Predict: The Next Generation Console Tech

[Entropy]

When was the last time someone was prepared to pay £50 for a Cell phone game though? Who do you know with 30 to 40 games all bought for their phone? People keep talking about how Cell phones have such a huge user base, they do, but so do vacuum cleaners. Of course I'm exaggerating for comic effect before anyone shouts at me for comparing a phone to a vacuum cleaner But my point is phones are used mostly for phoning people, playing music, or as a camera, gaming is way down the list of uses.

And £50 games are necessarily a sign of health for the games industry?

That aside, cell phones have a huge and growing user base and it is significant. With both your pricing comment and your number of games comment, (although, how many console buyers have 30-40 £50 games?) you seem stuck on the fact that "phones are not consoles". No they are not, but then again this is an advantage and an opportunity, not a liability. Unless, of course, you are emotionally or financially dependent on traditional stationary consoles.

Also, cell phone software does not use the old publisher/retail distribution model. If you sell 50000 £50 games or 500000 £5 games or 5000000 £0.50 games is pretty much a wash since you have removed the physical media and associated production and distribution costs. Since smartphones are not bought for the explicit purpose of gaming the average game attach rate is going to be lower than for consoles, but the flip side of this is that many who wouldn't want the hassle and expense of carrying around a dedicated games device, will carry their phone around, and these people represent an expansion of the user base. And that is arguably very notable and important.
Games capable devices, large user base comprised both by people who already frequently game and gamers of convenience, plus development environment and most importantly distribution model already in place. Of course this will be a growing segment - how could it not be? Can anyone here make a convincing case that the trend towards portable communication and entertainment will reverse? That the portable devices won't become ever more capable? What can they do in 2 years, or 5 or 10 years?

As far as this thread is concerned, no, the stationary console market won't attract any new competitors, and the reason isn't because the barrier of entry is prohibitively high.

 

[Entropy]

I think the Zune model may be the best indicator of MS possibly looking in that direction.

The service and software is MS, but the hardware is Toshiba (I think). Whether this would work in the console business remains to be seen.

MS has done vastly better from Xbox 1 to Xbox 360, with every new iteration I expect them to gain experience and avoid the same pitfalls. I fully expect MS to do a lot better at product testing the console for Next gen.

The consoles are getting very good at bringing all your entertainment into one box. Music, video, DVD, Blu Ray, games and pictures are already there. I would anticipate that next gen there should be at least IPTV, DVR, and web browsing. With all of the functionality that the console provides, why would you invest in a set top box like Roku or the Apple TV box if priced comparatively?

The problem is that nobody invests in the Roku or AppleTV as it is. They have a market window for as long as mobile bandwidth is below certain limits and portable storage is below certain limits/above certain cost levels. But even within that window, both devices have failed in the market place. They have no future. Saying that future consoles could do as well as these is hardly an endorsement.

What you outline for future consoles is that they will take over the role of the stationary computer in the home, another dying dinosaur. And this is a problem for Microsoft, not an opportunity. They got into the console business to a large extent as a preemptive strike against the threat they perceived from the PS1 and PS2, and Kutaragis visions that got too close to Microsofts home turf for comfort. They bought their way into the market at great expense, and by promoting this alternative to developers that traditionally catered to their Windows platform. But the return per invested dollar has been dismal, and luring developers to their console has weakened the appeal of Windows PCs (their real cash cow) for the home. The gains were bought at a high price and cannibalizes their main product. Double ouch.
Additionally, with the benefit of 20/20 hindsight, it is clear that the "living room trojan" simply doesn't exist as a threat anymore. People play video and music from their iPods and iPhones, they want to carry their entertainment around, and use it where they happen to be - the living room doesn't enjoy the same position as the center of entertainment in peoples lives anymore and there is really no doubt that this trend will continue. Portable devices that can connect to stationary equipment when desired and available is where the future lies. It's bleeding obvious by now that the PS4 (or whatever) isn't a significant threat to Microsofts software hegemony. Portable Cellphones/netbooks/whatever running non-Windows software however is.

If I had held a strong stake at Microsoft, I would drop stationary consoles entirely, promoted the hell out of Windows 7 as a gaming platform and bought the support of developers/publishers in order to try to strengthen the viability of Windows PCs as home use devices, and tried really, really hard to get Nokia et al to adopt their not-too-inspiring Windows for mobile devices if it isn't already too late. Microsoft has an awesome financial war chest. But spending it on stationary games consoles makes no sense at all.

 

[specwarGP2]

I'll answer that with 2 further questions;

1. With 2-3 distinct platform owners, why have the games console still only found its way into <200M homes in any given generation? (last gen from memory I think it was 170m approx, less in each previous gen). It is now the biggest entertainment industry in the UK, second in most other countries, why will the potential of the consoles to be installed in the majority of homes still be held back?

Are you suggesting that all these years, console companies have been unable to manufacture enough consoles to meet demand?

2. Does a software company such as MS have more to gain by taking advantage of other CE's capacity and more importantly brand name (Panasonic, Samsung, Toshiba e.t.c) when trying to dominate the console market, moreover, the living room?

Apparently, brand name hasnt helped Sony much. You haven't answered my question. If the hardware is profitable to manufacture, why wouldn't a platform owner manufacture all the hardware and keep all the profits?

 

[Nano]

Are you suggesting that all these years, console companies have been unable to manufacture enough consoles to meet demand?

It goes to show we all have different interpretations of words

I'm not suggesting that at all, in time they always have the capacity to supply for demand. I'm actually suggesting the industry will probably see some significant changes in format to really exploit the fact the games business is so big and growing by so much year on year.

The major difference between 'the console' and other entertainment devices is that each distinct console is only developed by one manufacturer. I quite strongly believe that the full input from other h/w manufacturers and the biggest consumer electronics companies will push the instalment of dedicated game systems in more homes. One company, one model, one marketing strategy, can be replaced by several. I can say now, Samsung or Pioneer would go about marketing & pushing their 'Xbox powered' system in different ways, and through different regions and different formats MS has little/no influence in.

Apparently, brand name hasnt helped Sony much.

No? If the Wii or the PS3 were made by Microsoft I doubt they would be selling more in Japan than a 360 that had either of the other names attached to it. Expanding on Sony, if their system were to overtake their HD rival this generation, IMO it would be almost solely down to superior brand recognition in the 3rd biggest games industry in the world.

You haven't answered my question. If the hardware is profitable to manufacture, why wouldn't a platform owner manufacture all the hardware and keep all the profits?

Well again its down to interpretations. Cos the question I posed had the answer in a roundabout way.

MS even if they own every bit of hardware they sell, they has a much lesser scope (at least where I'm looking at it) to push their Windows/DX hybrid into every living room, than if they make OS' and IPs for 3rd party companies to take it in their different directions and through their own unique, industry leading methods. I think there is a hell of a lot more money in it, to be quite frank if the best case scenario of 'my model' worked, they could potentially lead the market by miles and have a hell of a lot more licensees; be it developer, or consumer electronics company and subsequently more gamers buying software for that platform base; basically more $

 

[pc999]

... 5) Make a whole lot less money!

To counter those points:
1) Problems of manufacturing are not really problems if they are making you lots of money.

2) Install base and software sales do not grow in a linear fashion. Having a install base twice as big does not mean you sell twice as much software. The falloff would be effected even more if you had CE devices with wii integrated being sold to people with no intrest in gaming.

3) Dont know why there would be a price war to begin with if the product is selling well with nice profit margin. see Wii

4) Options to the consumer means nothing if it is not making you money.

5) We cant know that.

1) I think we can say that Wii does have a big manufacturing problem. Any new player (or console, like a portable XB or such, or even like someone said a new model for a old console) can also preffer to offload must of the production, being a lesser risck etc...

2) True, but a bigger instal base give you new opurtunitys for other kinds of SW (meybe with time for that same kind of SW), see what happened in cell phones or even Wii.

Educational SW, utilitys, casual (tetris like, casual, even if more complex) gaming could see a new life if your (eg) TV bring all the console HW.

3) If there are more than one producing the same (kind of) product there is a price(/value) war, the more alike the product the bigger it is. We only benefict from it.

4) It may give you a new costumer that you would not have.

Just to make a point that I dont think this will come true anytime soon (if ever), neither would I prefer (meybe I would, meybe I wouldnt), or anything like that.

But there is some advantages for it.

 

[jetrii]

I apologize if this has already been discussed, but is a POWER 7 chip likely for the next generation Xbox or is it simply too expensive? Early reports suggest that the chips may feature:

45 nm process
2 chips per module
8 cores per chip
4 threads per core (32 threads per chip)
4.0 GHz clock speed

Would a single 8 core Power 7 chip be too expensive for the Xbox 720?

 

[pc999]

I apologize if this has already been discussed, but is a POWER 7 chip likely for the next generation Xbox or is it simply too expensive? Early reports suggest that the chips may feature:

45 nm process
2 chips per module
8 cores per chip
4 threads per core (32 threads per chip)
4.0 GHz clock speed

Would a single 8 core Power 7 chip be too expensive for the Xbox 720?

To litle info, anyway if you want to pick the 256Gflops figure, that is a litle more than Cell, a reworked/evolved Cell to overcome the main problems wouldnt be better (I think that in 2011 they could use such design).

Imagine a four core PPE, with 12SPUs @ 3.2Ghz, improved branching prediction, latency and all those litle things for gp code (at least in the PPE), would be much smaler than a Power 6 and give more (just theorical?) performance in a already know architeture.

 

[jetrii]

To litle info, anyway if you want to pick the 256Gflops figure, that is a litle more than Cell, a reworked/evolved Cell to overcome the main problems wouldnt be better (I think that in 2011 they could use such design).

Imagine a four core PPE, with 12SPUs @ 3.2Ghz, improved branching prediction, latency and all those litle things for gp code (at least in the PPE), would be much smaler than a Power 6 and give more (just theorical?) performance in a already know architeture.

Didn't Microsoft initially go to IBM with plans for a chip which IBM simply couldn't deliver in time and eventually settled for the current Xenon? If so, when you combine that with the fact that Sony will probably have another powerful Cell processor, is it possible that Microsoft may try to outdo Sony with their own CPU?

I can see an 8-16 core Power7 in the next generation Xbox in order to try and win the game performance crown. However, the Power7 sounds like it will be fairly expensive.

 

[N00b]

I can see an 8-16 core Power7 in the next generation Xbox in order to try and win the game performance crown. However, the Power7 sounds like it will be fairly expensive.

I don't think a 16 core cpu is going to happen for game consoles, probably not even an 8 core. Sandia Labs think so, too, as does Mr. Amdahl.

 

[scificube]

I don't think a 16 core cpu is going to happen for game consoles, probably not even an 8 core. Sandia Labs think so, too, as does Mr. Amdahl.

I would reassess. Cell already has 8 cores. Amdahl's law does not apply to the set of all tasks and when it does apply it does not preclude real gains in performance for more cores in all cases. The Sandia Labs article is more directed at the contraints the memory wall imposes on execution time than multi-core chips designs themselves. Rambus amoung others are projecting memory chips capable of delivering TB's of bandwith with respectable latency so the memory wall could be knocked down a great deal more than it has been in a long long time.

That said an Intel I7 CPU with more 8+ cores would make for a rather hot and bothered console I would imagine...and no doubt cost an arm and a leg...and a half.

 

[N00b]

I would reassess. Cell already has 8 cores. Amdahl's law does not apply to the set of all tasks and when it does apply it does not preclude real gains in performance for more cores in all cases. The Sandia Labs article is more directed at the contraints the memory wall imposes on execution time than multi-core chips designs themselves. Rambus amoung others are projecting memory chips capable of delivering TB's of bandwith with respectable latency so the memory wall could be knocked down a great deal more than it has been in a long long time.

That said an Intel I7 CPU with more 8+ cores would make for a rather hot and bothered console I would imagine...and no doubt cost an arm and a leg...and a half.

Cell's 8 SPs are a special case. They have their own local memory and can be arranged as pipeline stages (DMA), which minimizes overhead. It's quite a clever design. The question is however how high the throughput is in the average case and how much is lost due to scheduling, syncronization.

Rambus has been projecting a lot of things for a long time (yawn). Wake me up when they actually deliver.

Right now I don't see any memory technology that could feed 8 or 16 cores when using large datasets, which games usually have, without a severe performance loss. Which means going beyond 4 cores, we will see diminishing returns. Right now the I7 with 8 cores is targeted towards server applications, which don't require as large datasets or have workloads that are better parallelizable than games.

People here falsely assume that workloads can be parallized further without loss of efficiency, so more cores is always better. But that's not the case most of the time (see Amdahl). You write the Amdahl does not apply to all tasks, but then these tasks have to be independent, which is not the case for most tasks in a game.

 

[eastmen]

So do you think its a smarter idea to go with say a 4-6 core chip with tons of cache on it ? Perhaps a waternoose with 12MBs of cache instead of the current 1MB of cache . Would you see a bigger performance doing that than adding a 4th core to teh waternoose ? I"ve noticed that the newer the intel chip the more cache it has

 

[AlNom]

People here falsely assume that workloads can be parallized further without loss of efficiency, so more cores is always better. But that's not the case most of the time (see Amdahl). You write the Amdahl does not apply to all tasks, but then these tasks have to be independent, which is not the case for most tasks in a game.

Indeed... From previous threads, it's pretty clear that there are improvements to be made without worrying about massively multicore designs. Branch prediction, op/load/store latencies, cache... After a certain number of cores, it becomes more of a headache than a boon.

I'd almost say that the folks who agree with you have already voiced that opinion long ago, and it's the new batch of folk with the misconception...

 

[pc999]

Didn't Microsoft initially go to IBM with plans for a chip which IBM simply couldn't deliver in time and eventually settled for the current Xenon? If so, when you combine that with the fact that Sony will probably have another powerful Cell processor, is it possible that Microsoft may try to outdo Sony with their own CPU?

I can see an 8-16 core Power7 in the next generation Xbox in order to try and win the game performance crown. However, the Power7 sounds like it will be fairly expensive.

Why would that be better than proper evolution of the Cell, sometimes a CPU is the best for somethings but a real bad choice for others.

So do you think its a smarter idea to go with say a 4-6 core chip with tons of cache on it ? Perhaps a waternoose with 12MBs of cache instead of the current 1MB of cache . Would you see a bigger performance doing that than adding a 4th core to teh waternoose ? I"ve noticed that the newer the intel chip the more cache it has

Well see the main diference between Phenon and Ph II, or between a C2D and a Pentium Core Duo, Gekko and Broadway, or the AMD investiment in Z-Ram.

Anyway cache is just part of it...

 

[N00b]

So do you think its a smarter idea to go with say a 4-6 core chip with tons of cache on it ? Perhaps a waternoose with 12MBs of cache instead of the current 1MB of cache . Would you see a bigger performance doing that than adding a 4th core to teh waternoose ? I"ve noticed that the newer the intel chip the more cache it has

Yes. This has the additional advantage that you don't have to worry about your software architecture as much and you will reach better average performance earlier in the console lifecycle.

Creating a software architecture that scales well over multiple threads is pretty hard and most developers don't have the skills for that. Considering that next-gen games will most probably have even larger development teams and will use even more middleware, balancing and optimizing a software architecture for 4 core is much easier than optimizing for 16 cores.
With 16 cores, fine-grained tasks and possibly a lot of 3rd party code a system can quickly become unpredictable, resulting in poor or unstable performance.

I'd almost say that the folks who agree with you have already voiced that opinion long ago, and it's the new batch of folk with the misconception...

Yes. But since we are a technical forum here, a little education now and then cannot hurt.

 

[AlNom]

Yes. But since we are a technical forum here, a little education now and then cannot hurt.

Oh certainly. Sorry, I didn't mean that you shouldn't do so!

 

[Megadrive1988]

Since the Power Processor Element in Xenon and CELL were derived from POWER4, I wouldn't be surprised if the main cores of the next-gen Xbox and PS4 CPUs are derived from the current POWER6 rather than the upcoming POWER7.

 

[3dilettante]

Since the Power Processor Element in Xenon and CELL were derived from POWER4, I wouldn't be surprised if the main cores of the next-gen Xbox and PS4 CPUs are derived from the current POWER6 rather than the upcoming POWER7.

POWER4 was a wide OOE processor with a complex decode method and no vector capability.
The PPE and Xenon are simple, dual-issue, narrow in-order cores with vector capability.

What exactly was kept?

 

[scificube]

Cell's 8 SPs are a special case. They have their own local memory and can be arranged as pipeline stages (DMA), which minimizes overhead. It's quite a clever design. The question is however how high the throughput is in the average case and how much is lost due to scheduling, syncronization.

Rambus has been projecting a lot of things for a long time (yawn). Wake me up when they actually deliver.

Right now I don't see any memory technology that could feed 8 or 16 cores when using large datasets, which games usually have, without a severe performance loss. Which means going beyond 4 cores, we will see diminishing returns. Right now the I7 with 8 cores is targeted towards server applications, which don't require as large datasets or have workloads that are better parallelizable than games.

People here falsely assume that workloads can be parallized further without loss of efficiency, so more cores is always better. But that's not the case most of the time (see Amdahl). You write the Amdahl does not apply to all tasks, but then these tasks have to be independent, which is not the case for most tasks in a game.

If I offer that memory will be faster and you say no it won't be fast enough there is little to discuss on the matter isn't it? The only thing we can do is see if the memory is delivered as promised. By the by Rambus is not the only player in the memory market. Rambus was merely a practical example. If it is any consolation I am as skeptical of Rambus as anyone because even if they do make good on their claims Rambus is likely to price themselves right out the market in typical Rambus fashion.

You are correct when you say most tasks in a game are not fully independent but there are quit a few things that are done asynchronously still.

As for your I7 example I would say it just as much a matter of use cases as anything else. What use is an 8 cores I7 to the average consumer? Null. What use is an 8 core I7 to the server market? That remains to be seen but the obvious expectation and hope from Intel is that its plenty.

Rather than having a protracted debate on the value of multi-core I'll go back to the original question and that is how likely is it that we will see more than 8 cores on console CPUs? I thinks it's likely given Cell's roadmap that we will see more than 8 cores from Sony's side. I don't think its likely well see CPUs with cores as nice as I7 with a multiplicity of 8 or more cores because of thermal/power/cost constraints. There is a chance MS may make a few wet dreams come true by using a couple Larabee's to do "it all" in their next console and if anything LRB is a multi-core design with far more than 8 cores.

This is what I was thinking of when I first responded to you but when I perused what you initially offered as support for what you were saying I didn't think what you presented was going to your point. The article you linked mainly spoke to issues with the memory wall as I previously noted and while I fully recognize Amdahl's law I feel there is much to be had with concurrent execution on multiple cores.

No - more cores is not always better and I never stated such. There is no worry that I would be promoting such an idea. However, I still see it as more likely we will more than 8 cores with console CPUs than we won't. Debating the value of more cores wasn't my original intent but none the less is fair game if you want. Speaking for myself I don't disagree with you except for at the point where we will certainly see diminishing returns. I don't think the cliff lies directly beyond 8 cores and I gather you do.

One thing you shouldn't assume is that you need to educate me on the subject matter especially when I truly said nothing to suggest you need to.

 

[Nano]

I apologize if this has already been discussed, but is a POWER 7 chip likely for the next generation Xbox or is it simply too expensive? Early reports suggest that the chips may feature:

45 nm process
2 chips per module
8 cores per chip
4 threads per core (32 threads per chip)
4.0 GHz clock speed

Would a single 8 core Power 7 chip be too expensive for the Xbox 720?

I could imagine an 8 core chip derived from that architecture for '720'.

I'd say the biggest factor will be price depreciation between release in 2010 and the launch of the 720 itself any where between then and 2012.