Predict: The Next Generation Console Tech

Status
Not open for further replies.
I'm just asking if there is a pot of gold out there in MMO,
Well, many people in the industry hope so, but it doesn't really seem to be the case. There are social interaction solutions on the net that are superior to game environments, and do not require the gaming fees. And as has been pointed out, the existance of WoW drains the pool of players for other games. (It was always thus, not only for no-skill MMOs, but earlier for MUDs or for the more skillbased multiplayer FPSs - there are only so many players, and they can't sustain multiple franchises at a high level.)

It is definitely not connected to future hardware though - all the platforms already have the hardware necessary for online MMOing. Consoles are typically about easy access - pick up and play. MMOs, or indeed online multiplayer games in most forms don't mesh well with the console no-hassle ethos. My guess is that they will never be a big item on consoles. Social "lounges" however may become popular. Or not, as they are already available on the net, and far more accessible than any solution tied to a console will ever be.
 
Instead of making a new thread I'm using this one (and maybe I'll also post this in that other next-gen GPU thread).

Lets think about the GPU for Wii 2 / Wii HD / Super Wii / Nintendo's next console.

Flipper / Hollwood is an outstanding architecture it's time. Technology that was originally designed in 1999/2000, released in 2001 in Gamecube. Given a speed boost and a few modest enhancements (i.e. 8 inputs to the 4 texel units instead of the original 4 inputs as in Flipper).

Should the next-gen GPU be:

1.) a totally new architecture that's a huge leap from Flipper/ Hollywood much like Flipper was a new architecture and huge leap over N64' RCP?
Flipper had about 10 times the fillrate of RCP (more like 20 times the realworld fillrate) and roughly 100 times the polygon rate. It was almost as big a leap as PS1 graphics to PS2 graphics. 45nm or 32nm

2.) or just a greatly upgraded Hollywood with more functional units, geometry engine, pixel pipelines, TEVs, texel units, faster clock etc (providing 5-10x the power of Hollywood) and added features. HD capable. 45nm
 
If Nintendo has any plans for HD resolutions then they're going to have to get a new GPU in there. Shrinking the current core and o/c'ing it AGAIN isn't going to cut the mustard this time.
 
I think it'll be AMD's latest unified-shader GPU available with a 1T-SRAM frame buffer.
So not much different from Xenos. But it has to meet these conditions:
1) It's a one-chip solution.
2) It's fully backward-compatible with Wii.
3) Supports HD.

They didn't launch Wii too quick even though it was technologically conservative, with Wii's success in hand they'll wait until all of conditions are met.
 
Here is what Rambus thinks they will have.
2wo9ge0.jpg
 
PS1: year 1994 - 132 MB/s (non-Rambus)

PS2: year 2000 - 3.2 GB/s (Rambus RDRAM)

(24x increase over PS1, not counting the massive eDRAM bandwidth)

PS3: year 2006 - 25.6 GB/s (Rambus XDR-DRAM)

(8x increase over PS2, not counting the RSX's bandwidth to GDDR3)



It's too difficult for me to compare PS2's 48 GB/sec eDRAM bandwidth to PS3's 22.4 GB/sec external graphics memory bandwidth. It is apples & oranges.
 
Last edited by a moderator:
I think the figures are just illustrative in a neat and readily observable way of the scope of performance increase. eg. PS3's RAMBUS BW isn't the 50 GB/s listed, but the system BW is approximately, and most importantly for RAMBUS it could have been. Which raises the important issue about reaching 1 TB/s - is it obtainable in a cost effective manner so the consoles get that performance?
 
They have to achieved 1 TB/s by going with higher clocking technology for their memory, I doubt even next gen consoles are going to use 256 bit bus. I think 200+ GB/s is what we should be expecting. And hope cache and local store can take care of it.
 
Agreed, 200GB is a good ball park.

>....Pure speculation....>

Ive said it before and Ill say it again. I think the graphics performance of the nextbox / ps4 will be around an SLI'd gt200 config (assuming its a 1.5 billion transistor part) or to put it another way, a gt200 with higher a clock and some added functionality. I cant see how it would be any more than that. And I cant see the mem bandwidth being much more than stated above.

Basically by the end of 2008 we will have a PC configuration albeit a very expensive one which will simulate the performance of a 2011 console.

Does that sound a reasonable ball park to you?
 
Given that this time they will have to avoid the "Diminishing returns" trap , i 'd expect dual GPUs with 2x300 GB/s at the least.
 
Agreed, 200GB is a good ball park.

>....Pure speculation....>

Ive said it before and Ill say it again. I think the graphics performance of the nextbox / ps4 will be around an SLI'd gt200 config (assuming its a 1.5 billion transistor part) or to put it another way, a gt200 with higher a clock and some added functionality. I cant see how it would be any more than that. And I cant see the mem bandwidth being much more than stated above.

Basically by the end of 2008 we will have a PC configuration albeit a very expensive one which will simulate the performance of a 2011 console.

Does that sound a reasonable ball park to you?

No...

Purely because although the GPU side might be comparable in performance, the CPU side of the next generation of consoles will likely be vastly superior to what any desktop could offer by the end of this year..
 
No...

Purely because although the GPU side might be comparable in performance, the CPU side of the next generation of consoles will likely be vastly superior to what any desktop could offer by the end of this year..

Superior is a difficult word. In what respect(s)? Cost effectiveness? Bandwidth to GPU? Bandwidth to main memory? Vector floating point capabilities? Cache size? Integer IPC?

I think the only thing that can be assumed is that the CPU will be tailored to the application, and not (like the original Xbox) be a typical low end PC chip of the time.

In the spirit of the thread title, as far as prediction goes, generally speaking I'd assume that keeping costs and power draw down from the outset will be more highly prioritized by Sony and Microsoft, given their respective issues, and the success of the Wii. But what would that mean?

I'd guess that the PS4 will do a relatively straightforward scaling of the BE, maximizing retention of assets and tools. It is designed for scaling, after all. GPU - difficult to predict, but it will be capable of rendering and outputting full HD with antialiasing.

The Wii, in all likelyhood, will get a small, cheap, capable processor. Quite likely still a single core, or possibly dual. There are candidates, but it could also be an upcoming core. GPU - impossible to predict, except to express my doubt that it will be a scaling of the existing GPU. A completely new design is likely, but just as this generation, I'd guess that Nintendo will go for a cost and power effective design.

The difficult one is the XboxNext. The CPU doesn't really lend itself to extended parallelism, an 8+core SMP UMA system is a horror to get reasonable utilization from. I'd assume that they would be conservative on the CPU (increase clock somewhat, possibly add another core, beef up IPCs somewhat, improve data paths), but they could also scrap the existing design completely and start over. Backwards compatibility may not be counted for much.
The GPU is another issue again. I'm not terribly knowledgeable as far as the internals of GPUs go, but as with the PS3, supporting rendering in, and outputting full HD with AA seems a given. Will they go the external ROP+memory route again? Will they coordinate with the Direct3D flavour of the day (probably)?

Just as SD resolution was a fixed target for the previous generation of consoles, the next generation will all go for HD, and by that time it will be more readily accomodated using relatively inexpensive hardware. I doubt that any of the players will be inclined to push the envelope in terms of hardware. That makes for less exciting fantasy systems, of course. But I just can't see that there will be any strong market forces that encourage the manufacturers to move strongly in that direction. There will be strong forces in favour of cool, quiet and inexpensive.
 
Superior is a difficult word. In what respect(s)? Cost effectiveness? Bandwidth to GPU? Bandwidth to main memory? Vector floating point capabilities? Cache size? Integer IPC?
I mean't superior with respect to game related processing performance..
Sorry I should have been more specific..

PS4
I mainly made this point with respect to the next Cell in the PS4 as it's probably the only area we can pretty accurately estimate it's direction for the next generation. Even with 24-32 SPUs & 1 or 2 PPUs in a Cell 2 at 3.2GHz, Sony will still have a relatively cheap monster of a chip which I suspect will benefit from the developer mindshare & experience gained from this generation significantly. Aside from Larabee (is this even coming out in the fall?) I can't see any desktop consumer chip competing in game-related processes this side of xmas 08..

XboxNext
Granted we know little of the direction they'll be going this time around but it's pretty clear that MS already know where Sony might go & so will likely want to compete reasonably enough but on a tighter budget than they did with the 360 maybe..?

Wii2
Who knows..? Likely not uber powerful & may not even hit current Cell real worldperformance numbers..

Just as SD resolution was a fixed target for the previous generation of consoles, the next generation will all go for HD, and by that time it will be more readily accomodated using relatively inexpensive hardware. I doubt that any of the players will be inclined to push the envelope in terms of hardware. That makes for less exciting fantasy systems, of course. But I just can't see that there will be any strong market forces that encourage the manufacturers to move strongly in that direction. There will be strong forces in favour of cool, quiet and inexpensive.
To be honest I'm not sure..

While the Wii has certainly generated huge success for small inexpensive hardware, software & hardware sales for PS3 & Xbox360 still provide enough of an incentive to push the envelope with respect to high end hardware platforms but i'm sure MS & Sony will be looking to strike an interesting balance next time around in a bid to find a sweet spot between performance hardware, low cost, innovative peripheral UI & software sevices as their differentiators..
 
PS4
I mainly made this point with respect to the next Cell in the PS4 as it's probably the only area we can pretty accurately estimate it's direction for the next generation. Even with 24-32 SPUs & 1 or 2 PPUs in a Cell 2 at 3.2GHz, Sony will still have a relatively cheap monster of a chip which I suspect will benefit from the developer mindshare & experience gained from this generation significantly.

Fully agree.

While the Wii has certainly generated huge success for small inexpensive hardware, software & hardware sales for PS3 & Xbox360 still provide enough of an incentive to push the envelope with respect to high end hardware platforms but i'm sure MS & Sony will be looking to strike an interesting balance next time around in a bid to find a sweet spot between performance hardware, low cost, innovative peripheral UI & software sevices as their differentiators.
Exactly, the aim the next time around will probably not be to outgun the competition in terms of technology, but to offer a tasty whole. Graphics technology is a part of that of course, and I expect to see improvements from Sony and MS both, because I suspect that their customers share that expectation, and S&M need to cater to expectations to some extent.
That said, neither of those two are probably happy to eat billions in losses in the first couple of years one more time when eventual success is not guaranteed. It has turned out OK for Sony and MS both - Sony managed to get Blu-Ray entrenched, and MS got a huge early lead in installed base that ensured that the 360 would do better overall than the Xbox. But the irony of the Wii outselling both of them on the merits of being more accessible, less imposing and having its emphasis on fun, can't really escape anyone.
Selling "more of the same" is a sure recipe for some sales, but it must be difficult to look at the DS vs. the PSP and the Wii vs. PS3/360, and not see a message there.

PS. While Microsoft shareholders are meek, the company has eaten some terrible losses on their XBox adventures, and for what - 20 million game systems? If I were a shareholder, I would be emphatic that Microsoft should withdraw from this disaster of a market, and spend that money bullying in on a market where the company can hope to see some real growth and revenue - cell phones. Consoles really are small fry, and the fears that Sony would muscle in on their computing turf has been shown to be unfounded. That game is over. Cell phones taking the place of PCs both for many personal computing needs and entertainment is much more likely. It's already happening.
 
Last edited by a moderator:
I just prefer to be a bit optimist.
After all this is a "predictions" thread, its not like i ll commit suicide if i am way-off from my predictions :smile:

Dont Lie, I know youll be slashing your wrists when you dont get your dual GPU console... lol :)
 
No...

Purely because although the GPU side might be comparable in performance, the CPU side of the next generation of consoles will likely be vastly superior to what any desktop could offer by the end of this year..

Ok well my post focused more on the GPU side than CPU, but having said that...

For raw FP performance your right cell2 will be better performing than any intel chip in 2008, but for complex serial code, you can forget it... Not even close. (if the PPU is in order rather than OoO.

And taking that in to account I still think a high spec PC from 2008 will be in the ball park of a 2011 console. I know we would all like the consoles to BATTER PC offerings, but its not likely to happen.

If a high end PC at the end of 2008 with SLI'd top end cards can run Crysis in 1080p or there abouts in high detail with 2xAA and 16x AF then that for me is about the right target performance for a next gen console, give or take.
 
If a high end PC at the end of 2008 with SLI'd top end cards can run Crysis in 1080p or there abouts in high detail with 2xAA and 16x AF then that for me is about the right target performance for a next gen console, give or take.

Next gen consoles would have to do ALOT better than Crysis at 120fps, in HD & no jaggies to get my vote..

& I feel it's likely they will..
 
Status
Not open for further replies.
Back
Top