What does AP stand for?Latest rumor from Korea (machine translation):
RDNA? I found it hard to believe. But this is the same leaker who posted Switch Lite info in June 2018.
What does AP stand for?Latest rumor from Korea (machine translation):
RDNA? I found it hard to believe. But this is the same leaker who posted Switch Lite info in June 2018.
Even if Nintendo is not contractually bound to stick with Nvidia, they're invested deeply enough in Nvidia's ecosystem that they're going to need a really compelling reason to leave it, and none of Nvidia's competitors in the area seem to have anything to offer on that front right now.Nintendo never signed a 10 year contract. Jenson pointed out that Nintendo worked with IBM and AMD(formerly ATI) for 3 consoles (Gamecube, Wii and WiiU) and suddenly people started saying Nvidia has a multi year contract.
Here is Jen-Hsun Huang’s statement regarding the Nintendo Switch:
It's likely supposed to be APU, which is a term AMD likes to use for their SoCs.What does AP stand for?
Application processor, which is basically a synonym for SoC.What does AP stand for?
A look back at this Korean leaker's track history (all machine translated):Latest rumor from Korea (machine translation):
RDNA? I found it hard to believe. But this is the same leaker who posted Switch Lite info in June 2018.
It is from Japan.
A tentatively named Switch Lite will be released in 2020. 3DS sequel role.
only 1 citation. One-piece, the controller cannot be separated.
Excluding TV and connection function. Improved portability by 5 inches.
Performance has not been released yet, but it is highly likely that the performance will decrease considering that the battery life will be greatly increased and a new tentative Switch Lite game lineup will be released.
The game lineup is divided into two for Switch and one for Switch Lite. Currently, it is said that the core third party, who has already heard about the specifications, is strongly opposed to the division of the lineup.
p.s There is a misunderstanding, but I wrote it above, but it is a guy who plays the role of a 3ds sequel, and the 2nd generation of Switch is in the previous Chirashi, but it is separate.
It is from Japan.
tegra x1 t210 processor
4GB LPDDR4
32GB EMMC 5.1
4.98 inch HD resolution LCD
Remove standalone mode
The gamepad is not a Joy-Con style, but an integrated body, and it can be operated with an additional Joy-Con accessory in games that require a Joy-Con. However, there is a separate type among the samples, so judgment is withheld.
ac not supported in wifi
HD vibration not supported
Initial sale price 18,980 yen
The battery capacity cannot be confirmed in the current situation, but there is information that the operating time has increased by 20-30% compared to the previous one.
It is said to be a replacement for the 3DS.
It is from Taiwan.
It is not known whether this is for the tentatively named Switch Pro, but a recent rumor says that Nintendo of Japan, like Google, reviewed the feasibility of Samsung's custom architecture development request in the first quarter of this year, and received a positive response from Samsung. Samsung is said to have responded that it is possible to design and mass-produce an AP with the specifications requested by Nintendo in the shortest possible time.
Nintendo requested a relatively low price, low power consumption, etc. for the mobile AP specifications with graphics processing performance equivalent to the current generation (the rumored basis for the PS4 initial model).
Samsung has presented a custom chip for Nintendo using cortex A76, which can be mass-produced at an early time and is available at a low price, and an option using cortex x1, which has good performance although the mass production time is delayed. The answer says that it is possible to provide an option.
However, talk of GPUs is not included in this rumor.
It is from Taiwan.
Recently, it is rumored that a new Tegra x1 sample has been released through TSMC's 7-nano process. It was taken from the existing 1.9Ghz to a whopping 2.52Ghz, and TSMC and Nvidia had a negative position that it would be difficult to exceed the 1.9Ghz limit of the Cortex A57. However, it is said that the actual clock is estimated to be lower than this because the power consumption is extreme than expected, but it is still unknown because Nintendo has overclocked the old architecture like crazy and put power in the sequel.
p.s It's a rumor... but it might be possible if Nintendo considers the precedent of the New 3DS... (Achieving 804Mhz quad-core, which has tripled the clock with the unprecedentedly old ARM11...)
But if that's the case, I think it would be better to just use a Cortex A73...
It is from Taiwan.
Currently, information is mixed on whether it is a Samsung 5nm process or a tsmc process.
The reason is that Nvidia has ordered both for mobile-related chipsets and is preparing for production.
However, both chipset specifications are different. One of the two production chipsets is possibly Nvidia's other purpose chipset
There is information that the GPU clock that will go into the Switch Pro is almost 1Ghz. Existing switch clock corrects. 768 MHz.
It is said that it can realize up to 720p 120 frames in Switch Pro portable mode.
However, it is said that it will come out with a compromise by fixing 60 frames at 720p due to the maintenance problem.
Performance in stationary mode is currently known to be optimized for 1440p resolution.
There was talk that Nvidia's latest technology would be included, but due to various problems, it may be applied only in stationary mode or may come out sealed at the time of release.
There are rumors that third-party Nintendo developers have already received development kits.
Nintendo has been trying to develop with AMD chipsets instead of Nvidia since 3 years ago, but there are rumors that they chose Nvidia after considering their own development because of various circumstances.
And now, there are rumors in Japan that Microsoft's Xbox Game Pass for Switch will be launched at the same time as the Nintendo Switch Pro launch.
yes. hell, Intel even made an Atom processor that had better perf/watt than ARM (at the time. of course it meant ditching a lot of legacy.Just to play along here, is it even possible for X86 to be shrunk down enough to fit into a Nintendo Switch successor, if possible im sure it would make porting games easier across all platforms correct?
The only thing that is gonna change in regards to porting are how willing companies want to be. If they do not wanna invest the resources to porting their games it honestly doesn’t really matter how strong a Nintendo machine is.Just to play along here, is it even possible for X86 to be shrunk down enough to fit into a Nintendo Switch successor, if possible im sure it would make porting games easier across all platforms correct?
Actually ARM is still faster at executing instructions that x86 due to its simplified nature (and also more efficient). Also I don't recall Intel Atom being any good in the netbook space, either.yes. hell, Intel even made an Atom processor that had better perf/watt than ARM (at the time. of course it meant ditching a lot of legacy.
and no, it wouldn't make porting easier. x86 vs ARM doesn't matter anymore
Calling those Switch Lite rumors "fairly accurate" seems a bit generous. The 2018 one in particular seems pretty wildly off.A look back at this Korean leaker's track history (all machine translated):
2018-06-20, about Switch Lite
2019-02-23, about Switch Lite
2020-07-08, about Switch Pro CPU (Samsung proposal)
2020-10-28, about TSMC testing 7nm TX1
2021-06-06, about Pro performance with Nvidia chipset (and Game Pass!)
- The leaker claimed to have two sources, one in Japan and one in Korea.
- The Japanese source gave them fairly accurate info regarding Lite, one year before its announcement.
- But the "rumors in Japan" (not sure if same source) about Game Pass is wild.
- The info from their Taiwanese source has been all over the place—veracity unknown.
- If I have to guess, their sources might be from the assembler side, hence the detailed Lite specs but vague SoC info.
- Should there be a sliver of truth in the latest AMD rumor, I can think of a few explanations:
- The source conflated the development of a future Exynos (after 2200) with the new Switch SoC.
- It's a Samsung SoC with Nvidia GPU, not unlike Exynos 2200 with AMD GPU, but the message got distorted during the game of telephone.
- The "AMD" SoC is not for a new Switch model but a future Nintendo product (e.g., AR or "quality of life").
- Note that "AMD" doesn't automatically mean x86. Both Lisa Su and Devinder Kumar (CFO) indicated that AMD would build ARM chips if their customers demand it.
- Lastly, I don't know how accurate the machine translation is.
- "it doesn't use Nvidia chipsets": It may not be an "Nvidia" SoC, but it could be a custom Nintendo SoC that still includes Nvidia GPU cores.
- "it is said that [...] there is a high probability that the RDNA series will enter the GPU side": This sounds like a conjecture of their source.
- Of course it could all be bull, or overblown/inaccurate hearsay.
It starts really well, but point 4 and 5 completely fucks it up.Calling those Switch Lite rumors "fairly accurate" seems a bit generous. The 2018 one in particular seems pretty wildly off.
![]()
Elden Ring performance: what to expect on PS5, Series X/S and PC
Elden Ring's launch has been marked by strong reviews praising its gameplay and art direction, but also questions over …www.eurogamer.net
Definitely a reason why Nintendo could be interested in having the DLSS model* support HDMI 2.1 and VRR, unless Nintendo plans to use Nvidia G-Sync or AMD Freesync to support VRR via HDMI 2.0b, which the OLED model supports, at least for TV mode. However, adding support for VRR for handheld mode is more complicated, unless Nintendo decides to use a 1080p display that has a refresh rate of 120 Hz and supports VRR, similar to the displays used for the iPhone 13 Pro and the iPhone 13 Pro Max, which used customised 1080p displays.
Have they said anything about how Eldin Ring performs on last gen consoles? We might be able to gauge how an Eldin Ring port could perform on Switch 2.
Wonder why PS5 has a better framerate...
Click to expand...
Yeah I was gonna say, that kinda makes their credibility plummet pretty early on. The 2019 Lite stuff was fairly accurate so by that time it was likely much closer to first hand info for their source.Calling those Switch Lite rumors "fairly accurate" seems a bit generous. The 2018 one in particular seems pretty wildly off.
My speculation is that outside of Microsoft sending out Xbox Series X|S devkits much later than Sony has for the PlayStation 5 devkits, the PlayStation 5 having a single RAM configuration vs a dual RAM configuration of the Xbox Series X and the Xbox Series S in terms of RAM bandwidth could explain why Elden Ring runs at higher framerates on PlayStation 5 than on the Xbox Series X|S.Wonder why PS5 has a better framerate...
The Wii u did not have an soc. It had an IBM power pc cpu and AMD gputhe last time AMD supplied them with an SoC design, there was a big lack of documentation for its architecture. I doubt Nintendo would want another WiiU scenario on their hands.
To your first point, I doubt it matters anymore. It explained some of the disparity on launch window games.My speculation is that outside of Microsoft sending out Xbox Series X|S devkits much later than Sony has for the PlayStation 5 devkits, the PlayStation 5 having a single RAM configuration vs a dual RAM configuration of the Xbox Series X and the Xbox Series S in terms of RAM bandwidth could explain why Elden Ring runs at higher framerates on PlayStation 5 than on the Xbox Series X|S.
It had an mcm.The Wii u did not have an soc. It had an IBM power pc cpu and AMD gpu
A look back at this Korean leaker's track history (all machine translated):
2018-06-20, about Switch Lite
2019-02-23, about Switch Lite
2020-07-08, about Switch Pro CPU (Samsung proposal)
2020-10-28, about TSMC testing 7nm TX1
2021-06-06, about Pro performance with Nvidia chipset (and Game Pass!)
- The leaker claimed to have two sources, one in Japan and one in Korea.
- The Japanese source gave them fairly accurate info regarding Lite, one year before its announcement.
- But the "rumors in Japan" (not sure if same source) about Game Pass is wild.
- The info from their Taiwanese source has been all over the place—veracity unknown.
- If I have to guess, their sources might be from the assembler side, hence the detailed Lite specs but vague SoC info.
- Should there be a sliver of truth in the latest AMD rumor, I can think of a few explanations:
- The source conflated the development of a future Exynos (after 2200) with the new Switch SoC.
- It's a Samsung SoC with Nvidia GPU, not unlike Exynos 2200 with AMD GPU, but the message got distorted during the game of telephone.
- The "AMD" SoC is not for a new Switch model but a future Nintendo product (e.g., AR or "quality of life").
- Note that "AMD" doesn't automatically mean x86. Both Lisa Su and Devinder Kumar (CFO) indicated that AMD would build ARM chips if their customers demand it.
- Lastly, I don't know how accurate the machine translation is.
- "it doesn't use Nvidia chipsets": It may not be an "Nvidia" SoC, but it could be a custom Nintendo SoC that still includes Nvidia GPU cores.
- "it is said that [...] there is a high probability that the RDNA series will enter the GPU side": This sounds like a conjecture of their source.
- Of course it could all be bull, or overblown/inaccurate hearsay.
RSR is just FSR at driver-level. It isn't a DLSS competitor. Their DLSS competitor will probably debut alongside RDNA 3 products and is being tentatively-called FSR 2.0 right now. It has no AI/NN/ML components but is Temporal like UE5 Temporal Super Resolution and seemingly works by extracting the motion vector data from games. It's being said that it won't need developers intervention to work, but developers who want to add in their games will be free to do so, just like with FSR.AMD does have their upcoming DLSS-contender "RSR", but if Nintendo were to switch to AMD, game compatibility would probably be an issue.
I do hope it is driver-level or at least open-source enough for Lossless Scaling to add it somehow XDRSR is just FSR at driver-level. It isn't a DLSS competitor. Their DLSS competitor will probably debut alongside RDNA 3 products and is being tentatively-called FSR 2.0 right now. It has no AI/NN/ML components but is Temporal like UE5 Temporal Super Resolution and seemingly works by extracting the motion vector data from games. It's being said that it won't need developers intervention to work, but developers who want to add in their games will be free to do so, just like with FSR.
I see, thanks for the info! It's good that AMD is taking a less intervening method for upscaling. DLSS is great but it takes time for Nvidia to support a game.RSR is just FSR at driver-level. It isn't a DLSS competitor. Their DLSS competitor will probably debut alongside RDNA 3 products and is being tentatively-called FSR 2.0 right now. It has no AI/NN/ML components but is Temporal like UE5 Temporal Super Resolution and seemingly works by extracting the motion vector data from games. It's being said that it won't need developers intervention to work, but developers who want to add in their games will be free to do so, just like with FSR.
This Korean rumor being accurate would either mean essentially every other rumor we've heard is inaccurate or it's referring to another device.Why RNA2? I thought dev kits have already been sent out. Wouldn't it be too late to change SOCs?
And the nvidia job postings mentioning Orin and game consoles, were not for Nintendo.This Korean rumor being accurate would either mean essentially every other rumor we've heard is inaccurate or it's referring to another device.
Good point. There's always the chance Nvidia has another Shield but that's pretty solid evidence they're likely involved with a Nintendo product.And the nvidia job postings mentioning Orin and game consoles, were not for Nintendo.
I don't think the TDP of high-end consumer GPUs are necessarily an indication of how efficient a GPU architecture is, considering that Nvidia could have set the TDP of a particular high-end consumer GPU to 450 W to ensure performance competitiveness against a competitor's similar offerings.Reading rumors that the next Nvidia GPU’s are really thirsty ( low end needs a whopping 450w power supply) doesn’t bode well for that
there's currently no method of exposing motion data to the driver level. if there is a driver-level TAA equivalent, it'd be exclusive to AMD unless all vendors have access to the methodI see, thanks for the info! It's good that AMD is taking a less intervening method for upscaling. DLSS is great but it takes time for Nvidia to support a game.
The thing is: All architectures/SoCs have a perf/W curve. When you stay within that perf/W curve, consumption doesn't go out of the window. But if you go towards the performance extreme of the curve, consumption goes extra-high. That's what's happening with Nvidia next GPUs. They will need to clock them super high to combat AMD. Dane/Orin will stay within the lower bounds of this perf/W curve, in the sane power consumption clocks(Hell, knowing Nintendo they could even downclock it even more to save more energy and produce less heat). No need to worry about that.Only reason Nintendo would go with AMD is if Orins gaming performance is subpar. Reading rumors that the next Nvidia GPU’s are really thirsty ( low end needs a whopping 450w power supply) doesn’t bode well for that
That’s pretty much what everyone is saying. We are just discussing it due to the newest “leak”.Nintendo is NOT going to AMD, or ending their partnership with Nvidia. Stop it.
The point of the matter is AMD still screwed Nintendo over with the WiiU, at least compared to the deals they've given Microsoft and Sony, by handing them a rather outdated GPU. I don't even think it was GCN compatible, either. (To be fair, one could argue Nintendo screwed themselves over by sticking to Power PC, but Iwata was adamant on that backwards compatibility so...)The Wii u did not have an soc. It had an IBM power pc cpu and AMD gpu
AMD did not screw over Nintendo. The Wii U was never designed to compete with the PS4 and Xbox.The point of the matter is AMD still screwed Nintendo over with the WiiU, at least compared to the deals they've given Microsoft and Sony, by handing them a rather outdated GPU. I don't even think it was GCN compatible, either. (To be fair, one could argue Nintendo screwed themselves over by sticking to Power PC, but Iwata was adamant on that backwards compatibility so...)
Add to the fact that Nvidia also provides documentation and developer tools for the Switch, documentation which the Wii U sorely lacked, and you can see how going back to AMD would be another lose-lose situation. Sure, AMD then isn't AMD now...but what kind of good deals could they provide Nintendo that they wouldn't do for the other two competitors?
Dream onThe Wii U was designed first and foremost for computability with a dead architecture.
Nintendo could have gotten something off the shelf from amd, and gotten better performance per watt/ dollar.
No need to be passive agressive here. The Wii U uses MCM (something you claimed the SNES and PSX used but never did) which made it exceedingly expensive.Dream on
To clarify, this is based on what Chipworks told a few of us back in the NeoGAF days, when they supplied us with the die shots for the Wii U's CPU and GPU. I still have the original email, and the specific quote is:No need to be passive agressive here. The Wii U uses MCM (something you claimed the SNES and PSX used but never did) which made it exceedingly expensive.
@Z0m3le had previously noted the PS4 chips AMD supplied to Sony cost about as much as the Wii U MCM package did.
It sounds like you know this but there is no way that Nintendo would pick the architecture based on 3rd developers. The tail doesn’t wag the dog.The only feasible(But very, very unlikely) reason why Nintendo would switch(heh) to AMD would be if 3rd-parties developers asked them to have a common ground with others consoles.
But again, not only is that very unlikely, it doesn't even matter as everything is abstracted these days.
3rd-parties are very important to Nintendo(To the point they even change some hardware specifications due to feeback) and their share have been rising on Switch, but yeah. They wouldn't change architectures because developers asked them and frankly, that was just baseless conjecture from my part, as no developer would ask for that anyway.It sounds like you know this but there is no way that Nintendo do would pick the architecture based on 3rd developers. The tail doesn’t wag the dog.
Where have you seen any evidence that they're being run more "committee-like"? From everything I've seen they appear to be following the same tenets set out during the Iwata era.The odd factor here is that whatever the Switch successor is, it would be the first console without Iwata's involvement.
All the talks about Nintendo in the present being more committee like do have me concerned about the "DNA" that makes them unique. At the same time, it's a double-edged sword that has also driven the ire of critics who often call them "backwards" on a number of things. And this wouldn't be the first time, either.