Wii U Speculation Thread of Brains Beware: Wii U Re-Unveiling At E3 2012

Status
Not open for further replies.
lunchwithyuzo said:
Nintendo launched Wiimotion+ with 3rd party games from EA, 2K and Sega, their own marquee title (Wii Sports Resort) didn't come until a month later. Can you imagine that happening on N64 or even GC?
Yes. The N64 Expansion Pak was released in 1998 so games like Turok 2 and Rogue Squadron could use it. Nintendo didn't use it until Donkey Kong 64 a year later.
 
BurntPork said:
And the original 360 didn't have an internal HDD either. And Wii U has 1/3rd the volume of the 360.

In the best case scenario, Nintendo could maybe stuff a 50W GPU in there, and even that may be pushing it.
Is that true? If so, then wow.

wsippel said:
See, that's the great thing about 470S: It can. With four cores at 1.6GHz, it's slightly faster than Xenon. And it's much easier to develop for, too. Only aspect where it doesn't outperform Xenon is floating point performance, which doesn't really matter if you have an OpenCL compliant GPU. Also, with a CPU consuming only 6W and a nice and efficient SoC, you have a significantly higher thermal budget for the GPU side all of a sudden.
The Xenon has bottlenecked several 360 games. Let's say Nintendo go with your CPU choice, how do you expect it not to massively bottleneck the whole machine?
 
Mr_Brit said:
The Xenon has bottlenecked several 360 games. Let's say Nintendo go with your CPU choice, how do you expect it not to massively bottleneck the whole machine?
OpenCL, basically. Leave it to the GPU.

Also, Xenon is shit. The whole PPE concept is terrible. In-order, long pipeline, not enough cache.
 
StevieP said:
Battlefield 3, Witcher 2, and Metro 2033 are the only recent examples that I can think of where games started on the higher spec'd platform and worked their way down.

I just meant higher end specs in general, not specifically PC to ps360 for example. This applies to the console only area as well. Picture how bad CoD would have turned out if the Wii was the lead platform and was ported up to the ps360. Instead it was ported down to the Wii and still looked great for that platform. Same thing last gen with games like Splinter Cell. Sure the xbox version looked the best, but when ported down to the PS2, it still looked very good for that platform. That series wouldn't have looked nearly as nice on the xbox if it were initially designed for the PS2 and ported up on the xbox.
 
We still have time until the launch of the Wii U but since software for a home console takes a long time to develop, and these aspects are strong especially from both the perspectives of content and quantity, we needed to disclose information about the Wii U to many developers so we can have a solid lineup at the time of its launch. On the other hand, there was great public interest in "what Nintendo will propose after the Wii." We were very thankful for this but also experienced a tough challenge from an information management perspective. Unless we officially informed the public of the overall configuration of the system, we knew that fractions of information would spread in a way we did not intend, and as a result, we would not be able to present the product as intended. Therefore, it was slightly early but we announced the basic configuration of this system at E3.

You mentioned, "good feedback and bad feedback" but there seems to be a great difference in how it's viewed between someone who has actually had hands-on experience at E3 and someone who saw our presentation far away using the Internet. Of course, maybe not 100% of the people who were at the show said the system was excellent or not everyone who saw it from far away had negative comments, but I have the impression that, clearly, there is a great difference in consumers' evaluations depending on whether someone has had actual experience with the system or not. We learned a key lesson at the most recent E3, that it is always important to consider what is occurring inside the exhibition and presentation, but for people who are watching the show through the Internet, the scale of these people is much larger and now that their discussions using social media spread out in society, unless we effectively present our message to these people, even if the people who were there say, "It was great!," the presentation will not be a great success. In fact, I was interviewed by the media many times at E3, and as I recall, over half of the reporters said, "congratulations" at the beginning or end. I think we can say that "congratulations" is great praise for a presentation and I only experience this once every few years when our show is good. I do not have any close relationships where reporters would say such a thing just to flatter me, so I think the response at E3 this year was very good. As I said, however, since some people outside there had a completely different impression, this is something we need to think deeply about.

Also, I just mentioned "home console" unwittingly, but although the Wii U was announced as the successor of the Wii, I believe it will be positioned slightly different than the past home consoles. In the past, there were two categories of gaming systems, handheld devices and home consoles, and there was thought to be a great wall in between those two categories; however, the Wii U will be connected to a home TV, but it is not a simple home console, meaning it is not something "playable only when facing the TV." This is because, the controller has its own screen, so even if you are not facing the TV or the TV is occupied for some other reason, you can still use the game system. Therefore, in our release, we have written "a new game console" and not "a new home console." In English, both are translated as "console" and therefore, this nuance will not be expressed but, in the Japanese releases, we would like to propose the Wii U without using the term "home console." For your information, for game systems that have been sold as handheld devices, many consumers do not actually use them outside of their homes. Given this fact, we proposed the idea of carrying the Nintendo 3DS outside by enhancing the communication functionality and challenging consumers to change their usual routine but, on the other hand, there is a very high demand for playing a game system inside the house without using the TV, and we would like to answer that demand with the Wii U. In other words, since our lifestyles are changing minute by minute, I think that a play style where a person plays only when the person sits in front of the TV and has to occupy the TV is becoming less and less fitting with the times. Therefore, I would like to make the Wii U Nintendo's new answer to the upcoming change of lifestyle.

In addition, I would like to use this occasion to talk about something I was hoping to talk about someday. That is, what is going to happen to the TV? I think there are people in here who know more about this topic than I, but when you look at things like electronics magazines, they basically state that "TVs will have multiple screens," or "TVs will work with smartphones." One of the Wii U's proposals is, "Even if you don't buy the latest TV, the existing TV in your house will become multi-screen." In households in Japan, I believe many families have recently replaced their TVs to watch terrestrial digital media broadcasting as the transitional period to which ended just the other day, and I do not imagine that mass amounts of those new TVs will be replaced in the near future. One of Wii U's proposals is that the TV will turn into a multi-screen TV that works very closely with the Internet, without replacing the TV. In September 2006, we announced the Wii and said, "The Wii will change the relationship between games, the TV, the Internet and a family," but we are considering a proposal of a higher dimension this time. Of course, we do not intend to take the aspect of giving the user the best game experience lightly at all. It is an extremely important factor. But it alone will not be able to make the Wii U something relevant for everyone in the family. Such other elements will be important to make the Wii U something relevant for everyone in the family. Therefore, please think that the Wii U is a product that is a combination of a proposal for a "new play style that is not restricted to the TV," a proposal for a "new kind of entertainment using multi-screens" and also "the future of TV." This year is very important period for us because we can say it is a challenge for the Nintendo 3DS, and since we have announced that the Wii U will not be launched by the end of March next year, I do not intend to speak about it anymore in detail, but I would like to speak about it more when the appropriate time comes, but the Wii U is not simply a Wii with a different user interface and we are not proposing to society a performance-enhanced Wii, rather we are proposing hardware with a totally different concept and vision.

As for the current situation for the Nintendo 3DS, as I said earlier, if each specific topic that I mention, for example, "this is what was wrong" or "that is what was wrong," is picked up, it will not be beneficial for future Nintendo 3DS business. Therefore, I will not mention any particular topics here, but I will say that we learned many lessons, and from these lessons, we would like to make a well-prepared situation by the year-end holiday season, and propose it to society to change the situation drastically. - Satoru Iwata

source http://www.**********.com/?mode=viewstory&id=158119

smart
 
Mr_Brit said:
Is that true? If so, then wow.

Yeah, its close to that. From the Nintendo website Wii U: 1.8x6.8x10.5=128.52 in^3

Xbox Slim:: 10.6x2.95x10.39=324.89 in^3

WiiU ~ 39% of the size of the Slim.

IMO, Nintendo would probably need some pretty expensive (for a console) cooling for that form factor to house significantly more powerful CPU/GPU than the 360/ps3. That's why I think a 470s-derivative CPU, that's very power efficient, and something along the lines of what B3D expected as a reasonable GPU is what we should expect.

With those types of specs, WiiU should be a pretty cool, efficient, and affordable system.
 
Mr_Brit said:
The Xenon has bottlenecked several 360 games. Let's say Nintendo go with your CPU choice, how do you expect it not to massively bottleneck the whole machine?
Not trying to give my prediction about WiiU's cpu here, but when talking about Xenon being a bottleneck, one needs to also know what kind of bottleneck it is. If it is a FPU bottleneck, then a 470 may or may not pose a similar FPU bottleneck, based on (a) how many cores there actually are, and (b) what is the SIMD block per core (as Xenon has really potent SIMD). Furthermore, some kind of FPU work can indeed be offloaded to a GPU, if the spare GPU cycles are there. For anything outside the FPU domain, a 4-6 core 1.6GHz 470, particularly one equipped with a massive edam pool, would run circles around Xenon (which was L2-starved, and generally of not particularly good performance per single thread).
 
Hiltz said:
So we'll have to wait until next year to know of the Wii U's price and release date.

It wouldn't surprise me if there was an event in January like the 3DS event last January 19th. That's less then half a year from now.
 
Hey all. Been out of the U.S. for a good portion of the summer and from the looks of it, we don't really have much new info? Only thing that puzzles me is what has led to people abandoning the >3Ghz Tri-core PPC rumors. Have they been discredited in some way? Or people just can't believe that some custom IBM processor (using technology found in Watson to some unknown degree and not necessarily any POWER7 variant out on the market today) would run cool enough to fit in the case? Didn't the final cases for the 360 and/or the PS3 gain an almost comical ammount of ventilation in comparison to their prototypes.

Actually, I wouldn't mind if the changed the case design entirely. It's completely uninspired and overly conservative in aesthetic.
 
KageMaru said:
I just meant higher end specs in general, not specifically PC to ps360 for example. This applies to the console only area as well. Picture how bad CoD would have turned out if the Wii was the lead platform and was ported up to the ps360. Instead it was ported down to the Wii and still looked great for that platform. Same thing last gen with games like Splinter Cell. Sure the xbox version looked the best, but when ported down to the PS2, it still looked very good for that platform. That series wouldn't have looked nearly as nice on the xbox if it were initially designed for the PS2 and ported up on the xbox.

Examples like yours are the exception, not the rule. Starting from the high end and going down, I mean. And for all intents and purposes, PC from 2006 -> forever will probably be the high end platform, thanks to the way the GPU has evolved.

Also, Xenon and Cell are shit. The whole PPE concept is terrible. In-order, long pipeline, not enough cache.

Fixed. The PPE (in both machines) is a generally poor processor. Anandtech had originally compared it to a 1.6ghz P4 or something along those lines. While it was hyperbole (and got pulled, from what I recall) it wasn't like they were talking out of their ass either.

What bothers me about the PPC 470 series being discussed here, is that it flies in the face of what IBM has already stated about its CPU.
 
StevieP said:
What bothers me about the PPC 470 series being discussed here, is that it flies in the face of what IBM has already stated about its CPU.

imo, it's not really contradictory as they haven't said anything other than it being manufactured on 45nm SOI and using EDRAM.

A custom design CPU design can be based of the 470s cores paired with EDRAM manufactured on 45nm SOI.
 
Fourth Storm, as far as Wii U news goes, Nintendo did confirm that it will be allowing DLC and microtransactions for third-party games. However, Nintendo itself is generally not interested in charging for add-on content in its own games. Although, Nintendo is only open to the idea if it can offer the right amount of content that gamers would find acceptable to pay for. Also, Nintendo may have prematurally revealed that Wii U may offer downloadable GameCube titles during an interview with Nintendo of America's Amber McCollum. Nintendo responded to her comment by saying that "To clarify the capabilities of the Wii U system: As correctly stated, Wii U will not play Nintendo GameCube discs, however Nintendo has not made any announcements regarding downloadable content." Most of us see it as a backhanded confirmation.
 
Mr_Brit said:
Is that true? If so, then wow.
I just double checked.

Original 360: 403.3 sq. in.
Wii U: 128.5 sq. in.

So, yeah. It's actually slightly less than a third. That's why I feel that it can't possibly have more than 480SPs at 40nm, and 28nm isn't an option unless they delay it into early-mid 2013. (Granted, it already looks like they've internally delayed it to holiday 2012.) It's also why the CPU can't possibly exceed Xenon by a significant amount.

Anyone hoping for 800SPs under any circumstance will be disappointed, and yes, this does mean that PS4/XB3 games would be maxed-out compared to Wii U on minimum settings at a lower resolution and possibly an unstable frame rate on many games. If Nintendo doesn't make some major changes, it'll have to sell gangbusters to get AAA games during the later half of the generation not because they can't be ported, but rather because the graphical difference will be so huge that gamers will not feel like they're getting the full experience on Wii U.
 
McHuj said:
imo, it's not really contradictory as they haven't said anything other than it being manufactured on 45nm SOI and using EDRAM.

A custom design CPU design can be based of the 470s cores paired with EDRAM manufactured on 45nm SOI.

I assure you that Watson hadn't a single 476SP in its thousands of processors.
 
StevieP said:
I assure you that Watson hadn't a single 476SP in its thousands of processors.
Certainly (Blue Gene, however, is based on the predecessor). And the comparison to Watson was extremely vague. "Same technology" can mean anything, from very basic (POWER7 and 470S are both POWER - same technology) to similar features (eDRAM) to "basically the same chip". My money is on option two: They share a few design details, but Wii U doesn't use POWER7.
 
wsippel said:
Certainly (Blue Gene, however, is based on the predecessor). And the comparison to Watson was extremely vague. "Same technology" can mean anything, from very basic (POWER7 and 470S are both POWER - same technology) to similar features (eDRAM) to "basically the same chip". My money is on option two: They share a few design details, but Wii U doesn't use POWER7.

I'm of the belief that they wouldn't have even mentioned the word "Watson" if this were a 400 series PPC. They would've just said "Power PC" like they did with the Wii. Although I highly doubt that there is a Power7 chip in the Wii U (it's simply overkill) I might be inclined to believe that it still may be based on it in some fashion, which is why they mentioned those specific things (Watson, large EDRam pool, 45nm SOI).
 
StevieP said:
I'm of the belief that they wouldn't have even mentioned the word "Watson" if this were a 400 series PPC.
You're giving a throwaway headline way, way too much credit.

People don't know wtf PPC is. They've heard of Watson.
 

No word on Wii U price and release date this year

Nintendo president Satoru Iwata concerned 3DS price drop could damage early adopter trust, explains approach to microtransactions, mobile gaming, and more.




Last week, Nintendo announced that it would badly miss its profit guidance and would be dropping the price of the 3DS by one-third less than half a year after the system's launch. Understandably, shareholders had some questions in hand for the company's postearnings investors Q&A session with president Satoru Iwata.


The Wii U is still a ways off.
While Iwata played the company's specific plans close to the vest, he did tell investors not to expect news of a Wii U release date or price for some time yet.

"Since the Wii U we showed you at the E3 show in June was still in the development phase without very specific proposals on the software titles, we are going to announce the release date and the price next year when we are able to explain the specific proposals," Iwata said.


Iwata offered that bit of information in an explanation of how the 3DS price drop would impact plans for the Wii U. Specifically, he was concerned with how it impacted the trust early adopters had placed in Nintendo.

"What we have to take most seriously is that the price markdown could damage the trust of the consumers who bought the Nintendo 3DS just after the launch," Iwata said. "I feel greatly accountable for it. Our decision of the price markdown this time has a side effect that, at the launch of the Wii U, people may feel that the price might drop in the near future if they wait."

Iwata also fielded investors' questions about a pair of emerging markets Nintendo has not been hasty to embrace: smartphones and microtransactions. He began by explaining Nintendo's focus on its own portable systems instead of the latest and greatest cell phone platforms.

"In the past, when games were becoming available with Java technology on cell phones, before smartphones appeared on the scene, we were often told that no one would buy handheld game systems once people could play games on their cell phones," Iwata said. "As a result, however, I believe that we have gone above the limits with Game Boy, Game Boy Advance, and Nintendo DS by continuing to try new things which games for cell phones cannot realize and to offer unique software in connection with hardware. Therefore, we are willing to do the same things in the future--we think that we can keep the high significance of our own platforms' existence by creating and offering software which cannot be realized on the other common platforms and by receiving the consumers' appreciation of them."

As for microtransaction-based games, Iwata stated that the free-to-play model doesn't fit with Nintendo's vision of the value in its games, but he said it's entirely possible the company would offer additional stages to games for a small fee. He also noted that regardless of Nintendo's own microtransaction plans, the 3DS and Wii U will support downloadable content for games should third parties decide they want to implement it.

"I'd like to emphasize that this is only me talking from Nintendo's point of view of what we want Nintendo to do," Iwata said, "and we do not intend to comment on whether another company is right or wrong, and I would like to avoid any misunderstandings on this point, but I would also like to mention that, under Nintendo's set of values, 'charging money just for changing the parameters to unlock something or to allow some large advantage' is a totally different earnings structure that is not compensation for creative work and, while pursuing this may create short-term profits, Mr. Miyamoto and I discussed that we should not use this type of billing system since we think that we will not be able to make long-term relationships with our consumers."

http://www.gamespot.com/news/6327000/no-word-on-wii-u-price-and-release-date-this-year
 
Furianshi said:
I would be very happy with 360 keeping it's lead platform status until next gen. Most 360 games seem to be built for better hardware anyway, they're not consistently 1080p, and 30fps seems to be the standard. So if 360 remains the lead platform (which it will), hopefully we'll enter a period between Wii U release and next gen where the average 3rd-party game on Wii U will be 1080p/60fps..

Sir, you are setting yourself up. Don't hate Nintendo when this is very far from reality.
 
There is strong chance that WiiU will get bigger console casing if they stumble upon heat distribution issues. This could be also the reason for keeping the console somewhere behind in slightly modified Wii cases and without any clear view of the console itself.

I think that tablet will stay mostly the same, but console itself could get facelift.
 
DrM said:
There is strong chance that WiiU will get bigger console casing if they stumble upon heat distribution issues. This could be also the reason for keeping the console somewhere behind in slightly modified Wii cases and without any clear view of the console itself.

I think that tablet will stay mostly the same, but console itself could get facelift.

Its more likely every chip will be down-clocked.
 
Luckyman said:
Sir, you are setting yourself up. Don't hate Nintendo when this is very far from reality.

Dude, can you at least come up with a more effective troll scheme? This hit and run thing is getting boring. I want drohne back :(

There is strong chance that WiiU will get bigger console casing if they stumble upon heat distribution issues.

They are already having heat issues with the current kits, and kits are generally way larger than final hardware.

Its more likely every chip will be down-clocked.

Not with Backwards compatibility in play, if the chip is a 400 series doing hardware-level emulation, it will likely be a multiple of Broadway.
 
Log4Girlz said:
Its more likely every chip will be down-clocked.
Possibly, but serious heat issues would require quite underclock and I think that developers would be pretty pissed about this.

They can mess with console case until it gets into production, around 4 to 6 months before launch.
 
StevieP said:

For the last time, this is the engadget quote:

IBM tells us that within the Wii U there's a 45nm custom chip with "a lot" of embedded DRAM (shown above). It's a silicon on insulator design and packs the same processor technology found in Watson, the supercomputer that bested a couple of meatbags on Jeopardy awhile back.

No where is there a direct quote with the word Watson used by IBM in the Engadget article or IBM PR.

Who was it that told this to Engadget? An engineer working on the chip or an over zealous PR person trying to create hype? Did the writer mangle what was said?

You continue to put way to much stock into one tech blogs hype.
 
DrM said:
Possibly, but serious heat issues would require quite underclock and I think that developers would be pretty pissed about this.

They can mess with console case until it gets into production, around 4 to 6 months before launch.
They could always just use better cooling. People overestimate the cooling in modern consoles and how much it could be improved for not a much more cost.
 
StevieP said:
I'm of the belief that they wouldn't have even mentioned the word "Watson" if this were a 400 series PPC. They would've just said "Power PC" like they did with the Wii. Although I highly doubt that there is a Power7 chip in the Wii U (it's simply overkill) I might be inclined to believe that it still may be based on it in some fashion, which is why they mentioned those specific things (Watson, large EDRam pool, 45nm SOI).
What makes Watson Watson? That's the question. What's the most important, defining feature? According to some IBM press material, the eDRAM alone is an incredibly important part. The other big features are the SMP interface (fastest in the industry, irrelevant for a console), the eight channel memory controller (overkill) and VSX. eDRAM is there, we know that much, even if it's not necessarily used as L3 cache. VSX might be there as well.

But the most important thing is that we have absolutely no idea who Engadget's source actually is. If it's a marketing guy... IBM's press release tells us exactly three things about the Wii U CPU: It's 45nm, it's SoI, and it uses eDRAM. That's it. And all of that is true for POWER7 - or Watson - as well. The statement was quite possibly hyperbole by some marketing guy. You know, similar to how AMD supposedly said they could do Avatar in realtime in two years... ;)
 
wsippel said:
What makes Watson Watson? That's the question. What's the most important, defining feature? According to some IBM press material, the eDRAM alone is an incredibly important part. The other big features are the SMP interface (fastest in the industry, irrelevant for a console), the eight channel memory controller (overkill) and VSX. eDRAM is there, we know that much, even if it's not necessarily used as L3 cache. VSX might be there as well.

But the most important thing is that we have absolutely no idea who Engadget's source actually is. If it's a marketing guy... IBM's press release tells us exactly three things about the Wii U CPU: It's 45nm, it's SoI, and it uses eDRAM. That's it. And all of that is true for POWER7 - or Watson - as well. The statement was quite possibly hyperbole by some marketing guy. You know, similar to how AMD supposedly said they could do Avatar in realtime in two years... ;)

Yeah, and with an APU at that.

I understand it could've been a marketing guy, but it would pain me to see a generally reputable site like Engadget succumb to sheer marketing. (Aside from anything Apple lol)
 
StevieP said:
Examples like yours are the exception, not the rule. Starting from the high end and going down, I mean. And for all intents and purposes, PC from 2006 -> forever will probably be the high end platform, thanks to the way the GPU has evolved.

You're missing my point. I don't care if the PC will forever be the most powerful platform from this point on. That is irrelevant to what I'm saying. I know what I'm saying is more the the exception and not the rule, especially if you want to include the PC. Doesn't have to stay that way though, at least within console development.

Basically PS4/720/PC to Wii-U would be preferred than Wii-U ported up to PS4/720/PC.

StevieP said:
Fixed. The PPE (in both machines) is a generally poor processor. Anandtech had originally compared it to a 1.6ghz P4 or something along those lines. While it was hyperbole (and got pulled, from what I recall) it wasn't like they were talking out of their ass either.

To be fair, both the Xenon and especially Cell were probably the best bang for buck MS and Sony could have gotten for systems that had their designs finalized in ~2004.
 
KageMaru said:
You're missing my point. I don't care if the PC will forever be the most powerful platform from this point on. That is irrelevant to what I'm saying. I know what I'm saying is more the the exception and not the rule, especially if you want to include the PC. Doesn't have to stay that way though, at least within console development.

Basically PS4/720/PC to Wii-U would be preferred than Wii-U ported up to PS4/720/PC.

That's usually not how the industry has worked in the past.

To be fair, both the Xenon and especially Cell were probably the best bang for buck MS and Sony could have gotten for systems that had their designs finalized in ~2004.

AMD would've likely been far higher bang-for-the-buck than a PPE. But they did build these machines when people still actually thought that GHZ matter (thanks Intel!). 3.2ghz PPE sounds a lot nicer than a 2ghz Power 5, despite one running circles around the other.
 
StevieP said:
Yeah, and with an APU at that.

I understand it could've been a marketing guy, but it would pain me to see a generally reputable site like Engadget succumb to sheer marketing. (Aside from anything Apple lol)
Wii U's brain is probably an APU as well - Avatar graphics confirmed...!

I just noticed that the SoC design has another huge benefit: If there's only one chip, you can get away with a much less complicated PCB - or use a more complex memory interface instead. IBM's quad channel DDR3 memory controller is quite impressive at 50GB/s. Should do the trick. Add 32MB eDRAM clocked at, say, 729MHz, and bandwidth definitely wouldn't be an issue. To summarize:

Quad core 476FP @ 1458MHz
R700 with either 400 or 640 shader units @ 729MHz
32MB eDRAM, shared between CPU and GPU ("MEM1") @ 729MHz
1GB quad channel DDR3 ("MEM2")

Overall, the system wouldn't look very impressive on paper, but it would be incredibly efficient and a joy to develop for. I'm probably still overshooting, though. If such a design even makes any sense at all...
 
wsippel said:
Wii U's brain is probably an APU as well - Avatar graphics confirmed...!

I just noticed that the SoC design has another huge benefit: If there's only one chip, you can get away with a much less complicated PCB - or use a more complex memory interface instead. IBM's quad channel DDR3 memory controller is quite impressive at 50GB/s. Should do the trick. Add 32MB eDRAM clocked at, say, 729MHz, and bandwidth definitely wouldn't be an issue. To summarize:

Quad core 476FP @ 1458MHz
R700 with either 400 or 640 shader units @ 729MHz
32MB eDRAM, shared between CPU and GPU ("MEM1") @ 729MHz
1GB quad channel DDR3 ("MEM2")

Overall, the system wouldn't look very impressive on paper, but it would be incredibly efficient and a joy to develop for. I'm probably still overshooting, though. If such a design even makes any sense at all...

Most 476FP implementations seem to have a dual-channel DDR3 controller, though.

And what about running the clock speed at 3x Broadway multiplier instead of 2x? Putting it at just under 2.2ghz. It could explain the heat issues some devs are having, because I highly doubt a GPU based on a customized 4770 would be causing them. The stock 476 is only meant to go to about 2ghz, and many implementations are running it at 1.6 or 1.8 just fine.
 
StevieP said:
Most 476FP implementations seem to have a dual-channel DDR3 controller, though.

And what about running the clock speed at 3x Broadway multiplier instead of 2x? Putting it at just under 2.2ghz. It could explain the heat issues some devs are having, because I highly doubt a GPU based on a customized 4770 would be causing them. The stock 476 is only meant to go to about 2ghz, and many implementations are running it at 1.6 or 1.8 just fine.
They could use another "Watson" part for the memory controller... ;)

Running the CPU at more than 2GHz might be possible. Power consumption and heat increases rapidly with higher clocks, though - six cores at 1458MHz might make more sense than four cores at >2GHz in that regard. Dunno. Four cores at 1.5GHz should be sufficient either way. It would only run stuff that absolutely has to run on a CPU after all. Less cores and a lower clockspeed leaves more headroom for shader unites.
 
wsippel said:
What makes Watson Watson? That's the question. What's the most important, defining feature? According to some IBM press material, the eDRAM alone is an incredibly important part. The other big features are the SMP interface (fastest in the industry, irrelevant for a console), the eight channel memory controller (overkill) and VSX. eDRAM is there, we know that much, even if it's not necessarily used as L3 cache. VSX might be there as well.
I guess I misunderstood you somehow, but how is the SMP fabric irrelevant for anything that has more than one core? Btw, the 476fp has some pretty spiffy SMP fabric as well.
 
blu said:
I guess I misunderstood you somehow, but how is the SMP fabric irrelevant for anything that has more than one cpu? Btw, the 476fp has some pretty spiffy SMP fabric as well.
The SMP fabric is certainly very important, but I was specifically talking about SMP with multiple sockets, not multicore.
 
Aside from the fact that they have to, have to, have to make 3DS the focus this year, they probably don't want to show the Wii U again without games they plan to release.

Aside from an otherwise unexplainable thirst to know more, I can't possibly imagine how not knowing the release date until next year is really all that bad a thing.
 
Mr_Brit said:
They could always just use better cooling. People overestimate the cooling in modern consoles and how much it could be improved for not a much more cost.
What kind of cooling would Nintendo use in case of WiiU current dimensions? This is not my field, so any information would be very welcome.
 
It is good to that Nintendo is staying far away from nickel and dime strategies with DLC at least. I've read it somewhere else and that is the only real positive news I've heard about them this week so far. It is nice to see that much reiterated in that article.
Alrus said:
Because you won't have a release date this year? Why do you care so much about that?
It is mostly because we have nothing to talk about and the article does nothing to change that.
 
wsippel said:
Wii U's brain is probably an APU as well - Avatar graphics confirmed...!

I just noticed that the SoC design has another huge benefit: If there's only one chip, you can get away with a much less complicated PCB - or use a more complex memory interface instead. IBM's quad channel DDR3 memory controller is quite impressive at 50GB/s. Should do the trick. Add 32MB eDRAM clocked at, say, 729MHz, and bandwidth definitely wouldn't be an issue. To summarize:

Quad core 476FP @ 1458MHz
R700 with either 400 or 640 shader units @ 729MHz
32MB eDRAM, shared between CPU and GPU ("MEM1") @ 729MHz
1GB quad channel DDR3 ("MEM2")

Overall, the system wouldn't look very impressive on paper, but it would be incredibly efficient and a joy to develop for. I'm probably still overshooting, though. If such a design even makes any sense at all...
There's a 0% chance of Nintendo being dumb enough to use DDR3 in the final unit. It'll be GDDR3 or GDDR5.

ShockingAlberto said:
Aside from the fact that they have to, have to, have to make 3DS the focus this year, they probably don't want to show the Wii U again without games they plan to release.

Aside from an otherwise unexplainable thirst to know more, I can't possibly imagine how not knowing the release date until next year is really all that bad a thing.
But if we don't get any info by the end of the year and we have to wait all the way until GDC or worse E3, all hype will die and nothing short of a mind-blowing launch line-up and price will prevent another 3DS. They need to say something this year.


Also, there's not exactly much that needs to be said for 3DS. The possible early 2012 titles seem fine; they just have to confirm which of those are coming. The rest can be saved for events next year. I don't think they're panicking as much as you guys say they are, and even if they are, what else can they do at this point?
 
Status
Not open for further replies.
Top Bottom