scales from 10W to 60W. If Nintendo truly plans a "4k" successor as Bloomberg and others have surmised over the past year (I have doubts), that would likely require scaling up clocks much higher than the current Switch (relative to handheld mode) plus using an aggressive form of DLSS or some other upscaling method. With respect to cooling, you'd be looking either at a chunkier, less portable Switch (not ideal) or perhaps an innovative new dock to assist with cooling a 20-40W docked Switch.
Backwards compatibility could very well end up being out of reach for the new system without a pure hardware solution ...
Anti-emulation code is a pretty effective way to prevent all translation layers from running software including a console vendors own in-house developed one. The reverse engineering community is largely fortunate in that the recent systems their targeting don't need to be "cycle accurate". Making an emulator robust against hardware timings isn't feasible with remotely modern systems ...
Anti-emulation code is a pretty effective way to prevent all translation layers from running software including a console vendors own in-house developed one. The reverse engineering community is largely fortunate in that the recent systems their targeting don't need to be "cycle accurate". Making an emulator robust against hardware timings isn't feasible with remotely modern systems ...
Silent_Buddha
Legend
Denuvo wouldn't necessarily interfere with any backwards compatibility plans for Nintendo. Publishers and developers regularly remove Denuvo DRM from their games after X period of time has passed.
The ability for a publisher or developer to remove Denuvo DRM at some point is one of the key selling points of their DRM scheme.
Regards,
SB
The ability for a publisher or developer to remove Denuvo DRM at some point is one of the key selling points of their DRM scheme.
Regards,
SB
D
Deleted member 11852
Guest
And the relative ease to do this is why some games get cracked really quickly. Not all, but many. I think part of the resilience may be linked to how much effort the dev spends initially ingraining that crap in there. An simple #include "shitdrm.c" and a few calls the library in the main gate loop probably doesn't afford you much protection.The ability for a publisher or developer to remove Denuvo DRM at some point is one of the key selling points of their DRM scheme.
All DRM has to do in this case is to prevent translation layers from implementing accurate timings or cache behaviour which are unreasonably expensive to emulate even on the most powerful hardware and some developers don't ever seem to remove DRM from their games ...And the relative ease to do this is why some games get cracked really quickly. Not all, but many. I think part of the resilience may be linked to how much effort the dev spends initially ingraining that crap in there. An simple #include "shitdrm.c" and a few calls the library in the main gate loop probably doesn't afford you much protection.
They aren't removing it from the carts it ships on.Denuvo wouldn't necessarily interfere with any backwards compatibility plans for Nintendo. Publishers and developers regularly remove Denuvo DRM from their games after X period of time has passed.
The ability for a publisher or developer to remove Denuvo DRM at some point is one of the key selling points of their DRM scheme.
Regards,
SB
Then there goes not requiring an internet connection.
tongue_of_colicab
Veteran
Sure but nobody is emulating from a cart so that doesn't really matter. As long as the DRM doesn't have any performance impact etc. when running on actual hardware it doesn't really matter if its there or not.They aren't removing it from the carts it ships on.
Then there goes not requiring an internet connection.
Sure but nobody is emulating from a cart so that doesn't really matter. As long as the DRM doesn't have any performance impact etc. when running on actual hardware it doesn't really matter if its there or not.
Wouldn't a person with a switch 2 who owned the game on the switch in cart form need to emulate it off a cart ?
Regardless, games and updates have to pass certification from Nintendo anyways. If Nintendo is worried about it, all they have to do is fail the certification and tell developers to take this out if they want to pass. It is 100% within Nintendo's right to govern DRM on their platform.Denuvo wouldn't necessarily interfere with any backwards compatibility plans for Nintendo. Publishers and developers regularly remove Denuvo DRM from their games after X period of time has passed.
The ability for a publisher or developer to remove Denuvo DRM at some point is one of the key selling points of their DRM scheme.
Regards,
SB
I'm confused why you guys aren't discussing T239 codename "Drake" that leaked on March 1st this year in the big Nvidia data breach, Drake is what Nvidia's/Nintendo's NVN2 API calls for and is Ampere based GA10F and comes with 12SM or 1536 cuda cores, has Tensor cores and RT cores.
I do not believe those numbers were actually in the leaked Nvidia documents, but instead were speculation by another Nvidia leaker on what a cut down Orin chip might look like. This was before Nvidia announced the Orin NX variants. When you look at the specs and power consumption traits of the Orin NX, it seems like a good fit for Switch 2. What we don't know is just how big the Orin NX chip is. The standard Orin is very big, I believe its around 350mm. The Orin NX has half the GPU cores and down to 6 CPU cores with a 128bit memory bus. Orin NX should hit the market towards the end of the year. If its size is under 150mm then its damn near a lock that it will power the Switch 2.
I've looked at the NVN2 API code from the leak dated Feb this year, it indeed calls for 12SM and 128bit memory bus, the chip is also identified as GA10F, so it is Ampere with 1536 cuda cores, which means that at Switch's portable 460MHz clock, it would have 1.41TFLOPs, and with 768MHz clock from Switch's docked performance, it would be 2.36TFLOPs, considering PS4 Pro doesn't do something like DLSS, that is probably enough for Drake to do 4K, but considering the size of the GPU, it is possible that it has a 1GHz GPU clock for 3.07TFLOPs, or even higher if Nvidia offered Nintendo all their excess TSMC 4N node (Nvidia's custom 5nm process with TSMC), that could push it to something like 1.3GHz or 4TFLOPs. Ampere in a closed system and with Tegra's common cache setup for CPU/GPU, it could achieve higher performance per clock than Ampere's desktop GPUs achieve.I do not believe those numbers were actually in the leaked Nvidia documents, but instead were speculation by another Nvidia leaker on what a cut down Orin chip might look like. This was before Nvidia announced the Orin NX variants. When you look at the specs and power consumption traits of the Orin NX, it seems like a good fit for Switch 2. What we don't know is just how big the Orin NX chip is. The standard Orin is very big, I believe its around 350mm. The Orin NX has half the GPU cores and down to 6 CPU cores with a 128bit memory bus. Orin NX should hit the market towards the end of the year. If its size is under 150mm then its damn near a lock that it will power the Switch 2.
Also, because it wouldn't use A78AE CPUs, I'd assume they are using A78C CPUs, they haven't been used in any device yet, but they were created by ARM, seemingly to fill an existing order, these CPUs are specifically marketed for dedicated gaming devices, and come in both 6 cores and 8 cores, either of which would be a huge improvement over last gen consoles, as A78 cores are 3 times more performant than A57 per clock and actually out perform Ryzen 2 per clock as well, though 8nm would limit the clock quite a bit, 4N could achieve over 2GHz per core for even an 8 core SoC, especially because the OS core could be clocked much lower.
Nintendo just announced that the Legend of Zelda: Tears of the Kingdom will launch May 21, 2023... what are the odds Nintendo launches a Switch Pro (or 2 or whatever) in March 2023 around the 6th anniversary of the original launch? Or am I just dreaming?
There's a very real possibility of that.Nintendo just announced that the Legend of Zelda: Tears of the Kingdom will launch May 21, 2023... what are the odds Nintendo launches a Switch Pro (or 2 or whatever) in March 2023 around the 6th anniversary of the original launch? Or am I just dreaming?
Do you have a link to this? Everything I find when searching still references speculation that Kopite7kimi made. I just dont see them fitting that many cores into an SOC that will work within the Switch form factor. I would love to be wrong, more performance the better, but until I see concrete evidence saying otherwise, I believe its going to essentially be the same as the Orin NX with 1024 cores and 6 A78 CPU cores.I've looked at the NVN2 API code from the leak dated Feb this year, it indeed calls for 12SM and 128bit memory bus, the chip is also identified as GA10F, so it is Ampere with 1536 cuda cores
Seems like everyone instantly thought it would be March 2023 when Nintendo said Spring 2023, but spring last into June. Nintendo didnt fully reveal the Switch until January 2017 and then released the first week of March. I could very much see a reveal in January. I am starting to believe Nintendo will want this successor to be more similar to the Gameboy to the Gameboy Color, where it feels like one long generation despite the newer hardware having exclusives for it. What better way to roll out new hardware than to launch aside Nintendo's most anticipated game? I like the chances.Nintendo just announced that the Legend of Zelda: Tears of the Kingdom will launch May 21, 2023... what are the odds Nintendo launches a Switch Pro (or 2 or whatever) in March 2023 around the 6th anniversary of the original launch? Or am I just dreaming?
Question is what price point? I feel like it can launch at $399 no problem and potentially even $499. Seems like consumers are conditioned to spend more money on gaming hardware these days.
I generally agree with the comparison to the Gameboy series of hardware, though Nintendo should be wary of the Wii/Wii U transition, in which many appeared to believe the Wii U was simply an accessory to the Wii.Nintendo didnt fully reveal the Switch until January 2017 and then released the first week of March. I could very much see a reveal in January. I am starting to believe Nintendo will want this successor to be more similar to the Gameboy to the Gameboy Color, where it feels like one long generation despite the newer hardware having exclusives for it. What better way to roll out new hardware than to launch aside Nintendo's most anticipated game? I like the chances.
Question is what price point? I feel like it can launch at $399 no problem and potentially even $499. Seems like consumers are conditioned to spend more money on gaming hardware these days.
In my opinion, Nintendo should launch a Switch Pro that is fundamentally the same machine (i.e. uses compatible game cards, joycons, etc.) but with a higher-end feel (i.e. design, buttons), an improved screen (maybe simply the OLED from Switch OLED), and the ability to play games at a higher resolution/frame, especially when docked. Discontinue the OG Switch and have a tiered pricing system of Switch Pro (high-end), Switch OLED, and Switch Lite. Game cards generally can be shared (though the newer hardware will necessarily have some exclusives), local multiplayer works, etc. Create a situation where older gamers with disposable income upgrade to the latest and greatest, or maybe families can pass the OG Switch to the young sibling, and the older sibling gets the Switch Pro for his next birthday. Kinda sorta like the Gameboy back in the day.
If this is the plan, given the current Switch lineup's ongoing momentum, it makes sense that Nintendo would not announce a new model until after this holiday season. And, yeah, I think a $399 (Pro), $279-$299 (OLED), and $199 (Lite) is feasible.
A key difference with the Wii to Wii U transition was the Wii's popularity has plunged by the time Wii U launched. Switch is still going strong so consumer awareness should be better. The reason I don't really like the idea of a "Pro" version is because that means the games still must target the OG Switch. Third parties need to have a more capable machine without being shackled to the OG Switch. Gameboy Color had tons of exclusives even though it was a direct evolution of the Gameboy and played OG Gameboy games. Nintendo worries about the risk of transitioning to new hardware and wants it to be as seamless as possible. Hell, Nintendo can call it a Switch Pro if they want, but it needs to offer developers the option to target that hardware. They will need a new logo to help consumers understand that a given game is for the new hardware and wont work on the old system.I generally agree with the comparison to the Gameboy series of hardware, though Nintendo should be wary of the Wii/Wii U transition, in which many appeared to believe the Wii U was simply an accessory to the Wii.
In my opinion, Nintendo should launch a Switch Pro that is fundamentally the same machine (i.e. uses compatible game cards, joycons, etc.) but with a higher-end feel (i.e. design, buttons), an improved screen (maybe simply the OLED from Switch OLED), and the ability to play games at a higher resolution/frame, especially when docked. Discontinue the OG Switch and have a tiered pricing system of Switch Pro (high-end), Switch OLED, and Switch Lite. Game cards generally can be shared (though the newer hardware will necessarily have some exclusives), local multiplayer works, etc. Create a situation where older gamers with disposable income upgrade to the latest and greatest, or maybe families can pass the OG Switch to the young sibling, and the older sibling gets the Switch Pro for his next birthday. Kinda sorta like the Gameboy back in the day.
If this is the plan, given the current Switch lineup's ongoing momentum, it makes sense that Nintendo would not announce a new model until after this holiday season. And, yeah, I think a $399 (Pro), $279-$299 (OLED), and $199 (Lite) is feasible.
At $399, it would have to have OLED screen and rock solid 1080p60 performance.
At $499, who knows. I think the price is a big ask since a lot of Nintendo users are kids. Not saying parents wouldn't buy them toys at such prices but there are a lot fewer of them who would compared to those who would buy them the $299 Switch.
At $499, who knows. I think the price is a big ask since a lot of Nintendo users are kids. Not saying parents wouldn't buy them toys at such prices but there are a lot fewer of them who would compared to those who would buy them the $299 Switch.
I agree about $499... I think that price exceeds the market Nintendo is looking to hit. But, $399 is more feasible provided Nintendo continues to offer lower priced entries into the Switch ecosystem.
At $399, I'd expect at least the same quality OLED screen, especially if the OG Switch is phased out and the Switch OLED becomes the mid-tier option at $299 (reduced from the $350 it currently sells at). It'd be unusual to have the mid-tier have a nicer screen than the top-tier. The other scenario is the OG Switch remains the mid-tier, the Switch OLED is discontinued, and you'd expect the Switch Pro to have at least as good a screen as the discontinued Switch OLED had.
As a point of comparison, you don't get an OLED with the $399 or even pricier versions of the Steam Deck. But, a Switch Pro is also likely (in my mind) to feature lower end innards than the Steam Deck (as well as less functionality being a game console versus a full-fledged computer).
To play devil's advocate with myself... if we assume the Pro's specs look similar to the Orin NX 8 GB module (Orin), Nvidia is looking to sell those for $399 without internal storage, a screen of any kind, Joycons, or a dock. We know Nvidia likes to make a very healthy margin on everything it sells, but so does Nintendo. Can Nintendo sell what amounts to the Orin NX 8GB with an OLED, internal storage, Joycons, and a Dock for $399 and still turn a profit per unit? When I put lay it out like that, I have my doubts. Oh well, that's what makes speculation so fun!
At $399, I'd expect at least the same quality OLED screen, especially if the OG Switch is phased out and the Switch OLED becomes the mid-tier option at $299 (reduced from the $350 it currently sells at). It'd be unusual to have the mid-tier have a nicer screen than the top-tier. The other scenario is the OG Switch remains the mid-tier, the Switch OLED is discontinued, and you'd expect the Switch Pro to have at least as good a screen as the discontinued Switch OLED had.
As a point of comparison, you don't get an OLED with the $399 or even pricier versions of the Steam Deck. But, a Switch Pro is also likely (in my mind) to feature lower end innards than the Steam Deck (as well as less functionality being a game console versus a full-fledged computer).
To play devil's advocate with myself... if we assume the Pro's specs look similar to the Orin NX 8 GB module (Orin), Nvidia is looking to sell those for $399 without internal storage, a screen of any kind, Joycons, or a dock. We know Nvidia likes to make a very healthy margin on everything it sells, but so does Nintendo. Can Nintendo sell what amounts to the Orin NX 8GB with an OLED, internal storage, Joycons, and a Dock for $399 and still turn a profit per unit? When I put lay it out like that, I have my doubts. Oh well, that's what makes speculation so fun!
Similar threads
- Replies
- 14
- Views
- 1K
- Replies
- 21
- Views
- 6K
