Wii U hardware discussion and investigation *rename

Status
Not open for further replies.
It seems that feeding the framebuffers of a 1080p screen and 4 tablets was never in their plans, though...

Where are you getting this from? Because of the 1 tablet / WiiU thing during E3? This would be in fact a possibility, but they could have changed it because of the reactions after E3 and I guess there would have been enough time before the tapeout. We will see.
 
If I am not wrong the internal L1 Caches and L2 Caches of a GPU are SRAM... what if we change it for small pools of 1T-SRAM that gives us more storage. How it can impact the rendering performance?
 
What would be needed for textures to be included?

Most modern games require 256MB as a minimum requirement, most gaming cards start with 1GB of available graphics RAM and even the Vita has 128MB of dedicated memory for a 5" 540p screen..
And AFAIK the textures take quite a good chunk of the graphics memory. Just look at the amount of games getting post-release high-resolution texture packs that get released for PC "ports" because of the higher available graphics memory.



Yes, but he said never. GC and 3DS have already shown that's not absolute.
Apparently, the 3DS being sold at a loss was a translation glitch from someone at adriasang. I don't really believe the console costs more than ~140€ to make.



LOL. Take your pick.
As pointed out by Shifty Geezer, that's only one source and not really a very reliable one..




And the latter is a "stereotype" that I've never understood that I can only blame on bias. Nintendo goes underpowered once, and suddenly that's all they're know for. And it's those views about Nintendo that they, IMO, have virtually no shot of overcoming unless they did something stupid and made a $500 console that few people would buy. To me they would just find other reasons to justify their bias.

"Once"?!

First of all, I didn't say, underpowered, I said underspecced.
The stereotype is justified with all the consoles Nintendo released during the past 10 years.

2001 Gamecube: similar "power" to PS2 (even though it launches a whole year later) but cuts on the DVD/CD media capabilities, underpowered compared to XBox 1 launching in the same year.
-> underspecced.

2004 NDS: uses a 67MHz ARM9 + 33MHz ARM7 + 4MB RAM + very basic GPU with no texture filtering, in a time where we get the much more powerful PSP with dual 333MHz CPUs + DX7-class GPU + 32MB RAM.
There are also Windows Mobile devices using XScales @ >600MHz + PowerVR MBX with 64MB RAM and 16MB dedicated graphics RAM.
-> underspecced and underpowered

2005: Gameboy Micro: Well.. it's a small Gameboy Advance.. not very relevant either but nonetheless it's underspecced and underpowered.

2006 Wii: No need to describe this one -> both underspecced and underpowered.

2011 3DS: Dual-core ARM11 CPU @ 268MHz + OpenGL ES 1.1 GPU + 128MB RAM, in a time where the most basic smartphone has a 600MHz ARM11 + OpenGL ES 2.0 GPU + 256MB RAM.. and high-ends are getting dual-core Cortex A9 @ 1GHz + 512MB/1GB RAM + fairly powerful OpenGL ES 2.0 GPUs. -> Maybe not "underspecced" because of the 3D screen and other stuff, but definitely underpowered.

2012 WiiU: ?



What would that be based on?
Where are you getting this from? Because of the 1 tablet / WiiU thing during E3? This would be in fact a possibility, but they could have changed it because of the reactions after E3 and I guess there would have been enough time before the tapeout. We will see.


Yes, because during E3 they never showed the console doing more than 1 main screen + 1 tabletmote.
When questioned about the issue, Satoru Iwata claimed they were studying the possibility of connecting a second tabletmote because of the apparent public demand.

The option of having 4 tabletmotes + mainscreen hasn't even been pondered, from the looks of it. The tabletmote won't even be sold separately, you'll need 4 WiiUs in order to get 4 tabletmotes.
 
The option of having 4 tabletmotes + mainscreen hasn't even been pondered, from the looks of it. The tabletmote won't even be sold separately, you'll need 4 WiiUs in order to get 4 tabletmotes.
They'll have to sell replacements, and if Nintendo decide to support 2+ Wuumotes, they'll need to provide a means to buy a second. I don't think support for 4 Wuumotes is on the cards, but even if it is that won't change the specs Nintendo are looking into. It'll just mean less resources to go around when playing multiplayer.
 
I find it strange that Wii U can't support four or more Wuupad. You can netwerk DSes together to play games like DQ9 for example. So in term of multiplayer functionality Wuupad is somewhat inferior to DS. That's just wrong.

I am sure Nintendo will make 3DS as an alternative to Wuupad somehow, maybe using an additional add on to 3DS. The Wii U GPU should be able to handle it, afterall ATI does have eyefinity, which support six screens. Four Wuupad should be easy.

I still think it's better for Nintendo to cancel the Wuupad and release something like my propose DS tablet with its own CPU and GPU, if they're so afraid of the popularity of iPad. The Wii U unit itself is so small they might as well pick a lesser CPU and GPU and integrate everything into the Wuupad and make it into a standalone product. Nintendo third pillar if you will (that's what they called DS initially)

Then make a Wii HD with significantly higher spec CPU and GPU. Budget that don't go into what Ninty called expensive controller could go into better CPU and GPU for Wii HD and release it closer to the X720 and PS4 launched.
 
It may be the wireless communication that sets the limit.

If they stream the display data (display size/resolution/colour depth/frame rate), it´s quite a lot of information. Having several devices there is more risk for interference with other wireless devices. They need to design these things to work well under most conditions.
 
They'll have to sell replacements,

There are ways to guarantee replacements without selling the controller through the retail channels. For example, if you damage the tabletmote, you must send the console for repair or something.. At E3, Nintendo sent the message of the tabletmote as being "part of the new console" quite clearly..

I'm just stating what they said during E3 and later interviews.
I obviously think they should make the tabletmote a purchaseable peripheral with the WiiU being able to use up to 4 devices.



I find it strange that Wii U can't support four or more Wuupad. You can netwerk DSes together to play games like DQ9 for example. So in term of multiplayer functionality Wuupad is somewhat inferior to DS. That's just wrong.

Networked DSes exchange position and player actions through Wifi and each console renders its own game.
The tabletmote has no rendering or processing capabilities at all, it receives a video+audio stream from the main console and send back the remote input.
The amount of data between Wii U and tabletmote is probably thousands of times larger than the data transferred between DSes.

A fair comparison would be how many Wii Us you could connect through WiFi with one of them being used as a server. I'd say some dozens isn't unrealistic. But that's a moot point. Who would buy several Wii U's for one house?
 
Last edited by a moderator:
I find it strange that Wii U can't support four or more Wuupad. You can netwerk DSes together to play games like DQ9 for example. So in term of multiplayer functionality Wuupad is somewhat inferior to DS. That's just wrong.
Yep. The Wii U is basically a Gamecube HD with "connectivity" made even worse, somehow. I expect Gamecube-like success from the thing.
 
I find it strange that Wii U can't support four or more Wuupad. You can netwerk DSes together to play games like DQ9 for example. So in term of multiplayer functionality Wuupad is somewhat inferior to DS. That's just wrong.

I am sure Nintendo will make 3DS as an alternative to Wuupad somehow, maybe using an additional add on to 3DS. The Wii U GPU should be able to handle it, afterall ATI does have eyefinity, which support six screens. Four Wuupad should be easy.

Nintendo already said they were looking to use 3DS as a controller for the WiiU.
Then, they announced a peripheral that added buttons to makes the 3DS more like the WiiU tablet. So I do think they are going in that direction.




I still think it's better for Nintendo to cancel the Wuupad and release something like my propose DS tablet with its own CPU and GPU, if they're so afraid of the popularity of iPad. The Wii U unit itself is so small they might as well pick a lesser CPU and GPU and integrate everything into the Wuupad and make it into a standalone product. Nintendo third pillar if you will (that's what they called DS initially)

Well thats what I was wondering, what would they need to put into the controllers, for them to support of extra controllers? Just enough memory?


Then make a Wii HD with significantly higher spec CPU and GPU. Budget that don't go into what Ninty called expensive controller could go into better CPU and GPU for Wii HD and release it closer to the X720 and PS4 launched.

Why do people keep saying this? Nobody knows what moves Sony and MS are going to make with their future consoles. Especially if they are launching only a year after the WiiU.




But I think other reasons why Nintendo was not pitching multiple controllers was for the following reasons:

1. Transition: The first year they want Wii owners to understand that their wiimotes can be used with the new console. In other words, they want Wii owners, in particular their large 'casual gamer' market, to invest in the new console. A market that probably bought many accessories for "family" or "group" oriented games.

2. Value. By restricting the availability of the WiiUController they want to raise its value, toward the level of the iPad. In this way they can get away with charging plenty of $$$ for it. And this might mean they would be willing to sell the console as a LOSS LEADER.

3. No games: Nintendo are currently developing games that make use of ONE tablet controller for the launch. And probably will launch the additional controller with a game that supports 2 to 4 extra controllers a year or two later. Possibly with Smash Bros.

4. Attracting the "Core" crowd. They see "Core" players as persons who play single player games. So why would they initially need more than one controller?



So there are many reasons why the issue was not technical, but more due to the market.
 
Most modern games require 256MB as a minimum requirement, most gaming cards start with 1GB of available graphics RAM and even the Vita has 128MB of dedicated memory for a 5" 540p screen..
And AFAIK the textures take quite a good chunk of the graphics memory. Just look at the amount of games getting post-release high-resolution texture packs that get released for PC "ports" because of the higher available graphics memory.

I guess we might be talking from a different perspective based on your response. I'm looking at this similar to the Xenos daughter-die of eDRAM. The issue with it was that the 10MB wasn't enough for a 720p image correct? I'm looking strictly at 1T-SRAM being embedded. This is where I was trying to get better understanding at.


Apparently, the 3DS being sold at a loss was a translation glitch from someone at adriasang. I don't really believe the console costs more than ~140€ to make.

Whether that is the case or not all that matter is "never" can be used in relation to that.


As pointed out by Shifty Geezer, that's only one source and not really a very reliable one..

I responded with the correct term for it. It goes beyond just spending on hardware. A lot of people who like to look at teardowns forget all the other expenses that should be factored in. Sony and MS don't lose a billion in a year early on due to a few million consoles selling. The wholesale price vs retail price is obvious, but not enough to generate that kind of loss.

"Once"?!

First of all, I didn't say, underpowered, I said underspecced.
The stereotype is justified with all the consoles Nintendo released during the past 10 years.

2001 Gamecube: similar "power" to PS2 (even though it launches a whole year later) but cuts on the DVD/CD media capabilities, underpowered compared to XBox 1 launching in the same year.
-> underspecced.

2004 NDS: uses a 67MHz ARM9 + 33MHz ARM7 + 4MB RAM + very basic GPU with no texture filtering, in a time where we get the much more powerful PSP with dual 333MHz CPUs + DX7-class GPU + 32MB RAM.
There are also Windows Mobile devices using XScales @ >600MHz + PowerVR MBX with 64MB RAM and 16MB dedicated graphics RAM.
-> underspecced and underpowered

2005: Gameboy Micro: Well.. it's a small Gameboy Advance.. not very relevant either but nonetheless it's underspecced and underpowered.

2006 Wii: No need to describe this one -> both underspecced and underpowered.

2011 3DS: Dual-core ARM11 CPU @ 268MHz + OpenGL ES 1.1 GPU + 128MB RAM, in a time where the most basic smartphone has a 600MHz ARM11 + OpenGL ES 2.0 GPU + 256MB RAM.. and high-ends are getting dual-core Cortex A9 @ 1GHz + 512MB/1GB RAM + fairly powerful OpenGL ES 2.0 GPUs. -> Maybe not "underspecced" because of the 3D screen and other stuff, but definitely underpowered.

2012 WiiU: ?

I didn't mean to infer you were stereotyping them. It's the non-Nintendo fans that tend to do that.

You're doing the same thing Shifty did. Handhelds are irrelvant in this discussion. They are on their own path. I'd say the GC was closer to Xbox than PS2 and was hindered by other things than power. Not saying it was on the same level as Xbox, but if PS2 is 1 and Xbox is 10, I would say GC was 5-7 depending on the games. Anything lower was probably developer effort since Xbox was essentially a PC in a big black box giving some devs a more familiar setting.


Yes, because during E3 they never showed the console doing more than 1 main screen + 1 tabletmote.
When questioned about the issue, Satoru Iwata claimed they were studying the possibility of connecting a second tabletmote because of the apparent public demand.

The option of having 4 tabletmotes + mainscreen hasn't even been pondered, from the looks of it. The tabletmote won't even be sold separately, you'll need 4 WiiUs in order to get 4 tabletmotes.

Actually Miyamoto said that. Iwata claimed multiple controllers could be used, but that cost was the issue. Then this also came out not long after E3.

http://www.next-gen.biz/news/report-wii-u-can-stream-multiple-controllers

According to sources close to Japanese gaming website Game Watch - translated by Nintendo World Report - the Wii GPU, supplied by AMD, has two output modes, one of which supports the streaming of up to four "SD images".

Riccittello also seemed to let it slip that EA will be using multiple controllers for their games.

The real question IMO is how much will Nintendo use it, not if it's possible. Nintendo tends to put things in their system they don't use, but this is something the consumers would want so I doubt it goes unused.

I find it strange that Wii U can't support four or more Wuupad.

It can. See the article quote I posted above. Also AMD's press release said the GPU supported multiple displays. So we know the GPU has Eyefinity.
 
On a side note, with the announced delay of the next gen from AMD and nV to 2012, I'm finding it hard to believe they'll have 28nm supply for WiiU for launch.
 
On a side note, with the announced delay of the next gen from AMD and nV to 2012, I'm finding it hard to believe they'll have 28nm supply for WiiU for launch.

I still have some doubts as well as I didn't think any of the next consoles would be able to use 28nm due to possibly more (before this latest one was confirmed) delays, but if NEC is providing the fab like they did for GC and Wii it may be possible.
 
Yahbut Nintendo is spending on more than just chips. I should have been more specific when stating that number in that the amount was "$6 operating profit".
That still doesn't add up. Looking at the amount of silicon in Wii, you could buy way more in a PC graphics card for way cheaper. So unless Nintendo have the world's worst contract negotiators and their payments for outdated tech and 1TSRAM are exhorbitant, there is no way Wii was scraping by $6 at a $250 launch price. Another comparison is PS2. That was selling at a profit when $150 and cheaper. Same distribution and packaging and ancillary costs as shipping Wii. The accelerometer and Wiimote in a Wii controller doesn't come to $100, not even adding in 64 MBs GDDR3. :p

I cannot in any way reconcile the cost of Wii to not being ludicrously profitable at launch, and a look at Nintendo's financials seems to support that too. They made an absolutely killing this round and a good lot of that must have come from selling lucrative hardware.

There are ways to guarantee replacements without selling the controller through the retail channels. For example, if you damage the tabletmote, you must send the console for repair or something.. At E3, Nintendo sent the message of the tabletmote as being "part of the new console" quite clearly..
I would take that to mean intrinsic to the experience, like Wiimote. It would be an odd statement to say at E3 that you absolutely won't supply Wuumotes outside of a boxed product - you'd leave that to decision to final logistics when you've decided what to allow and not.
 
That still doesn't add up. Looking at the amount of silicon in Wii, you could buy way more in a PC graphics card for way cheaper. So unless Nintendo have the world's worst contract negotiators and their payments for outdated tech and 1TSRAM are exhorbitant, there is no way Wii was scraping by $6 at a $250 launch price. Another comparison is PS2. That was selling at a profit when $150 and cheaper. Same distribution and packaging and ancillary costs as shipping Wii. The accelerometer and Wiimote in a Wii controller doesn't come to $100, not even adding in 64 MBs GDDR3. :p

I cannot in any way reconcile the cost of Wii to not being ludicrously profitable at launch, and a look at Nintendo's financials seems to support that too. They made an absolutely killing this round and a good lot of that must have come from selling lucrative hardware.

But you're not factoring in marketing and distribution costs (which include retailer profits). The tech is only a part of the equation. On the profit side, you're forgetting licensing fees from third party games (54 games were million sellers), and first party games that have even a larger margin for profit (22 games were million sellers). And that still doesn't break it down to multi-million sellers.
 
But you're not factoring in marketing and distribution costs (which include retailer profits).
Distribution costs are no more for Nintendo than any other CE deivce, like PS2. If Sony can break-even on a $100 piece of kit distributing it worldwide, Nintendo can do the same for a similarly priced bit of kit. Marketing costs can't be factored into per-unit profit/loss because marketing is a sunk cost distributed across units. Let's say Nintendo spent $50 million on marketing. The first Wii sold would have lost them $50 million per unit if we count to cost of marketing in with the per unit losses! The second Wii sold would have brought that down to $25 million loss per unit. And so on, with a dwindling cost. Marketing is a business loss, but never measured per unit, so an analyst's evaluation that Nintendo makes $6 per Wii would mean out of $250, with a portion to sellers, a portion to distributors, and a portion on manufacturing and licensing, Nintendo have managed to spend $244 achieving what Sony achieved for <$100. Doesn't make sense!
 
Distribution costs are no more for Nintendo than any other CE deivce, like PS2. If Sony can break-even on a $100 piece of kit distributing it worldwide, Nintendo can do the same for a similarly priced bit of kit. Marketing costs can't be factored into per-unit profit/loss because marketing is a sunk cost distributed across units. Let's say Nintendo spent $50 million on marketing. The first Wii sold would have lost them $50 million per unit if we count to cost of marketing in with the per unit losses! The second Wii sold would have brought that down to $25 million loss per unit. And so on, with a dwindling cost. Marketing is a business loss, but never measured per unit, so an analyst's evaluation that Nintendo makes $6 per Wii would mean out of $250, with a portion to sellers, a portion to distributors, and a portion on manufacturing and licensing, Nintendo have managed to spend $244 achieving what Sony achieved for <$100. Doesn't make sense!

We're getting OT so I'll let you have the last response. There's no way you can believe that everyone has similar distribution costs. That by itself and in general makes no sense at all. All I'm doing is debating it from that perspective of the analyst who said "operating profits". That's totally different than how you are trying to compare the PS2 (I can practically guarantee that "break-even" doesn't factor operating costs) and factors in all the things you are trying to say are not included, plus some things I left out previously. It doesn't make sense to you because we don't see it taken from that perspective. By the way are you really trying to compare console that was near its end at that time to a console that was at its peak when this was done? And the way you achieved that conclusion is just weird. However it would have helped to see his total breakdown on how he reached that conclusion. As I mentioned before (which I should have been clearer), this doesn't seem to include money made from games. You said you were looking at the financials, and guess what those do include? And not just Wii games, but everything related to DS as well. No way you can support your view from looking at that.

In the end we'll probably never truly know, but just depending on tech costs is a flawed way of coming to a conclusion.
 
On a side note, with the announced delay of the next gen from AMD and nV to 2012, I'm finding it hard to believe they'll have 28nm supply for WiiU for launch.

I never believed it for a second.

Armchair message boarders are always massively too aggressive on process shrinks. Just go to any negoaf next gen speculation thread...they'll probably be penciling in PS4/XB720 for 18nm fall 2012 or something...
 
I never believed it for a second.

Armchair message boarders are always massively too aggressive on process shrinks. Just go to any negoaf next gen speculation thread...they'll probably be penciling in PS4/XB720 for 18nm fall 2012 or something...

I find it hard to believe, however it's not pure messageboard speculation. Unless we assume Urian is lying then the info comes from an investor at Mosys.
 
Status
Not open for further replies.
Back
Top