By using this site, you agree to our Privacy Policy and our Terms of Use. Close

Forums - Nintendo Discussion - Digital Foundry: Nintendo Switch CPU and GPU clock speeds revealed

Frogger on 19 December 2016
zorg1000 said:
spemanig said:

50m for a combined two platforms in the most lucrative market in gaming over the course of almost 6 years is textbook abysmal.

But out of curiosity - how did PS360 do in the west over the course of 6 years?

Its certainly not record breaking amazing by any means but "will absolutely die in the west" is certainly an exaggeration.

Not entirely sure about PS3/360, over 100 million though.

Handheld software competing in a home console marketplace absolutely will. It's like trying to sell iPhone games on a Mac(book).



Around the Network
Frogger on 19 December 2016

Terrel from Neogaf said:

I'll say it again, but the lessons we should have learned with GameCube clearly weren't learned.

GameCube CPU: 485MHz, FPU 1.9GFLOPs
GameCube GPU: 162MHz
GameCube RAM: 43MB

PS2 CPU: 294MHz, FPU 6.2 GFLOPs
PS2 GPU: 147MHz
PS2 RAM: 32MB

Xbox CPU: 733MHz, FPU performance unknown (?)
Xbox GPU: 233MHz
Xbox RAM: 64MB

At the time of their comparison in 2001, GameCube was labeled "garbage-tier" compared against the Xbox and just barely better than PS2, with its floating-point performance being regularly singled out.

And we all remember how things panned out that generation: PS2 was the weakest, naturally, but Xbox wasn't this massive unparalleled technology leap compared to any of them. How every component works with the total package in real-world performance is the only way to measure a console.

Nintendo clearly demonstrated its design philosophy, a philosophy that always gets overlooked because it's not something you can use as bait when trolling: Optimal RAM and cache for fewer wasted CPU/GPU cycles. I don't expect Switch to be any different in that regard. How optimized the design is as a whole will be the question, but as always, we'll have to wait until January to know for sure.

Thought this was an interesting point made on Gaf. Also me salvaging any hope left for a multiplat machine. Fingers crossed for "good enough."



superchunk on 19 December 2016
Shadow1980 said:
So, it's got a better CPU than the Wii U, but the GPU is about the same power as the Wii U's when docked and half as powerful when used as a portable.

Is that about the gist of it?

No.

Better CPU
Better RAM
Better GPU (while slightly less flops, better capability) when portable
2.5X (or more) better GPU when docked



Soundwave on 19 December 2016
spemanig said:

Terrel from Neogaf said:

I'll say it again, but the lessons we should have learned with GameCube clearly weren't learned.

GameCube CPU: 485MHz, FPU 1.9GFLOPs
GameCube GPU: 162MHz
GameCube RAM: 43MB

PS2 CPU: 294MHz, FPU 6.2 GFLOPs
PS2 GPU: 147MHz
PS2 RAM: 32MB

Xbox CPU: 733MHz, FPU performance unknown (?)
Xbox GPU: 233MHz
Xbox RAM: 64MB

At the time of their comparison in 2001, GameCube was labeled "garbage-tier" compared against the Xbox and just barely better than PS2, with its floating-point performance being regularly singled out.

And we all remember how things panned out that generation: PS2 was the weakest, naturally, but Xbox wasn't this massive unparalleled technology leap compared to any of them. How every component works with the total package in real-world performance is the only way to measure a console.

Nintendo clearly demonstrated its design philosophy, a philosophy that always gets overlooked because it's not something you can use as bait when trolling: Optimal RAM and cache for fewer wasted CPU/GPU cycles. I don't expect Switch to be any different in that regard. How optimized the design is as a whole will be the question, but as always, we'll have to wait until January to know for sure.

Thought this was an interesting point made on Gaf. Also me salvaging any hope left for a multiplat machine. Fingers crossed for "good enough."

That's not really a great comparison. The PS2 had a beefy CPU design because Ken Kutaragi was into that (PS3 also had the CELL design which a relatively powerful CPU for its time). 

The GameCube GPU was definitely better than the PS2. 



Frogger on 19 December 2016

Also, can someone please explain to me what SMs are and why the expectation is 2 SMs so I'm not barking like an idiot?



Around the Network
superchunk on 19 December 2016
Bristow9091 said:
How do these specs compare to PS4, Pro, and XBO? And how about other handheld like the 3DS and Vita?

I have a thread fo rthat.

http://gamrconnect.vgchartz.com/thread.php?id=222261

 

1/3 XBO (you can figure out rest)

many, many times better than 3DS or Vita.



Frogger on 19 December 2016
Soundwave said:
spemanig said:

Thought this was an interesting point made on Gaf. Also me salvaging any hope left for a multiplat machine. Fingers crossed for "good enough."

That's not really a great comparison. The PS2 had a beefy CPU design because Ken Kutaragi was into that (PS3 also had the CELL design which a relatively powerful CPU for its time). 

The GameCube GPU was definitely better than the PS2. 

Why is it not a good comparision?

Not a challenge. A question. The point of the post is that looking at flops alone can send a misleading image.



zorg1000 on 19 December 2016
spemanig said:
zorg1000 said:

Its certainly not record breaking amazing by any means but "will absolutely die in the west" is certainly an exaggeration.

Not entirely sure about PS3/360, over 100 million though.

Handheld software competing in a home console marketplace absolutely will. It's like trying to sell iPhone games on a Mac(book).

So Breath of the Wild & Dragon.Quest XI are handheld software competing in a home console marketplace?



When the herd loses its way, the shepard must kill the bull that leads them astray.

potato_hamster on 19 December 2016
Goodnightmoon said:
potato_hamster said:

Because console games are heavily optimized specifically for the hardware they're ran on. Games will run worse on the switch if game engines have to be designed to constantly have to check the operating mode of the device rather than just use hardcoded values. It's the same reason PS4 games run and look better on the PS4 than games on similarly spec'd PCs.

They can optimize it for both modes, find the way to optimize it the best possible way while docked and then try to downgrade different parts until it works fine undocked then set that configuration to be used automatically when the game is undocked.

Optimizing it for both modes isn't nearly as effective as optiizing it for one, and it's still a considerable amount of work vs the amount for work to optimize for say, a PS4 or Xbox One. Notice how many PS4 pro game modes are minimal improvements if developers decide to do them at all? That's because it's a lot of work to really take advantage of that extra hardware. Now with the Switch, that would either be mandtory - or the more likely option - developers will just develop for the undocked mode and not really bother with optimizing for the docked mode because that's extra work that isn't required.

What that effectively means is that most third party games will essentially be running on hardware that may be weaker than the Wii U.



Soundwave on 19 December 2016
spemanig said:
Soundwave said:

That's not really a great comparison. The PS2 had a beefy CPU design because Ken Kutaragi was into that (PS3 also had the CELL design which a relatively powerful CPU for its time). 

The GameCube GPU was definitely better than the PS2. 

Why is it not a good comparision?

Not a challenge. A question. The point of the post is that looking at flops alone can send a misleading image.

It's misleading because the PS2/PS3 are somewhat unusual designs in that they have an overpowered CPU relative to the GPU. Think of it like a bodybuilder with huge forearms but compartively average/small-ish biceps. 

The GameCube was better than the PS2 on the GPU side and that accounts for most of graphical capability.