Predict: The Next Generation Console Tech

[Blazkowicz]

Thats my point, it was the idea of Microsoft for their xbox360 to solve the ps2 edram problem, using edram only as a front buffer, and not at all for textures (using the 512 mo unified gddr for textures), but even this solution failed, even 10 mo for only basic front buffer operations was unsufficient to sustain any 720p+2AA image without tiling. they needed at least 14 mo of edram. but thats the idea : for edram bandwidth to solve rendering problems, you need a lot of edram which is too expensive and is against the logic of going with edram...

wild idea for a semi-cheap console : you put a relatively large quantity of edram such as 32MB, but limit the output resolution to 720p. well, feel free to do 1080p but the lower res is what's intended regarding the console specs.

I'm imagining an IBM soc made on 32nm, with the aforementioned edram, 64bit (or better) gddr5 : good one on a fast memory controller. a CPU from the PowerPC A2 ligneage (edram L2) and a GCN from the future Fusion APUs, about of radeon 7750 level.

four cores, 16 threads, or three and 12, with vector units, make it good enough. redundancy : a spare CPU core, spare edram, spare GPU units.

cheap flash storage, gigabit ethernet with some expectation you might use a NAS to store the games, two SATA ports (compatible with industry standard optical, HDD and SSD), the usual SD, USB, analog sound etc.

I would like the low footprint, low price and good enough rendering!
ask Valve and Steam how they manage to do DRM that doesn't piss people off too much, while half their users run everything as root under windows XP.

 

[-tkf-]

Question, and sorry if it was discussed before. In regards to the PS4, would Sony kill themselves by going with Cell again?

I see several options.

Cell as we know it, just with 8 SPU's instead of 7 or the PowerXCell 8i?

Wouldn't sony save a good chunk of money by going with the same CPU and instead using every cent and watt on a stronger GPU?

A bit like a reverse PS3, strong cpu weak gpu is turned around into a strong gpu "weak" cpu. Question is if the Cell isn't enough?

 

[Gubbi]

Sony could go with a CELL design again, but they need to beef up the general processing resources (PPU) or they are going to see a generation of multi platform games suck on PS4 with only their first party developers being able to get the most from the system.

Cheers

 

[N2O]

not sure,but i think this news is pretty hurt Sony for next gen
http://www.reuters.com/article/2012/04/10/us-sony-idUSBRE83907220120410?type=companyNews

 

[fehu]

Sony has bleeded money for the last ten years
thay know it and for this reason they are aiming for a low R&D, and to make the developers comfortable with hsa and gcn (and so fight twimtbp) I'm sure that amd will make sony a big discount

 

[McHuj]

wild idea for a semi-cheap console : you put a relatively large quantity of edram such as 32MB, but limit the output resolution to 720p. well, feel free to do 1080p but the lower res is what's intended regarding the console specs.

I'm imagining an IBM soc made on 32nm, with the aforementioned edram, 64bit (or better) gddr5 : good one on a fast memory controller. a CPU from the PowerPC A2 ligneage (edram L2) and a GCN from the future Fusion APUs, about of radeon 7750 level.

four cores, 16 threads, or three and 12, with vector units, make it good enough. redundancy : a spare CPU core, spare edram, spare GPU units.

That's been my prediction as well. IMO, it makes sense to go with an APU like solution as well.

On 32nm an 4 module bulldozer is 1.2B transistors and 315mm. The 4 core llano 1.45B transistors and 228mm. Two years after their introduction (2013), I think you could manufacture an SOC around 300mm and 2B transistors on a pretty mature 32nm process.

I'm expecting some sort of customized GPU consuming about 50-60% of the SOC with performance around a 777x and the rest of the die dedicated to the CPU and edram.

 

[Shifty Geezer]

I hadn't seen this report on Sony's managerial changes before. Of interest is this:

R&D also receives a renewed focus at Sony, as the company looks to find the next big technology. Kunimasa Suzuki has been put in charge of the new User Experience, Product Strategy, and Creative Platform division, charged with planning and designing a unified user experience and interface across Sony's entire product line. In addition, he will run the company's mobile business, which includes smart phones, tablets, and PCs.

If Sony are going to have a single unified experience across devices, then they'll want as few hardware variations to support as possible. Hence the potential choice of x86, no? ARM and x86 would be all they need for cameras, tablets, PCs and consoles.

 

[Rangers]

That's been my prediction as well. IMO, it makes sense to go with an APU like solution as well.

On 32nm an 4 module bulldozer is 1.2B transistors and 315mm. The 4 core llano 1.45B transistors and 228mm. Two years after their introduction (2013), I think you could manufacture an SOC around 300mm and 2B transistors on a pretty mature 32nm process.

I'm expecting some sort of customized GPU consuming about 50-60% of the SOC with performance around a 777x and the rest of the die dedicated to the CPU and edram.

I saw some B3Ders in one of these threads complaining about how bulldozer sucks, yet compared to this gen CPU's, and depending what MS goes with next gen, it'll be a dream comparatively. Could give Sony the easy programming edge anyway.

Not saying I believe the APU+discrete rumors, but it does kind of fit in with Sony's whacky engineering, we can say. They seem to like doing things differently.

 

[bgassassin]

Not saying I believe the APU+discrete rumors, but it does kind of fit in with Sony's whacky engineering, we can say. They seem to like doing things differently.

That would definitely be a big switch from the original target specs. Not saying it's impossible, but what IGN's source(s) told them doesn't match with what was said last year.

 

[GuestLV]

Thats my point, it was the idea of Microsoft for their xbox360 to solve the ps2 edram problem,

What are you talking about?x360 Edram and PS2 VRAM is a completely different architectures

but even this solution failed

IMO this solution shows advantages in many games,and call this ''failed'' is far from reality

 

[McHuj]

I saw some B3Ders in one of these threads complaining about how bulldozer sucks, yet compared to this gen CPU's, and depending what MS goes with next gen, it'll be a dream comparatively. Could give Sony the easy programming edge anyway..

I don't know the details of the performance of how well Bulldozer would fair as a console chip. My only point in bringing it up was more on the manufacturing side of things. It's a fairly high clocked and large chip, but readily available (likewise Llano [and Trinity soon]) so I think 32nm process is the likely candidate for a console chip next year.

When you think about a potential SOC/APU like solution for a console, I don't know how they cannot be an order of magnitude faster then current gen consoles.

The current XCGPU (without edram) is only 372M transistors with a 500MHz GPU. The rumored 6670 GPU for the next xbox is built with 716M transistors and 800MHz. Almost twice the transistors of the entire XCGPU and a higher clock.

Either we're drastically becoming more and more inefficient in designs or people really underestimate how old and dated the current consoles are. I think a console built with an SOC of about 2B transistors and very fast RAM will be easily an order of magnitude faster than the current generation. If they

 

[-tkf-]

Sony could go with a CELL design again, but they need to beef up the general processing resources (PPU) or they are going to see a generation of multi platform games suck on PS4 with only their first party developers being able to get the most from the system.

Cheers

Where is it that the Cell fails in the next generation? Even just the current Cell. With the SPU's free to do other stuff than helping a weak GPU , wouldn't it have enough power?

 

[TheWretched]

Well, to be honest, most PC games of today are console ports with no additional work done to the assets. So the only advantage PC games get is higher resolution <whatever> and not much else. So I can't fault Joe Sixpack for not seeing a difference, unless you show it to him side by side... because then (i.e. PS360 comparisons), console a or b (whatever he likes more) "clearly wins" because of "blurryness" or "framerate" or whatever.

For me, the upgrade is only marginally needed for better graphics, although I don't mind better looking games, I actually "prefer" them. It's more about the additional possibilities they might provide. I love interactive physics a lot. This generation brought a lot of nice additions to games (say GTA4 for example), but nothing earth shattering. I hope this trend continues with more processing power available.

 

[DoctorFouad]

I have a crazy idea to suggest for sony and microsoft engineers :

dont waste your time and money on R&D for your next gen hardware, let Nintendo do the dirty work of searching the most efficient/cheapest hardware solution they can find on earth,

just listen to Nintendo WiiU hardware presentation details this E3, copy/paste the architecture, overclock every chip from the 50-90 watt WiiU TDP to create a 250 watt TDP console (add the necessary cooling system, it will cost you 50$ per console), and you will get a monster next gen console, easy to program for, very cheap to produce, cheap to buy for the final consumer and you will beat the hell out of Nintendo WiiU hardware without spending a cent on R&D.

Please sony, microsoft, if by chance you read my post, just do it, please !

 

[DoctorFouad]

I don't know the details of the performance of how well Bulldozer would fair as a console chip. My only point in bringing it up was more on the manufacturing side of things. It's a fairly high clocked and large chip, but readily available (likewise Llano [and Trinity soon]) so I think 32nm process is the likely candidate for a console chip next year.

When you think about a potential SOC/APU like solution for a console, I don't know how they cannot be an order of magnitude faster then current gen consoles.

The current XCGPU (without edram) is only 372M transistors with a 500MHz GPU. The rumored 6670 GPU for the next xbox is built with 716M transistors and 800MHz. Almost twice the transistors of the entire XCGPU and a higher clock.

Either we're drastically becoming more and more inefficient in designs or people really underestimate how old and dated the current consoles are. I think a console built with an SOC of about 2B transistors and very fast RAM will be easily an order of magnitude faster than the current generation. If they

You mean this is the ps4 main CPU/GPU ?

I would like to see the faces of naughty dog programmers trying to achieve some new physics and animation tricks in this CPU compared to the super fast SPUs they would be shocked granted....

 

[liolio]

DoctorFouad I wonder if you're serious or trolling both way you're not making much sense.

 

[pjbliverpool]

You mean this is the ps4 main CPU/GPU ?

I would like to see the faces of naughty dog programmers trying to achieve some new physics and animation tricks in this CPU compared to the super fast SPUs they would be shocked granted....

I think you'll find Trinity offers vastly increased capabilities and options for developers compared with the PS3.

Certainly not what I would consider enough for PS3's successor but compared with PS3 itself there's no competition at all.

 

[DoctorFouad]

DoctorFouad I wonder if you're serious or trolling both way you're not making much sense.

it dosent matter if i am serious or joking

this is a thread about the prediction of next gen technology :

- lets suppose the WiiU ends up theoritically 2X more powerful than ps360, at the same time very cheap to produce, in a small form and very cool and silent consuming very low levels of electricity (50-90 watt TDP).

- why not microsoft or sony use the same WiiU architecture, but just adding more of the same chips and overclock everything ?

- for example : the WiiU GPU runs at 650 MHZ, and costs 30$ to produce, well just overclock it to 900 MHZ and use 3 instead of one.

If the WiiU CPU runs at 2.4 GHZ with 3 cores costing 25$ to produce well use the same CPU but with 6 cores and overclock to 3.6 GHZ and put 2 of them in the box....

with this you end up easily with the desired target of 10x more powerful next gen console than ps360, and developers would be very happy the console would be cheap to produce and consumers would be happy too.

 

[hoho]

I think you'll find Trinity offers vastly increased capabilities and options for developers compared with the PS3.

Certainly not what I would consider enough for PS3's successor but compared with PS3 itself there's no competition at all.

Compared to the whole platform, sure. But when comparing just the CPU part and ignoring that half or more of the Cell was occupied doing RSX work I'm not all that sure.

 

[pjbliverpool]

Compared to the whole platform, sure. But when comparing just the CPU part and ignoring that half or more of the Cell was occupied doing RSX work I'm not all that sure.

I don't see how Cell would be able to compete with 4 Piledriver cores running at 3.8 Ghz (rumored Trinity specs) in all but the most corner cases that have little relevance to gaming CPU workloads. Especially given that Cell wouldn't be needed to help out with graphics tasks in this theoretical PS4 where it's high throughput might actually come in handy.