• 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!)


Ya I saw this while looking up what the hall effect was. I doubt the manufacturer's ready to send 20 million per annum to Nintendo's product assembly line though. I wonder if other manufacturers could easily supplement?
 
0
If they do that I would bet on it being a completely silent revision.

If it hasn't already happened.

Personally, i expect DRake Joy-Cons to work with base Switch and vice-versa, so yeah, any improvements will be done within the current Joy-Con shell design.
 
they would not silently and expensively eliminate drift permanently

that shit would get marketed like crazy
Not a chance. That would be acknowledging the problem exists, which they have vehemently denied since 2017. In ongoing lawsuits too IIRC.

Count on it, it'll be totally silent.
 
they would not silently and expensively eliminate drift permanently

that shit would get marketed like crazy
They can't admit to drift being an ongoing issue with joycon, especially if they were to continue to sell joycon - a product they'd be simultaneously admitting is defective. That'd just be horrible PR and an admission of knowledgeable guilt in the courtroom.
 
Not a chance. That would be acknowledging the problem exists, which they have vehemently denied since 2017. In ongoing lawsuits too IIRC.

Count on it, it'll be totally silent.
they have acknowledged that it happens though

they've publicly acknowledged that they're always improving them

there's nothing admissible about announcing your peerlessly durable new controllers
 
I think it's possible for them to market hall effect sensors while not admitting that their previous controller had drift issues. "Improved responsiveness" or some other marketing fluff.
 
they have acknowledged that it happens though

they've publicly acknowledged that they're always improving them

there's nothing admissible about announcing your peerlessly durable new controllers
Right yeah they COULD say it's an improved controller in general, but I don't expect them to say anything in terms of marketing. They've said in interviews that they're continuously improving the joycons, and I do believe they're continuously iterating it silently. This will be a continuation.
 
If true, why wasn't this already common knowledge? Why did we assume that LPDDR5 was a better product overall than LPDDR4? Why learn that just now?
It is a better product overall than LPDDR 4. Did you read the tweet?

not always bring a bandwidth advantage and that some LPDDR5 configurations should be avoided for specific workloads.”
 
0
Right yeah they COULD say it's an improved controller in general, but I don't expect them to say anything in terms of marketing. They've said in interviews that they're continuously improving the joycons, and I do believe they're continuously iterating it silently. This will be a continuation.
That’s not going to be silent. The joycons is a big part of the Switch. They would definitely market the joycons have been improved (if it’s actually something tangible they can show people).
 


If true, why wasn't this already common knowledge? Why did we assume that LPDDR5 was a better product overall than LPDDR4? Why learn that just now?

It's true, but not really a big issue for a device like the Switch. The first part, that LPDDR5 only gives benefits if you clock it higher than LPDDR4 is hardly surprising. The second part, about the particular modes to use to maximise performance in different circumstances, is exactly the kind of thing that Nvidia would optimise for in the design of the memory controller and drivers, and that developers would optimise for in their games.

A games console like the Switch is in a good position for this because (a) the typical access patterns are well known from the start, (b) those access patterns are typically highly sequential, particularly on the GPU side which would account for most of the bandwidth and (c) developers can optimise around the specific properties of the memory interface they're using.

This is very much the same case for GDDR memory which is used in other consoles, which, being designed for graphics use, is built for maximum bandwidth on sequential access rather than good random access performance. I would suspect LPDDR5 would actually be better under random access than GDDR6 if anything, and with game engines built around keeping data access as sequential as possible, it's really not a big issue.
 
Yeah but Nintendo has stumbled onto a relatively gimmick-less concept that works with the Switch, so why would they suddenly go back to the financially unstable idea of gimmicks?
I really don't view the switch as a gimmick-less device, not even "relatively".

I have a bunch of friends who argue they don't have a reason to buy a switch because they can "emulate it", and depending on their overall tone and argument, I reply that not only emulation isn't perfect to this day, but the point of a switch besides running games natively without issues* is portable play with the added benefit of convenience when switching from portable to couch and the gimmick of sharing a controller with a friend without the need to buy a second one.
The discussion usually ends there. But if someone were to ask "what about just buying joycons and/or a pro controller separately to use with an emulator" I'd simply reply that not only this wouldn't fix the other issues that comes with emulating the switch, but it would also deviate from the point of playing those games portably and while yes, the steam deck or a chinese handheld pc would be a solution to that problem, one of the core aspects of handheld gaming is being able to play games for hours on end, and that CANNOT be achieved with the steam deck when emulating the switch let alone chinese handhelds that perform worse.

*Performance, crashes, vertex explosions, texture rendering issues, shader compilation lag, rebuilding shader cache and the amount of time it takes, finding people to play games online, no ACTUAL online support for ranked ladders and ordering gear in SP2 and SP3 as well as other NSO-only features.

So no, not only the switch has a bunch of gameplay gimmicks imported from previous consoles like the Wii and WiiU (motion controls, touch-screen, etc...) it also in of itself IS a gimmick considering how the console operates and how some (clueless) people pick emulation over it due to the current state of emulators for the device.
 
Yeah but Nintendo has stumbled onto a relatively gimmick-less concept that works with the Switch, so why would they suddenly go back to the financially unstable idea of gimmicks?
Because, as mentioned elsewhere, predictability leads to lack of imagination, complacency, stagnation, and ultimately, decline. Following one successwith another is hard, and while new controllers were needed to navigate 3D spaces after 2D ones, continuing with "the same, but more powerful" approach would've finished Nintendo, at least, in the traditional home console arena - the fact that their home consoles never outsold their predecessors until the Wii is proof of that. With the Wii U, we had more power. The GamePad had more in common with the Dual Shock controller variant, and that was just as true for the Wii U Pro Controller. Nintendo did the whole "Gamers Are Over. They Don't Have To Be Your Audience" thing well before games journalists wrote the articles in the Wii/DS era, and it was an unprecedented success. They tried to adopt elements of the PS/XBox model on the Wii U, but that is no model of stability or sustainability, and the fact that there are more "Wii and Switch Wannabes" out there than PS/XBox imitations and competitors should speak volumes.

Switch has plenty of gimmicks, and the 2017 Presentation leaned hard on them, from split controls to 1-2-Switch showing the Wii Remote user how motion controls have evolved into HD Rumble, while playing anywhere with whoever you want spoke to those who were familiar with phones, tablets and being on the move in an ever faster-living world. We had Labo and Ring Fit, as well as Mario Kart Live. We just need a stylus inspired by the S-Pen and motion control capabilities, and that would be sweet. Gimmicks aren't bad things. They aren't celebrated enough. They can become tomorrow's standards. They're needed to encourage creativity and breathe new life into the industry - We have more power than ever, but seen many beloved IPs in the low to mid-budget tiers, some in higher tiers, even, disappear. There's an over-reliance on the Indie Circuit to pay homage to some of those games, but bigger publishers have the power and resources, and yet these games are nowhere to be seen. Sony has all the 3rdP support in the world AND more power, but outside of Media Molecule, the overwhelming majority of their 1stP output is "The Generic Stubbly, Rugged, Pale, Male and Able Gun/Sword Axe-Wielding Protag Cometh" - Look no further than Today's PlayStation for what "No Gimmicks" looks like. Even when they introduce a gimmick, it's one that's been done elsewhere in some capacity. Then there's XBox - Kinect was a key differentiator. The moment they ditched it, they had little else that captured he imagination. So, the short of it is that gimmicks are important, and they don't mean instability. They bring new players to the table, and there can be rewards in the risk. The moment Nintendo stops trying new gameplay concepts is the moment the whole gaming industry becomes infinitely poorer for it.
 
That’s not going to be silent. The joycons is a big part of the Switch. They would definitely market the joycons have been improved (if it’s actually something tangible they can show people).
Again they've actually improved them already and haven't said anything about it outside of interviews.

Openly marketing new joycons with improved accuracy or whatever would open up a huge can of worms in terms of internet backlash (not super meaningful) and potential legal backlash (yes super meaningful). They'll err on the side of caution here and say nothing.
 
I really don't view the switch as a gimmick-less device, not even "relatively".

I have a bunch of friends who argue they don't have a reason to buy a switch because they can "emulate it", and depending on their overall tone and argument, I reply that not only emulation isn't perfect to this day, but the point of a switch besides running games natively without issues* is portable play with the added benefit of convenience when switching from portable to couch and the gimmick of sharing a controller with a friend without the need to buy a second one.
The discussion usually ends there. But if someone were to ask "what about just buying joycons and/or a pro controller separately to use with an emulator" I'd simply reply that not only this wouldn't fix the other issues that comes with emulating the switch, but it would also deviate from the point of playing those games portably and while yes, the steam deck or a chinese handheld pc would be a solution to that problem, one of the core aspects of handheld gaming is being able to play games for hours on end, and that CANNOT be achieved with the steam deck when emulating the switch let alone chinese handhelds that perform worse.

*Performance, crashes, vertex explosions, texture rendering issues, shader compilation lag, rebuilding shader cache and the amount of time it takes, finding people to play games online, no ACTUAL online support for ranked ladders and ordering gear in SP2 and SP3 as well as other NSO-only features.
I think we differ on definition of “gimmick.” When I say “gimmick” I mean things like “motion controls” or ”Glasses-free 3D Screen” or “IR Camera.” The hybrid nature of the Switch isn’t so much a “gimmick” as it is a core fundamental part of the system’s paradigm. It’d be like saying the Nintendo DS’s touch screen is a gimmick.

I guess what I’m saying is to me, a “gimmick” is a feature that a game developer could and usually does simply ignore. Some kinda funky lil feature that is brand new and developers will creatively struggle with or outright refuse to support.
 
Because, as mentioned elsewhere, predictability leads to lack of imagination, complacency, stagnation, and ultimately, decline. Following one successwith another is hard, and while new controllers were needed to navigate 3D spaces after 2D ones, continuing with "the same, but more powerful" approach would've finished Nintendo, at least, in the traditional home console arena - the fact that their home consoles never outsold their predecessors until the Wii is proof of that. With the Wii U, we had more power. The GamePad had more in common with the Dual Shock controller variant, and that was just as true for the Wii U Pro Controller. Nintendo did the whole "Gamers Are Over. They Don't Have To Be Your Audience" thing well before games journalists wrote the articles in the Wii/DS era, and it was an unprecedented success. They tried to adopt elements of the PS/XBox model on the Wii U, but that is no model of stability or sustainability, and the fact that there are more "Wii and Switch Wannabes" out there than PS/XBox imitations and competitors should speak volumes.

Switch has plenty of gimmicks, and the 2017 Presentation leaned hard on them, from split controls to 1-2-Switch showing the Wii Remote user how motion controls have evolved into HD Rumble, while playing anywhere with whoever you want spoke to those who were familiar with phones, tablets and being on the move in an ever faster-living world. We had Labo and Ring Fit, as well as Mario Kart Live. We just need a stylus inspired by the S-Pen and motion control capabilities, and that would be sweet. Gimmicks aren't bad things. They aren't celebrated enough. They can become tomorrow's standards. They're needed to encourage creativity and breathe new life into the industry - We have more power than ever, but seen many beloved IPs in the low to mid-budget tiers, some in higher tiers, even, disappear. There's an over-reliance on the Indie Circuit to pay homage to some of those games, but bigger publishers have the power and resources, and yet these games are nowhere to be seen. Sony has all the 3rdP support in the world AND more power, but outside of Media Molecule, the overwhelming majority of their 1stP output is "The Generic Stubbly, Rugged, Pale, Male and Able Gun/Sword Axe-Wielding Protag Cometh" - Look no further than Today's PlayStation for what "No Gimmicks" looks like. Even when they introduce a gimmick, it's one that's been done elsewhere in some capacity. Then there's XBox - Kinect was a key differentiator. The moment they ditched it, they had little else that captured he imagination. So, the short of it is that gimmicks are important, and they don't mean instability. They bring new players to the table, and there can be rewards in the risk. The moment Nintendo stops trying new gameplay concepts is the moment the whole gaming industry becomes infinitely poorer for it.
I guess what I mean by “play it safe” and “be relatively gimmick-less” is this:

I think Nintendo will have ADDITIVE gimmicks. Like a touch screen, or motion controls, or the IR camera. What Nintendo won’t return to is a Wii Remote, or forcing two screens by way of a GamePad. They want to keep the barrier for software development to be as low as possible, and the moment a third party suddenly has to consider “ok, what do we put on this other screen” or “how do we map the controls to this remote with 2 buttons” the possible answer of “maybe we just don’t” pops back up. They want to avoid that.

I want to go on-record and say an S-Pen/Apple Pencil-styled stylus???? I’m 100% all for it. It would, frankly, singlehandedly fix Super Mario Maker 2 haha
 
If true, why wasn't this already common knowledge? Why did we assume that LPDDR5 was a better product overall than LPDDR4? Why learn that just now?
The paper is pretty short, but I'll do a sum up. Basically, LPDDR4 was good, so LPDDR5 didn't have fat to trim, and depends on some very clever redesigns to get bandwidth wins and power reductions. tl;dr LPDDR5 is very good

LPDDR5 devices can operate at data rates up to 6400 MT/s1 compared to 4266 MT/s for LPDDR4. Since the clock signal would contribute a considerable amount to the total power at such high data rates, LPDDR5 devices use a slower continuously running command clock and a faster data clock that is disabled during idle times. Depending on the data rate the ratio between both clock signals can either be set to 2:1 (up to 3200 MT/s) or 4:1. To compensate for the slow command clock, commands are now transmitted at double data rate. While an LPDDR4 channel is always comprised of 8 banks, LPDDR5 introduces three different bank modes, which can be selected at power-up

Emphasis mine. The paper discovers two things. One, under some workloads LPDDR5 is just as good as LPDDR4, but not better

For sequential traffic an upgrade from LPDDR4 to LPDDR5 only brings an advantage from a bandwidth perspective if data rates above 4266 MT/s are targeted.

Emphasis mine, again. This isn't super surprising. One of the other things that this paper finds is that both standards are really good at getting close to their theoretical max bandwidth under sequential reads (where you start at one end of memory and go to the other without jumping around). If you're not utilizing LPDDR5's faster data transfer rate, you're not going to see bandwidth wins. That LPDDR5 keeps up at lower transfer rates shows how solid the design is. And of course, as I highlighted, there are other wins. Namely, power draw.

However, there was a case where LPDDR5 performance tanked. Highly random reads, where the data clock runs at the fastest configuration. This is the downside of LPDDR5, not that it runs slower, but that in order to get best performance under some workloads you have to use tuning knobs that didn't exist in LPDDR4.

For random traffic and a burst length of 16 the 4:1 clock ratio of LPDDR5 should be avoided

The memory system has two channels, one for data, and one for commands. LPDDR5 uses a setup where the data clock runs very fast, but turns itself off when idle to save power, and the command clock runs quite slow, but is always on. Random reads cause cache misses every read (there is always a cache, right?), and cache misses require additional work on the command channel. This puts LPDDR5's tradeoffs in the worst possible position

The designers of LPDDR5 almost definitely knew that, which is what the ratio of data clock to command clock speed is actually configurable. This is the most interesting part of the paper, because if you go by raw bandwidth numbers, this configuration seems like it should be the fastest. And it is, unless you have lots and lots of random reads. Which brings us back to why we this wasn't common knowledge before this paper.

Truly random reads are rare. This is not a common workload by any stretch of the imagination. The authors of this paper work under a grant for an institution that specifically invests in simulation based research.

This work was supported within the Fraunhofer and DFG cooperation programme (Grant no. 248750294) and supported by the Fraunhofer High Performance Center for Simulation- and Softwarebased Innovation.

So they built a simulation framework that allowed them to run artificial workloads against LPDDR4 and 5 where the only difference was the standard rather than any manufacturing details between hardware. Their conclusion doesn't seem in any way surprising based on the designs of LPDDR4 vs 5 (not that I am an expert) but might be to engineers going entirely from manufacturer's data sheets, which seems to be their primary driver

However, since the standard comes with lots of changes including a completely new clocking scheme, the memory controllers also have to be largely redesigned. Thus, many companies are faced with the decision of whether to upgrade their existing system designs or continue with the well-established predecessor for their target applications. However, taking such a decision on the basis of performance numbers by DRAM vendors is difficult as they usually only publish peak bandwidths
 
Not really news, but an ex-coworker of mine works at Zynga and said that all he knows of is the mobile/NSW version of Star Wars hunters, no other versions. And then he said there are other mobile games he's privy to but can't share details.
 
I think we differ on definition of “gimmick.”
Yes.
It’d be like saying the Nintendo DS’s touch screen is a gimmick.
But it is a gimmick. Try to imagine the core physical aspects of a gaming handheld, now strip down any unnecessary change that isn't a must for a developer or the player. Now look back at a console like the DS/3DS: how many games could be re-developed without the use of a secondary display?
For any given feature on a game's UI that requires touch input, a simpler implementation through buttons and an analog stick could be viable.

Of course, pokemon ranger or kirby canvas curse for instance wouldn't be the same if you had to press 'A' and move around the dpad or an analog stick to draw.
And I'm glad they went for a touch screen (I really am), but all I'm saying is that it wouldn't be a physical impediment to develop such games if they didn't have access to a secondary (touch) screen. Are there games that would indeed be impossible to develop without the second screen? Yes. But generally, games that extensively use a console's gimmicks aren't the majority (because often it's not something convenient for developers to do).

Even some dictionary definitions of the word "gimmick" I don't agree with
eg: "gimmick" - noun
"Something that is not serious or of real value that is used to attract people's attention or interest temporarily, especially to make them buy something".

I heavily disagree with the thought that the DS touchscreen doesn't add real value to its game library. Like, imagine playing the aforementioned games the way I specified a couple lines ago? shit would suck, HARD.

I think by now you get the point. The switch has many features that could be stripped down without making the device non-functional*, the switch lite is the best example of this. But having said features is a key selling point for the more expensive device (regular switch, switch OLED).
*Again, for a wide majority of games on the platform, but not all.
 
It's true, but not really a big issue for a device like the Switch. The first part, that LPDDR5 only gives benefits if you clock it higher than LPDDR4 is hardly surprising. The second part, about the particular modes to use to maximise performance in different circumstances, is exactly the kind of thing that Nvidia would optimise for in the design of the memory controller and drivers, and that developers would optimise for in their games.

A games console like the Switch is in a good position for this because (a) the typical access patterns are well known from the start, (b) those access patterns are typically highly sequential, particularly on the GPU side which would account for most of the bandwidth and (c) developers can optimise around the specific properties of the memory interface they're using.

This is very much the same case for GDDR memory which is used in other consoles, which, being designed for graphics use, is built for maximum bandwidth on sequential access rather than good random access performance. I would suspect LPDDR5 would actually be better under random access than GDDR6 if anything, and with game engines built around keeping data access as sequential as possible, it's really not a big issue.
Be blessed! Thanks for your accurate, well-worded, easy to understand responses.

The paper is pretty short, but I'll do a sum up. Basically, LPDDR4 was good, so LPDDR5 didn't have fat to trim, and depends on some very clever redesigns to get bandwidth wins and power reductions. tl;dr LPDDR5 is very good



Emphasis mine. The paper discovers two things. One, under some workloads LPDDR5 is just as good as LPDDR4, but not better



Emphasis mine, again. This isn't super surprising. One of the other things that this paper finds is that both standards are really good at getting close to their theoretical max bandwidth under sequential reads (where you start at one end of memory and go to the other without jumping around). If you're not utilizing LPDDR5's faster data transfer rate, you're not going to see bandwidth wins. That LPDDR5 keeps up at lower transfer rates shows how solid the design is. And of course, as I highlighted, there are other wins. Namely, power draw.

However, there was a case where LPDDR5 performance tanked. Highly random reads, where the data clock runs at the fastest configuration. This is the downside of LPDDR5, not that it runs slower, but that in order to get best performance under some workloads you have to use tuning knobs that didn't exist in LPDDR4.



The memory system has two channels, one for data, and one for commands. LPDDR5 uses a setup where the data clock runs very fast, but turns itself off when idle to save power, and the command clock runs quite slow, but is always on. Random reads cause cache misses every read (there is always a cache, right?), and cache misses require additional work on the command channel. This puts LPDDR5's tradeoffs in the worst possible position

The designers of LPDDR5 almost definitely knew that, which is what the ratio of data clock to command clock speed is actually configurable. This is the most interesting part of the paper, because if you go by raw bandwidth numbers, this configuration seems like it should be the fastest. And it is, unless you have lots and lots of random reads. Which brings us back to why we this wasn't common knowledge before this paper.

Truly random reads are rare. This is not a common workload by any stretch of the imagination. The authors of this paper work under a grant for an institution that specifically invests in simulation based research.



So they built a simulation framework that allowed them to run artificial workloads against LPDDR4 and 5 where the only difference was the standard rather than any manufacturing details between hardware. Their conclusion doesn't seem in any way surprising based on the designs of LPDDR4 vs 5 (not that I am an expert) but might be to engineers going entirely from manufacturer's data sheets, which seems to be their primary driver
Thanks as well. So this is new results based on never before done testing? I can understand why this result is not surprising for the initiate but it sure for the profane. What was heard in the media during last decade was that DDR5 had twice the speed of DDR4, without any caveats. So reading this article after so much time sure feels awkward.

But hey, the superiority of DDR5 (or LPDDR5 in this case) for Switch is established so it's ultimately a nothing burger.
 
Last edited:
0
Yes.

But it is a gimmick. Try to imagine the core physical aspects of a gaming handheld, now strip down any unnecessary change that isn't a must for a developer or the player. Now look back at a console like the DS/3DS: how many games could be re-developed without the use of a secondary display?
For any given feature on a game's UI that requires touch input, a simpler implementation through buttons and an analog stick could be viable.

Of course, pokemon ranger or kirby canvas curse for instance wouldn't be the same if you had to press 'A' and move around the dpad or an analog stick to draw.
And I'm glad they went for a touch screen (I really am), but all I'm saying is that it wouldn't be a physical impediment to develop such games if they didn't have access to a secondary (touch) screen. Are there games that would indeed be impossible to develop without the second screen? Yes. But generally, games that extensively use a console's gimmicks aren't the majority (because often it's not something convenient for developers to do).

Even some dictionary definitions of the word "gimmick" I don't agree with
eg: "gimmick" - noun
"Something that is not serious or of real value that is used to attract people's attention or interest temporarily, especially to make them buy something".

I heavily disagree with the thought that the DS touchscreen doesn't add real value to its game library. Like, imagine playing the aforementioned games the way I specified a couple lines ago? shit would suck, HARD.

I think by now you get the point. The switch has many features that could be stripped down without making the device non-functional*, the switch lite is the best example of this. But having said features is a key selling point for the more expensive device (regular switch, switch OLED).
*Again, for a wide majority of games on the platform, but not all.
I think arguing over the semantic meaning of the term gimmick ultimately misses the point. Derachi provided how they define the term of themselves, making a long post trying to argue they're using it wrong isn't actually arguing against what they're saying, just how it's being said.
 
making a long post
So you didn't read the post, got it.
I think arguing over the semantic meaning of the term gimmick ultimately misses the point. Derachi provided how they define the term of themselves
If you actually read the post, you'd have noticed I provided my own definition of what I see as a gimmick in terms of console hardware (specifically the portable kind). But I never pointed that they're using the term wrong while trying to convince them of using my own definition, but merely tried to explain my own.
trying to argue they're using it wrong isn't actually arguing against what they're saying, just how it's being said.
I did argue against what they said. In fact, I even stated that I disagree with the general definition of the word (when applied to the context of a console, specifically the nintendo portables we were discussing).
 
It’s really exciting to consider that, after that direct and heading into fall, we’re really in the final act now. After all this time.

Since we’re mostly in agreement that this device isn’t launching any later than 5/12/23, we have a countdown. And now we truly are closer than ever to the supply chain gearing up, industry chatter becoming more frequent, and leaks from either of those.

And the next time we hear from Nintendo in a major way (like a presentation, etc.) it will most likely include Drake.

Exciting!!!
 
It’s really exciting to consider that, after that direct and heading into fall, we’re really in the final act now. After all this time.

Since we’re mostly in agreement that this device isn’t launching any later than 5/12/23, we have a countdown. And now we truly are closer than ever to the supply chain gearing up, industry chatter becoming more frequent, and leaks from either of those.

And the next time we hear from Nintendo in a major way (like a presentation, etc.) it will probably be about Drake.

Exciting!!!
I think there will still be a direct or two before the announcement.
 
So you didn't read the post, got it.

If you actually read the post, you'd have noticed I provided my own definition of what I see as a gimmick in terms of console hardware (specifically the portable kind). But I never pointed that they're using the term wrong while trying to convince them of using my own definition, but merely tried to explain my own.

I did argue against what they said. In fact, I even stated that I disagree with the general definition of the word (when applied to the context of a console, specifically the nintendo portables we were discussing).
I did read your post. That won't happen again.
 
All this talk of gimmicks and I can't help but think, what is there that can be a gimmick? There doesn't seem to be anything new that can be a serious game changer, so to speak. Touch screens and motion controls are the peak of changing up user interfaces. Otherwise there doesn't seem to be anything out there have can provide new gaming experiences without being complex or expensive, like how VR is.

If the Switch 2 doesn't have some major new feature beside additional processing power, I think is because there isn't really any game changing gimmick that would draw attention.
 
0
What I'm more concerned about is how much the iGPU on drake that shares it's memory pool with system RAM benefits from the jump from LPDDR4 to LPDDR5.
I remember back in a couple months looking at a (different) scenario where cards with their own dedicated memory (the MX450 under GDDR5 vs the MX450 under GDDR6) and the difference was noticeable.
 
Last edited:
I think there will still be a direct or two before the announcement.
Really? Interesting. The next two Directs will be (roughly) February and E3 so I think two is definitely out.

I'll grant you they could do a Jan/Feb Direct and then reveal Drake soon after, I suppose that's not impossible. Although I'd have to imagine if they're revealing a model that improves IQ and performance so soon after a games showcase they'd probably repurpose said showcase into a reveal for the hardware. My guess, at least.
 
I think it's possible for them to market hall effect sensors while not admitting that their previous controller had drift issues. "Improved responsiveness" or some other marketing fluff.
Yeah, I think that's exactly the sort of language I expect. I'm reminded of how Apple handled the transition away from their much-maligned and unreliable butterfly keyboard to the newly branded "Magic Keyboard" on their notebooks. Commentator Daniel Jalkut expressed it best: "it concedes everything while confessing nothing." That's going to be the needle Nintendo's going to need to thread, if they do indeed comment on any changes to the sticks — they'll want to signal to enthusiasts that the stick mechanism is different while not admitting that there was a reliability problem on the old sticks.
 
It’s really exciting to consider that, after that direct and heading into fall, we’re really in the final act now. After all this time.

Since we’re mostly in agreement that this device isn’t launching any later than 5/12/23, we have a countdown. And now we truly are closer than ever to the supply chain gearing up, industry chatter becoming more frequent, and leaks from either of those.

And the next time we hear from Nintendo in a major way (like a presentation, etc.) it will most likely include Drake.

Exciting!!!
this is all I've ever wanted: a deadline
 
Really? Interesting. The next two Directs will be (roughly) February and E3 so I think two is definitely out.

I'll grant you they could do a Jan/Feb Direct and then reveal Drake soon after, I suppose that's not impossible
Nintendo has at least 9 2023+ games for base Switch.

Release date set: Fire Emblem: Engage, Kirby's Return to Dreamland Deluxe, Tears of the Kingdom
No release date: Metroid Prime 4, Pikmin 4, Advance Wars 1+2 Reboot Camp
Widely leaked to be done: Metroid Prime 1 Remakester, Twilight Princess HD, Windwaker HD
Leaked in development, probably exist and support base Switch: Style Savvy 5, Detective Pikachu 2, EPD Donkey Kong, Untitled 3D Remaster by Bamco,
Plus big 3rd party games without set release dates that would likely make a Direct: Hollow Knight: Silksong, Outer Wilds, Persona 3 Portable, Persona 4 Golden

FE:E, LoZ:TotK
likely both have DLC, Splatoon and Mario Kart both have ongoing support and DLC waves, NSO+ has a schedule of N64 games into the middle of next year, and both GBC and GBA support for NSO has leaked along with support for it in datamines.

That is a lot of base Switch stuff, and while a decent chunk of it could be last-year-of-a-dying-console stuff (Style Savvy) there is some meaty, unannounced games in that list. It seems likely that Nintendo will be supporting base Switch in some capacity through 2024.

I don't know what message they want to send, but if this thing launches in H1 2022, then they're not sending an ordinary "this is our successor" message. And if it's not that, then you probably want to make clear how much support base Switch gets before dropping any Drake exclusives or talking about enhanced Drake support. Mixing them up risks a very muddle message both on what the hardware can do and what level of support all the various hardware lines are getting.

That's why I expect* at least one Direct that is Switch focused before Drake is announced, possibly a Direct and an Indie World, to really hammer the pitch home, before the first "merged" Direct around E3.

In that scenario, E3 sets the template of how they'll differentiate "legacy" games on base switch and "next gen" games, says that the Switch will have a long tail, but the writing will be on the wall. All the really exciting 2024 announcements are listed as Drake exclusives, and the cool upcoming stuff plays better on Drake. September brings the last Direct with some meaty Switch stuff in it. February 2024, Switch owners get some announcements targeting Lite owners - a couple casual games/another 2D Kirby - and by E3 2024, there is some cross gen stuff in the Direct, but it's all 3rd party stuff, and you can only know it's coming if you look real hard at the fine print on the screen. Some people start calling Indie World the "Switch Direct" because indies can ride the combination of large install base, modest specs, and lack of first party competition for a lot longer than AA-AAA development can.

* But not strongly, I've given up confident predictions :)
 
What I'm more concerned about is how much the iGPU on drake that shares it's memory pool with system RAM benefits from the jump from LPDDR4 to LPDDR5.
I remember back in a couple months looking at a (different) scenario where cards with their own dedicated memory (the MX450 under GDDR5 vs the MX450 under GDDR6) and the difference was noticeable.
It’s so funny that you bring this up because me and @Z0m3le were having a discussion very recently, like yesterday, about how the system will be sharing its memory bandwidth with all parts of it, it’s not just the GPU, but the CPU also gets its own necessitated memory bandwidth allocation. The PS4 for example is rated for 176 GB per second but apparently it’s more like 140 to 120 depending on the CPU load, the more the CPU is utilized the lower the actual memory bandwidth is for the system is it seems?



Cortex a78C is rated to have a maximum bandwidth of 60 GB per second.


 
Minor nitpick, but this isn't a leak. The game was officially announced years ago.
Ah, yes, indeed! The recent indications are just that it hadn't been canceled but was nearing release. That's just a boatload of solid Switch support. I can see how it might look like the sort of last hurrah stuff (cross-gen Zelda, pokemon spinoff, Style Savvy) but even if this is the Switch's closing slate, it is an excellent one.
 
0
Nintendo has at least 9 2023+ games for base Switch.

Release date set: Fire Emblem: Engage, Kirby's Return to Dreamland Deluxe, Tears of the Kingdom
No release date: Metroid Prime 4, Pikmin 4, Advance Wars 1+2 Reboot Camp
Widely leaked to be done: Metroid Prime 1 Remakester, Twilight Princess HD, Windwaker HD
Leaked in development, probably exist and support base Switch: Style Savvy 5, Detective Pikachu 2, EPD Donkey Kong, Untitled 3D Remaster by Bamco,
Plus big 3rd party games without set release dates that would likely make a Direct: Hollow Knight: Silksong, Outer Wilds, Persona 3 Portable, Persona 4 Golden

FE:E, LoZ:TotK
likely both have DLC, Splatoon and Mario Kart both have ongoing support and DLC waves, NSO+ has a schedule of N64 games into the middle of next year, and both GBC and GBA support for NSO has leaked along with support for it in datamines.

That is a lot of base Switch stuff, and while a decent chunk of it could be last-year-of-a-dying-console stuff (Style Savvy) there is some meaty, unannounced games in that list. It seems likely that Nintendo will be supporting base Switch in some capacity through 2024.

I don't know what message they want to send, but if this thing launches in H1 2022, then they're not sending an ordinary "this is our successor" message. And if it's not that, then you probably want to make clear how much support base Switch gets before dropping any Drake exclusives or talking about enhanced Drake support. Mixing them up risks a very muddle message both on what the hardware can do and what level of support all the various hardware lines are getting.

That's why I expect* at least one Direct that is Switch focused before Drake is announced, possibly a Direct and an Indie World, to really hammer the pitch home, before the first "merged" Direct around E3.

In that scenario, E3 sets the template of how they'll differentiate "legacy" games on base switch and "next gen" games, says that the Switch will have a long tail, but the writing will be on the wall. All the really exciting 2024 announcements are listed as Drake exclusives, and the cool upcoming stuff plays better on Drake. September brings the last Direct with some meaty Switch stuff in it. February 2024, Switch owners get some announcements targeting Lite owners - a couple casual games/another 2D Kirby - and by E3 2024, there is some cross gen stuff in the Direct, but it's all 3rd party stuff, and you can only know it's coming if you look real hard at the fine print on the screen. Some people start calling Indie World the "Switch Direct" because indies can ride the combination of large install base, modest specs, and lack of first party competition for a lot longer than AA-AAA development can.

* But not strongly, I've given up confident predictions :)
Interesting - thanks for sharing, it's super fun to speculate about this stuff! I apologize if the answer to this in your post (having Friday-afternoon-brain right now), but I'm not sure I see your need for one more Direct before Drake's announced. Why couldn't the timeline be shifted forward? The messaging they communicate at E3 after Drake's announced could just be their messaging for a March Direct after Drake's announced.

I also feel like, to the majority of their install base (which I agree is who they need to carefully message to), one last Switch-focused direct before Drake is announced won't feel any different to them than the Direct we just had a few weeks ago. The average person doesn't know Drake is coming so I don't feel like that person needs a final Switch-only blowout for N to pull off the Drake message.

I wonder if it's a Drake reveal first that shows 60fps TotK, Arkham/RDR2/whatever, that's then followed up by a Direct in Feb/Mar some time later. Even with continued Switch 1 support, they are never going to show a game running on Switch 1 in their promotions again once Drake is announced. I figure that first Direct we get next year gets the benefit of showing everything running on Drake if the announcement will be so proximate anyway.
 
Cortex a78C is rated to have a maximum bandwidth of 60 GB per second.


Hold up; I'm not necessarily sold that's referring to CPU to RAM bandwidth.
Extremetech's version of the news thinks that it's CPU to L3 cache bandwidth.
I'm actually doubtful that there was change in CPU to RAM bandwidth. The A78C announcement makes no mention of any changes to the # of AMBA CHI ports (should still be 1x 256-bit like with 2017 DynamIQ?). In contrast, the DSU-110 update specifically mentions the ability to scale up to 4x 256-bit ports.
 
So this is definitely above my level of understanding. I (and I'm sure others in this thread) would appreciate ELI5's of the salient points.
I just want to emphasize to the general audience that while this roadmap is not so much applicable to the relatively near future (Drake), it is pertinent to the big picture of a longer term future (next ~decade, or decade and a half).

Also, the PPA targets for node scaling every 2-3 years are depressingly low :unsure:
 
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