StarTopic Future Nintendo Hardware & Technology Speculation & Discussion |ST| (New Staff Post, Please read)

Reader mode
Dekuman said:
@MattAgain can you verify this guy is full of shit? Claims to have access to 'old' Switch 2 devkit, had Paper Mario 6 installed, claims it only outputs 1080p docked/portable, no DLSS, C button is to close software.

Click to expand...

he's probably full of shit, but to clarify, he says the unit is 1080p handheld and is only outputting 1080p docked because his dev kit is using a Switch 1 dock.
 
Dekuman said:
@MattAgain can you verify this guy is full of shit? Claims to have access to 'old' Switch 2 devkit, had Paper Mario 6 installed, claims it only outputs 1080p docked/portable, no DLSS, C button is to close software.

Click to expand...

It's so bad. The idea of C opening a Wii-like HOME Menu isn't terrible, perhaps it could switch between software like Quick Resume.

But I don't believe this for a second.
 
I might be wrong, but I'd also expect it to say 'Ounce' instead of 'BEE' since the latter is a product code, the former is the alias of the NS2.
 
Letā€™s all just agree to ignore the obvious troll post

Nintendo isnā€™t leaving a fully completed Paper Mario game on a dev kit, be for real
 
Dekuman said:
@MattAgain can you verify this guy is full of shit? Claims to have access to 'old' Switch 2 devkit, had Paper Mario 6 installed, claims it only outputs 1080p docked/portable, no DLSS, C button is to close software.

Click to expand...

Why would you have a dedicated button to close software when the Home button already suspends it very well?
 
Dekuman said:
@MattAgain can you verify this guy is full of shit? Claims to have access to 'old' Switch 2 devkit, had Paper Mario 6 installed, claims it only outputs 1080p docked/portable, no DLSS, C button is to close software.

Click to expand...

(Mods, please do not delete this post, i cannot share hardware photos cuz of glossy screen that might be bad for anonymity)
Click to expand...
Now call me crazy but I think someone able to get their hands on something like a dev kit would be intelligent enough to take a picture at an angle and then check the picture to make sure their face isn't in the picture.
 
Serif said:
I might be wrong, but I'd also expect it to say 'Ounce' instead of 'BEE' since the latter is a product code, the former is the alias of the NS2.
Click to expand...
No kidding. Dev units of Nintendo Switch didn't even HAVE HAC on them. BET, ONC, MUJ, etc. would be what I'd expect on a Switch 2 dev system product code. I think we'll find out around launch, much like HAT.
 
NZZN said:
Taiwanese Commercial Times reports NVIDIA Thor is on TSMC 4N, which would imply Tegra as a whole isn't sticking with Samsung.
Important to note that automotive applications will value energy efficiency much more than low cost.

SJvUGxi.png

https://autotech.news/nvidias-drive-thor-chips-set-for-robust-growth-2025/
https://www.ctee.com.tw/news/20241202700027-430501
Click to expand...
Commercial Times is wrong about Orin using TSMC's 7 nm* process node.

Nvidia said Orin is using Samsung's 8 nm* process node.

And the first letter in the second row of the labeling on the Orin X die confirms Samsung is used for fabrication.
PdUurgaY8p.png

* ā†’ a marketing nomenclature used by all foundry companies
miakx said:
So, according to this, and the picture earlier of a Thor SoC, the S letter would not mean necessarily it is a Samsung fab? Could change some things over all the speculations of the leak of Switch 2 mobo
Click to expand...
The first letter on the second row of labelling in the Thor die implies TSMC is used for the fabrication.
Ggx007la4AEFnte
 
Crandriel said:
NVIDIA has stated that, in addition to requiring 4 times more training time, the neural network has twice the number of parameters.


Knowing this, it is understood that twice the amount of time is needed to obtain the output from the neural network, or in other words, double the power is required to get the response in the same amount of time.
Click to expand...
They said it takes 4x as much compute at inference. So quadruple, not double.
Dekuman said:
C button is to close software.
Click to expand...
Ahahaha.
 
Hartmann said:
The physical mockup might have been created solely to match the shape.
In the Genki mockup in commercial movie, it is shown with a bezel-less display.
gxeYr9Y.jpg

Click to expand...

Your understanding is correct.
 
oldpuck said:
Assets might be compressed a couple of times on the cartridge. Once in a lossy format - that's where you sacrifice texture quality in order to get a really small file size. Then again in a lossless format - that's where you don't sacrifice any quality, but you don't get great compression either. You do this to get as many assets on the limited space of the cartridge as possible.

You need to do the lossless decompression before you can use the asset. But GPUs natively speak some lossy compressed formats. So you don't need to decompress those. Which is good, because otherwise you'd only fit a few assets in VRAM at a time.

Neural texture compression essentially gives each individual texture a custom format. Obviously the GPU doesn't speak those, so you need a special neural shader to decompress them.


The FDE speaks the lossless compression formats natively. So during loading times, the assets can be read off cartridge extremely quickly, with the FDE decompressing them, instead of the CPU having to do the work.

As best as we can tell, the FDE doesn't do anything special when it comes to lossy compression.
Click to expand...
at the risk of pulling us into uninteresting minutia, I'm curious to know what sorts of lossy and lossless compression algorithms have been used and what their relation is to more familiar user-facing compression schemes like JPEG/MPEG, LZMA, etc. could you point me to any resources you have on this topic?
 
Crandriel said:
I've said it very clearly. Whether or not it is the most efficient point of Odin is fine, but that is not the frequency that Nintendo has chosen, nor is the most efficient point the criterion that Nintendo has taken to choose the frequency at which the console will go. The point that it has chosen is to maintain exactly the same decibels in the fan that the switch 1 v1 had in its day.

I can't say it any clearer.
Click to expand...
It's very possible that the handheld clock speed is lower than 470MHz, and it's very possible that you have knowledge of this.

But if it is, it's because the peak efficiency clock of T239 is lower than Orin. Nintendo and Nvidia did not design a chip to run it below the peak efficiency clock. Full stop. And while there are people out there with knowledge of the clock speeds, no one but a select few at Nintendo and Nvidia know the peak efficiency clock of T239. It's not knowledge one could have just by having a dev kit.
 
darthdiablo said:
What is that image I'm looking at? I'm not sure what is being shown
Click to expand...
I would say supposedly a Switch 2 with the screen on. Look at the joycons and curves.
Bezel looks bigger than the rendering, more close to the 3D print from Genki. See the gap between display on and joycons.
 
lanamoon said:
Your understanding is correct.
Click to expand...

Y & X are in the correct spots. Must be real.

Anyway, to me that looks kinda like what the NHH person was trying to explain on reddit a while back. Maybe this is another of theirs that was shared privately. (though by now I'm sure others have access in some form)

 
If it's real, it's an 8" and not an 8.4" display and it would pretty much match with the Genki 3d print screen area.

At this point it's more realistic it's real than not with the floodgates open too šŸ˜…
 
thegodsend said:
If it's real, it's an 8" and not an 8.4" display and it would pretty much match with the Genki 3d print screen area.

At this point it's more realistic it's real than not with the floodgates open too šŸ˜…
Click to expand...
Yeah, I don't believe a 8.4 rumor. That I don't think would reconcile with what we saw with previous PCB leaks (there being USB on bottom). The 8.4 render is with no USB on bottom right?

Edit: And yeah, I copied that image Lanamoon shared in the event the post is ninja-edited later. Not sure if it's anything real, so just in case.
 
Crandriel said:
I've said it very clearly. Whether or not it is the most efficient point of Odin is fine, but that is not the frequency that Nintendo has chosen, nor is the most efficient point the criterion that Nintendo has taken to choose the frequency at which the console will go. The point that it has chosen is to maintain exactly the same decibels in the fan that the switch 1 v1 had in its day.

I can't say it any clearer.
Click to expand...
What is your proof to back up this claim?
 
0
Escuse me What?!
We got an image of a switch 2 turned on.
Nintendo buddy, Friend whats going on?
How is this system the most leaked video game system of all time.
Its Nintendo! They are crazy when it comes to this stuff.
 
Hosermess said:
Y & X are in the correct spots. Must be real.

Anyway, to me that looks kinda like what the NHH person was trying to explain on reddit a while back. Maybe this is another of theirs that was shared privately. (though by now I'm sure others have access in some form)

Click to expand...

ah yes, .json scripts.

It does seem to match the general description of internal diagnostic software, though.
 
Factory setup information from Switchbrew:

Factory Setup - Nintendo Switch Brew

switchbrew.org switchbrew.org
At the factory, a minimal version of the Switch OS is installed. A modified version of the boot2 title (boot2.manuBoot) is installed that launches an additional "Manu" sysmodule, and the system config title specifies to launch "Test Application Launcher" instead of qlaunch.
Test Application Launcher is used to launch a number of tests, "CAL0" calibration data is written to NAND, and retail firmware is installed.
Click to expand...
0100000000002000BoardFunctionBoard testing.
Click to expand...
1000000000000001SystemInitializStrings internally refer to this as "SystemInitializer". See here.
Click to expand...
 
darthdiablo said:
Haha. While I don't believe the 8.4 screen - to be fair it's possible they are measuring the thickness of Joycon with buttons extending into the background. Not flush with the back of the tablet.
Click to expand...
Doesn't make sense either way. They did not change dimensions of the NS2 depending on the buttons extending, which now makes it look like the NS2 will be 3x thicker than the OG Switch, looking at their chart.
 
Reader mode
You must log in or register to reply here.
Please read this new, consolidated staff post before posting.

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


Back
Top Bottom