• 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!)

Random thought: I'm like 3 hours into Scarlett and really noticing just how bad the game performs and looks, especially in the main city. It's a noticable step down from Arceus Legends, and I don't recall it looking ever this bad in trailers or screenshots.

I think someone brought this up before, but it really does feel like it was made with stronger hardware in mind first, and then down ported on Switch. I recall a video someone made with Mariko switch with boosted clocks, and saw it ran at 60 fps.

So yeah the game is unoptimized, but would be funny if it was made initially on Drake or Mariko switch and then down ported. This game definitely has to be a top contender for a patch on Drake. 4k 60fps, with or without DLSS.
This is probably too tinfoil for most of the thread, but I thoroughly agree. Remember that our only consistent rumor of launch timing from NateDrake was shooting for the end of this year with possibility of 2023
 
the problem with a Drake lite is that the Mariko lite didn't prove very popular, probably due to the lack of switching. finding another way to artificially lower the price would be the name of the game for Drake. it would also allow them to not need a die shrink like with the Lite model. if they can make 8nm last (or maybe make a smaller efficiency jump), that would mean they can squeeze out more for their next system, whenever that is
The issue here is, with the SoC being the most expensive component in any build, the path to a substantial cost reduction after launch is a die shrink. That’s why Mariko ended up in both the Lite and the standard Switch, because the primary cost savings was to be found there. There is, frankly, not the kind of cost savings one would hope for by sticking to the same node for less than 10 years, because the real expense of a node comes from foundry R&D/building new fab lines and the silicon wafer. More efficient nodes generally use less wafer for the same performance (although the silicon savings benefit is getting harder and harder to achieve the closer things get to “1nm”, as I understand it). So one way or another, a die shrink is going to happen.
On yields, there's some important context required about the type of yield. SemiAnalysis reported that, for the Exynos 2200 (the first 4LPE SoC), "the parametric yields were horrendous even though catastrophic yields were fine". Catastrophic yields here means "is the chip physically functional or not?", whereas parametric yields mean "can I clock the chip to my expected speed with my expected power draw?". Effectively, it looks like Samsung Foundry were producing chips which were fully functional, but not meeting their customer (Samsung LSI's) expectations in terms of clocks and power draw. This is a big difference from a manufacturing line that is pumping out massive numbers of useless chips, and speaks more to Samsung Foundry overselling the performance of the process rather than anything else.
So, even leaving out the suggestion that Samsung LSI had targeted too high of a clock which resulted in low parametric yields, bad parametric yields on Drake just screams “new Nvidia Shield coming in hot”.
 
Wait, wait. Random thought, I don’t know if it’s been answered yet: Could the Switch Lite only exist if it was on Mariko? I mean, I know it was “lite” because it couldn’t dock, but it was also smaller. Was that something that only Mariko could provide?
 
Wait, wait. Random thought, I don’t know if it’s been answered yet: Could the Switch Lite only exist if it was on Mariko? I mean, I know it was “lite” because it couldn’t dock, but it was also smaller. Was that something that only Mariko could provide?
Also the greater battery life Mariko allowed made a permanently handheld Switch more viable.
 
Re: the Drake talk, the biggest issue for a Drake Lite is the price. Not being able to die-shrink is a big limitation on many fronts for a console.

It is probably the biggest reason why I believe Samsung 8nm is unlikely.
Even if it is the better financial deal day one, it is doubtful to be the case during the whole lifecycle of the platform.
Die shrinking from Samsung's 8N process node to a newer process node is definitely possible. Just don't expect the process to be a trivial one, considering that the SoC probably needs to be redesigned with EUV lithography in mind instead of DUV lithography in mind, which was used with Samsung's 8N process node, and with different IP if Nintendo and Nvidia decides to switch from Samsung to TSMC or Intel.

7LPP/6LPP - 7LPP was Samsung's first process using EUV, and was used in a variety of smartphone SoCs and more recently for some IBM CPUs. Outside of IBM it hasn't seen any new chips in several years, and IBM are a bit of a special case, both in terms of their longer time-to-market and their R&D involvement (I believe Samsung's 7LPP leverages IBM R&D). With EUV tools in short supply, Samsung would have also been eager to move production lines over to 5nm/4nm processes, which would bring in more revenue. 6LPP was announced as an improved version of 7LPP, but later disappeared off roadmaps with no chips ever manufactured on it. I would consider 7LPP or 6LPP very unlikely at this stage.

5LPE/5LPP - Samsung considers 5LPE and 5LPP to be their second generation EUV process, and lists them alongside 7LPP as part of the same family (see here). Here 5LPE is the early version of the process, and 5LPP the improved version. They're largely compatible with 7LPP designs, which would explain why they replaced 7LPP so quickly with the exception of IBM. The 5nm family is a possibility for Drake, but the lines between Samsung's "5nm" chips and their "4nm" chips is a little blurry, as I'll mention below.
IBM mentioned that Telum's fabricated using Samsung's 7HPP process node, which I assume is IBM's custom variation of Samsung's 7LPP process node, which I imagine is similar to how TSMC's 4N process node is allegedly Nvidia's custom variation of TSMC's N5P process node.

And IBM seems to have recently implied starting the transition from Samsung's 7HPP process node to Samsung's 5 nm** process node with the IBM Artificial Intelligence Unit (AIU).

This is not a chip we designed entirely from scratch. Rather, it's the scaled version of an already proven AI accelerator built into our Telum chip. The 32 cores in the IBM AIU closely resemble the AI core embedded in the Telum chip that powers our latest IBM's z16 system. (Telum uses transistors that are 7 nm in size while our AIU will feature faster, even smaller 5 nm transistors.)

I am curious what Nvidia are using 3GAE/3GAP for, though. It's not likely to outperform 4N, so it doesn't really make sense for a successor to Hopper, Ada, or Grace.
I was thinking Nvidia could plan on using Nvidia's custom variation of Samsung's 3GAP process node to fabricate entry level consumer Blackwell GPUs, with the rest of the consumer Blackwell GPUs being fabricated using Nvidia's custom variation of TSMC's N3E process node, which is similar to how Nvidia used Nvidia's custom variation of Samsung's 14 nm** process node to fabricate entry level consumer Pascal GPUs (here and here), with the rest of the consumer Pascal GPUs being fabricated using Nvidia's custom variation of TSMC's 16 nm** process node.
** → a marketing nomenclature used by all foundry companies

Wait, wait. Random thought, I don’t know if it’s been answered yet: Could the Switch Lite only exist if it was on Mariko? I mean, I know it was “lite” because it couldn’t dock, but it was also smaller. Was that something that only Mariko could provide?
Mariko's considerably, if not significantly, more power efficient than Erista. So although I think equipping Erista in the Nintendo Switch Lite is possible, that probably comes with ~25% to ~40% lower battery life (from comparing the Nintendo Switch (2019)'s battery life with the Nintendo Switch Lite's battery life), so probably around ~1.5 hours to ~4.875 hours for a Nintendo Switch Lite equipped with Erista.
 
Wait, wait. Random thought, I don’t know if it’s been answered yet: Could the Switch Lite only exist if it was on Mariko? I mean, I know it was “lite” because it couldn’t dock, but it was also smaller. Was that something that only Mariko could provide?
Erista gets a bit warm when playing a heavy title for long periods of time. they can't just slap a bigger fan since the Lite is smaller, and it'd be louder. a die shrink solves the heat issue
 
They'd have to make sure Drake has really good battery life.. like near Mariko levels.
DLSS will help a ton with this but of course not every (especially older games) will support that technology. I’m personally expecting something in between the launch Switch and the OLED Switch for Drake in handheld mode so around 5 hours for games pushing 3D visuals / open Worlds. I think that’s fine when you compare that with the sub 3 hours you get on the likes of Steam Deck when running demanding games.
 
I would absolutely loved to replay lots of Switch games in 4K with stable framerates on the next Nintendo hardware. I don't think it will happen though. Would be nice if Nintendo at least patched some of their toy tier games at the very least but I'm expecting nothing personally.
 
Random thought: I'm like 3 hours into Scarlett and really noticing just how bad the game performs and looks, especially in the main city. It's a noticable step down from Arceus Legends, and I don't recall it looking ever this bad in trailers or screenshots.

I think someone brought this up before, but it really does feel like it was made with stronger hardware in mind first, and then down ported on Switch. I recall a video someone made with Mariko switch with boosted clocks, and saw it ran at 60 fps.

So yeah the game is unoptimized, but would be funny if it was made initially on Drake or Mariko switch and then down ported. This game definitely has to be a top contender for a patch on Drake. 4k 60fps, with or without DLSS.
There are a ton of much more impressive open World games versus the latest Pokemon game on Switch. Games like Breath of the Wild, Xenoblade 2, The Witcher 3 and Xenoblade 3 all push more geometry, better textures, better lighting, better models and more models / objects on screen while having much more stable framerates. This isn’t a hardware issue.

Pokemon looks and runs the way it does because Game Freak are slowly ramping up their modern rendering experience aswell as manpower / budget with each passing game on Switch. It’s also more ambitious than Arceus. It was also in dire need of 3 more months of development time but because of the way their marketing / merchandising push works they just didn’t want to delay it and miss the Christmas shopping window. It really didn’t matter considering it sold what 10 million copies in a weekend… not that it makes it right just that they were correct by predicting most of the market wouldn’t care about videos of bugs and framerate drops on Reddit and twitter.

I bet that by the end of March there will be several more patches for the game which target performance improvements and the game will hit 30fps much more consistently aswell as the lower framerate lod improvements for characters in the distance.

Game Freak will then have a good base for the next game and with the added power of Drake we might actually see the first fully open World modern Pokemon game which will have good image quality and run at 60fps.
 
This is probably too tinfoil for most of the thread, but I thoroughly agree. Remember that our only consistent rumor of launch timing from NateDrake was shooting for the end of this year with possibility of 2023
Yeah it might be tin foil enough that it originally ran on Drake/Orion, but GF seemed to clearly run on better hardware. Besides Totk, this game, Bayonetta 3 and Xenoblade 3 will get. performance updates on them.


DLSS will help a ton with this but of course not every (especially older games) will support that technology. I’m personally expecting something in between the launch Switch and the OLED Switch for Drake in handheld mode so around 5 hours for games pushing 3D visuals / open Worlds. I think that’s fine when you compare that with the sub 3 hours you get on the likes of Steam Deck when running demanding games.
I would be very very surprised if we didn't get a node shrink on Drake.

There are a ton of much more impressive open World games versus the latest Pokemon game on Switch. Games like Breath of the Wild, Xenoblade 2, The Witcher 3 and Xenoblade 3 all push more geometry, better textures, better lighting, better models and more models / objects on screen while having much more stable framerates. This isn’t a hardware issue.

Pokemon looks and runs the way it does because Game Freak are slowly ramping up their modern rendering experience aswell as manpower / budget with each passing game on Switch. It’s also more ambitious than Arceus. It was also in dire need of 3 more months of development time but because of the way their marketing / merchandising push works they just didn’t want to delay it and miss the Christmas shopping window. It really didn’t matter considering it sold what 10 million copies in a weekend… not that it makes it right just that they were correct by predicting most of the market wouldn’t care about videos of bugs and framerate drops on Reddit and twitter.

I bet that by the end of March there will be several more patches for the game which target performance improvements and the game will hit 30fps much more consistently aswell as the lower framerate lod improvements for characters in the distance.

Game Freak will then have a good base for the next game and with the added power of Drake we might actually see the first fully open World modern Pokemon game which will have good image quality and run at 60fps.
Absolutely. They really need help from developers Monolithsoft on open world games. Breath of the wild on Wii U looks and runs better. The game feels really unoptimized. I did say that in my original post. And I wouldn't be surprised if it was made on much stronger hardware. The trailers and shots kinda prove that. Just saying that Drake could give it a 4k 60fps boost and should without too much trouble.

I'm somewhat pessimistic on GF patching this game with significant performance updates. We got one already and it wasn't really significant. MS did with xb2 at least. But you can only do so much, if the engine also happens to be the limitation.This game does feel rushed though.
 
I expect that Drake will offer the same battery life as the OG switch... when playing patched games. Depending on how the backwards compat solution works, I can totally see Drake trouncing Mariko on classic Switch games in BC mode.
 
I expect that Drake will offer the same battery life as the OG switch... when playing patched games. Depending on how the backwards compat solution works, I can totally see Drake trouncing Mariko on classic Switch games in BC mode.
What makes you think this? Not challenging you, just totally ignorant and watching this thread from the backseat.
 
Nvidia claims 50% power draw, iso performance for Ada vs Ampere. I'm wondering if that is against Desktop Ampere or the A100 (which is often the baseline they use when talking about Ampere performance)
It’s vs the GA102 die GPU. If I’m looking at the right chart you are referring to that is.

GA100 isn’t really for graphics, or matches the specs listed by NVidia. Again, assuming I’m looking at the same graph.
 
0
Yay, more doom.

If Nintendo's plugged the leaks there won't be leaks. Not only that but there's already BEEN so many leaks. Obviously they can't release it until they announce it, but come on. They can announce it any time before launch and it'll still sell out.
I wouldn't be surprised if this is the real reason Nintendo reorganized their EU branches into one entity last month. 🤔
 
0
I would absolutely loved to replay lots of Switch games in 4K with stable framerates on the next Nintendo hardware. I don't think it will happen though. Would be nice if Nintendo at least patched some of their toy tier games at the very least but I'm expecting nothing personally.
Smash 3DS got C-nub and amiibo support when the new 3DS launched.
 
Smash 3DS got C-nub and amiibo support when the new 3DS launched.

Anything is possible. Still I wouldn't be surprised if Nintendo did no 4K patches and instead waited years and then released $70 "remasters" instead. Kind of like Sony is doing these days.
 
Either
  1. Nvidia and Nintendo are spending extra money to be on a process node with no other shared products, or...
  2. They're on the same node as Lovelace, releasing after Lovelace, but still not running Lovelace ,or...
  3. Drake has power saving magic that is not on any other device in the Nvidia product line, including their top of the line cards, but is so simple that it's still the same Ampere architecture, or...
  4. The wattage numbers in the test don't mean what we think they mean.
1) Has happened before, when Nvidia went with 20nm for the TX1, but it's worth noting they planned on moving Maxwell over to it, and were very unhappy with the yields and didn't make the move. 2) Is insane, but I've theorized it was possible myself. 3) I simply do not believe. 4) Seems likely

For point 1., Do we know for sure what process Atlan was supposed to be on? Or Thor, for that matter.
 
Imagine the industries reaction if they not only got DLSS & RT but also DLSS frame generation into a portable device 😂

You’d theoretically only need to render a game at 240p/15fps to get a decent experience on the go. Now that’s battery management 😝
 
0
For point 1., Do we know for sure what process Atlan was supposed to be on? Or Thor, for that matter.
Atlan is dead and was replaced by Thor. Given the intended release and arch, 4nm is likely for both

I wonder if nvidia can customize the rt core to have twice the rt intersection testing like with Ada.

Would help the low power device.
I doubt it would do much. Intersection testing isn't the heaviest part of RT, nor the part where "more = better". Shading pixels is still the costliest aspect as is refitting and rebuilding data structs. Solving that and you get a lot of performance back. But for power management, I would think you would do as little data structure rebuilding as possible. So maybe a hybrid of ray tracing static objects and dynamic objects using raster effects
 
0
Drake will almost certainly use a different colour to differentiate it from the OLED model. The base console is Black, OLED is White how about Nintendo Red for Drake?

80e28418604f41714224b8510a5f64c7.jpg


I could see all black - as in black console with black joy cons.
 
Last edited:
This is probably too tinfoil for most of the thread, but I thoroughly agree. Remember that our only consistent rumor of launch timing from NateDrake was shooting for the end of this year with possibility of 2023
To tinfoil for me at least. I take Nate's explanation.
 
0
There were several errors in this calculations. @Zomble made a better estimation of Orin GPU power consumption. It is higher than what I estimated.

TDLR : 8 nm is dead. Just let it die just like the TX1. Embrace 5 nm.

It basically has to be that. Looking at APUs that are in the Aya Neo Pro, the Drake SOC needs to be no larger than 200mm2, and honestly probably much smaller for cost reasons. I guess the question is if a low margin SOC like Drake is being designed with 5nm in mind, why would the rest of the Orin lineup remain at 8nm?
 
It basically has to be that. Looking at APUs that are in the Aya Neo Pro, the Drake SOC needs to be no larger than 200mm2, and honestly probably much smaller for cost reasons. I guess the question is if a low margin SOC like Drake is being designed with 5nm in mind, why would the rest of the Orin lineup remain at 8nm?
Purchased capacity + Orin doesn’t have the same thermal and power constraints as Drake.
 
0
80e28418604f41714224b8510a5f64c7.jpg


I could see all black - as in black console with black joy cons.
I actually like that but with maybe glossy black joycons and the same red trim around the console and red dock. Red pro controller.
 
0
It basically has to be that. Looking at APUs that are in the Aya Neo Pro, the Drake SOC needs to be no larger than 200mm2, and honestly probably much smaller for cost reasons. I guess the question is if a low margin SOC like Drake is being designed with 5nm in mind, why would the rest of the Orin lineup remain at 8nm?
Orin was designed for automotives and then scaled down from there. and automotive certification takes a long time given Orin is supposed to power safety features and autonomous driving
 
0
I'm really hoping that SwitchDrake (though I wouldn't know how) replicates the qualities seen with Switch in portable mode.
During this holiday period I had a chance to try out Steamdeck in this mode, and you can't understand how much this made me appreciate Nintendo's console even more.
Lightness, handling, quietness (Steamdeck's fan is a nightmare) and battery life are things you notice when you 'lose' them.

I hope Nintendo manages to strike the right balance despite the rumoured jump in specifications.
 
from my understanding in his talks with developers john heard about something like a mid-gen refresh that no longer seems to be happening and thinks 2023 is off the table
I wonder if this is the “heavy” news Nate heard?
 
well I sure am glad I pawned off the direct thread

I've got no interest in following switch news for the foreseeable future
 
0
I'll just say I find DF's reasoning poor. at least Oliver acknowledges the T239. John talking about some scrapped refresh without details like it's common knowledge is some shit though.
 
Well if it isn't released in 2023 I'd definitely say it puts the nail in the coffin of the next hardware being anything but a successor at that point.

Still sticking to my guns though
 
0
I now feel pretty bad for NateDrake

he presumably heard its future was uncertain at some point this year and has been trying to find any definitive word on it since
 
Unless I'm wrong, but hasn't DF been off the mark regarding the Drake for a bit? Plus if they're still mentioning the mid-generation refresh, this is obviously not what the Drake is (regardless if Nintendo markets it as such). There's also the fact Nintendo has had huge bursts of Raw Materials between Summer and Winter this year in their inventory that still hasn't been explained (on top of the hiring for new hardware, accessory testing and distribution). Not sure what that would be unless it was the Drake.

That said, take what I say with a grain of salt, but I feel DF's been pretty negative about the idea for a bit.
 
Like pretty much everything we've seen in this thread in the past 3 months this is far from conclusive information.
 
Now we wait for articles to pop up saying: "Switch Pro is dead?!?" ,"No hardware refresh till 20XX?", "You might have to get use to your crusty switch games for a while." Or videos talking about it.
 
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