Wii U 'Has A Horrible, Slow CPU' Says Metro Last Light Dev

I doubt you'll find a single person with that fear today.

 

In order for the consumer to care Nintendo would need to educate them about it. Is Nintendo advertising their low power consumption as a feature? Buy a board game it's 33w greener than the WiiU.

 

Remember, in business there is the 20/80 rule.
Secondly, its common sense they will make their console for the Japanese
market because they are Japanese living in a Japan.
Most of their developers are there, as well as their HQ.
They will not make a product that wont be successful at home.

 

If there was no fear that Nintendo will have the PS2 equivalent for next gen, people would not care about how powerful it is. I dont see anybody getting hung up on the OUYA specs.
The number of threads focusing on the WiiU clearly shows people fear WiiU's influence on MS and Sony if its popular. Wii has done this already.

 

I seriously do admire your ability to find scenarios in which these design decisions will either somehow be beneficial to the WiiU or at least not to its total detriment.

Really? A lack of AAA games, or said another way, a market comprised mostly of shit games, will help the WiiU because it will have the ability to compete on that playing field.

And Indie developers? From everything I've seen, indie developers already have a haven in the 360 ecosystem. Who is paying $350 to buy a WiiU for low-powered indie games when the indie developers are already putting their product out on the 360, available for $99?

But yes, I suppose if the stars align just right the WiiU could be seen as boon for those developers that don't have the time or money to compete in the AAA arena and for some reason don't want to target the already existing 360 and PS3 user bases.

 

Which speaks to the ridiculousness of the entire idea, let alone that it was a marketing decision Nintendo based their design decisions on. But again, I give Lumpy credit for attempting to find any bright spot, or any potential angle that makes the WiiU design a logical, intelligent decision.

Video game consoles are not requirements. There's a huge difference between buying a "green" video game console or buying an electric car. Those buying electric cars still see the vehicle as a necessity, for whatever reason, as public transportation isn't an viable option.

There is no necessity in video game consoles. If you're concerned about energy consumption, you simply don't buy one. And while you're at it, turn off the TV it would be plugged into as well.

 

What are you basing that on?

The threads focusing on the WiiU exist because it's new, and because it's so shockingly poorly designed that people are beside themselves trying to figure out what exactly Nintendo was thinking.

This isn't fear. This is rubbernecking at a traffic accident.

And what exactly was Wii's influence on MS and Sony? From the current rumors, it certainly doesn't appear that MS decided they want to copy Nintendo's strategy of a cheap, underpowered console with limited on-line capabilities for their next console. It appears that MS is very happy with their 360 success and plans to build on it with their next, high powered, console.

If anything, the WiiU demonstrates that Nintendo has finally realized that you need a online network and entertainment system integration and can't be "just" a gaming console anymore. Of course, this happened a generation too late as both MS and Sony have huge leads in this area.

 

I disagree, I wish consoles would just go back to being exactly that..

That's why I loved PS1 & PS2 so much, no frills, no stupid gimmicks....just put your disks in and play...

Now it seems that if your console doesn't have the ability to flash advertising in your consumers face it's not doing its job properly.

 

Apparently there's no need to worry about that situation.


By the way, energy is just getting a lot cheaper with new gas sources. Industrial production is going to move back to the United States in the next years because it'll be cheaper than making it in China and shipping it around the world.

Not that we should start wasting energy again, but in this case the marginal gains far outweigh the negative consequences of underpowered hardware.

 

Also, when did GPGPU became a magic bullet? I think a lot of people have absolutely no idea what it can be used for, especially when there's also graphics work to do, and they're just repeating it to avoid facing the truths about the Wii U...

 

Yeah and we didn't have a console launch in years

 

Yep I agree.
GPGPU is anything but a magic bullet.
There are just certain classes of problem that are not well suited to it.
For it to save WiiU the the class of problems it is good for would have to dominate the CPU workload of modern games and I don't think that's true.
I spoke to the Havok guys recently despite the demo they did a few years ago, they don't think rigid body dynamics is a good fit on the GPU and they are looking for other "physics" effects to use the GPU for instead.
And all of this is dependent on WiiU actually having a significant amount of compute performance on the GPU to begin with.

 

Oh I don't know, but I heard there's this "motion control" thing that was rushed into the PS3's launch and eventually made its way into a X360 peripheral.

Not the cheap & underpowered, but google for "xbox smartglass" and you'll see how at least some of Nintendo's ideas keep being.. adapted by Microsoft.

Yap, had the Wii U been launched until late 2010, it'd be awesome.
The Wii had already faded by then and a 4-year cycle would've been more than appropriate for a cheap-ass, underpowered console.
2 years later than that, on the verge of the real next-gen systems coming up, is just dumb.

 

Indeed the lack of CPU is extremely bothering, especially as the other next gen system might push the market further. I expect most PC gamers to move (even one on a budget which some are still playing on AMD dual/tricore) to quad cores configurations as a basis.
Anyway the system is ill designed, it is a waste of silicon imo.
There are no offshelves CPUs* now that would have been the perfect target for Nintendo, I give that to them, still... what did they spend their R&D budget on?

(* in my opinion, proper CPU cores would have been either Jaguar and A15, IBM has nothing in that range for now, Intel is not an option. Jaguar are not ready and it seems that A15 hard design are meant for 32/28nm lithography which means a premium in costs. The closest would have been Power a2 but it is not optimal in anyway and requires the use IBM more expansive process, ultimately as Nintendo used that process that should have been their choice).

Well Patcher was right, which happens quiet often no matter companies may prove him right by their decisions, his analysis are more than often sound.
With regard to hardware, the CPu would still suck even in 2010, though it would have been less unacceptable as a trade off at that time.

 

Wouldn't the poor memory bandwidth be an issue with GPGPU anyway?

 

So it's starting to look like the WiiU GPU might not even be faster than Xenon in every way.

But that's not really the most worrying thing IMO - a video Ika linked to in the Digital Foundry thread shows Darksiders 2 taking much longer to load than the Xbox installed version, and it even manages to be slower than the vanilla Xbox DVD version.

http://www.youtube.com/watch?v=5nha4XiXnSg

Anyone want to take a punt on the speed of the BluRay (but not) optical drive in the WiiU? 2X CLV like the PS3 and most BR players? I'd expected something like a 4X CAV drive, but that should be faster than the Xbox DVD drive iirc. Presumably it's not a data layout and seeking issue when they've got so much more space on the drive.

 

Yes, a lot of the challenge with GPGPU is trying not to be memory bound.
Though it's an area the EDRAM might help with if it can be used for intermediate buffers.

 

First off, you seem to have pigeon holed me as some kind of WiiU sympathiser. I'm not, I'm just trying to have a rational conversation without simply saying "its screwed, end of". We all said that last time and we were all wrong - so forgive me for trying to look at things from all angles.

Secondly you haven't read my post properly or are ignoring the context. I didn't say a lack of triple A games would be a benefit, did I? No. I said the only saving grace is that for the next year or so a least, there aren't going to be hundreds of triple A titles which the WiiU won't be able to receive a port of ( although there will inevitably be some it misses out on, as we've seen already). After that, it's going to go the same ways of the Wii most likely.

Thirdly: My comment on Indie devs was based on the remarks by some developers that the lack of charges for patches/dlc and the ease at which they can get their games on the shop, set the prices themselves and update their games is reminiscent of Steam and bodes well for the Indie ecosystem. I didn't say 360 wasn't already a good place for Indie devs. I didn't even comment on another platform I just said WiiU seems to be a good fit for small developers so far. No need to get defensive when no one is attacking, dude.

Again, read my post. I'm not trying to defend their decisions, I'm just trying to find the justification. Unless you think there is none? Maybe you think this is just Iwata and Myamoto in a room spitballing ideas and not a rather large global company that researches and designs it's hardware thoroughly to meet its specification, whatever that specification be.

I suspect the WiiU design brief was for a small, low power machine which could receive ports from 360, BC with wii, incorporate the lag free second screen etc and without being sold at a loss. And that's what they've come up with.

Whether its a crap decision or not, there's always a rationale. They're obviously quite proud of their power consumption, that's the first thing Iwata bragged about when detailing the specs isn't it?

And besides (if you'd read my post), you'd see that I mentioned in response to Shifty that it's not so much a USP as a convenient 'string to its bow'. Or at least that's how I assume Nintendo sees it.


"Quit straw manning me, man!"

 

You're putting words in my mouth there. I haven't said it was a magic bullet. I was just saying what Nintendo was probably banking on: the other two new consoles being heavily GPU dependent aswell rather than CPU dependent. Looks like they were wrong, or if the weren't then they've gimped their CPU far too much anyway to remain competitive for more than the next year or so.

As far as I knew GPGPU are designed to take the reliance off of CPUs and allow the GPU to be used for some of the processes the CPU would usually be required for and more fully take advantage of parralel processing (using OpenCL) In theory this negates the need somewhat for a more powerful CPU. My thinking was that if the other two new consoles also went down the same route of using a GPGPU based system then porting from them would be more straight forward than if the the games being ported were based on platforms with a heavy reliance on a powerful CPU.

Not sure if that's correct, but that's what I was saying referring to. In any case it seems from the leaked dev kits for Durango/PS4 that they will both have substantially more capable CPUs (aswell as mahoosive GPUs) and games will of course take full advantage of that. Nintendo will have gimped their CPU way, way too much.

That's probably all wrong

 

On the contrary, offering backwards compatibility purely with hardware makes it harder to enhance the games. Nintendo's modus operandi has always been to provide the same hardware as previous gen, either augmented or ran in slightly different modes to facilitate some new feature or cordoned off from the new hardware entirely. I don't think Nintendo has ever pushed previous gen software emulation except for the Zelda port on Gamecube.

How MAME emulates something and how XBox 360 probably emulates XBox are pretty different. The latter is going to be more comparable to user mode emulation with a lot of OS level HLE. If the games don't access hardware directly then a bulk of the CPU emulation can be reasonably fast, since memory accesses don't have to be guarded or trapped and are probably happy being located wherever. The mention of kernel knowledge for the task seems to reinforce this.

Xenon's cores suck clock for clock vs XBox's CPU as compilation targets, but for emulation purposes the situation isn't that bad. PPC has a lot of registers compared to x86 so you don't have to worry about dynamic allocation schemes, plus you have extra space to perform some amount of renaming, memory caching, different views of partial registers, etc. Having multiple condition code registers is also nice because it means you can do emulation-additional processing without overwriting the emulated flags, and it can make combining partial flags updates easier. And they probably could get away with doing almost nothing but CPU emulation on one Xenon core. So that gives about 4 cycles for one emulated cycle, which in this case is very aggressive but not totally unrealistic (also considering that not all XBox games would have run the CPU anywhere close to full speed, a lot were probably bandwidth limited with the GPU)

Because XBox only had 4-wide float only SSE that executed over two cycles and 64-bit MMX you could capture it fairly reasonably in Xenon's SIMD. You need to aggressively schedule to make everything else play nice, but having all the extra registers at least lets you unroll and rename pretty aggressively in software.

Given that they didn't get great compatibility they could have taken a lot of shortcuts, like assuming the ABI is followed across function calls and returns.