Predict: The Next Generation Console Tech

[bgassassin]

It's "real" FLOPs. It means that in 1 second, at 1920x1080p, the demo needs 2.5 trillion of operations on average. So much more than what a single GTX580 and a 6970 can provide, and more like what Tahiti and Kepler can do. (Giving that Tahiti seems to have an higher efficiency, and Kepler will have a much larger throughput)
I don't think the demo is a good example of what that number of pixel operation can do. If i remember they used an incredibile heavy boken filter, without any much advantage on image quality. The Samaritian demo is just a technology demo for DirectX11 and Unreal Engine 3, and it was put together fairly quickly by few persons, not Unreal Engine 4

I understand that, which I thought/hoped my posts should have shown.

Also considering the 580 was was not really blown away by the 7970 stock) in some benchmarks, and held its own in other tests, I'd be left to believe there was no improvement in efficiency with Tahiti continuing the trend of AMD/ATi GPUs still not doing well with utilization.

 

[TheChefO]

I understand that, which I thought/hoped my posts should have shown.

Also considering the 580 was was not really blown away by the 7970 stock) in some benchmarks, and held its own in other tests, I'd be left to believe there was no improvement in efficiency with Tahiti continuing the trend of AMD/ATi GPUs still not doing well with utilization.

Much of the die budget in Tahiti went into making the core more processing flexible (GPGPU) and modularly scalable.

Prior to GCN, AMD was doing much better in utilization/efficiency than Nvidia in performance per mm2. But now that they are both GPGPU centric designs, that AMD advantage is gone. Should be interesting to see how Keplar stands up to GCN.

 

[Shifty Geezer]

So you're predicting next gen tech to be powered by cloned savant brains?

This is very OT now.

woah there. I didn't say nextgen,

It was a joke, hence the smilie. Anyway, brain discussion moved here.

 

[GBPack]

The way you bring kinect2 with the new platform is by adding it onboard (piggyback technique ... see ps3 w BRD). The initial push of console gamers are hardcore. They will be the new console at a higher price (assuming the specs are worth it). This enables MS to amortize the cost of the initial run with a higher price, and present a nice base of kinect2 enabled consoles to developers.

After a year of this, introduce a kinectless xb720 for a lower price.

All the while, kinect+xbox360 will still be on the shelf.

All the investment MS is making in kinect will pay off in the future, and they can monetize the advantage today and in the near future with xb720 sitting on the shelf next to xb360's.

But as I said, sacrificing the advantage that brought them into the livingroom in the first place is a good way to get shown the door. It isn't so much about winning the console space as it is about securing an influential push into the livingroom. A big key to that for MS all the way up to the point they introduced Kinect, is the hardcore gamer.

They demand high end specs, not middle of the road.

Wiiing out in this manor especially to sacrifice for casuals will not be looked upon with understanding eyes from the hardcore market.

They will be leaving themselves open for a huge competitive advantage for Sony (or another).


As to the notion of xb360 limiting Kinect, that's true. Just as xb360 is limiting graphics, physics, interaction, etc. That's what happens with hardware from 2005.

Having said that, if you want cool, quiet, small ... then older spec makes sense. It's perfect for that purpose, and for casuals that don't care much about graphics, physics, interaction, they just want the interface ... well xb360+kinect is perfect for them. For the ones that want more, xb720.

There's no rule that says a new box can't sit on the shelf next to the old box. Nintendo proved that point rather well with Wii next to xb360/ps3. And for the staggered Multi-SKU approach, in varying degrees, Sony and MS proved this can work too.
_____________
As for editing posts, I think it takes a few days or posts for that to pop up for whatever reason.

.... Hardcore gamers demand high end specs? Where did you get that? That's not true at all. And releasing a system without Kinect is counter-productive towards their goal.

I also don't think you understand what I'm talking about about having two consoles on the shelf. Wii vs 360/PS3 is not an example at all. Wii is a current generation console, the same as the 360 and PS3. It replaced the Gamecube. You don't see Gamecubes selling in stores anymore, do you? Of course not, so why would you have two Xboxs. Selling 360 with Kinect and 720 with Kinect 2.0 at the same time doesn't make any sense. It's either one or the other, and I highly doubt they will drop Kinect for next-gen.

Xbox is far more than a gaming brand, and that is being reflected in their direction recently. Assuming they're going to continue designing their hardware around the needs of your definition of a "hardcore" game doesn't make sense either. Middle of the road will be perfectly fine, especially when that becomes the standard for next-gen. If people really cared about graphics so much, they could just buy a PC.

Thanks for answering the question about editing posts. It's a shame you have to wait to edit, though.

 

[Rangers]

Microsoft knows they dont want to be like the Wii, with a rapidly fading and short life.

They're not dumb, they know their foundation is the hardcore graphics heavy gamer. They can build from there but the foundation must be in place otherwise you end up like the Wii currently. So I am sure the graphics capabilities will be solid.

Kinect was all about building on the solid 360 hardcore base, and there's no way they foolishly abandon that chasing a casual market that may or may not even exist post-Wii. The problem with Wii is it had no base of hardcore gamers, and that's why Nintendo is back to trying to gain hardcore gamers back with Wii U (albeit, a all wrong way to go about it imo).

It's the PS4 I'd more worry about having a power deficiency because of Sony's financial troubles and recent credit downgrading. It's looking to me like they simply cant lose a lot of money on PS4, and that will likely mean compromises. But then again those gamers that care about graphics can just "buy a PC" so obviously it wont be a problem

 

[GBPack]

Microsoft knows they dont want to be like the Wii, with a rapidly fading and short life.

They're not dumb, they know their foundation is the hardcore graphics heavy gamer. They can build from there but the foundation must be in place otherwise you end up like the Wii currently. So I am sure the graphics capabilities will be solid.

Kinect was all about building on the solid 360 hardcore base, and there's no way they foolishly abandon that chasing a casual market that may or may not even exist post-Wii. The problem with Wii is it had no base of hardcore gamers, and that's why Nintendo is back to trying to gain hardcore gamers back with Wii U (albeit, a all wrong way to go about it imo).

It's the PS4 I'd more worry about having a power deficiency because of Sony's financial troubles and recent credit downgrading. It's looking to me like they simply cant lose a lot of money on PS4, and that will likely mean compromises. But then again those gamers that care about graphics can just "buy a PC" so obviously it wont be a problem

They're also not going alienate any base of theirs. I don't understand why you keep thinking I'm saying Microsoft will go the Wii route. The Wii was a completely different situation. All I'm saying is that I think it won't be incredibly powerful. I was seeing speculation about Tahiti and 2k+ ALUs, and other insane predictions like that. I'm saying it will be a moderate increase, with Sony following as well.

 

[Rangers]

Well, a "modest" specced console (lets say, something from a 6870 to a 6970) by current PC GPU standards would still be a huge increase on 360, given Moore's law has been applying for nearing 7 years now. The lengthened between gen time frame means even a "modest" "leap" will be relatively huge.

I still think something near Tahiti's level might be possible, on the fact we're still looking at nearly two years until release best case imo (late 2013) and possibly even as late as late 2014.

Tahiti may look awesome now, but we'll be looking at 1-2 more GPU generations before than, depending. It's not going to look so awesome in two years. Any more than a 5870 or GTX 480 does today.

 

[ultragpu]

It's the PS4 I'd more worry about having a power deficiency because of Sony's financial troubles and recent credit downgrading. It's looking to me like they simply cant lose a lot of money on PS4, and that will likely mean compromises. But then again those gamers that care about graphics can just "buy a PC" so obviously it wont be a problem

I doubt going cheap on the PS4 could save Sony for the long run though, yeah they may not loose as much but the benefit would only be short lived. Since a new console launch is considered the next biggest thing for the company, you really don't wanna rest the fate of their ten years plan on a miserable 6670 based gpu. Now considering Sony wouldn't go a fully casual route if their massive core game fan base means anything, they should invest big in the hardware as they would eventually recuperate the initial loss years down the road.
My point being launching an underpowered system would only hurt them more in the long run, just look at ps3 now, they're the only division that's bringing Sony profit. You know what they say, nothing is gained if nothing is ventured.

 

[AlphaWolf]

The only reason you'd need to include something as low as a 6670 in console hardware launching in 2013 or beyond is because you're targeting a $200 or lower. I think they could afford more than $25 (or whatever that chip would cost in 2013) for the GPU. (you can get a DDR3 6670 card for $50 with rebate right now).

<edit>And I think the 7750 results should put the 6670 rumor to bed for good. ~30-50% more performance while using quite a bit less power.

 

[french toast]

They're also not going alienate any base of theirs. I don't understand why you keep thinking I'm saying Microsoft will go the Wii route. The Wii was a completely different situation. All I'm saying is that I think it won't be incredibly powerful. I was seeing speculation about Tahiti and 2k+ ALUs, and other insane predictions like that. I'm saying it will be a moderate increase, with Sony following as well.

I really do think we are going to see a 2000 ALU x720 gpu, ive said it all along, its perfectly doable, on a mature 28nm process, VLIW 5, 64 TMU's 32 ROP's...using the 6870 cut down shaders as a guide, with some more advanced units from GCN, a 4 core 4xSMT PPC cpu @3.2 ghz. 256bit bus, 4-6 gb gddr5. 500mm2 die budget 250w tdp. £299 RRP. Holiday 2013.

Kinect 2, wireless hdmi, bluetooth 4+, NFC, wifi N, small flash storage on every device (16gb?), proprietry flash storage for games, with DRM (second hand games) Some advanced DVR functionality + custom w8 operating system.
A funky new 720 pad with a touch screen...expanding on what the Dreamcast pioneered all those years ago.

Microsoft to make a loss on every console, but still make a profit form the whole entertainment & games division from launch.:smile:

 

[Rangers]

I doubt going cheap on the PS4 could save Sony for the long run though, yeah they may not loose as much but the benefit would only be short lived. Since a new console launch is considered the next biggest thing for the company, you really don't wanna rest the fate of their ten years plan on a miserable 6670 based gpu. Now considering Sony wouldn't go a fully casual route if their massive core game fan base means anything, they should invest big in the hardware as they would eventually recuperate the initial loss years down the road.
My point being launching an underpowered system would only hurt them more in the long run, just look at ps3 now, they're the only division that's bringing Sony profit. You know what they say, nothing is gained if nothing is ventured.

All this applies to MS as well, but people dont seem to see it that way. I fully agree though.

I do think both companies will lose less money on hardware next time, but you can still easily do that while launching a powerful system. There's no need to rush headlong the other way.

Yeah, and if MS really is going with a 6670, miserable is the word. People dont seem to understand that, what looks weak today, is going to look, unimaginably pathetic in 5 years when next gen is still supposed to be going strong. I mean, you literally wont be able to find an integrated GPU anywhere as weak as the 6670 in a few years. The iPad 7 will be 10X as powerful. Think what a joke the 720 will be. People have to be smoking some ganja to think that's gonna happen, yet they're portraying themselves as the logical ones.

That's why I'm pretty sure that report was nonsense.

 

[Shifty Geezer]

Yeah, and if MS really is going with a 6670, miserable is the word. People dont seem to understand that, what looks weak today....

Only it won't look weak next to the existing consoles, which is what most console gamers have as a reference point. A 6670 XB3 will look a generational advance on the consoles, gamers will buy it as such, and it doesn't need to be the flagship box 5 years later. The current consoles look unimaginably pathetic next to what modern tech makes possible these days, but they still sell.

...yet they're portraying themselves as the logical ones.

There's nothing illogical about it. It's a perfectly valid strategy. It might not be the one used, but don't go calling those with a different observation to you as illogical. There are multiple logical arguments that can be made for different strategies.

 

[MarkoIt]

All this applies to MS as well, but people dont seem to see it that way. I fully agree though.

I do think both companies will lose less money on hardware next time, but you can still easily do that while launching a powerful system. There's no need to rush headlong the other way.

Yeah, and if MS really is going with a 6670, miserable is the word. People dont seem to understand that, what looks weak today, is going to look, unimaginably pathetic in 5 years when next gen is still supposed to be going strong. I mean, you literally wont be able to find an integrated GPU anywhere as weak as the 6670 in a few years. The iPad 7 will be 10X as powerful. Think what a joke the 720 will be. People have to be smoking some ganja to think that's gonna happen, yet they're portraying themselves as the logical ones.

That's why I'm pretty sure that report was nonsense.

What about Microsoft switching to the "Ipad/Iphone/Ipad" model? A cheap updated Xbox every year, and developers are free to decide the models they want to support.. selling their games with the Windows Store (20% to MS) or for AAA-titles on Blu-ray (royalties).
I don't see the 7-8 years generation as viable anymore.

 

[Shifty Geezer]

There are plenty of options. This isn't really the thread to discuss them. This is the thread to decide what hardware selection is valid for a given model, and to gradually hone in on the real specs as rumours become facts.

6670 is an option that is workable within the right business model.

 

[Rangers]

There's nothing illogical about it. It's a perfectly valid strategy. It might not be the one used, but don't go calling those with a different observation to you as illogical. There are multiple logical arguments that can be made for different strategies.

I find it pretty illogical. Again, when you look at what a 6670 is going to look like in 3 years, 5 years, 8 years...for a console not even coming out for at least a year, probably two. I also find it pretty illogical again if you look at MS strategy with the Xbox 360, which they added on to with the Kinect, not replaced. To suggest they're now going to willy nilly go casual only (when the Wii is dying quickly) is silly. It frustrates me so many seem to actively push for this path or act as if it's inevitable when there's just no evidence. I dont mind if a small tiny minority, a stray post here and there, held this view, that would seem reasonable, but it actually seems to be the majority in this thread.

Consoles are not the only thing people are going to judge other consoles against. If games start looking better on the iPad, or countless other devices, they will notice (they already do notice console approaching graphics with stuff like Infinity Blade, but at least in IB's case the embarrassing "it actually looks better" scenario isn't actually realized).

A 6670 XB3 will look a generational advance on the consoles, gamers will buy it as such, and it doesn't need to be the flagship box 5 years later. The current consoles look unimaginably pathetic next to what modern tech makes possible these days, but they still sell.

I'm not sure a 6670 is a generational advance at all. The best I can figure is it may be around 3X Xenos (born out by shader calcs and gigaflops). I really question if 3X is enough to bother at all, you need a significant leap to get people to upgrade, multiple devs have stated this. I believe people in this very thread have stated 2X isn't enough to see a major difference in onscreen graphics. That ignores how it would look next to the presumable next gen competition anyway, which is the real issue. Would you be happy if you found out for a known fact tomorrow that PS4 would use a 6670? I think you'd find it appalling.

Current console GPU's are pathetic, but again what people always ignore is they were fairly state of the art at the time. Imagine how they'd look today if they were whatever the 2005 equivalent of the 6670 was, that's the real issue.

 

[Shifty Geezer]

Would you be happy if you found out for a known fact tomorrow that PS4 would use a 6670? I think you'd find it appalling.

I would indeed, same as I found Wii to be appalling. But logic isn't about finding an explanation that fits one's preferences. As we have no context for what the next box is giong to be or what it'll cost or what its intended lifespan will be, then we cannot conclude that selection of a low-performance part is illogical. If the next box is confirmed to be in the same traditions of the existing console industry than it would be illogical to use a low-performance part. As it is, without too many unknown variables, we are faced with contemplating a branch in the prediction of the hardware: 1) Either the rumours are wrong and the console will fit our expectations, or 2) the rumours are right and we have to change our understanding of the approach to the business that the next Xbox will take. You are considering path one exclusively because it fits your preferences. There's perfectly allowed, but the alternative isn't illogical; just different in this case.

 

[Rangers]

If you were going the low power console route, and please dont think I'm endorsing this view as I recoil at it , but still being fun to kick things around, the new Cape Verde chips are out


The 7750 certainly seems like it would fit in a console with absolute ease, for those who harp about such things. 75 watt typical power. 7770 is 100 watts and runs at 1ghz. I would think in a console that clock might be slightly problematic, maybe downclock it to 900? The 1.28 teraflops would handle Epic's Samaritan at 720P requirments, and would be about 5X over Xenos. still way way too low imo, but anyway.

Looking at some tech power up benches, the 7750 actually compares well to the 5850 in games like Crysis 2 and Metro 2033, despite a seeming total lack of grunt. The 5850 is a 1440 SP card! Tech Power Up's benches seem wonkey though, at Anand the 7750 consistently is mostly a little below, occasionally above, the 800 SP 5770, still not bad.

Only thing I'd question is if those efficiency gains are real, or only a product of sloppy PC code, and in a console a 5850 would show a huge advantage over the 7750. Anyway, I think this chip would be very nice for Wii U

 

[fehu]

after reading anand review i'm very confused about the utilization of gcn architecture in a nextgen console
considering amd as the only option, we must expect a revamped vliw4 or a gcn?

 

[Blazkowicz]

radeon 7750/7770 shows what a good enough GPU on 128bits bus is like. it's nothing too low for a console I believe, if you double the specs you only get a console incrementally better but noisy and expensive as the original PS3 is.

such a console would be good enough to get rid of low res textures and shimmering/bad texture filtering, while better CPU and more memory gives us better framerates and complex game environments.
I would like it enough. I didn't have any interest in current gen, but now got bored about PC gaming (I got rid of windows that's why). I might get interested in a newer console, even though I don't like DRM'd or walled garden computers that much, but I don't want it to have a big footprint. consoles used not to have a fan afterall.

 

[GBPack]

When I talk about moderately powered for nextgen, I don't think of 6870-6970. I was thinking of somewhere along either that 6770 or even 7750/7770 now. Those recently released numbers for the 7750/7770 look nice, and fall close (maybe a little under) to what I was thinking for nextgen.

I really do think we are going to see a 2000 ALU x720 gpu, ive said it all along, its perfectly doable, on a mature 28nm process, VLIW 5, 64 TMU's 32 ROP's...using the 6870 cut down shaders as a guide, with some more advanced units from GCN, a 4 core 4xSMT PPC cpu @3.2 ghz. 256bit bus, 4-6 gb gddr5. 500mm2 die budget 250w tdp. £299 RRP. Holiday 2013.

Kinect 2, wireless hdmi, bluetooth 4+, NFC, wifi N, small flash storage on every device (16gb?), proprietry flash storage for games, with DRM (second hand games) Some advanced DVR functionality + custom w8 operating system.
A funky new 720 pad with a touch screen...expanding on what the Dreamcast pioneered all those years ago.

Microsoft to make a loss on every console, but still make a profit form the whole entertainment & games division from launch.:smile:

I'm not saying it's not doable. What I'm saying is that my view on Microsoft's console business plan is less is more. I just have a hard time seeing Microsoft go big and power-hungry again.