• 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!)

Hi everyone! Didn't knew that you all had moved to Famiboards. Nice to see everyone!
Was doing some reading on AnandTech Apple A15 breakdown and one thing that impressed me was Apple 32MB SLC. What everyone think are the chances that Nvidia and Nintendo go for a full Arm implementation of L3 cache(8MB) instead of cutting down like Qualcomm and Samsung does? The biggest point against that would be area, given that Dane is 8nm based.
 
Hey everybody! Feels good to be back. Haven't posted on the other site for ages.

Anyway, here's hoping if Dane is Switch 2, Nintendo decides to use a 5nm node at least. 8nm is too damn old for 2022-2023 tech. Will be interesting to see if Nano Next is Dane, or perhaps a lite/lower wattage version of Dane, like for a Switch 2 lite model.:unsure:

I think since Steam Deck is coming next year on 7nm, 8nm for the next Switch should be similar to the difference with Ampere and RDNA 2 desktop cards. Question is will Nintendo and Nvidia opt for a higher docked TDP to have equal or better performance to the Steam Deck?
 
I think since Steam Deck is coming next year on 7nm, 8nm for the next Switch should be similar to the difference with Ampere and RDNA 2 desktop cards. Question is will Nintendo and Nvidia opt for a higher docked TDP to have equal or better performance to the Steam Deck?
they won't need to match match Steam Deck's TDP so much since Steam Deck will have to power through efficiency losses of the OS and Proton and whatnot (though that might still be better than running windows straight up). optimized-for-dane games can bleed out a little more performance
 
they won't need to match match Steam Deck's TDP so much since Steam Deck will have to power through efficiency losses of the OS and Proton and whatnot (though that might still be better than running windows straight up). optimized-for-dane games can bleed out a little more performance

Yeah that checks out, Steam Deck aside though it will be interesting to know what the base metric will be to meaningfully have a DLSS like solution work in low power form.
 
0
I'm not expecting a top of the line chip like 3nm for 2022-2023 release. Anything but 8nm, which would have made sense for 2021 release, but seems far past that now. I'm not expecting a top of the line chip like 3nm for 2022-2023, just something between 5 and 7. But we'll see.
I think TSMC's N6 process node or Samsung's 7LPP process node used for the Exynos 990, Snapdragon 765/765G/768G, etc., being used to fabricate Dane would be the best case scenario. However, kopite7kimi has been adamant on Dane being based on Samsung's 8N process node since last year. But I also think there's a possibility, albeit not a high possibility, that Dane may be fabricated using Samsung's 7LPP process node used for the Exynos 9825, especially if Nintendo and Nvidia has plans to refresh Dane, similar to how the Tegra X1 was refreshed, considering that there are very little differences between Samsung's 7LPP process node used for the Exynos 9825 and Samsung's 8LPP process node outside of the type of lithography being used.

I think since Steam Deck is coming next year on 7nm, 8nm for the next Switch should be similar to the difference with Ampere and RDNA 2 desktop cards. Question is will Nintendo and Nvidia opt for a higher docked TDP to have equal or better performance to the Steam Deck?
Just to ILikeFeet's post, Valve mentioned that there's no performance boosts when connecting the Steam Deck to a dock.
 
Yeah so far that's one of the few things that are more or less confirmed (as confirmed as a rumor can get).
Off-topic: the dad jokes get every single time lol


On-topic: the difference, aside from DLSS, with the SD and Dane, is that Dane has a docked boost mode that lets it go higher and it is a fixed scenario, SD would vary clocks and doesn’t need to dock for that, and battery life is a bigger issue for SD than Dane or even the GPD Win 4(speculation on my part).
 
Off-topic: the dad jokes get every single time lol


On-topic: the difference, aside from DLSS, with the SD and Dane, is that Dane has a docked boost mode that lets it go higher and it is a fixed scenario, SD would vary clocks and doesn’t need to dock for that, and battery life is a bigger issue for SD than Dane or even the GPD Win 4(speculation on my part).
portable pcs have teh flexibility to limit TDP if battery is a priority. though for the demographic, battery life is the first thing chucked out of the window. however, if devs do bother making a "steam deck" quality setting, then Dane would only benefit
 
portable pcs have teh flexibility to limit TDP if battery is a priority. though for the demographic, battery life is the first thing chucked out of the window. however, if devs do bother making a "steam deck" quality setting, then Dane would only benefit
In those scenarios the CPU gets pretty limited, in which Dane would edge out mult-core perf-wise.

in either case Dane sits comfortably next to the SD and whatever GPD Win 4 ends up (again, speculating)
 
0
I'm not expecting a top of the line chip like 3nm for 2022-2023 release. Anything but 8nm, which would have made sense for 2921 release, but seems far past that now. I'm not expecting a top of the line chip like 3nm for 2022-2023, just something between 5 and 7. But we'll see.

Yeah it makes sense it would be a binned version of Dane.
I don't know the feasibility of Samsung or TSMC (does TSMC have a 6nm?) 6nm.

If not 5nm, they can go to Samsung 7nm, which while not having too much improvement density/efficiency-wise over 8nm, does jump the DUV-EUV gap allowing Dane to be die-shrunk to 5nm or 3nm. etc in the future without completely redesigning the SoC.
 
Yes, they do have. 18% higher logic density compared to N7 DUV and same design rules and toolset(So you can port a 7nm DUV design like PS5/Series SoC or Zen/RX 6000 CPU/GPUs to TSMC N6 without needing to re-design everything).
Well I think I know what they will use for the PS5 Slim and the Series X"2"

Have we actually considered the plausibility of them making Dane for Samsung 8nm, then porting it to TSMC 6NM? Which likely would be an Efficiency/Density boost of above 20% at least? As Samsung 8nm is less Efficient/Dense than TSMC 7nm?
 
Well I think I know what they will use for the PS5 Slim and the Series X"2"

Have we actually considered the plausibility of them making Dane for Samsung 8nm, then porting it to TSMC 6NM? Which likely would be an Efficiency/Density boost of above 20% at least? As Samsung 8nm is less Efficient/Dense than TSMC 7nm?
It makes more sense that a silent revision for PS5/Series will use TSMC N6 rather than Slim because N6 does not offer performance or power-efficiency increases compared to N7 DUV. Slim consoles usually are meant for form-factor reduction and energy-efficiency, thus allowing them to drive costs down. If there's a Slim model for both, it will use TSMC N5 imo.
Regarding your second point, given that Samsung and TSMC are different fabs, too much hassle. You need to see what parts of the IP are owned by the fabs, you would need a new and separete team to take the design and implement on the new fab, etc, etc, etc. It makes more sense that you will continue to manufacture within the same company(Either Samsung or TSMC).
 
0
IMO Dane being on Samsung 8N/LPA/U process is a rather smart move by Nintendo, regardless if they're launching next year or in 23. TSMC N5/Samsung 5N output are going to be taken by multiple companies for the foreseeable future and both processes are going to be long-lived. A die-shrunk Dane can be done at Samsung 5LPP/4LPP in 2025/6, when costs are down, yield are much better and capacity is freed up as companies move to 3nm/2nm processes.
 
IMO Dane being on Samsung 8N/LPA/U process is a rather smart move by Nintendo, regardless if they're launching next year or in 23. TSMC N5/Samsung 5N output are going to be taken by multiple companies for the foreseeable future and both processes are going to be long-lived. A die-shrunk Dane can be done at Samsung 5LPP/4LPP in 2025/6, when costs are down, yield are much better and capacity is freed up as companies move to 3nm/2nm processes.
Well problem is DUV and EUV.

A DUV Desgin can't be ported to a EUV Process.
So to port Dane to anything Samsung 7nm or smaller, they would have to outright redesign it for EUV.

So if they want to get a die-shrunk Dane out at some point without pretty much going through the hassle of re-designing Dane, a hassle that could've been put to just making the next SoC for the Switch 3/Switch 2 Pro.etc, they need to make Dane on a EUV process and launch it on that.
 
Last edited:
Have we actually considered the plausibility of them making Dane for Samsung 8nm, then porting it to TSMC 6NM? Which likely would be an Efficiency/Density boost of above 20% at least? As Samsung 8nm is less Efficient/Dense than TSMC 7nm?
I think it depends on how TSMC's N6 process node is priced and if Nintendo and Nvidia can secure enough capacity from TSMC for TSMC's N6 process node.
 
0
Well problem is DUV and EUV.

A DUV Desgin can't be ported to a EUV Process.
So to port Dane to anything Samsung 7nm or smaller, they would have to outright redesign it for EUV.

So if they want to get a die-shrunk Dane out at some point without pretty much going through the hassle of re-designing Dane, a hassle that could've been put to just making the next SoC for the Switch 3/Switch 2 Pro.etc, they need to make Dane on a EUV process and launch it on that.
When you're doing a die-shrunk, you're already redesigning things. I don't see it as a hassle but more of a normal thing. Tegra X1 went from TSMC 20nm planar to TSMC 16/12nm FF(FinFET). The Exynos 9820 was ported from Samsung 8LPP DUV to Samsung 7LPE EUV(Exynos 9825). Regarding the next SoC point, we need to remember that Atlan is years off(2025 target), Switch Next Next would be 26/27. Nintendo/Nvidia can do Dane in Samsung 8LPU/A, die-shrunk to Samsung 7LPP/5LPP and launch the next next SoC using TSMC/Samsung 4/5nm process.
 
When you're doing a die-shrunk, you're already redesigning things. I don't see it as a hassle but more of a normal thing. Tegra X1 went from TSMC 20nm planar to TSMC 16/12nm FF(FinFET). The Exynos 9820 was ported from Samsung 8LPP DUV to Samsung 7LPE EUV(Exynos 9825). Regarding the next SoC point, we need to remember that Atlan is years off(2025 target), Switch Next Next would be 26/27. Nintendo/Nvidia can do Dane in Samsung 8LPU/A, die-shrunk to Samsung 7LPP/5LPP and launch the next next SoC using TSMC/Samsung 4/5nm process.
Yeah but Samsung themselves made the Exynos and the process so that is a bit of an undeserved comparison point.

And yeah they can redesign it, but redesigning isn't the norm, usually, die-shrinking is not too complicated most of the time.

DUV to EUV requires a fundamental looking at how the SoC is designed as the lithography is different.

So again, it would be likely just better for them to use the time to just make a "Dane 2/Dane +" on 5nm from scratch with a better GPU/CPU Config rather than waste the time to redesign Dane as-is.

Because cost-wise, it would cost about as much to do that due to the DUV-EUV switch.
 
Looking at previous generations of console lifespans and sales of switch. I would say 2023 is the earliest we will get a new switch. looking back many consoles have had a lifespan of 6-7 years. Ned Nintendo saying that Switch will have a Longer Life Cycle Than Previous Consoles. It is up in the air if that means that Switch will be around long after a new version is out or it means we will have to wait longer the previous gens to get a new console.
 
0
Yeah but Samsung themselves made the Exynos and the process so that is a bit of an undeserved comparison point.

And yeah they can redesign it, but redesigning isn't the norm, usually, die-shrinking is not too complicated most of the time.

DUV to EUV requires a fundamental looking at how the SoC is designed as the lithography is different.

So again, it would be likely just better for them to use the time to just make a "Dane 2/Dane +" on 5nm from scratch with a better GPU/CPU Config rather than waste the time to redesign Dane as-is.

Because cost-wise, it would cost about as much to do that due to the DUV-EUV switch.
I think you're ovestating the whole DUV-EUV. It's isn't uncommon. See the Exynos example, Ryzen 5000 -> 6000(Rembrandt), etc. Yes, the tools change, your library and design rules aren't the same, but it's totally feasible and I quite doubt it's as expensive as 5nm process are right now. Like, I would also like to see Dane on newer EUV based process, but Nvidia/Nintendo choosing 8N isn't a bad choice when you consider the current enviroment.
 
0
Sorry for thread spamming. I'm still getting a little accostumed with this forum tools and didn't knew that you couldn't add a reply while editing.
I actually wonder how much of that is due to Exynos 9825 being a simply die-shrink rather than a design from the ground-up for Samsung 7LPP EUV. Because 7LPP and 8LPP are vastly different processes. Would be amazing to have more public info regarding this but alas.
edit: Ok, so was browsing for more info regarding the Exynos 9820 -> xx25 and it seems Samsung 7N has lightly usage of EUV. And were still using the same machinery that they used for the 8N process for 7N FEOL. That's quite interesting. I think will research more on the whole Samsung 7N x 8N topic.
 
0
die-shrinking is not too complicated most of the time.
we are way past the days of "32nm > 28nm". any sort of non "enhanced same node" is asking for more upfront. not impossible, and I expect Nintendo to actually do it, but things aren't so simple as before
 
0
I am total noob so I apologize if I get obvious things wrong with my question or if the answer was already given or if there is no answer to it.

Since we assume that Dane will be on a 8nm node and that the Apple A12 from 2018 is on a 7nm node, can we compare both? If yes, which chip would come out on top? Also, how far would the A12 soc would be from the one in the PS4?

I am sorry if I sound like an absolute idiot. But from what I gather from other sources online is that all GFlops are not equal and that core count doesn't tell the whole story. Hence, I imagine that comparing two mobile chips which sip little power with comparable nodes is a good starting point for me to start grasping things.
 
They will be behind what you think they should have. I feel confident saying that.
Yeah, outside of RAM, historically they have been behind in tech by at least 2 years. But considering x series x/s and PS5 are using 7nm nodes and they came last year in Q4 2020, would be weird to release 8nm in 2022/23, but w/e. 7nm Steam deck for next year was mentioned after my post.

At least Nintendo doesn't need 16GB LPDDR5 RAM. 8-12 would be perfect and 88-102GB/s bandwidth at least. I think they can afford to match the GPU and add DLSS, but with lower clocks for CPU. If they can go focus more in power and aim for V1 switch battery life (3-5 hrs), and can top out at 15-20 watts in docked, that would be fantastic. Then for the revision using a 5nm node or less, they can go back to matching V2 and OLED battery life. That's my hope at least. 720p or 900p screen would be fine to me.


Speaking of Steam Deck, anyone known to have done head to head performance comparisons of games on it vs base ps4? With Steam deck, it's hard to compare what graphic settings (high, medium, low) would be comparable to ps4. Still would be interesting to find the closest match as well as framerate differences.
 
Yeah, outside of RAM, historically they have been behind in tech by at least 2 years. But considering x series x/s and PS5 are using 7nm nodes and they came last year in Q4 2020, would be weird to release 8nm in 2022/23, but w/e. 7nm Steam deck for next year was mentioned after my post.

At least Nintendo doesn't need 16GB LPDDR5 RAM. 8-12 would be perfect and 88-102GB/s bandwidth at least. I think they can afford to match the GPU and add DLSS, but with lower clocks for CPU. If they can go focus more in power and aim for V1 switch battery life (3-5 hrs), and can top out at 15-20 watts in docked, that would be fantastic. Then for the revision using a 5nm node or less, they can go back to matching V2 and OLED battery life. That's my hope at least. 720p or 900p screen would be fine to me.


Speaking of Steam Deck, anyone known to have done head to head performance comparisons of games on it vs base ps4? With Steam deck, it's hard to compare what graphic settings (high, medium, low) would be comparable to ps4. Still would be interesting to find the closest match as well as framerate differences.
That’s the thing. This is Nintendo. I know people say you shouldn’t say that. I’m not dogging them. It just seems they always find a way to do something that makes no sense. That jumanji meme of “what year is it”
 
0
Yeah, outside of RAM, historically they have been behind in tech by at least 2 years. But considering x series x/s and PS5 are using 7nm nodes and they came last year in Q4 2020, would be weird to release 8nm in 2022/23, but w/e. 7nm Steam deck for next year was mentioned after my post.
what option do they have that won't blow out the budget? Nvidia always planned for this to be 8nm. maybe if they have though about using someone in addition to Samsung from the start, but we'd probably hear about that by now.
 
what option do they have that won't blow out the budget? Nvidia always planned for this to be 8nm. maybe if they have though about using someone in addition to Samsung from the start, but we'd probably hear about that by now.
But we are just basing this off of what Kopite7 said from Twitter, and not directly from Nvidia. Plans change all the time, which can make information outdated pretty fast. I'm not betting it won't be, but hoping it isn't.

And Tron1 you are right it would be a Nintendo move to do so. We'll see.
 
But we are just basing this off of what Kopite7 said from Twitter, and not directly from Nvidia. Plans change all the time, which can make information outdated pretty fast. I'm not betting it won't be, but hoping it isn't.

And Tron1 you are right it would be a Nintendo move to do so. We'll see.
I’m not sure if plans change all the time when it comes to decisions like that, since they would have to scrap a lot of work and probably delay if they change node.
 
0
But we are just basing this off of what Kopite7 said from Twitter, and not directly from Nvidia. Plans change all the time, which can make information outdated pretty fast. I'm not betting it won't be, but hoping it isn't.
you can't just change what node you're gonna be manufacturing on. especially if the design doesn't exist. that's like multiple years of work. and depending on how close you are to launch, manufacturing contracts would have to be written up months or a year (maybe more) in advance
 
But we are just basing this off of what Kopite7 said from Twitter, and not directly from Nvidia. Plans change all the time, which can make information outdated pretty fast. I'm not betting it won't be, but hoping it isn't.
Why would Nvidia need to comment on a product that hasn't been finalised yet, considering Dane's rumoured to not have been taped out yet? And I don't think companies can very easily change which process node to use for fabricating chips very quickly, especially nowadays with demand outpacing supply to the point where lead times are extended.
 
0
Feeling late this year because the Cars that Orin is supposed to be in are entering production now and are slated to release Early-Mid 2022 iirc.
If that's the case for Dane too, which hopefully is, we may be getting more concrete info on it by maybe December then. Though I guess we may get something nice to discuss once Nate's podcast is up and when Thugs comments on what he has heard. At least one of those things will probably happen next week, so we will have some info soon to think about.
 
0
I think TSMC's N6 process node or Samsung's 7LPP process node used for the Exynos 990, Snapdragon 765/765G/768G, etc., being used to fabricate Dane would be the best case scenario. However, kopite7kimi has been adamant on Dane being based on Samsung's 8N process node since last year. But I also think there's a possibility, albeit not a high possibility, that Dane may be fabricated using Samsung's 7LPP process node used for the Exynos 9825, especially if Nintendo and Nvidia has plans to refresh Dane, similar to how the Tegra X1 was refreshed, considering that there are very little differences between Samsung's 7LPP process node used for the Exynos 9825 and Samsung's 8LPP process node outside of the type of lithography being used.


Just to ILikeFeet's post, Valve mentioned that there's no performance boosts when connecting the Steam Deck to a dock.

I think people's fears when it comes to the 8nm process is from what's currently on the market with Ampere cards. Samsung seems adamant on continuing to further develop the process (which could only yield better density/efficiency gains that are much closer to TSMC's 7nm DUV process) in the end. A mobile Orin/Dane SoC using 8nm though can be scalable to meet both demands of being efficient enough in handheld, while also having room to get decent performance docked...

Side note: How come there's minimum information on GA107 die size/density specs? This would probably give us a good idea of what Orin could look like(minus the Ai added logic of course).

Yeah but Samsung themselves made the Exynos and the process so that is a bit of an undeserved comparison point.

And yeah they can redesign it, but redesigning isn't the norm, usually, die-shrinking is not too complicated most of the time.

DUV to EUV requires a fundamental looking at how the SoC is designed as the lithography is different.

So again, it would be likely just better for them to use the time to just make a "Dane 2/Dane +" on 5nm from scratch with a better GPU/CPU Config rather than waste the time to redesign Dane as-is.

Because cost-wise, it would cost about as much to do that due to the DUV-EUV switch.

I definitely think the level of importance for Nintendo and Nvidia using 8nm are not only in having a mature manufacturing process, but also establishing a starting place of meeting such a high level demand(with good yields) for "Dane" Switch production numbers (with Samsung probably being very aggressive to keep that volume after they see it take place).
But we are just basing this off of what Kopite7 said from Twitter, and not directly from Nvidia. Plans change all the time, which can make information outdated pretty fast. I'm not betting it won't be, but hoping it isn't.

And Tron1 you are right it would be a Nintendo move to do so. We'll see.
My guess is if something changes, Kopite or Greymon55 will be the first to notify the public about it...
 
Probably because GA107 is only used in laptops. And I imagine separating GA107 from the motherboards of laptops would be extremely difficult without destroying the laptops in the process. Fortunately, kopite7kimi does offer a rough estimation of GA107's die size.
GA106 and 104 mobile variations both have their specs out in the wild, but going by Kopite's 190+mm2 die size does give some perspective on possibilities (if they can achieve better transistor density with Dane targeting lower power draw over the Ampere mobile/desktop parts).
 
people are just jumping into UE5. I expect Dane Switch to be way more optimized for tools like UE. there seems to be a limitation in the current switch that really hobbles UE4's potential unless you got time to root around the engine

 
0
GA106 and 104 mobile variations both have their specs out in the wild, but going by Kopite's 190+mm2 die size does give some perspective on possibilities (if they can achieve better transistor density with Dane targeting lower power draw over the Ampere mobile/desktop parts).
I don't think there are architectural differences between the desktop and laptop variants of GA104 and GA106. My impressions is that laptops are using the highly binned variants of GA104 and GA106 targeted towards a certain amount of performance under a strict set of TDPs.
 
0
I am total noob so I apologize if I get obvious things wrong with my question or if the answer was already given or if there is no answer to it.

Since we assume that Dane will be on a 8nm node and that the Apple A12 from 2018 is on a 7nm node, can we compare both? If yes, which chip would come out on top? Also, how far would the A12 soc would be from the one in the PS4?

I am sorry if I sound like an absolute idiot. But from what I gather from other sources online is that all GFlops are not equal and that core count doesn't tell the whole story. Hence, I imagine that comparing two mobile chips which sip little power with comparable nodes is a good starting point for me to start grasping things.
From what I remember when Apple was talking about the iPad Pros at the time, the A12X in them was around Xbox One level of gpu performance. The A12 had half the cores of the A12X.
I don't think the chips will be designed the same though.
 
Cool little blog that makes Mesh Shaders more clear to understand:



Ampere, Turing, RDNA2, Lovelace, RDNA3 and later feature mesh shaders.

If anyone very knowledgeable about this can comment on a thing I wonder: What is the difference between primitives and mesh shaders? Mesh is an evolution of primitives, right?




Should be pretty useful nonetheless for a device like Dane, for those that don’t feel like answering the top question, this has to be implemented and incorporated into game engines to support the feature.



Funny how we go back to the PS2 after all these years ;P few will get this reference and joke
 
From what I remember when Apple was talking about the iPad Pros at the time, the A12X in them was around Xbox One level of gpu performance. The A12 had half the cores of the A12X.
I don't think the chips will be designed the same though.
Thank you! Alright then. According to sources, the A12X has a die size of 122 m2 slightly bigger than 118m2 of the Switch Soc. Also, it should have a power draw of 15W according one of the sources I found (although this could not be corroborated by others) and by the time Dane launches, five years would have passed since the launch of the A12X.

So reasonably, before even touching questions such as costs and such, can we assume that Dane can compete with the A12X? If yes, it would be outstanding because we would be having the capability of a Xbox One S in the Switch Form factor. I hope that would be enough power for the devs as well.
 
Thank you! Alright then. According to sources, the A12X has a die size of 122 m2 slightly bigger than 118m2 of the Switch Soc. Also, it should have a power draw of 15W according one of the sources I found (although this could not be corroborated by others) and by the time Dane launches, five years would have passed since the launch of the A12X.

So reasonably, before even touching questions such as costs and such, can we assume that Dane can compete with the A12X? If yes, it would be outstanding because we would be having the capability of a Xbox One S in the Switch Form factor. I hope that would be enough power for the devs as well.
No, sadly. Couple caveats to consider.

A12X was on the 7nm process at the time, Dane will be on 8N, which is supposed to be an improved version of 8LP(x). Node wise, it is already outclassed in logic.

CPU that the A12X has is clocked pretty high, and it curbstomps the Dane model CPU, it being 150% the multi-core performance that the Dane model would ever be. Unless Nintendo goes ballsy and clocks it pretty high, or just high enough to at least trade decent blows with the A12X. Single core score? Forget it. Though, this doesn’t really matter to be honest on a game console since they are always doing multi-threaded applications anyway. Single core matters more for phone things

GPU-wise, well this part I do expect it to be pretty competitive, but keep note its on a lesser node (presumably). What the iPad does is on a more efficient process, and does so at 15 watts portably while the switch would not be using 15W in portable, likely around half of that. But it doesn’t matter since the GPUs in both are better than the GPU in the XB1 on paper specs.

The Dane will likely use a 4310mAh battery, while the iPad can use something like twice of this right? It’s less constrained by the power draw and can power higher yet still manage the battery life of the Mariko unit on the longer end.


HOWEVER! A couple points of contention, the iPad is not meant for long gaming sessions unlike the Dane model. It has no active cooling, GPU being better on paper does not mean it’s better in real life performance, the thing can throttle down, though mobile devices have gotten pretty good with this regard.

Dane will be at a unique place, it will be more modern yet old at the same time and have a prettty heavy lean on the GPU side with ML hardware that allows it to push well above us weight.


Nintendo can possibly have a denser battery for the same size, like in the Samsung phones, that allows them to have a long battery life, slightly higher clocks which makes it more performant before the docked mode boost.

Current mariko aims to draw only 6-7W I think, Erista drew like 9-11W or so portably. iPad draws 15W to perform its tasks (if I understood what you said correctly) at peak performance, but likely clocks down for energy saving reasons. Dane doesn’t need to do that clock switching, it’s a constant device.


It’s why I do hope they lean more onto the CPU being clocked modestly well at the sake of the GPU speeds. Should tide them over until the next platform rears it’s ugly head and we have the 7th speculation thread ;)


A main takeaway is that this leans more into Apples to Oranges comparison.
 
Last edited:
Cool little blog that makes Mesh Shaders more clear to understand:



Ampere, Turing, RDNA2, Lovelace, RDNA3 and later feature mesh shaders.

If anyone very knowledgeable about this can comment on a thing I wonder: What is the difference between primitives and mesh shaders? Mesh is an evolution of primitives, right?




Should be pretty useful nonetheless for a device like Dane, for those that don’t feel like answering the top question, this has to be implemented and incorporated into game engines to support the feature.



Funny how we go back to the PS2 after all these years ;P few will get this reference and joke
In a way, mesh shaders and primative shaders aim to accomplish a similar task (triangle efficiency). One of which is through culling. And unlike primative shaders, mesh shaders is now standardized.

This is about as far as I understand it, anyway. Also, UE5's nanite is similar to mesh shaders, but takes a different path

Still really like this demo

 
Last edited:
In a way, mesh shaders and primative shaders aim to accomplish a similar task (triangle efficiency). One of which is through culling. And unlike primative shaders, mesh shaders is now standardized.

This is about as far as I understand it, anyway. Also, UE5's nanite is similar to mesh shaders, but takes a different path
I don’t believe ps5 actually has mesh shaders. They have some custom “geometry engine” with similar functionality.
 
Please read this staff post before posting.

Furthermore, according to this follow-up post, all off-topic chat will be moderated.
Last edited:


Back
Top Bottom