I could see Wii2 rivaling the best that the 360 and PS3 has to offer now, which techwise would be pretty cheap in 2011.
Next-Gen Wii HD Due by 2011
- Thread starter ziutek
- Start date
I could see Wii2 rivaling the best that the 360 and PS3 has to offer now, which techwise would be pretty cheap in 2011.
Well Nintendo has the money to do whatever the hell they want, but in truth they did before this gen got underway as well. These guys are just truly committed to making money, so I think that whatever Sony and MS do, Nintendo'll stick to offering very basic silicon in relation to the other guys, if only to maintain a very high margin on goods sold.
I mean it wasn't always like this for Nintendo of course (in terms of being content with weaker hardware), but I think they've found a rhythm that works well for them. That's not to say that I think it'll just be GameCube Mk III - that move made sense for their context at the time, but they can keep that ideology intact while still moving forward in technology this next gen.
I think only Nintendo knows that Wii like device won't work again in the future. But I don't think they'll go all out and take losses on their hardware either or price it much above $300 or whatever it is after inflation in 2011.
I think they might go with Apple route and go with Intel, since the politics had changed since Gamecube. If Larrabee actual performance is like what they simulated, I can see a single chip solution for Wii HD. 16-24 cores @ 1.5-2 GHz Larrabee with the southbridge integrated to it and paired with 64 bit GDDR5 or DDR3 controller with around 1-4 GB worth of memory.
archangelmorph
Veteran
We should at least be trying to think about Nintendo's strategy. A key part of disruption theory is that you start moving your new users upmarket, tracking with consumer expectations instead of overshooting them. I suspect that Nintendo is looking for significantly upgraded processors that will give a significant boost in graphics while remaining backward compatible. It should be in keeping with their focus on "fun" rather than "cinematic." I'm expecting power more on the order of midrange DX9 hardware from 2004, and possibly an incentive for developers to make cross-compatible titles that would feature (for example) upgraded shaders and textures on the Wii HD, much like Half-Life 2 would run on a huge range of video cards and take advantage of them to varying degrees.
According to Beyond3D forums.. Everyone's going with LRB in the next round of consoles..
That's the direction where everything is going. AMD and NV will probably have LRB like GPU too around the time Intel is out with LRB.
Anyway LRB is a good candidate for a single chip solution, which I'm sure what Nintendo is after, so they can spend the rest of the budget on a new gimmick.
archangelmorph
Veteran
That's because it's a GPU..
If they did decide to use it as an all-in-one chip then they could easily stick one/two GP IO core(s) on die & i'm pretty sure it would run rings around the Wii..
Uh, even the 1GHz LRB can run GoW...I don't see why it wouldn't be able to run a lowly GC/Wii game.
You're referring to LRB as GPU. fehu was referring to LRB as the *only* processor in the system (only compute-intense processor anyway). There's a big difference between having to run graphics-only code and having to balance general purpose code with graphics (and possibly other tasks).
Megadrive1988
Veteran
Whatever Nintendo packs into Wii HD or Wii 2 or whatever, I hope it's enough for Nintendo's key franchises (Mario, Zelda Metroid, Pokemon, etc) to not only be in HD resolutions (including 1080p) but for the actual graphics themselves to be much closer looking to offline CGI. Now I don't mean photorealistic at all, I mean like still Nintendo-esque, closer to Pixar level
(notice I didn't say as good as Pixar, I mean closer to it) with much higher model/background complexity, per pixel (or even sub pixel) lighting, much better image/rendering quality w/ more AA, filtering, motion-blur, post processing, etc.
I don't want just 'Wii graphics in HD'. I won't even be happy with 360 / PS3 level graphics since those consoles were both a disappointment to me as far as graphics -- It seems advancement in graphics & framerate were sacrified for HD resolutions with 360/PS3.
Now I don't expect Nintendo to have a cutting-edge main CPU processor like current-gen CELL or next-gen 32-SPE CELL, or Intel's upcoming Larrabee, but I do expect a major, major leap in graphics beyond GCN/Wii, nevermind the resolution which is a seperate thing.
I'd be happy if the pre-rendered Space World 2000 demos (Metroid, Wave Race, and most of Rebirth) could be done in realtime @ 60fps @ 1080p in next-gen Nintendo games. That would be noticable step beyond 360/PS3 graphics, although would not need hardware as advanced as the high-end technology that will exist in 2010/11/12 for PCs, that *could* potentially be used in XB3/ PS4.
p.s. I think that some version of AMD's Fusion is much more likely to be in Nintendo's next console than Larrabee. I am not saying that Nintendo will use AMD Fusion, but it's not as unlikely as Larrabee.
(notice I didn't say as good as Pixar, I mean closer to it) with much higher model/background complexity, per pixel (or even sub pixel) lighting, much better image/rendering quality w/ more AA, filtering, motion-blur, post processing, etc.
I don't want just 'Wii graphics in HD'. I won't even be happy with 360 / PS3 level graphics since those consoles were both a disappointment to me as far as graphics -- It seems advancement in graphics & framerate were sacrified for HD resolutions with 360/PS3.
Now I don't expect Nintendo to have a cutting-edge main CPU processor like current-gen CELL or next-gen 32-SPE CELL, or Intel's upcoming Larrabee, but I do expect a major, major leap in graphics beyond GCN/Wii, nevermind the resolution which is a seperate thing.
I'd be happy if the pre-rendered Space World 2000 demos (Metroid, Wave Race, and most of Rebirth) could be done in realtime @ 60fps @ 1080p in next-gen Nintendo games. That would be noticable step beyond 360/PS3 graphics, although would not need hardware as advanced as the high-end technology that will exist in 2010/11/12 for PCs, that *could* potentially be used in XB3/ PS4.
p.s. I think that some version of AMD's Fusion is much more likely to be in Nintendo's next console than Larrabee. I am not saying that Nintendo will use AMD Fusion, but it's not as unlikely as Larrabee.
Last edited by a moderator:
And? You don't think one or two of the cores out of the 20+ cores in that 1GHz LRB can handle what the single 700MHz PPC in Wii does?
Last edited by a moderator:
And? You don't think one or two of the cores out of the 20+ cores in that 1GHz LRB can't handle what the single 700MHz PPC in Wii does?
Except we're not talking about a Wii, we're talking about a next generation console capable of much more than the Wii. Also, I don't see the Wii running GoW so your "what Wii already does" statement is clearly not even remotely based in reality.
Also, Larrabee's cores are not well-suited to single-thread IPC outside of vector code. It would make a HORRIBLE central processor for anything compute-intensive (again, outside of vector code).
Actually we were talking about Wii BC and what is needed as minimum. A single LRB with 20+ cores at 1GHz is MORE than enough for BC. As for WiiHD...uh add more cores if you WANT to duh.
Actually according to people with credibility like NOA, a console with only LRB cores sounds like what developers may want. If you disagree then first establish some credibility first, then we can talk.
Ok, even in that context it's still easy to see why LRB is not well-suited to this task.
1) it's based on the *ancient* P54 Pentium core - IPC is laughable compared to a modern OoO CPU.
2) you fail to account for the fact that LRB's ISA is different from PPC, hence any software written for the Wii would need to be either ported to Wii2 or emulated - both result in a performance loss compared to running the application on a CPU with the native ISA.
nAo is a programmer, he wants flexibility above all else because he believes himself capable of overcoming any limitation, given the proper tools. Understand people's motives before using them as an "end-all be-all" reference in a debate.
1. It doesn't matter all that much how man IPC it can process vs an OoO CPU when you have two 1GHz cores at your disposal vs one 700MHz core. Also you're not restricted to only two cores if you need more processing power.
2. See point 1.
I know he's a programmer that's why he's more credible when it comes to what architectures has promise from a development/performance perspective. Not only that, he's actually worked on difficult to program achitectures like PS3.
His motives and yours are irrelevent to me. I'm not a mind reader so motive doesn't even factor into the equation. Any statement made about one's motives is speculation, conjecture and has little real value in this specific debate.
That said what do YOU propose WiiHD should be since God himself told you LRB would suck if used as a base for WiiHD?
The number of cores is irrelevant, especially in this context. It's not like you can just take an existing serial workload and split it up into two equal chunks and get twice the performance.
It's common sense that programmers main desire from any given hardware set is the ability to use it as they see fit (i.e. flexibility/programmability). No speculation or mind reading is necessary.
That said what do YOU propose WiiHD should be since God himself told you LRB would suck if used as a base for WiiHD?
I've said no such thing. All I've said is that LRB makes for a poor central processor compared to its contemporaries. This is a commonly acknowledged precept (namely due to it's use of "meager" in-order P54-derived cores).
LRB may well be the next revolution in executing graphics code, by no means is it that for general purpose code.
Understand this before you respond again:
LRB is:
a great programmable processor of vector code
LRB is not:
a great processor of serial code
