• 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.
  • Do you have audio editing experience and want to help out with the Famiboards Discussion Club Podcast? If so, we're looking for help and would love to have you on the team! Just let us know in the Podcast Thread if you are interested!

StarTopic Future Nintendo Hardware & Technology Speculation & Discussion |ST| (Read the staff posts before commenting!)

Reader mode
mjayer said:
Ya.. it's in there.. at the end.. in a list. So technically you could say they said that, but I still don't believe it was a marketing message they pushed. Too me, this is similar to the PS5 having 8k slapped on the box.
Click to expand...
They list it on their site as a 1440p device.
 
mjayer said:
Ya.. it's in there.. at the end.. in a list
Click to expand...
That list is literally the only information about the console in the launch trailer.

mjayer said:
. So technically you could say they said that, but I still don't believe it was a marketing message they pushed. Too me, this is similar to the PS5 having 8k slapped on the box.
Click to expand...
What are you talking about? I don’t mean that facetiously. Why is it relevant if 1440p was part of the marketing message - and why is every single marketing message insufficient? 1440p support was mentioned as often as ray tracing in the initial marketing and in the same way.

The core of the series S message at launch was “for players without a 4K TV.” The specs were on the surface the same as the One X, and the core driver of the Series X was 4K support. Because of that they mentioned 1440p support constantly, to differentiate it from its predecessor (1080p) and its big brother (4K).

Anecdotally, “series s supports 2k” was a thing I knew, and I didn’t play games at all when it came out
 
0
Terrell said:
And, as far as storage density, there is still room for these ROM chips to be manufactured on a smaller non-FinFET process node, which will inherently increase storage while maintaining similar costs to existing chips but with larger capacities. The ROM chips used in Game Cards were manufactured at 45-48nm at launch (and maybe still are), but they can go as low as 28nm without raising costs, meaning a potential 71% density improvement at the same price they're paying for the current chips at their current available storage sizes (and also meaning, if a game does not require more storage, an incredibly significant cost savings per Game Card).
Click to expand...
XtraROM is currently manufactured at 32nm process.
在ROM唯讀記憶體部份，我們已開始量產32奈米XtraROM®產品。
Click to expand...

旺宏電子 - 公司簡介

www.mxic.com.tw www.mxic.com.tw
 
Hartmann said:
XtraROM is currently manufactured at 32nm process.

旺宏電子 - 公司簡介

www.mxic.com.tw www.mxic.com.tw
Click to expand...
Dakhil said:

Macronix - Company Overview

www.mxic.com.tw www.mxic.com.tw
Click to expand...
It's important to recall that Nintendo is not Macronix's only customer.
The change in process node happened in 2019, 2 years after Switch's launch. And without disassembling modern Game Cards and checking the chip sizes in comparison to titles shipped in those first 2 years, we have no means to confirm that Nintendo is ordering chips on that new process node. If Nintendo struck the deal with Macronix in 2016 to supply ROM chips for Game Cards, it was done on either then 45 or 48nm process node and, depending on the contract terms, a die shrink may not have been on the table when the new process node rolled out and was intended for another customer. I even requested help identifying the planar node from this inside look at the ROM chip in a launch title, but no one seemed able to assist.

Additionally, even if Nintendo is using the new 32nm process node, a shrink to 32nm would not provide the means to achieve the necessary storage density for 64GB Game Cards without increasing the physical size of the Game Card itself (since they take up almost all the space inside the Game Card chassis at 32GB as it is). It would allow them to increase the number of Game Card ROMs produced per silicon wafer and there is still an option for higher capacity at that node, but getting high enough density to reach 64GB at least is what should be the goal for Drake.

CORRECTION: It was found on a report dated for 2019, but the 32nm process was available in 2014, my mistake. So there's a chance Nintendo is already using the 32nm node, but again... we don't know until someone analyses the chip guts... so, any takers?
But that also means that, if it's already at this smaller process, the push to get a die shrink will be even greater than it was already, since it would mean that Nintendo has used the smallest node currently available for ROM production, so any increases will require a new process node AND may require larger Game Cards, depending on how high the capacity will need to go.
 
Last edited:
Stinky Horse said:
NAND density improvements mostly come from vertical stacking, you aren't going to see physically bigger cards.
Click to expand...
Vertical stacking adds physical height to the chip, and there's not enough room to expand the height of these chips inside the Game Card chassis, either.
 
My Tulpa said:
I still think there is as much a chance as there is a Feb announce and March release, yes
Click to expand...

We would have much more leaks and infos from multiply sources if Drake hardware would be releasing in around one month,
also, Nintendo would want to announce it at least around 2-3 months before launch like they were doing with every revision before,
especially now because its basically next gen Switch hardware so they would want people know about it.

At this point, we can say that launch or even announcement this year is not realistic option any more.

Counting that launch is 1st half of next year, announcement most likely will be January-February.
 
Terrell said:
It's important to recall that Nintendo is not Macronix's only customer.
The change in process node happened in 2019, 2 years after Switch's launch. And without disassembling modern Game Cards and checking the chip sizes in comparison to titles shipped in those first 2 years, we have no means to confirm that Nintendo is ordering chips on that new process node. If Nintendo struck the deal with Macronix in 2016 to supply ROM chips for Game Cards, it was done on either then 45 or 48nm process node and, depending on the contract terms, a die shrink may not have been on the table when the new process node rolled out and was intended for another customer. I even requested help identifying the planar node from this inside look at the ROM chip in a launch title, but no one seemed able to assist.

Additionally, even if Nintendo is using the new 32nm process node, a shrink to 32nm would not provide the means to achieve the necessary storage density for 64GB Game Cards without increasing the physical size of the Game Card itself (since they take up almost all the space inside the Game Card chassis at 32GB as it is). It would allow them to increase the number of Game Card ROMs produced per silicon wafer and there is still an option for higher capacity at that node, but getting high enough density to reach 64GB at least is what should be the goal for Drake.
Click to expand...
I think this is the most recent video about that

 
0
Stinky Horse said:
You're telling me you've measured the height of the silicon in micrometres and determined there isnt enough room?
Click to expand...
No, but I understand industrial engineering well enough to know that, as a means to keep cartridge sizes small, the chassis was not engineered with a lot of headroom, which you can see when it's removed from its chassis.
Beyond that, vertical stacking is only employed when it is more economical than a die shrink. If there are still planar process node shrinks in play (which, by all accounts, there are), they are absolutely more economical. 3D NAND is a cost-benefit analysis solution, as vertical stacking is the better choice economically and technologically than die shrinks into smaller and smaller FinFET territory; most NAND manufacturers were dancing right at the edge of Moore's Law with 20nm and 19nm nodes for production, while the first 3D NAND products most of these manufacturers released allowed them to go back up to the 30s and 40s before jumping back down to 20nm over time and better production expertise. But planar node shrinks before resorting to FinFET are the economical choice for density improvements and always will be. Lastly, such 3D NAND solutions also require firmware development and memory controllers that 2D planar solutions usually do not and, most importantly, are generally applicable to flash memory, which is not what's being discussed.
The only reason there has not been further die shrinks to Macronix XtraROM is likely delays due to COVID. We already know that they've struggled to get 64GB Game Cards, and that is likely both due to expense and space considerations. And that means that, to get larger sizes, Macronix needs to get them a die shrink. If they're currently using 45-48nm, that's all they'll need. If Nintendo is already buying ROMs from them at 32nm, however... that's likely going to mean a size change in Game Cards for Drake. And that can't be answered until we know for sure what Nintendo is buying currently from Macronix.
 
Last edited:
Mr Swine said:
“ALL SYSTEMS, FULL POWER!”
Click to expand...

Terrell said:
No, but I understand industrial engineering well enough to know that, as a means to keep cartridge sizes small, the chassis was not engineered with a lot of headroom, which you can see when it's removed from its chassis.
Beyond that, vertical stacking is only employed when it is more economical than a die shrink. If there are still planar process node shrinks in play (which, by all accounts, there are), they are absolutely more economical. 3D NAND is a cost-benefit analysis solution, as vertical stacking is the better choice economically and technologically than die shrinks into smaller and smaller FinFET territory; most NAND manufacturers were dancing right at the edge of Moore's Law with 20nm and 19nm nodes for production, while the first 3D NAND products most of these manufacturers released allowed them to go back up to the 30s and 40s before jumping back down to 20nm over time and better production expertise. But planar node shrinks before resorting to FinFET are the economical choice for density improvements and always will be. Lastly, such 3D NAND solutions also require firmware development and memory controllers that 2D planar solutions usually do not and, most importantly, are generally applicable to flash memory, which is not what's being discussed.
The only reason there has not been further die shrinks to Macronix XtraROM is likely delays due to COVID. We already know that they've struggled to get 64GB Game Cards, and that is likely both due to expense and space considerations. And that means that, to get larger sizes, Macronix needs to get them a die shrink. If they're currently using 45-48nm, that's all they'll need. If Nintendo is already buying ROMs from them at 32nm, however... that's likely going to mean a size change in Game Cards for Drake. And that can't be answered until we know for sure what Nintendo is buying currently from Macronix.
Click to expand...

I think that compression will probably play a really big role with Drake. If Nintendo can’t get prices down fast enough for the game cards then they and developers will resort to heavy compression. FDE block is probably what they have put on the SoC.
 
0
RealUncleNintendo said:
When is the rumoured date for the successor? Most likely next year?
Click to expand...
i’d say next year, we know more than enough about this chip, more than what we should know about such a chip, for it to be several years out (read: +2 years).


We knew very explicit details about the PS5 and the Xbox series a year or so before it released, the first time we had any explicit details about Drake was in March of this year.
 
coincidentally, xbox and ps5 also came from stolen info

Sol said:
The Alan Wake Switch port doesn't even come close to 720p when docked.

And it runs at sub 400p when in handheld mode.

Nintendo, please release some new hardware. This is getting pitiful.
Click to expand...
I'm seeing folks say that the xbox and ps4 version weren't the best performing either
 
0
unknown.png


I forgot to mention or bring this up, but if they went with the highest config it would increase the CPU perf by 9%

@Look over there
 
Sol said:
The Alan Wake Switch port doesn't even come close to 720p when docked.

And it runs at sub 400p when in handheld mode.

Nintendo, please release some new hardware. This is getting pitiful.
Click to expand...
Unfortunately no amount of tensor cores is going to fix people who manage to make shit ports of X360 games to more capable hardware.
 
Sol said:
The Alan Wake Switch port doesn't even come close to 720p when docked.

And it runs at sub 400p when in handheld mode.

Nintendo, please release some new hardware. This is getting pitiful.
Click to expand...
Just recently we had Nier Automata and No Man's Sky which are fine conversions from their PS4/XBO counterparts. Not to mention the dozens of 360/PS3 conversions which run superior to their original versions.

Stronger hardware doesn't remove the possibility of bad ports, it just makes it easier to brute force minimum acceptable performance. It's odd to me that this port is cherrypicked as some fuel for demanding new hardware, when it's a remastered version of a 360 game - just like Skyrim, which runs excellently (even the last performance-degrading update was just fixed).

The Alan Wake port could be patched or redone like the Ark port, no new hardware needed.
 
Sol said:
The Alan Wake Switch port doesn't even come close to 720p when docked.

And it runs at sub 400p when in handheld mode.

Nintendo, please release some new hardware. This is getting pitiful.
Click to expand...
That’s not really the switch’s fault here tbh lol

It’s a 360 game, even the switch handles that majestically in something like Dark Souls, Skyrim, Need for Speed, Assassin’s Creed well enough.


Alan Wake’s issues aren’t hardware related I think.
 
Alan Wake is a bad port.

Better hardware would make bad ports run better.

Bad ports are signs that publishers are looking at the large install base of a system and thinking, man, if we do a port cheaply enough, the attach rate doesn't need to be great to make money.

Bad ports are things you can hide in the tail end of a system's life because you figure it won't taint your reputation, and the hardware will get blamed.

This the path of the DS, where the success of the system was almost a detriment to Nintendo. People being satisfied with their Switch and the Switch continuing to sell can weirdly lead to a dead-end.
 
0
ReddDreadtheLead said:
unknown.png


I forgot to mention or bring this up, but if they went with the highest config it would increase the CPU perf by 9%

@Look over there
Click to expand...
Aside: Between the increase in L3 cache and improvement in L3 bandwidth within the DSU, it does seem like not much of the A710's alleged +10% IPC comes from the changes to the core itself (reminder: that +10% claim comes from comparing A710 with 8 MB L3 cache versus A78 with 4 MB L3 cache)

...new thought: is the A78C closer to the 2021 set of updates than I previously thought?
Cortex-A78.jpg

Emphasis on the "Optimized timing support" section; 'increased snoop filter', 'configurable slicing for timing/floorplan flexibility'
Then for the DSU-110 update from last year:
arm-dsu-110-cache.png


IIRC, 2017 DynamIQ's L3 cache could be logically partitioned, but physically it's one chunk. So then, did the A78C's DSU get updated to slice up the L3 cache and thus start the ramping up of L3 bandwidth? How far apart are the A78C's DSU and the 2021 DSU-110 update, really? :unsure:
 
Look over there said:
Aside: Between the increase in L3 cache and improvement in L3 bandwidth within the DSU, it does seem like not much of the A710's alleged +10% IPC comes from the changes to the core itself (reminder: that +10% claim comes from comparing A710 with 8 MB L3 cache versus A78 with 4 MB L3 cache)

...new thought: is the A78C closer to the 2021 set of updates than I previously thought?
Cortex-A78.jpg

Emphasis on the "Optimized timing support" section; 'increased snoop filter', 'configurable slicing for timing/floorplan flexibility'
Then for the DSU-110 update from last year:
arm-dsu-110-cache.png


IIRC, 2017 DynamIQ's L3 cache could be logically partitioned, but physically it's one chunk. So then, did the A78C's DSU get updated to slice up the L3 cache and thus start the ramping up of L3 bandwidth? How far apart are the A78C's DSU and the 2021 DSU-110 update, really? :unsure:
Click to expand...
It’s possible that A78C is that “Mid generation” CPU modification on the way to the A710.

This is what Android Authority had to say on the matter:

The Cortex-A78C core is also a little different from the standard Cortex-A78. It implements instructions from newer Armv8.X architecture revisions, such as Armv8.3’s Pointer Authentication and other security-focused features. As a result, the CPU can’t be paired up with existing Armv8.2 CPUs, such as the Cortex-A55 for a big.LITTLE arrangement. We’re looking at six or eight big core only configurations. This wouldn’t be a good fit for mobile, but small core power efficiency is not so important in the laptop market.
Click to expand...
www.androidauthority.com

Here's how Arm's latest CPU targets laptop and handheld console performance

Arm's latest Cortex-A78C CPU is built to offer always-connected laptops and handheld console class performance.
www.androidauthority.com www.androidauthority.com

ReddDreadtheLead said:
I’d definitely see this as being the precursor for it, either it was backported to the A78, or the A78C was the genesis that led to the ARM v9 changes.
Click to expand...
Sorry, this got wonky so I have to modify it to see what I say and it’s not from the article above lol
 
mjayer said:
Ya.. it's in there.. at the end.. in a list. So technically you could say they said that, but I still don't believe it was a marketing message they pushed. Too me, this is similar to the PS5 having 8k slapped on the box.
Click to expand...

I gotta disagree on this, they literally introduced it as “the same games, but at 1440p,” and anyone who could divide immediately became skeptical.
 
Raccoon said:
I'm gonna say it: I think the next 3D Mario should be exclusive and come out in the first six months
Click to expand...
They are not gonna make a potential +20m seller exclusive to Drake, let alone in its first year or with the Mario IP.

Same way Sony didnt make Horizon Forbidden West, GT7 or God of War PS5 exclusives, but was okay with smaller games ala Demon Souls or Returnals being exclusive.

Xenoblade X or Astral Chain 2 or something among those lines makes more sense imo.
 
something that's fun is going around and seeing all the ray tracing experiments people have done that run on low end hardware




scaling this up to a larger game might prove difficult, but it does show that RT doesn't have to mean crazy Pixar-level visuals
 
cw_sasuke said:
They are not gonna make a potential +20m seller exclusive to Drake, let alone in its first year or with the Mario IP.

Same way Sony didnt make Horizon Forbidden West, GT7 or God of War PS5 exclusives, but was okay with smaller games ala Demon Souls or Returnals being exclusive.

Xenoblade X or Astral Chain 2 or something among those lines makes more sense imo.
Click to expand...
let me be clear: I'd like the next 3D Mario to not even be able to run on Switch
 
ILikeFeet said:
something that's fun is going around and seeing all the ray tracing experiments people have done that run on low end hardware




scaling this up to a larger game might prove difficult, but it does show that RT doesn't have to mean crazy Pixar-level visuals
Click to expand...

To me, the best examples of rt are the ones where it’s applied to simple graphics. Quake 3, Minecraft, recently portal.

I wonder if Nvidia will update the portal collection for Drake.
 
0
Raccoon said:
what is new hardware for if not to enable more interesting games?
Click to expand...
Honestly interesting and quality games arent an issue on Switch - P5R/Sparks of Hope just dropped, Bayonetta 3 is next week, Pokemon releasing soon - next year we have Fire Emblem, BotW 2etc.

All these games running at 60 and in a higher resolution would already been great. If the OG Switch can handle Tears of Kingdom, it very likely can also handle the next 3D Mario unless they just add a bunch of ridiculous effects for no reason. If there is one company that can justify good native version of their games on OG Switch and Drake, its Nintendo.
 
cw_sasuke said:
Honestly interesting and quality games arent an issue on Switch - P5R/Sparks of Hope just dropped, Bayonetta 3 is next week, Pokemon releasing soon - next year we have Fire Emblem, BotW 2etc.

All these games running at 60 and in a higher resolution would already been great. If the OG Switch can handle Tears of Kingdom, it very likely can also handle the next 3D Mario unless they just add a bunch of ridiculous effects for no reason. If there is one company that can justify good native version of their games on OG Switch and Drake, its Nintendo.
Click to expand...
something more ambitious than odyssey would be nice
 
0
Raccoon said:
yeah, because they're both

A) miles ahead
and
B) not interested in any fundamental/unscalable changes to how games work
Click to expand...
Because they’re both in no hurry to leave behind the last gen install base, and they are moneymaking corporations first and foremost.

Like Nintendo.
 
Reader mode
You must log in or register to reply here.
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