• 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!)

Felt the same... The Hori Split Pro Controller is a real game changer for handheld play.

On the subject of VR... they are definitely interested and most probably internally trying different things especially around Mario Kart I'd guess.
Having played the touring Mario kart VR arcade machine, it is incredibly fun, but that hardware is surprisingly low res.
 
0
Given the Switch Lite still needed a fan, to get to something like this or smaller, the SoC needs another die shrink to run at docked clocks and probably a good heat sink solution.

Strangely Nintendo was in the Smart TV space well before anyone else. Wii was increadibly popular as a Netflix box in the 2nd half of the 2000s. And they simply did not consider it a market worth going after and let Roku and Amazon fill in the space. I don't see a TV box like this as something Nintendo will pursue. Switch is ubiquitous and will likely remain in production for several more years. A TV box like that is expected to sell in the sub $150 range and add in the need for controller support and it's not really a high priority for them. They probably rather redirect the parts to higher margin Switch consoles or the successor.

This has always been my position with TV only Switch, it's serving a market that's already served. The cost savings of not needing a screen and battery is offset by an equally large drop in perceived value. They can't sell a dockless Switch for the same price as a hybrid model, so the $50-80 they save will be eaten up by the price difference.
Its not about selling tonns.
If its break even and reaches an audience the normal switch doesnt then its already a win, since they earn 33% on every game you buy in the e shop. With the screen and the battery removed its pretty easy to have it all in a tiny housing (and optimize for airflow since it doesnt need to be portable). Since it also doesnt need the big dock the packaging can be almost half the size. Yeah, im confident that they can get it to 150$, and that would be a big win, because it would be way cheaper then the regular switch, but not in direct competition since you cant play portable. The only thing that would cost more: having the charging joy con grip instead of the passive one.
 
….why wouldn’t it bring any such advantage? The feature is built into the hardware, it’s not an add-on.


Drake already has hardware support for what would be the equivalent to Foveated Rendering and Eye-Tracking
Because my assumption is that Nintendo is not going to want to release a $400+ headset, and the requisite headset hardware for inside-out tracking, eye tracking, etc does not come cheap.

nb: wrote while walking the dog, may have come across more tersely than intended
 
Last edited:
0
Its not about selling tonns.
If its break even and reaches an audience the normal switch doesnt then its already a win, since they earn 33% on every game you buy in the e shop. With the screen and the battery removed its pretty easy to have it all in a tiny housing (and optimize for airflow since it doesnt need to be portable). Since it also doesnt need the big dock the packaging can be almost half the size. Yeah, im confident that they can get it to 150$, and that would be a big win, because it would be way cheaper then the regular switch, but not in direct competition since you cant play portable. The only thing that would cost more: having the charging joy con grip instead of the passive one.
Or just bundle a pro controller.
 
Or just bundle a pro controller.
I thought about that, but then you are excluding some games like Mario party and removing the option to play multiplayer with the base bundle ... and by that kinda ruining the benefit of the low entry price to a degree. And i dont really think that the charging grip is that much more expensive then the regular one.
 
I thought about that, but then you are excluding some games like Mario party and removing the option to play multiplayer with the base bundle ... and by that kinda ruining the benefit of the low entry price to a degree.
That's already the case with the Nintendo Switch Lite.
 
That's already the case with the Nintendo Switch Lite.
Different case. Lite was clearly a more personal gaming experience focused consoles, and its design dictated that it cant have the normal joy cons.

A dock only switch would be less of a personal console -> more use for multiplayer, + there is no technical/physical reason to not use the general joy cons.
 
I'm going to say some things, feel free to call them dumb, but:

Wii Mini came out alongside Wii U, stripping key features to meet a price point.
New Nintendo 2DSXL came out alongside Nintendo Switch, stripping key features to meet a price point.
Xbox 360E came out alongside Xbox One.
GameBoy Micro alongside DS, SNES JR. alongside N64.

A at-home only Switch with no Game Card slot, using Mariko, could make a lot of sense. Digital only, pack in a Pro Controller (cheaper to manufacture than two Joy-Con and a Joy-Con Charging Grip). No screen, no battery, no Joy-Con, no speakers, no headphone jack. Heck, 32GB of eMMC storage while we're at it, the bottom of the barrel.

I'd say that device could come very close to Nintendo's targets for the lowest end, with Wii Mini starting at $99.99 and 2DS $79.99, Nintendo is no stranger to price cutting a console below 100 dollars at the end of a generation. Even getting it to the 2DS' introductory price would be a good position for it, at around 120-130USD.

We know Microsoft think there's room in the console space for a sub-120 dollar "ultra low end" device, a niche that has historically been readily filled. The Wii Mini was a success, the 2DS was a success and New 2DS was the last Nintendo DS lineage device ever manufactured because it too succeeded.

The niche exists, the technology exists, and on a software level... Calcio... Technically exists!
Every successful home video game system from Nintendo received a sub-100$ revision at the end of its lifespan.

With us looking down the barrel of a next generation Switch, be it Pro, Super, 2, or what have you, I think a home console only Switch could still happen next year.

I realise this is annecdotal, but I've also spoken to quite a few people who want to play Switch games but are put off by the fact that to get the one that connects to a TV it costs 300 dollars, but if they got a TV only one for half that they'd probably get it. Given the history of the segment, including Nintendo's, Microsoft's assertion that the segment still exists, and the fact I've met people who would fall into this niche, I'd say it's possible! Maybe not likely, but very possible!
 
Combine a TV only Switch with NVidia’s AI upscalling and let me install TV apps. Then I’d get rid of my Nvidia Shields and would save the additional Switch docks.
 
Why else would they add t239 drivers to Linux though?
To play devil's advocate, for development purposes. Nintendo can't just make drivers for the chip ab initio methinks. Linux is probably the development platform for the actual Switch system software (it's far more similar to Switch architecturally than Windows, and is better at virtualization).

However I think the real answer is a new Shield TV with a focus on streaming- just a device that's really good at it. I mean, the TX1+ went into the Shield, and the Shield is still on shelves and by far the most respected Android TV box. There's a new one coming! The segment isn't dead! And I think a Nintendo Switch TV-Mini wouldn't be in the segment to begin with, it'd be a microconsole dedicated to gaming, a niche Nintendo could have all to themselves even with the X1+ in a cheap plastic shell. There are so many people I know who would jump at a 100-150 dollar device that plays Smash Bros. Ultimate, Mario Kart 8 Deluxe, Persona 5 Royal, Nier Automata, Witcher 3, etc. People in that segment are unlikely to care about the low resolution, I mean, they're in the market for a console in a sub-180 dollar price range.
 
0
The segment is all but dead. Otherwise they would have followed up on the shield
I mean...its not? if you're not a google or amazon or apple guy, its the Shield, still sold, still up to date with 4k streaming and great upscaling. could they improve it with a new soc? (drake?) sure.
But streaming did not really change that much since 2016. There are no 8k streaming services, there is no 120Hz movies. There will be a new one, but the only benefit as a TV media box will be... better AI upscaling?
This will mainly happen to move production to a newer node when they phase out the old one.
 
The segment is all but dead. Otherwise they would have followed up on the shield

I bought a pretty expansive TV back in 2016 ( Sony 75x9405) and I’m still loving it’s screen. The OS and the SoC however is outdated and slow. The shield tv is brilliant for me. I’d rather upgrade that one than buying a new TV.
 
A dock only switch would be less of a personal console -> more use for multiplayer, + there is no technical/physical reason to not use the general joy cons.
To play devil's advocate, I can argue people are probably not buying a TV mode only model for local multiplayer. I think if people want local multiplayer, people are probably going to buy a Nintendo Switch or the OLED model. (But I can see people buying a TV mode only model for online multiplayer.)

Speaking of a T239 equipped Shield TV model, I think Nvidia could use a T239 equipped Shield TV model as the platform to expand upon GeForce Now's resolution scaling, especially the "AI-enhanced mode".
 
To play devil's advocate, I can argue people are probably not buying a TV mode only model for local multiplayer. I think if people want local multiplayer, people are probably going to buy a Nintendo Switch or the OLED model. (But I can see people buying a TV mode only model for online multiplayer.)

Speaking of a T239 equipped Shield TV model, I think Nvidia could use a T239 equipped Shield TV model as the platform to expand upon GeForce Now's resolution scaling, especially the "AI-enhanced mode".
Oh yeah, it wont be the main audience (thats probably the "a cheap switch for a handfull of cool exclusives", but there will be a sizable price concious crowd. I could se either joy cons or pro controller be an option. But joy cons do open up more gameplay options while them still being able to have the pro as an value add after the people are hooked on the platform.

---------
Yeah, better ai upscaling for gforce now is probably the only real benefit an x2 would have.

(People forget that the x1 was capable to do 4k60 hdr for streaming)
 
To play devil's advocate, I can argue people are probably not buying a TV mode only model for local multiplayer. I think if people want local multiplayer, people are probably going to buy a Nintendo Switch or the OLED model. (But I can see people buying a TV mode only model for online multiplayer.)

Ll
Why, because table top mode is so great?

Imo the vast majority of local multiplayer gaming on Switch happens in docked mode. Even though I don’t have the statistics, it would surprise me if tabletop holds a significant percentage.
 
Last edited:
0
---------
Yeah, better ai upscaling for gforce now is probably the only real benefit an x2 would have.

(People forget that the x1 was capable to do 4k60 hdr for streaming)
The tx2 ship sailed years ago, we were talking about t239 Drake.

Doesn’t change your point though.
 
The segment is all but dead. Otherwise they would have followed up on the shield
I wouldn’t say it’s dead, but it most certainly does not mandate the kind of cycles of iteration we see with other consumer electronics, being able to exist for up to 4-5 years without hardware renewal (and most of the time, hardware renewal is more about the benefit of re-purposing silicon made for other devices on more efficient nodes but only when it’s cost-efficient to do so, which almost always means binned chips.)
 
The tx2 ship sailed years ago, we were talking about t239 Drake.

Doesn’t change your point though.
i mean yeah, i meant x2 as in the next chipset. Drakes a long word compared to that (and i forgot that there was an actual x2)

But thx for correcting.
 
0
I wouldn’t say it’s dead, but it most certainly does not mandate the kind of cycles of iteration we see with other consumer electronics, being able to exist for up to 4-5 years without hardware renewal (and most of the time, hardware renewal is more about the benefit of re-purposing silicon made for other devices on more efficient nodes but only when it’s cost-efficient to do so, which almost always means binned chips.)
Gonna be blunt, this implies Nvidia is gonna gear up to sell a new Shield TV soon after Drake releases, which tracks.
 
To play devil's advocate, I can argue people are probably not buying a TV mode only model for local multiplayer. I think if people want local multiplayer, people are probably going to buy a Nintendo Switch or the OLED model. (But I can see people buying a TV mode only model for online multiplayer.)
Are you thinking system-to-system multiplayer? Because a TV-only Switch would absolutely get a lot of same-screen Mario Kart and Smash going on.
 
To play devil's advocate, I can argue people are probably not buying a TV mode only model for local multiplayer. I think if people want local multiplayer, people are probably going to buy a Nintendo Switch or the OLED model. (But I can see people buying a TV mode only model for online multiplayer.)

Speaking of a T239 equipped Shield TV model, I think Nvidia could use a T239 equipped Shield TV model as the platform to expand upon GeForce Now's resolution scaling, especially the "AI-enhanced mode".
For the record, OFA can be used for media content to give good video at a higher framerate.
 
T239 is quite a bit beefier than the binned A15 Bionic that is in Apple TV 4k. I can imagine that Nvidia might make a new Shield out of it, but it would absolutely be a premium product. There is a market for such a thing, but integrated voice assistants and media libraries need to come along with it, which is why Apple and Amazon are in good places there and Nvidia isn't.

Nvidia is also hampered because I'm sure that whatever arrangement they made with Ninty involves not directly competing against them with their own hardware. Which limits the ability to use the set top box as a games console, a direction Amazon and Apple would definitely like to take.

I find it somewhat surprising that NuSwitch devkits would use Linux, but there are some strong indications that is the case. The question becomes why would a definitely-never-set-for-release version of Linux wind up in public repositories? It still smells like Nvidia plans on releasing a Drake based product of their own.
 
T239 is quite a bit beefier than the binned A15 Bionic that is in Apple TV 4k. I can imagine that Nvidia might make a new Shield out of it, but it would absolutely be a premium product. There is a market for such a thing, but integrated voice assistants and media libraries need to come along with it, which is why Apple and Amazon are in good places there and Nvidia isn't.

Nvidia is also hampered because I'm sure that whatever arrangement they made with Ninty involves not directly competing against them with their own hardware. Which limits the ability to use the set top box as a games console, a direction Amazon and Apple would definitely like to take.

I find it somewhat surprising that NuSwitch devkits would use Linux, but there are some strong indications that is the case. The question becomes why would a definitely-never-set-for-release version of Linux wind up in public repositories? It still smells like Nvidia plans on releasing a Drake based product of their own.
I don't think the dev kits for making games necessarily use Linux, though they might, rather that building and testing drivers for it on Linux makes some sense. Having Linux support for it makes it easier for Nintendo to develop Horizon. I have no doubt that the new console will continue to use Horizon, like the Xbox Series and One families using Windows 10 HyperV (sorta, I know that's not entirely accurate.)
 
T239 is quite a bit beefier than the binned A15 Bionic that is in Apple TV 4k. I can imagine that Nvidia might make a new Shield out of it, but it would absolutely be a premium product. There is a market for such a thing, but integrated voice assistants and media libraries need to come along with it, which is why Apple and Amazon are in good places there and Nvidia isn't.

Nvidia is also hampered because I'm sure that whatever arrangement they made with Ninty involves not directly competing against them with their own hardware. Which limits the ability to use the set top box as a games console, a direction Amazon and Apple would definitely like to take.

I find it somewhat surprising that NuSwitch devkits would use Linux, but there are some strong indications that is the case. The question becomes why would a definitely-never-set-for-release version of Linux wind up in public repositories? It still smells like Nvidia plans on releasing a Drake based product of their own.
Hence binned chips. If they set the threshold for non-faulty cores in the CPU and GPU low enough, Nvidia could (potentially) achieve near-100% use of chips produced per wafer (minus the incomplete chips at the edge of the wafer).

And this would be downright necessary, as Nintendo will not get a full yield on most wafers, there would be defect waste if they were the only use case that would drive the cost per SoC up, so any agreement with Nvidia likely includes their right to use defective chips that don’t meet Nintendo’s yield requirements however they see fit to. Be that new Shields, consumer micro-PCs, a new Jetson Nano variant, or all of the above; Nintendo likely won’t get a say in how binned chips are used and likely don’t care, as whatever it might be won’t compete with Drake (Nvidia is smart enough to not shit where they eat, so a native-software console of any variety is likely not in the mix, and would be less performant than Drake anyways), so as long as it keeps SoC costs as low as possible, anything is fair game.
 
I don't think the dev kits for making games necessarily use Linux, though they might, rather that building and testing drivers for it on Linux makes some sense.
What I've gleaned from job postings and the couple of reasonable reports I've read, the devkit indeed runs Linux. You boot into a Linux environment, and then launch either a virtualized or "mini" HorizonOS which skips a homescreen and loads the mounted game image directly.

This allows them to use various debugging tools that launch from the Linux environment. This kind of makes sense, because Horizon is really designed for all that, and likely doesn't have kernel support for instruction counters and the like
 
Gonna be blunt, this implies Nvidia is gonna gear up to sell a new Shield TV soon after Drake releases, which tracks.
Question is .. Does Nvidia really need something as powerful as Drake for Nvidia Shield? TV streaming, playing games from Google play store, certain PC games to play download, and Nvidia GeForce now to stream games is what the Nvidia Shield/Pro did. Well I guess it doesn't hurt. I wonder how Drake could benefit then.. Maybe we'll see Nvidia GeForce Now and get more PC ports.

Kind of surprised Switch didn't get Metal Gear Rising Vengeance when Shield did..
 
I hope that with the next console, call of duty or battlefield make a return to nintendo. I def think there is a big market for more hardcore fps games. Playing Crysis on the go was such a lovely time, id love to experience cod or bf on the go too.

FPS and Horror games are really a blind spot right now so i hope the next console manages to address that : )
 
0
Question is .. Does Nvidia really need something as powerful as Drake for Nvidia Shield? TV streaming, playing games from Google play store, certain PC games to play download, and Nvidia GeForce now to stream games is what the Nvidia Shield/Pro did. Well I guess it doesn't hurt. I wonder how Drake could benefit then.. Maybe we'll see Nvidia GeForce Now and get more PC ports.

Kind of surprised Switch didn't get Metal Gear Rising Vengeance when Shield did..
That's what I keep saying... if they use binned chips, it won't be. They could technically disable half the CPU and GPU cores and it'd still be a T239-powered Shield.
The main benefit would be hardware-accelerated AV1 decode, which current Shield models lack, which offers 4K Netflix streaming with less bandwidth use and buffering, and it would be one of (if not THE) first set-top box to have hardware-accelerated AV1 decode.
It featuring an A78C CPU instead of an A78AE CPU makes it the better choice for chip in a set-top box, as well.
 
Are you thinking system-to-system multiplayer? Because a TV-only Switch would absolutely get a lot of same-screen Mario Kart and Smash going on.
I was thinking about games that require only Joy-Cons for multiplayer. But then I've remembered that local multiplayer for Nintendo Switch Sports on the Nintendo Switch Lite isn't really possible. So a TV mode only model is definitely beneficial for Nintendo Switch Sports.

And games, such as Super Smash Bros. Ultimate and Mario Kart 8 Deluxe, do have online multiplayer, which I did mention I can see people buying a TV mode only model for.

For the record, OFA can be used for media content to give good video at a higher framerate.
Thanks for the reminder. I do wonder if the Optical Flow Accelerators (OFA) on Orin and Drake (Orin and Drake share the same OFA, which I believe are different from the OFA in consumer Ampere GPUs) are at least almost as fast as the OFA on Ada Lovelace GPUs.

T239 is quite a bit beefier than the binned A15 Bionic that is in Apple TV 4k.
I agree, GPU wise. But CPU wise, the Apple A15 Bionic on the Apple TV 4K probably still outperforms Drake, comparing a Geekbench 5 single-core score of the Apple A15 Bionic (iPhone 13 Mini) with a Geekbench 5 single-core score of AGX Orin.

That's what I keep saying... if they use binned chips, it won't be. They could technically disable half the CPU and GPU cores and it'd still be a T239-powered Shield.
The main benefit would be hardware-accelerated AV1 decode, which current Shield models lack, which offers 4K Netflix streaming with less bandwidth use and buffering, and it would be one of (if not THE) first set-top box to have hardware-accelerated AV1 decode.
It featuring an A78C CPU instead of an A78AE CPU makes it the better choice for chip in a set-top box, as well.
And as ReddDreadtheLead mentioned, the OFA on Drake can theoretically aid with video playback, assuming the OFA on Drake performs at least almost as fast as the OFA on Ada Lovelace GPUs.
 
Question is .. Does Nvidia really need something as powerful as Drake for Nvidia Shield? TV streaming, playing games from Google play store, certain PC games to play download, and Nvidia GeForce now to stream games is what the Nvidia Shield/Pro did. Well I guess it doesn't hurt. I wonder how Drake could benefit then.. Maybe we'll see Nvidia GeForce Now and get more PC ports.

Kind of surprised Switch didn't get Metal Gear Rising Vengeance when Shield did..
I’d like to think the Portal collection was the beginning of Lightspeed Studios bringing their AAA Android ports to Switch
 
Question is .. Does Nvidia really need something as powerful as Drake for Nvidia Shield? TV streaming, playing games from Google play store, certain PC games to play download, and Nvidia GeForce now to stream games is what the Nvidia Shield/Pro did. Well I guess it doesn't hurt. I wonder how Drake could benefit then.. Maybe we'll see Nvidia GeForce Now and get more PC ports.

Kind of surprised Switch didn't get Metal Gear Rising Vengeance when Shield did..
Shield hasn't had an upgrade since Mariko. While not competing with Apple, the new Apple4K TV box is more capable. It makes sense that they'll just piggyback off Drake for their SheildTV designs and save on the cost of the R&D and maintain that segment.

I'm actually thinking ShieldTv could launch considerably later than the new Switch, if for example, they need every silicon they can get for the launch. The woul deprioritize ShieldTV.
 
What I've gleaned from job postings and the couple of reasonable reports I've read, the devkit indeed runs Linux. You boot into a Linux environment, and then launch either a virtualized or "mini" HorizonOS which skips a homescreen and loads the mounted game image directly.

This allows them to use various debugging tools that launch from the Linux environment. This kind of makes sense, because Horizon is really designed for all that, and likely doesn't have kernel support for instruction counters and the like
you mean regular Switch dev kits?
 
Just curious has anyone stayed on top of the Linux code searches?
I went just last night to see what would turn up and there have been 4 new Tegra entries that also include support for t239.
Interesting enough the Linux code is also building support for a Tegra th500, which Google says is the next Tegra Hopper SoC naming.
 
Thanks for the reminder. I do wonder if the Optical Flow Accelerators (OFA) on Orin and Drake (Orin and Drake share the same OFA, which I believe are different from the OFA in consumer Ampere GPUs) are at least almost as fast as the OFA on Ada Lovelace GPUs
Considering the segment that ORIN occupies, which is meant for automotive, the Optical Flow Accelerators in it has to be incredibly precise in that area.

Optical flow was used by robotics researchers in many areas such as: object detection and tracking, image dominant plane extraction, movement detection, robot navigation and visual odometry. Optical flow information has been recognized as being useful for controlling micro air vehicles.

The application of optical flow includes the problem of inferring not only the motion of the observer and objects in the scene, but also the structure of objects and the environment. Since awareness of motion and the generation of mental maps of the structure of our environment are critical components of animal (and human) vision, the conversion of this innate ability to a computer capability is similarly crucial in the field of machine learning.



According to nVidia the Optical Flow Acceelrators in the Ada Lovelace architecture are made to be “significantly faster” than their previous version version:

DLSS 3 overcomes this problem by running optical flow calculations using the updated Ada architecture Optical Flow Accelerator, which significantly speeds up optical flow processing compared to prior generation OFA units.



And this is what Tom’s Hardware had to say about this speed up on the OFA:

Nvidia's DLSS 3 technology relies on the Ada's fourth generation Tensor cores and the new Optical Flow Accelerator (OFA) to work. OFA isn't some new invention that Nvidia recently cooked up with Ada. OFA has been present since the Turing days. The difference is that Nvidia has tremendously improved it in Ada versus Ampere, and it now delivers higher performance and better quality. The OFA in Ada is reportedly 2 to 2.5 times faster than on Ampere. Nvidia has seemingly made some algorithmic changes as well.


Ampere and Turing can theoretically leverage DLSS 3, but they won't yield the same benefits. Catanzaro said that DLSS 3 likely won't boost frame rates on Ampere or Turing — on the contrary, owners would probably experience laggy gameplay and bad image fidelity.



Which nVidia also said:

The NVIDIA Ada Lovelace Architecture has a new optical flow accelerator, NVOFA, that is 2.5x more performant than the NVIDIA Ampere Architecture NVOFA. It provides a 15% quality improvement on popular benchmarks including KITTI and MPI Sintel.




Further information on the OFA of ORIN:


And here’s the OFA in the generations:




And with that GitHub thing about how T23x have the same or similar drivers for their OFA, where T234 is meant for automotive and T239 is not, it’s likely that T239 shares the OFA with T234 and due to its use case and the field it is meant for, T239 and T234 have an OFA that is closer in performance to that of the Ada line in offering incredible precision or to be incredibly fast for its usage similar to the Ada OFA.

If nVidia deemed the OFA in regular ampere to not be fast enough but the one in Ada to be fast enough, it’s possible they used that for ORIN to make its use case much quicker for the automotive vehicles. Being slow to make a decision would have its own detriment in the cars with moving objects and detection of moving objects. Imagine the car hit someone? Like a SMALL CHILD??




The OFA in Ampere and Turing seems as though it is more suited for video playback.



Funnily enough, Optical Flow is technically possible on older hardware according to nvidia, but it was only for video and nothing more. However, these were in the NVEC/NVDEC blocks of the silicon (the TX1 has these). And it seems as though if Ampere and Turing used the OFA in other use cases they would actually slow down and lag behind.



So, it is possible that due to it having specialized drivers for the OFA, that Drake and Orin have OFAs closer to that of the Ada generation than the Ampere generation.


Here’s extra information of interest:







And for the record, the Tegra have always had something completely different from how it’s found in the desktop architecture than the architecture they have inside…. because reasons.
 
Last edited:
I just don’t understand why would Nvidia use the T239 over a variant of Orin in a new Shield TV?
I would assume that the production volume for Drake is going to be very, very high compared to Orin, and that consequently there are going to be a lot of binned chips that aren’t good enough for Nintendo hardware but are plenty powerful enough for something like a Shield.

No guarantee they’re using binned chips, of course, but it seems like a good way to get more value per wafer.
 
I don’t think that the single core would matter too much for the case of games these days that rely a lot more on multithreading than single threaded.

But I could be wrong on that when it comes to games running on ARM.
Looking at the clock speeds, number of cores, and the scaling of the multi-core scores, the A15 could very well beat out Drake on this count as well.
 
Considering the segment that ORIN occupies, which is meant for automotive, the Optical Flow Accelerators in it has to be incredibly precise in that area.





According to nVidia the Optical Flow Acceelrators in the Ada Lovelace architecture are made to be “significantly faster” than their previous version version:





And this is what Tom’s Hardware had to say about this speed up on the OFA:





Which nVidia also said:






Further information on the OFA of ORIN:


And here’s the OFA in the generations:




And with that GitHub thing about how T23x have the same or similar drivers for their OFA, where T234 is meant for automotive and T239 is not, it’s likely that T239 shares the OFA with T234 and due to its use case and the field it is meant for, T239 and T234 have an OFA that is closer in performance to that of the Ada line in offering incredible precision or to be incredibly fast for its usage similar to the Ada OFA.

If nVidia deemed the OFA in regular ampere to not be fast enough but the one in Ada to be fast enough, it’s possible they used that for ORIN to make its use case much quicker for the automotive vehicles. Being slow to make a decision would have its own detriment in the cars with moving objects and detection of moving objects. Imagine the car hit someone? Like a SMALL CHILD??




The OFA in Ampere and Turing seems as though it is more suited for video playback.



Funnily enough, Optical Flow is technically possible on older hardware according to nvidia, but it was only for video and nothing more. However, these were in the NVEC/NVDEC blocks of the silicon (the TX1 has these). And it seems as though if Ampere and Turing used the OFA in other use cases they would actually slow down and lag behind.



So, it is possible that due to it having specialized drivers for the OFA, that Drake and Orin have OFAs closer to that of the Ada generation than the Ampere generation.


Here’s extra information of interest:







And for the record, the Tegra have always had something completely different from how it’s found in the desktop architecture than the architecture they have inside…. because reasons.

Assuming it's too late to change the OFA to support dlss3, would improving it to support it on a pro model be possible with minum compatibility issues?
An easier version of the question, maybe: is it possible to swap dlss2 .dlls for dlss3 ones on Windows?
 
I would assume that the production volume for Drake is going to be very, very high compared to Orin, and that consequently there are going to be a lot of binned chips that aren’t good enough for Nintendo hardware but are plenty powerful enough for something like a Shield.

No guarantee they’re using binned chips, of course, but it seems like a good way to get more value per wafer.
It's a solid assumption. One way or another, those binned chips from T239 fabrication will be used for something, a new Shield just makes sense because it's been a long time since it got a hardware refresh and new models seem to align with new chips being manufactured for Nintendo (like new Shield models coming out following Mariko production in October 2019).
 
It's a solid assumption. One way or another, those binned chips from T239 fabrication will be used for something, a new Shield just makes sense because it's been a long time since it got a hardware refresh and new models seem to align with new chips being manufactured for Nintendo (like new Shield models coming out following Mariko production in October 2019).
I am very much just a thread-junkie with zero industry expertise, so I appreciate the second opinion!
 
0
Assuming it's too late to change the OFA to support dlss3, would improving it to support it on a pro model be possible with minum compatibility issues?
An easier version of the question, maybe: is it possible to swap dlss2 .dlls for dlss3 ones on Windows?
It does take extra work to implement frame generation. as its really a completely seperate module thats only callled DLSS for marketing reasons. In other words, it is not an improved version of DLSS, it is a seperate technology that can work in concert with DLSS, or FSR for that matter.
 
Looking at the clock speeds, number of cores, and the scaling of the multi-core scores, the A15 could very well beat out Drake on this count as well.
The Drake that Nintendo will use? Yes. Because they’ll clock it low.

Drake as a whole? No.

At 3GHz all cores, all 8 of the performance cores, it would trounces the A15


Apple silicon is good don’t get me wrong, but it wouldn’t be that much in its favor in this scenario.


Apple A15 Bionic is a 6-core processor, only 2 of those cores are huge performance cores and they can clock to 3.23GHz

And it is accompanied by 4 “efficiency” cores that don’t go above 2GHz.

The A78 at max can clock to 3.3GHz. And has 8 of them. And they are performance cores.


If the comparison was to the M1 or the M2? Yes it would be bested by those.

Assuming it's too late to change the OFA to support dlss3, would improving it to support it on a pro model be possible with minum compatibility issues?
An easier version of the question, maybe: is it possible to swap dlss2 .dlls for dlss3 ones on Windows?
DLSS 2 and DLSS3 are two different techniques with the same branding.

And it is too late to change it. But I don’t think that the change is greatly needed….
 
Last edited:
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