• Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.

Ventura Beat: Nintendo Switch graphics are based on Nvidia's Maxwell Architecture

ggx2ac

Member
Which is exactly what Nvidia officially stated in their press release.

Does the article provide any actual new information?

They did? It's vague because it only says their top of the line GPUs or whatever it was. They never specified Maxwell or Pascal.
 

120v

Member
Maybe it's Nvidia "custom" tegra as they refer.

i think that's the brass tacks here. how well it can scale, how it can handle games about 3-5 years out

a rough example is the 3DS, it's basically a gimped ps2 but can do other things far better like surface rendering. i expect a hazy middle ground with this thing
 
If they don't provide reasons why Nintendo would be using a 20nm process when 16nmFF is much better to manufacture more chips per wafer and to reduce power consumption than this article should be taken with a lot of salt until we can find out more from elsewhere than can corroborate info.

Also, isn't it more EXPENSIVE to use an older abandoned process versus a newer and in production process?
 
It's a bust! The article so poorly written that they are going into damage control.

If they don't provide reasons why Nintendo would be using a 20nm process when 16nmFF is much better to manufacture more chips per wafer and to reduce power consumption than this article should be taken with a lot of salt until we can find out more from elsewhere than can corroborate info.

What's especially bad is that we still know nothing about the CPU or RAM either, it's always been about the GPU...

Well the reasoning they give is that "it's not ready yet" but based on what I've read here I don't see how that's true.

Which is exactly what Nvidia officially stated in their press release.

Does the article provide any actual new information?

Nvidia never mentioned 20nm when talking about the Switch.
 

Vic

Please help me with my bad english
If this is true then what the FUCK is Nintendo doing.

Why would you use a older architechure when Pascal is right there. Only thing I can think is cost but the loss in power and espically efficency is going to be huge.
As a lot of people have mentioned, the design of the GPU could be based on the Maxwell architecture, but on a 16nm process. We just don't know at this point.
 
That leak make no sense.... Pascal is cooler and consumes less power than Maxwell at the same speed. #teamfake.

Not much difference if they use Maxwell with 16nm.

Nonetheless the technical details are superfluous, as even some third parties stated Switch can run current games. Nintendo made some of the most graphical impressive games of the current gen on WiiU Hardware...
 

SpacLock

Member
I don't know what this means, but judging by everyone's reactions I'm glad I kept my Wii U for Zelda.

Just one more misstep along the path that will lead Nintendo to becoming third party.

I am 100% behind this.

Super Mario Run money will keep them in the business for two more console generations. You have no idea how much they're about to rake it in...
 

Polygonal_Sprite

Gold Member
It now specifies a custom Maxwell Tegra that uses a 20nm process

I bet they made these changes based on Thraktor's post lol

I hope they give Thraktor a cut of their click bait ad revenue ;).

In all seriousness a site changing their story in real time, piece by piece as people poke holes in their points should be all the proof someone needs that they're full of shit. At least in this particular case.

They probably heard that the Switch dev kit uses Maxwell and saw dollar signs. I still believe the final retail unit will use Pascal.
 
This sucks if it's 28nm. The jump to 16nm with pascal was huge for peformace/watt and a smaller die as well.

I hope it's not true. 16nm is a must for portables now. Damn.
 
Just one more misstep along the path that will lead Nintendo to becoming third party.

I am 100% behind this.

I don't know what this means, but judging by everyone's reactions I'm glad I kept my Wii U for Zelda.

Guys... This article states that they expect 1TFlop from the Switch GPU which is far above any of our expectations. Reading this thread title and assuming bad without seeing how the info in the article makes no sense in either direction is just silly.

You guys were expecting a portable PS4 Pro at $200? Lol expectations are out of control for this thing.

Literally nobody here expected anything close to that. Think before you post please.
 

ggx2ac

Member
Also, isn't it more EXPENSIVE to use an older abandoned process versus a newer and in production process?

It's not abandoned exactly, it's just that there's a lot less 20nm chips being manufactured which is what makes it more expensive compared to 28nm or 16nmFF.

Lots of chips are still manufactured on 28nm which makes the process still cheap compared to 20nm.

Lots of chips have moved from 28nm -> 16nmFF or 20nm -> 16nmFF which again makes 16nmFF chips cheaper to manufacture compared to 20nm because there are a lot more 16nmFF chips being manufactured compared to 20nm.
 

wapplew

Member
Who care what tech switch use when it can run games like Zelda and that pretty grass game.
Finally result is good, no need to know the nitty gritty.
 

thefro

Member
This sucks if it's 28nm. The jump to 16nm with pascal was huge for peformace/watt and a smaller die as well.

I hope it's not true. 16nm is a must for portables now. Damn.

It's not going to be 28 nm... it's either 20 nm (what the Tegra in the Shield TV is) or 16 nm.
 
Disappointing considering this thing isn't even out yet. Imagine what the gap will look like in 3 years. Third parties will continue to ignore Nintendo.
 

TLZ

Banned
I have a few immediate thoughts after reading through the article:


  • Firstly, it's worth noting the difference between Maxwell and Pascal is almost entirely down to the manufacturing process. Maxwell was made on 28nm (and in the case of the TX1, 20nm) whereas Pascal is made on 16nm. The actual architectural difference between the two is minimal, and aside from improved color buffer compression, largely irrelevant for a device like the Switch.
  • Despite that, the article never makes any mention of the manufacturing process. I find that extremely strange, as it's obviously the defining difference between the two sets of GPUs.
  • In fact, the article gets the difference between the two completely the wrong way around, saying "Nintendo’s box is relatively small, and so it has to fit into the heat profile of a portable device, rather than a set-top box. That’s another reason that explains the older Maxwell technology, as opposed to the Pascal’s state-of-the-art tech." Pascal is literally a more power efficient version of Maxwell, so the incentive would be the other way around.
  • The author says "we expect the Nintendo Switch to be more than 1 teraflop in performance", which is notably higher than even those of us who were expecting Pascal were considering (I literally posted earlier today with a 500-750 Gflop estimate). If this is a Maxwell chip, then that would mean at least 4 SMs (512 "CUDA cores") at 1GHz, as they're not going to be able to push much past that on 28/20nm. This is a much larger GPU than most people would have been expecting.
I see a few different scenarios here:


  1. The Switch SoC uses Maxwell at 20nm, and simply has a much larger GPU than anticipated to account for the performance.
  2. Nintendo looked at the feature-set planned for Pascal when design started, realised that the new features were largely irrelevant, and decided that they would save time and just use a straight-forward die shrink of Maxwell to 16nm instead. That would technically be a Maxwell GPU, but would be almost completely indistinguishable from Pascal in terms of performance.
  3. The sources are wrong about Maxwell, the 1 Tflop performance, or both.
Basically, if you're to take the article as being accurate, then the only worthwhile takeaway is this quote:



A Maxwell Tflop is identical to a Pascal Tflop, and it's largely irrelevant to us whether they achieved that by using a larger Maxwell GPU on 20nm/28nm at a lower clock or a smaller Pascal GPU on 16nm at a higher clock.

Thank you mister.
 
I don't know what this means, but judging by everyone's reactions I'm glad I kept my Wii U for Zelda.

The article is full of technical errors, the author seems a bit naive. It seems some of the questionable passages got already "corrected" or deleted. In the end it matters not much if its Maxwell or Pascal. Switch will be much more powerful than WiiU anyway.
 

Inuhanyou

Believes Dragon Quest is a franchise managed by Sony
Only read the first page but man, everybody jumped in believing this, but I bet if he said its on par with PS4 nobody would believe him

I was skeptical to begin with about the conclusions made because we already know Nvidia is releasing a brand new Shield TV, presumably at 16nm process

So why they would charge Nintendo without convincing them to use a lower process node as well made no sense...but whatever.

Again, it generally doesnt even matter. Maxwell at 16nm gets rid of most of he inhibitions to using its previous 28nm form anways.

And regardless of what it uses, its still going to be far more powerful architecturally advanced than Nintendo's previous unit, which is the main thing. And all the other components of the machine are far more powerful than Wii U too.
 

BitStyle

Unconfirmed Member
Who care what tech switch use when it can run games like Zelda and that pretty grass game.
Finally result is good, no need to know the nitty gritty.

Well the nitty gritty would be interesting to know, but I don't have confidence that the author of this article even knows the nitty gritty. Just a lot of assumptions off of a bit of info they received, which I would guess is either the Maxwell architecture or the 1 Tflop. I'm guessing the former.
 

Nanashrew

Banned
bumping Thraktor's post for more to see. Also since it's not in the OP like it should be.

I have a few immediate thoughts after reading through the article:


  • Firstly, it's worth noting the difference between Maxwell and Pascal is almost entirely down to the manufacturing process. Maxwell was made on 28nm (and in the case of the TX1, 20nm) whereas Pascal is made on 16nm. The actual architectural difference between the two is minimal, and aside from improved color buffer compression, largely irrelevant for a device like the Switch.
  • Despite that, the article never makes any mention of the manufacturing process. I find that extremely strange, as it's obviously the defining difference between the two sets of GPUs.
  • In fact, the article gets the difference between the two completely the wrong way around, saying "Nintendo’s box is relatively small, and so it has to fit into the heat profile of a portable device, rather than a set-top box. That’s another reason that explains the older Maxwell technology, as opposed to the Pascal’s state-of-the-art tech." Pascal is literally a more power efficient version of Maxwell, so the incentive would be the other way around.
  • The author says "we expect the Nintendo Switch to be more than 1 teraflop in performance", which is notably higher than even those of us who were expecting Pascal were considering (I literally posted earlier today with a 500-750 Gflop estimate). If this is a Maxwell chip, then that would mean at least 4 SMs (512 "CUDA cores") at 1GHz, as they're not going to be able to push much past that on 28/20nm. This is a much larger GPU than most people would have been expecting.
I see a few different scenarios here:


  1. The Switch SoC uses Maxwell at 20nm, and simply has a much larger GPU than anticipated to account for the performance.
  2. Nintendo looked at the feature-set planned for Pascal when design started, realised that the new features were largely irrelevant, and decided that they would save time and just use a straight-forward die shrink of Maxwell to 16nm instead. That would technically be a Maxwell GPU, but would be almost completely indistinguishable from Pascal in terms of performance.
  3. The sources are wrong about Maxwell, the 1 Tflop performance, or both.
Basically, if you're to take the article as being accurate, then the only worthwhile takeaway is this quote:



A Maxwell Tflop is identical to a Pascal Tflop, and it's largely irrelevant to us whether they achieved that by using a larger Maxwell GPU on 20nm/28nm at a lower clock or a smaller Pascal GPU on 16nm at a higher clock.
 

bomblord1

Banned
Guys... This article states that they expect 1TFlop from the Switch GPU which is far above any of our expectations. Reading this thread title and assuming bad without seeing how the info in the article makes no sense in either direction is just silly.

Wait, that's insane that's not even saying just 1 teraflop that's saying ABOVE 1 teraflop.

The system should be beating the Xbox One across the board with that at least while docked.
 

AgeEighty

Member
Look, I have a lot of doubts about the OP article, but I just can't follow this line of defending it if it's true.

Article - Switch uses Cortex A57 instead of A72 despite the latter being better in every characteristic, power draw, performance, and transistor budget

Defence - it's unlikely the A72 would have been used for higher performance because it's battery limited.


Well, thing is, being more efficient means you can turn up either performance, battery life, or both in various ratios. More efficient is of course always better given a fixed power budget.

I never said Pascal isn't "better". I said anyone thinking this will mean a major difference up or down in Switch's graphics—either docked or not—isn't understanding what the relationship between Maxwell and Pascal actually is. As multiple other people in this thread have also said.

I even posted earlier that if it is Maxwell, people would be justified in disappointment at possible battery issues. But I disagree that the extra power would have been used for major graphical boosts in docked mode, because I don't think Nintendo would want to create any major differentials between the two modes. Full 1080p play on TVs I see them doing, but nothing beyond that... and Pascal's power isn't necessary to pull that off.

Additional power isn't required to drive 1080p?

Not what I wrote.
 
I hope they give Thraktor a cut of their click bait ad revenue ;).

In all seriousness a site changing their story in real time, piece by piece as people poke holes in their points should be all the proof someone needs that they're full of shit. At least in this particular case.

They probably heard that the Switch dev kit uses Maxwell and saw dollar signs. I still believe the final retail unit will use Pascal.

Ok, this is confirmed clickbait bs then.

Disappointing considering this thing isn't even out yet. Imagine what the gap will look like in 3 years. Third parties will continue to ignore Nintendo.

Not in the age of PS4 Pro, Scorpio, and Nintendo being known to upgrade their handhelds, they could very well just do that with the Switch in 2/3 years.
 
How many tablets have built-in fans for cooling, have a dock that has built-in fans for cooling while docked. I have not seen anyone carrying a "Tablet", with any moving parts or been in any store where a dock is available for purchase to increase cooling and performance.

I keep reading it being labeled as a "Tablet" as if it's exactly that and should be lumped together with Ipads, Galaxy Tabs, or Microsoft Surface.

I don't know what to think about Switch when it comes to power or performance, but calling or labeling it a Tablet seems a little misguided.
 

guybrushfreeman

Unconfirmed Member
Can't wait for the speculation to be over. I see a lot of people here referencing rumours that have had literally no backing to them at all. I feel like the published articles paint a pretty clear picture of the device, be wary of twitter and Neogaf rumours telling you what you wanted to hear. Either way I'm glad we'll finally be getting the facts soon
 

bomblord1

Banned

EDarkness

Member
I'd like to point all you guys out to that one thread were I called it that the Switch would be no more powerful than a Wii U and was literally assaulted by nay-sayers.

http://www.neogaf.com/forum/showthread.php?p=222824819&highlight=#post222824819

Where are you now, OrbitalBeard, Mariolee, Doctre81, KingSnake, Seik, etc..??

How about some of that crow?

I don't blame you for not reading the article but you probably should've read at least one page of the thread
 

Richie

Member
I'd like to point all you guys out to that one thread were I called it that the Switch would be no more powerful than a Wii U and was literally assaulted by nay-sayers.

http://www.neogaf.com/forum/showthread.php?p=222824819&highlight=#post222824819

Where are you now, OrbitalBeard, Mariolee, Doctre81, KingSnake, Seik, etc..??

How about some of that crow?

...Are you being ironic or something? Pretty sure that even the worst case scenario posits the Switch's innards as notoriously more powerful than the Wii U's.
 
Top Bottom