Predict: The Next Generation Console Tech

[AlNom]

Overall I'm with Erick, I believe the most powerful system (if there is a clear advantage, bigger than PS3/360 or PS1/N64 for example) will generally triumph in the hardcore space

I'm not sure what the point of your arguments are then because this is a thread about predicting what the console companies will make, not what the hardcore want.

Wii though it still lost to the HD twins as a whole

While true, it's rather dubious to combine the two because each company is looking out for their own individual shares of the market. Or you may as well combine the losses sustained from both companies and put it up against the profits of nintendo alone.

 

[AlphaWolf]

I'm not sure what the point of your arguments are then because this is a thread about predicting what the console companies will make, not what the hardcore want.

I certainly believe maintaining the hardcore base is important.

While true, it's rather dubious to combine the two because each company is looking out for their own individual shares of the market. Or you may as well combine the losses sustained from both companies and put it up against the profits of nintendo alone.

And how many units do you think that MS (or Sony) would have sold if they had Wii level graphics and waggle while lacking the Nintendo first party library vs a Sony (or MS) with a much more powerful console? MS and Sony are certainly competing much more with each other than with Nintendo.

And as for losses/profits. I wouldn't be surprised if MS is in the black this gen. If you have some actual (console only) numbers I'd love to see them. People keep saying how much money MS lost, but their E&D division has been profitable for quite a while and was surrounded by money sinks (zune, kin etc) in that division.

 

[AlNom]

I certainly believe maintaining the hardcore base is important.

Yes, but I don't believe in astronomical wishlists.

And how many units do you think that MS (or Sony) would have sold if they had Wii level graphics

The whole point is trying to come up with a reasonable machine. I'm not suggesting they should stick to the stone ages for heaven's sake.

I wouldn't be surprised if MS is in the black this gen

Perhaps, but it's a heavy risk if you're going to come out with a heavy loss-leader and you're telling investors just to wait years before the machine/materiel breaks even. Do you think MS/Sony want to do a repeat of this gen in the face of what they've seen with Nintendo?

Again, I'm not saying they should copy them 1:1 with some unit that's two 5-year-old consoles duct-taped together. That's just short-sighted in the face of the rest of the market they're targeting. I'm just struggling to make any sense of the craziness that certain specs you folks are throwing around that have big implications beyond just "zomg double dimension textures", or "people care that much about resolution and texture fidelity"

Spoiler

(which is TOTALLY why they go for consoles in the first place, right)

.

 

[AlphaWolf]

Yes, but I don't believe in astronomical wishlists.

The whole point is trying to come up with a reasonable machine. I'm not suggesting they should stick to the stone ages for heaven's sake.

There's certainly a lot of room for debate about what's reasonable. Was the 360 reasonable in 2005? It was certainly more reasonable than the PS3 in 2006, (but that's more about other choices that I don't really want to debate here).

Perhaps, but it's a heavy risk if you're going to come out with a heavy loss-leader and you're telling investors just to wait years before the machine/materiel breaks even. Do you think MS/Sony want to do a repeat of this gen in the face of what they've seen with Nintendo?

Again, I'm not saying they should copy them 1:1 with some unit that's two 5-year-old consoles duct-taped together. That's just short-sighted in the face of the rest of the market they're targeting.

It's a lot easier sell to the shareholders on a loss leader when they can show massive growth over the last 3 fiscal years (all of which have been profitable, $100M, $600M, $1.3B for E&D). If MS can avoid a major misstep (ie RROD) a hardware loss leader shouldn't be too hard to sell, especially as they push to grow in non gaming markets. The lengthened cycle may have given them some room if not for the fact that they now have to compete against Onlive (and other cloud gaming), tablets, smartphones and $350 PCs that can play games reasonably. So there is a still a need for them to present some significant performance, or build their niche elsewhere like Wii:U.

As much as the Wii was a recipe for success for Nintendo, I actually believe it could have been a recipe for disaster for Sony or MS. Neither had a Super Mario to pull that casual crowd and a weaker/casual box could have left the other 2 dividing their share.

 

[erick]

OK guys, since I cannot address each of your points on a personal basis because it would be time consuming, let's just get some things out of the way.






1. 8GB vs 4GB of RAM does not make sense due to multiplatform environment/dev costs.

Have you heard of "detail levels" as in low, medium, high detail? That's right, something that nearly ALL devs use RIGHT NOW for any game that is ported to the PC!

The fact that most of the developers do this as we speak and manage to somehow take advantage of the fact that PCs have 4x to 8x as much RAM + VRAM without seemingly incurring significant extra costs should tell you something basic about game developement:

Developing a game always starts at the highest fidelity, which is then optimized or watered down to fit on particular platforms. Even if you do an exlusive, you'll want your source material as high fidelity as possible in the hopes that not all of it will be stripped away during optimization process and you will still have a looker.

So, in fact, nearly all AAA devs have those high fidelity textures and materials in the first place. They are f*cking happy if they can use them on the final product, but the sad reality is that significant sacrifices have to be made for the final product.

More RAM equals less time spent on painstaking optimization equals less money burned equals a better looking game equals happier devs who choose your platform to be the lead platform and the rest get a stripped-down port.

And since one of the 3 lead platforms already does not have a RAM issue (PC), then it would make sense anyway because you can make TWO good-looking versions with minimal porting costs and then only one "optimized" version.






2. Stop with the Llano references, please. It simply does not apply. Llano is an example of the lowest end hardware applied on the lowest common denominator (4GB 1333MHz DDR3, can be bought at less that €20). Naturally there will be a bandwidth problem if the GPU is twice as powerful (regarding shader power) as that on the X360 while having access to *less* bandwidth. However, bandwidth will not be a problem for the next gen boxes.

Remember, we are speaking about 8GB GDDR5 here? The fastest available GDDR5 right now was manufactured already in 2008 - it runs at 7GHz and one 32bit chip can put out 28GB/s.

On a 128bit bus that would be 112GB/s, on a 256bit bus it would be 224GB/s. Remember, this is 3 years old!

There is no reason to assume that by 2014 there won't be significant advances which double that bandwidth while providing 4-8x more density ... 8 of those chips made 1GB in 2008, and 8GB by 2014 is reasonable.

As for the GPU power to harness all that RAM? This winter we will have 4GB single-gpu cards with texture fill rates reaching 100 000+ Mtex/s (compare this to PS3-s 13 200 Mtex/s or Llano's 8 900 Mtex/s). By 2014, this number will be doubled, and then nearly doubled again. Think about that. By 2014 8GB VRAM will be as mainstream in the PC world as 1GB VRAM is now, and the GPU power to use that will be there.

 

[Gubbi]

You are talking about MicroSoft; they wanted to gimp 20cent of DVD lens shielding on advertising, knowingly having this damage the discs due to slight vibrations @ 20 speed dvd rotation. And then deny repairs for months until the EU forced them to.
You can expect them to want to save every cent they can, let alone 15 dollar.

That is one of two examples of the MS beancounters getting it horribly wrong. The other one was being cheap on the cooling solution resulting in RROD.

Both attempts to reduce cost ended up costing a lot. And it is not just the direct cost of the replacement 360s. Loss of mind share ( remember all the RROD jokes? ), and not being able to pursue an aggressive pricing strategy earlier. The 360 would have killed PS3 outright if not for RROD.

Lessons were learned after the first XBOX:
1. Own the IP for the ICs so that they can shop around for the cheapest source.
2. Don't include devices that won't cost-reduce over time, that aren't absolutely necessary (optical drive necessary, HDD not)

It would be naive to think that they wouldn't learn from their mishaps this gen.

Cheers

 

[fehu]

we are still waiting the 4Gbit density, and in 2 years you see an 8x increase

 

[Shifty Geezer]

Actually, I've explained it on numerous occasions.

The factors you have brought out are simply shortsighted. If $15 buys you 8GB instead of 4GB, you simply cannot argue this advatage away by saying that that money would be better spent elsewhere - if we're talking bang for buck it is THE BEST investement seeing as $15 won't exactly buy you a faster CPU or GPU which are both custom made anyway and cost many, many times more initially.

You are arguing solely in terms of hardware resources per dollar. Yes, if a company is trying to get as much hardware for their BOM as possible, doubling the RAM can give great long term benefits. However, the bit you seem unable to follow as that maximising console sales and profits is not all about hardware, and you have just brushed the alternative options aside without giving them any thought. How can securing key exclusives or better services be considered short term?! A strong start escalates word-of-mouth and mindshare and sets up the rest of the generation - it's the very opposite of a short-term option!

How do you know that faced with two consoles, XB3 and PS4, one with 8GBs and one with 4GBs, that Joe Gamer is going to pick the 8Gb console when there are so many other factors in play? Looking back through the history of consoles and pre-PC computers, it's not always the highest performing machine that wins out. All to often the lower spec machine ends up the more popular choice because of non-performance reasons?

As long as you cannot see the world with eyes other than for hardware costs, you cannot make the right decisions for a whole business strategy.

 

[fehu]

when it come to marketing all can happen
i have no difficulties to imagine some pr that pulls out another cell mantra repeating that double the memory will give you a quantum leap difference even if the remaining hardware is inferior, and microsoft or sony finance a game with the only scope to exploit the bigger memory hiding the other deficiencies

btw GB and MHz are the only thing that average users can understand and use to compare different products
you can see it in current generation with the mass memory used to differentiate and compete

 

[Heinrich4]

I remember this same discussion in 2002/2003 about next xbox and ps3 how kind memory, which would amount to 256MB (4GB now) or 512 (8GB), which will reach the XDR bandwidth 25.6GB/sec (50/70GB today / sec) or 51.2/sec (100/150GB/sec) and so these would be enough etc..

Despite the Wii business model have worked well until 2008/2009 (not today) and the growth of smart phones and tablets have emerged as competitors or withdraw the gamer market and today there is much greater pressure from shareholders to lose billions until reach break even costs, I resist to believe that sony and ms will not be able to offer in 2013/2014 hardware with at least the graphics capabilities of today's high end pcs*.



*based in hardware 2009 and 2010 see Radeon HD 5870(2009) still offer one of the best option 3D cards.


http://techreport.com/articles.x/20126/7

 

[erick]

Shifty, maybe because we're discussing the next gen console tech, not marketing or software?

In this instance, I view tech separately from marketing costs. However, I'm pretty sure that guys at Epic would die of pleasure at the thought of producing Gears of War 4 on NextBox using 8GB of GDDR5, while the best that the competition can use for Killzone 4 is 4GB (that's assuming that both boxes will have roughly comparable amounts of processing power like last gen).

 

[liolio]

I would not tone down a difference in RAM amount between system competing for core gamers.
It only a part of the picture, but a significant one.
The 360 had a head start, the Rrod hurt it badly, if it would have come with more RAM it would mostly have been a game over for the PS3. The ps3 was vouch from scratch by the gaming press, internet, etc has a significantly stronger system, on top of impressive Sony games catalog it helped a lot to settle quiet some gamers minds. Now if you put economic considerations aside (which I aknowledge is impossible) and Ms added even more RAM to the system, clearly it would have a significant impact, there is no dismissing it, even 768 would have made a significant difference, imho enough of a difference for MS to stealth from Sony a lot more of the core gamer segment.

OK, there are economic realities but I would not underestimate the effect on sales on extra RAM. Clearly RAM doesn't sell games or systems but clearly with multi platform games taking the crown in gaming sales it would have been a great competitive advantage.
I'm confident that neither Sony or Ms want to be in a situation where they ship a system with a significant difference in the amount of ram available to developers as most likely power consumption concerns will ensure that the difference in processing power won't be much.
RAM would have a really significant effect on visual even to the average guy looking at the system side by side at say best buy. The "word on the street" would help too, gamers would clearly notice and say about it, no need for DF, lens of truth and all that kind of site, they exist because the systems were a wash imho. If they were not people could easily tell and so it would have sale implications.

 

[Arwin]

I do really look forward to the next generation though. It has been coming a long time and although somewhat mitigated by the previous gen launching with too expensive hardware, combined with next gen perhaps wanting to be cheap even earlier. But 8-9 years is still more than long enough to see big advances in hardware, I'd say! Hoping for XDR2 or something similar, as I do feel it is a good solution for RAM that needs to be accessed in paralel by multiple threads.

Crucially, all those advances can go to more information per pixel, as the resolution that most games will run on isn't going to chance significantly. Here too, there is still a chance that we'll lose some if games target 1920x1080x60fps in general, but I think considering the distance in years, that should be possible while still having more information per pixel.

I think this would make for a very satisfying picture, with not too many compromises this time around. Though of course you never know - it's so far away! I can't believe it takes this long. By the time next-gen is coming, my son who was born a year after the PS3 came out in Europe, will be old enough to beat me at half of the games we play!

 

[hoho]

Have you heard of "detail levels" as in low, medium, high detail? That's right, something that nearly ALL devs use RIGHT NOW for any game that is ported to the PC!

You do realize that bigger textures will also need higher bandwidth and more processing power to be delivered on screen, right? If someone would boost XB360 RAM and texture sizes 4x and not change BW or CPU/GPU speeds then the games will become several times slower.

After you're already at maximum throughput for your GPU and memory channel pretty much the only thing where more RAM helps is with lowering the need for streaming data in and would remove some loading screens. It definitely won't give you better image quality from higher resolution textures or from less reuse of them.


Another thing is as long as we are stuck on 1080p using textures with more than 1k pixels per side is rather pointless. Anything that is even just a littel bit farther from the player will be occupyng relatively little space on screen and for every output pixel the GPU will blend together several texels. You'd be far better off to do that offline by simply using decent filters when downsizing the textures.

 

[Shifty Geezer]

Shifty, maybe because we're discussing the next gen console tech, not marketing or software?

The whole platform is built around a budget which can't be ignored and, prior to this RAMathon, hasn't been. Every discussion this thread has had to date regards feasible GPUs and CPUs and storage mediums etc. has been with an eye of producing a balanced system without anyone suggesting more CPU or GPU will hands-down win the next console-war battle. Indeed, console-warring hasn't been a factor at all, and we've just discussed reasonable probabilities of particular architectures. The whole marketing and services discussion hasn't needed to be raised because everyone who has contributed to this thread over the past several years has already understood that.

4 GBs fast RAM is a nice fit for the next gen boxes based on speculations of how much content devs want to make, overall system design and costs, how much actual difference 8 GBs would make versus 4GBs considering a very probable 1080p maximum resolution, how much processing power there'll be to drive those pixels and use that RAM, etc. 8 GBs is an option, but it'd come with compromises elsewhere to the platform. 2 GBs is a cheaper option, that'd save money that could be spent elsewhere. There's also the possiiblity of less RAM and more eDRAM, with eDRAM being a significant cost. You can repeat all you like that an 8 GB console would be better than a 4 GB console, but that won't change those options.

 

[function]

I would not tone down a difference in RAM amount between system competing for core gamers.
It only a part of the picture, but a significant one.
The 360 had a head start, the Rrod hurt it badly, if it would have come with more RAM it would mostly have been a game over for the PS3. The ps3 was vouch from scratch by the gaming press, internet, etc has a significantly stronger system, on top of impressive Sony games catalog it helped a lot to settle quiet some gamers minds. Now if you put economic considerations aside (which I aknowledge is impossible) and Ms added even more RAM to the system, clearly it would have a significant impact, there is no dismissing it, even 768 would have made a significant difference, imho enough of a difference for MS to stealth from Sony a lot more of the core gamer segment.

OK, there are economic realities but I would not underestimate the effect on sales on extra RAM. Clearly RAM doesn't sell games or systems but clearly with multi platform games taking the crown in gaming sales it would have been a great competitive advantage.
I'm confident that neither Sony or Ms want to be in a situation where they ship a system with a significant difference in the amount of ram available to developers as most likely power consumption concerns will ensure that the difference in processing power won't be much.
RAM would have a really significant effect on visual even to the average guy looking at the system side by side at say best buy. The "word on the street" would help too, gamers would clearly notice and say about it, no need for DF, lens of truth and all that kind of site, they exist because the systems were a wash imho. If they were not people could easily tell and so it would have sale implications.

I don't think the case for "double ram" (compared to a competitor) would necessarily go like that though.

For instance, imagine a system with 8GB of RAM that can be filled at 25 MB/s (or much less) from a Bluray drive. Now image a system with 4GB of RAM that can fill it at a guaranteed minimum of 100 MB/s from a section of flash memory (and no penalty for random access). The board is smaller and less complex, the flash memory would cost reduce too, and you may be able to justify using faster memory because you're using less of it.

Which is going to better support the model of data access that games are moving to? I'd guess it's the latter, but developer types would have a better idea than me. Getting data into memory seems to be almost as big an issue as how much memory you have these days.

 

[Arwin]

Getting data into memory seems to be almost as big an issue as how much memory you have these days.

It has always been like that. You can do with very little RAM as long as you can stream the data into it efficiently. Of course there are limits, but in theory your memory needed for graphics should be fine with 8x what a framebuffer needs, as long as you can stream in new data fast enough.

Of course in the end it all amounts to data, and what you need at what speeds where and when. As has been mentioned repeatedly, the amount of RAM is just a small piece of that picture. I think Rage's use of megatexture is a very good example of looking at the whole pipeline rather than just that small piece.

 

[function]

The disparity between memory size and storage transfer rate is growing though.

Using Playstation as an example to try and create a Loading Badness Index:
- PS1 was 3.5 MB of ram and 300 KB/s peak so call that 12
- PS2 was 40 MB and about 2.4 MB/s peak so call that 17.
- PS3 is 512MB and 9MB/s (I think*) constant so call that 57.

Now taking into account that PS1 and PS2 used CAV drives (iirc) it won't be quite as bad as this (although game data seemed to take up very little space on Saturn games and so I assume PS1 also), but things do seem to be getting pretty bad now on the Loading Badness Index.

A 1GB 360 would have had a minimum LBI of 64, getting increasingly more foul as you moved towards the inner edge of eternity.

*http://forum.beyond3d.com/showthread.php?t=42157

 

[erick]

function,

but how does the data get to flash? o_o

Do you mean to exhange optical media for flash cartridges? Too expensive, maybe?
Do you mean that data will be downloaded from Live!/PSN onto the SSD inside a console? Expensive and too constrained.

Surely you cannot mean that the data will be loaded from disc to flash to RAM, thus negating all possible upsides from flash cache, since loading straight to RAM and keeping it there would be faster (if system has enough RAM to hold all data ofc)?



Ho Ho, regarding your comment I don't think a GPU with the texturing capability of ~250 000 Mtex/s (a mid-high range GPU in 2014, roughly extrapolating from 84 480 MTex/s for 6970 in 2010) will prove unable to render the quality of textures that could fit in 8GB alongside with other content.

In fact, if we were to extrapolate that Xenos is powerful enough to render 256MB worth of textures without issues @ 8000 MTex/s, then mathematics tells us that to render 8GB worth of textures at a similar speed we would, in fact, need 256 000 MTex/s of texture fillrate. Which should be completely normal by 2014.

 

[liolio]

I don't think the case for "double ram" (compared to a competitor) would necessarily go like that though.

For instance, imagine a system with 8GB of RAM that can be filled at 25 MB/s (or much less) from a Bluray drive. Now image a system with 4GB of RAM that can fill it at a guaranteed minimum of 100 MB/s from a section of flash memory (and no penalty for random access). The board is smaller and less complex, the flash memory would cost reduce too, and you may be able to justify using faster memory because you're using less of it.

Which is going to better support the model of data access that games are moving to? I'd guess it's the latter, but developer types would have a better idea than me. Getting data into memory seems to be almost as big an issue as how much memory you have these days.

I don't mean strictly doubling the RAM, simply have more RAM that what you expect your competitors to ship with.
By PS360 time between 512MB and 1GB or RAM the load time might have been longer but not significantly, the difference in what the costumers would have perceived would have been greater than the trade off. Now if we speak 4GB vs 8GB, populating the RAM become more and more problematic and doubling already consistent loading may not be a good idea. So I agree with you, some pages ago I was actually considering 2GB back with enough flash storage a better tradeoff than 4GB without it. Obviously both is better, 8GB with really good SSD would be even greater, but I completely agree that economic considerations won't let this happen.
I was reacting about what sound a bit like a denial of the extra RAM (even without doubling) as a competitive advantage. Back to the ps360 that would have a huge impact as devs are constrained, one you remove the cost of various render targets, a frame buffer, the os, devs are not let with much, 75%? (let take this as a gross figure).
75% of 2GB is 1.5GB which is not in the same ballpark as 384MB (<=which get filled really fast).