Predict: The Next Generation Console Tech

[zRifle1z]

This article is interesting.

Thanks for the good read Liolio. I for one have really stopped purchasing games simply because I"m not impressed with much that is out there. I'm more than ready for a new console with better graphics, new game concepts and other "vast" improvments to the game play expeirence.

 

[xexuxjy]

Out of interest what improvement to the game play experience would you expect a new console to bring? Just about every game I've seen has really been an incremental improvement rather then something new... (don't mind that, just find it odd that a new generation of consoles (or even pc's) will provide a way of making new game concepts etc.)

 

[Prophecy2k]

In terms of improvements in gameplay experiences and concepts, i'd say that we're still getting alot of that this current generation.

Also, as the generation continues and certain specific titles tend to dominate the popularity and sales fields of their respective genres, I think we'll see yet even more innovative and distinctive gameplay ideas and experiences before next-gen rolls around.

One thing i would like to see before next-gen rolls around however is the rejuvenation of classic genres like the "Colony Wars" space sim. We need more genre diversity in this current generation that is beginning to stagnate from way too many FPS & TPS.

As xexuxjy said however, I don't expect any difference to the types or even "scope" of games being developed next-gen unless next-gen brings about the total proliferation of "waggle" (or its ultimate refinement) across every hw platform. Something that would be rather interesting...

 

[MfA]

For the same transistor budget (+2 billions) as ATI they end with a bigger chip +500mm² against supposedly 330mm², and to reach the same FLOPS figure they need to clock the thing @ 2GHz which @45nm might very well be out of their reach (late hints were more about ~1GHz). That's a consistent disadvantage in costs and perfs.

Clock is a design target, a miss by that kind of margin is unlikely. The main reason why you don't want high clocks is simply power, a 2 GHz GPU can not be compared to a 2 GHz CPU of the same die size ... the average switching activity is many times higher.

By the way there is a lot of talk about CPU vs GPU right now (and right here) and how architecture will meet or not, how would look a super flexible GPU without going the larrabee route?

They're all GPUs in my book. As for an evolution of the current NVIDIA/ATI architectures though ... subdividable warps (lets say down to 4 wide), more sophisticated caching, internal message passing facilities and a fully general heterogenous architecture (for instance all units can function as texture units or shader units ... some are just better at one or the other).

 

[cbarcus]

Current performance per watt @ 45 nm?

Anyone care to calculate the current performance per watt and cost of available CPUs on a 45 nm process? Say, Cell and Intel's Core i7 965?

Granted, this is a somewhat contrived comparison, an embedded processor vs. a high performance desktop chip only focusing on single-precision, but I think some readers here will benefit.

 

[liolio]

Anyone care to calculate the current performance per watt and cost of available CPUs on a 45 nm process? Say, Cell and Intel's Core i7 965?

Granted, this is a somewhat contrived comparison, an embedded processor vs. a high performance desktop chip only focusing on single-precision, but I think some readers here will benefit.

I'm not sure that that's possible. The cell power consumption is impressive in regard to its throughput that's it.

 

[Rangers]

Saw this on Gamespot..looks like next gen consoles are at least two years away (which I guess is no surprise, but it means 2001 is probably out, meaning 2012 at earliest.

http://www.gamespot.com/ps3/action/...&om_clk=newsfeatures&tag=newsfeatures;title;2

Kotick added that publishers don't take advantage of the full capabilities of today's hardware and said it might be some time before the next generation arrives. Typically, console makers give publishing partners about two years' notice when they plan to introduce new hardware, Kotick said. But so far, the console makers haven't given him specs or white papers on new hardware nor have they consulted him on design decisions, leading him to believe their current priority is instead to reduce the cost of each system.

 

[grandmaster]

Anyone care to calculate the current performance per watt and cost of available CPUs on a 45 nm process? Say, Cell and Intel's Core i7 965?

One of my contacts in the Far East has just offered me a Gulftown 2.4GHz six-core i9 engineering sample, fabbed at 32nm. If Intel is capable of making these chips now, could the next gen launch at 32nm two years from now?

 

[liolio]

One of my contacts in the Far East has just offered me a Gulftown 2.4GHz six-core i9 engineering sample, fabbed at 32nm. If Intel is capable of making these chips now, could the next gen launch at 32nm two years from now?

Imho opinion easily they may have already move to 22nm for their most lucrative activities like high end CPU server.
The problem is even @32nm a 32 cores larrabee would be big. If we assume a perfect shrinking it would be above 250mm² while it would still have to run @2GHz to reach ~2TFlops which may still prove bothering in regard to power consumption and thermal disspation (if it's the problem faces right now).
Ati by this time will use 28nm and could provide twice the raw power or more for the same overall budget (die size, thermal disspation, power consumption).

 

[Rurouni]

Probably someone had already suggested it, but I think we will see an incremental upgrade on the current hardware architecture for the next gen consoles. Basically PS3.5 and Xbox540(?). This hardware can run at least current spec'd game but in 1080p and 60fps (or probably more if you want to run it in 3D while still keeping the res and fps). And Sony and MS should make it mandatory for the games to run on both gen... basically like current PC situation where beefier hardware could run games at higher res and better effects. By doing this they can achieve their (at least PS3) 10 years life cycle while still have a product than can showcase better graphic.

My question is how hard for a developer to support such console (PS3 and PS3.5)? I'm thinking if the hardware tech exactly the same it shouldn't be much of a problem... correct? but how about if PS3 replace their GPU with current GPU? is there anything that can break compatibility (since the level of optimization on console can run deeper than PC)? Can they make a PS3 but replace it with current GPU but it can still play current PS3 games without needing any emulation/compatibility patch?
As for the CPU.. hm.. I don't know much. The resolution shouldn't matter much.. but in PS3 case since CPU is used to assist GPU, I would imagine that it would need a larger upgrade compared to current PC architecture style (or Xbox360 to a certain extent) where I can swap GPU and increase performance easily.

 

[grandmaster]

The likes of Carmack and Crytek appear to be gearing towards an actual generational leap though. At the end of the day it's going to be about affordability versus the best tech available at the time.

 

[sevanig]

What will happen is that PS3 and PS4 will be sold as the same device...

Meaning that the Ps4 will have a upgraded cpu/gpu better GFX etc.

But the Ps3 will match those GFX with the use of the Playstation Cloud.

So Ps3 will have the same games as Ps4, So there will be no need to upgrade unless you don't have a good internet connection

 

[Grall]

When the next gen of consoles does arrive, MS will be forced to go blu-ray... DVD will simply not suffice all that much longer; certainly not for an entire new console generation.

At the same time, digital distribution of games will not be feasible either for the vast majority of customers, not for many, many years. Even the largest laptop harddrive can't hold more than a handful of blu-ray titles, and most gamers don't have the kind of internet connection needed to download dozens of gigabytes in any kind of reasonable timespan.

Besides, we're seeing a gigantic backlash in several western countries right now where ISPs are crying crocodile tears over how much network capacity their customers are using, and rather than invest in their networks they implement throttling controls and putting monthly download caps in place. Just downloading one or two blu-ray discs would break these caps wide open, and result in huge "data overdraft" penalty fees for customers.

So there's no other choice for MS than to again go for a physical format, and blu-ray is the only remaining choice there... Oh, the irony!

 

[sunscar]

When the next gen of consoles does arrive, MS will be forced to go blu-ray... DVD will simply not suffice all that much longer; certainly not for an entire new console generation.

At the same time, digital distribution of games will not be feasible either for the vast majority of customers, not for many, many years. Even the largest laptop harddrive can't hold more than a handful of blu-ray titles, and most gamers don't have the kind of internet connection needed to download dozens of gigabytes in any kind of reasonable timespan.

Besides, we're seeing a gigantic backlash in several western countries right now where ISPs are crying crocodile tears over how much network capacity their customers are using, and rather than invest in their networks they implement throttling controls and putting monthly download caps in place. Just downloading one or two blu-ray discs would break these caps wide open, and result in huge "data overdraft" penalty fees for customers.

So there's no other choice for MS than to again go for a physical format, and blu-ray is the only remaining choice there... Oh, the irony!

What I wonder is if even *standard* Blu-ray will be sufficient starage for next-gen. Aren't some games already filling up a 50GB Blu-ray? Would it be inconceivable for a game to fill 250gigs this next go-round? If so, I can't figure how feasible it'd be to stick with DVD, or to even go with digital distribution.

I realize that a generation from now, it won't be absolutely necessary to store insane numbers of LODed meshes, but texture res if anything is going to go up, and if we start adding high-res displacement maps to the equation, storage requirements *are* going to have to grow.

 

[AlNom]

Why should content balloon when targeting even 1080p resolution Pre-rendered movies take up the most space...

 

[idsn6]

MegaTextures, GigaVoxels. If "unique, high detail" takes off, it will need the space.

 

[AlNom]

So every game will automagically require those eh?

 

[sunscar]

Why should content balloon when targeting even 1080p resolution Pre-rendered movies take up the most space...

Not saying it should, only that it may. Kind of thinking from the perspective of RAGE here - ID's already saying they could easily over fill a Blu-ray, so what happens if we take that philosophy forward, I.E. they're using 256Kx256K texture layers for their environment, instead of 128Kx128K, and they're using a few new layer types, too. That could get very large very fast.

So every game will automagically require those eh?

Not at all, but it seems it will be the trend for *mainstream* games going forward to continue looking better and better. It's hard to argue that future game engines won't allow for higher and higher resolution textures, and it's clear enough we'll probably start adding a few new texture types in there before all is said and done, too.

 

[Acert93]

What I wonder is if even *standard* Blu-ray will be sufficient starage for next-gen. Aren't some games already filling up a 50GB Blu-ray? Would it be inconceivable for a game to fill 250gigs this next go-round? If so, I can't figure how feasible it'd be to stick with DVD, or to even go with digital distribution.

I realize that a generation from now, it won't be absolutely necessary to store insane numbers of LODed meshes, but texture res if anything is going to go up, and if we start adding high-res displacement maps to the equation, storage requirements *are* going to have to grow.

The problem with 250GB games isn't just storage but transfer. You have to get the data from the storage media to the system memory at some point. This will be a hurdle to hardware and software and will curtail some options.

 

[sunscar]

The problem with 250GB games isn't just storage but transfer. You have to get the data from the storage media to the system memory at some point. This will be a hurdle to hardware and software and will curtail some options.

Recalculating, I don't think things will get quite as bad as that, but I still see it may be common for games to weigh in over 100GB. Compression should be in the picture, tesselation + displacement may also have a fairly huge impact on over-all what has to be stored. If instead of storing hundreds or thousands of unique meshes, we had maybe one humanoid, one vase, one man hole cover, but allowed for dozens of permutations of those objects based on displacement maps, we could probably get away with even less storage.

This is ofcourse all just the gears in my head ticking away at what the picture will be a few years out - it could all be nothing.