PS4 to be based on Cell?

[mrcorbo]

I fail to see the logic in that. Ive argued the opposite of this for a while now. IMO, The fact it has blue ray and other great components should future proof the console over the 360. They have all the power and features to make good of the PS3. The only thing holding them back is software and machine price (Which a new console launch wont fix).

There's an opportunity, though, to fix a couple of architectural deficiencies the PS3 has that were forced by launch timing and pricing targets and still come out with a cheaper product in the end.

I could see a PS4 with:

2X the RAM
A higher-speed BR drive
Flash Memory replacing the Hard Drive (with HD add-on available)
More advanced Nvidia GPU

Even without significantly improving Cell (say just having a higher clockspeed), this would be a noticeable upgrade over the existing PS3 and could target a $299 launch price. Full BC with PS3 should also be easily achievable and sustainable.

It could be a smart move, actually, for Sony. It only sucks if you were a consumer who were counting on the "10 year" thing as a justification for purchasing a PS3.

 

[dobwal]

Interesting: http://pc.watch.impress.co.jp/docs/2008/0929/kaigai469.htm

"Goto Hiroshige has received PSX4 spec info from Japanese developer sources who received preliminary spec from SCEI for developer feedback(whom Goto refuses to name due to NDA).

- SCEI has sent rough PSX4 spec to 3rd party developers for feed back. Based on the description, PSX4 is a Wii-tized PSX3, no more than 2X as powerful.
- SCEI wants to beat Xbox 3 to launch. 2011 is the deadline, or sooner.
- SCEI pulled all its engineers from IBM Texas, and there is no new CELL architecture being developed. PSX4 will use same CELL architecture with improvements.
- SCEI will dump XDR and use standard JEDEC memory type(GDDR3/DDR3) for main memory. "

Stolen from Neogaf.

The spec info seems pretty lacking. You would think the preliminary specs would include some information that would allow for more robust feedback.

What kind of feedback to you get from "2X as powerful", "released 2011 or sooner to beat the xbox3", "same Cell" and "GDDR?" (GDDR 3 seems alittle old for 2011). Proposed feature set of the GPU would be more relevant in terms of guaging devs wants and needs, but there is no mention of the GPU.

This seems like material for marketing analysts not preliminary specs for devs as it revolves more around cost saving then its does tech features.

 

[Butta]

There's an opportunity, though, to fix a couple of architectural deficiencies the PS3 has that were forced by launch timing and pricing targets and still come out with a cheaper product in the end.

I could see a PS4 with:

2X the RAM
A higher-speed BR drive
Flash Memory replacing the Hard Drive (with HD add-on available)
More advanced Nvidia GPU

Even without significantly improving Cell (say just having a higher clockspeed), this would be a noticeable upgrade over the existing PS3 and could target a $299 launch price. Full BC with PS3 should also be easily achievable and sustainable.

It could be a smart move, actually, for Sony. It only sucks if you were a consumer who were counting on the "10 year" thing as a justification for purchasing a PS3.

I would hope for more than that 6 years after the release of the PS3... considering that console generations usually provide 8 x the power, RAM, etc. I would expect at least the following and think it could still be reasonably priced:

2 x the Cell performance
4 x the RSX performance
4 x the overall RAM
4 x the available bandwidth
BR operating at 8x speeds
flash drive with 128GB and optional HD add-on (internal)

This would provide about 1/2 the performance upgrade that we normally see from one console gen to the other, but would also provide far better performance increase than what happened with the Wii.

 

[RudeCurve]

I could live with a cheaper PS4 that's 2X as powerful as PS3.

 

[Carl B]

With how vauge the rumor is it is hard to say... if they change the memory controller, what else? How will they approach just shy of 2x performance without significantly ramping up the frequency? Seems like adding some more trannies makes sense, in which case why not the 1 TFLOPs Cell (2 PPE / 32 SPE) on 45nm IBM has slotted for 2010? That should hit 32nm by 2012 you would think, giving Sony some price mobility quite quickly and a chip like that would be, what, roughly half the size of the 90nm PS3 Cell in area (90nm vs. 32nm)? Time will tell, but as an enthusiest I hope we see at least 1 console take technology pretty seriously and I think this gen shows there is a market for early adopters who buy a lot of games and who want pretty graphics

Frequency/core-count increases would be a given though in a new console iteration, just because - at least the frequency increase - would be so easy. The memory controller seems a very important architectural change to make to the fundamental layout of Cell, given how core to the present architecture of the chip the Rambus I/O is. So, I was just surprised that a move like that would take place without IBM at all, since I just wouldn't see why IBM would be distanced from that shift.

But, One's improved translation offers a much clearer picture; after all having few engineers in Austin is very very different than having no engineers. Beyond that, most of the ideas Goto floats are things we've been discussing here for a couple of months anyway. New threads/posts just get folk that had been absent those other discussions newly involved/worked up.

 

[kyetech]

I could live with a cheaper PS4 that's 2X as powerful as PS3.

are you a ps3 owner?

 

[dobwal]

I agree that this is pure hyperbole speculation, it seems that article writers aka bloggers tend to use certain events during slow news months to generate hits for their web sites.

That said, it is fair to possibly have a negative outlook on PS3 with a PS4 being something that could happen sooner than expected because of recent events in Japan with XBox 360 hosting almost all of the "current gen" RPGs and its starting to reflect in Japan sales.

Sony seems to unlike Nintendo, made a console for the hardcore gamer and the hardcore game developer in the PS3, more so than Microsoft did with the Xbox 360 because when you boil things down, it won't take a colorful glowing article about Xbox 360's capabilities to know how limited it is by its API and how PS3 will always be a much more efficient use of game taking advantage of console hardware.

However it seems the hardcore gamer that the PS3 was targeting, the same "so called hardcore gamers" who were paying over $1000+ for PS2s in 2000 and well over that ammount for Xbox 360s in 2005 have basically turned their backs on Sony with their slow adoption and price complaints or just dropping Sony in favor of Microsoft and then there is the Nintendo casual factor just making a mockery of high technology and progress.

So basically graphics whores, tech spec geeks, and progressive futurists for consoles better not get all angry and mad and puffy faced once Sony releases the PS4Wii with casual games at 99%

One thing that seems left out though is that there was a recent article talking about IBM jumping Intel over chip engineering fabricating process with a hint of Cell BE being that chip that IBM wants to make into 22nm so it does lend credibility that one day a multi core Cell BE would be made for...

The challenge of Intel Larrabee and their intentions of offering it to Microsoft as another rumor that got on the interwebs and makes sense since Intel want a way to spread their chips everywhere.

Sony did not developed the PS3 for hardcore users. Sony simply misjudged its and its partner's manufacturing prowess and end up with tech initiallyfar too expensive for a game console. The current production cost of the PS3 today is probably in line with what Sony was expecting back in 2001-2003 to be the production costs for its consoles at launch.

 

[Mobius1aic]

With how vauge the rumor is it is hard to say... if they change the memory controller, what else? How will they approach just shy of 2x performance without significantly ramping up the frequency? Seems like adding some more trannies makes sense, in which case why not the 1 TFLOPs Cell (2 PPE / 32 SPE) on 45nm IBM has slotted for 2010? That should hit 32nm by 2012 you would think, giving Sony some price mobility quite quickly and a chip like that would be, what, roughly half the size of the 90nm PS3 Cell in area (90nm vs. 32nm)? Time will tell, but as an enthusiest I hope we see at least 1 console take technology pretty seriously and I think this gen shows there is a market for early adopters who buy a lot of games and who want pretty graphics

The expanded Cell theory (2 PPEs/32 SPEs) is my speculation. Although I'm going to expect 4 PPEs for added general purpose capabilities based on the latest Power Architecture by that time (Power 8?). The SPEs themselves I'm sure will be reworked a bit to be even more powerful per core as well, and I'll expect it to be 32 nm from the get go. The next problem is a GPU to go with it, be it goddamn Larrabee or another Nvidia product, it needs to be good, efficient (current Nvidia GPUs are power hogs) and something Sony can be proud of. Honestly Sony should look into DAMMIT considering ATi designed a great GPU (plus subsystem) for the 360 a year ahead of Sony's release.

Problem is just speculating what MS does this time around, as well as Nintendo. Honestly I think Nintendo can't release another severely underpowered system. Consumer expectations rise especially if they want people to drop two to four hundred dollars on a new game system.

 

[joker454]

It could be a smart move, actually, for Sony. It only sucks if you were a consumer who were counting on the "10 year" thing as a justification for purchasing a PS3.

It's not that bad for existing customers. A Cell based PS4 with beefier gpu would make backwards compatibility 'relatively' easy, so their existing games would still work. They can keep bluetooth on PS4, so existing controllers and peripherals would still work. The only thing that couldn't be reused when transitioning from PS3->PS4 is the main unit itself. However, the main unit is an excellent movie player and a pretty good media center. Just add the blu-ray remote control and move the PS3 to another room and it lives on perfectly fine doing other duties.

This would provide about 1/2 the performance upgrade that we normally see from one console gen to the other, but would also provide far better performance increase than what happened with the Wii.

What you describe would make for a very nice console upgrade at a reasonable cost. I'd buy that at launch for $299 Although I suspect they will get a bit more than just a 2x cpu performance jump...but we'll see.

 

[dobwal]

The expanded Cell theory (2 PPEs/32 SPEs) is my speculation. Although I'm going to expect 4 PPEs for added general purpose capabilities based on the latest Power Architecture by that time (Power 8?). The SPEs themselves I'm sure will be reworked a bit to be even more powerful per core as well, and I'll expect it to be 32 nm from the get go. The next problem is a GPU to go with it, be it goddamn Larrabee or another Nvidia product, it needs to be good, efficient (current Nvidia GPUs are power hogs) and something Sony can be proud of. Honestly Sony should look into DAMMIT considering ATi designed a great GPU (plus subsystem) for the 360 a year ahead of Sony's release.

Problem is just speculating what MS does this time around, as well as Nintendo. Honestly I think Nintendo can't release another severely underpowered system. Consumer expectations rise especially if they want people to drop two to four hundred dollars on a new game system.

They could probably go with the same Cell design, up the frequency and simply pair it with a more robust GPU. Thereby alleviating alot of the work the Cell is used for now to make up for some of the short comings of the RSX. Ramping up the speed of Cell should be no problem as now Cell is a pretty mature design and will be on a smaller process.

 

[Carl B]

The next problem is a GPU to go with it, be it goddamn Larrabee or another Nvidia product, it needs to be good, efficient (current Nvidia GPUs are power hogs) and something Sony can be proud of. Honestly Sony should look into DAMMIT considering ATi designed a great GPU (plus subsystem) for the 360 a year ahead of Sony's release.

Why the rage against Larrabee? Anyway it would definitely *not* be that chip in this scenario, for being a more acute reason of why nAo laments the combination to a 2011 GPU; they are treading on each others toes if extrapolated along their present trajectories, which means API hassles and/or wasted transistors.

To do it right, Cell should be slightly evolved and have the computational power scaled via clock and core, and the GPU should be designed to complement it; in 2011 that would be a very custom job, and is NVidia willing to take the time/effort to do that? It's not been their pattern, and unless it were to change, well... the result would be inelegant.

Problem is just speculating what MS does this time around, as well as Nintendo. Honestly I think Nintendo can't release another severely underpowered system. Consumer expectations rise especially if they want people to drop two to four hundred dollars on a new game system.

I think if you ask many Wii owners whether it is underpowered or not, they would say 'no.' That view is more held by those that resent the system than those that actually own/enjoy it... and that later group is the majority, for better or worse.

 

[wco81]

Are flash drives that much faster than regular drives?

Worth the higher incremental costs? They're great in iPods but in computers or consoles?

 

[Neb]

If PS3.5 isn't far enough ahead of PS3 in specs, they won't sell to the hardcore. If it doesn't have a paradigm-shift controller it won't compete in unique titles. Basically, there's nothing they could do with a PS3.5 in 2011 and they couldn't do with PS3 except a lot cheaper and at lower fidelity.

Basically, what would be the USP of PS4? 'It's PS3 with better graphics'?

There is much more money in the non-hardcore sector, look at Wii.

 

[nAo]

Why would they want to up CELL frequency? I don't see any reasons to do it.

 

[Carl B]

Are flash drives that much faster than regular drives?

Worth the higher incremental costs? They're great in iPods but in computers or consoles?

You're looking at it backwards - solid state drives are attractive because of their expense. Yes right now they are pricey, but they scale in cost with transistor density improvements, whereas mechanical drives will always have the pricing floor associated with their mechanical components to deal with. Whether substantial drive densities will be considered consumer attractive by 2011 (in the context of storage needs) I personally am not sure of, but wen people suggest a solid state option for future consoles, they're doing so from a cost competitiveness position.

 

[dobwal]

Why would they want to up CELL frequency? I don't see any reasons to do it.

Aren't there plenty of reasons to up the speed of the processor, if all possible without reducing stability and reliability, as its the cheapest way to get more performance out of Cell.

 

[Butta]

What you describe would make for a very nice console upgrade at a reasonable cost. I'd buy that at launch for $299 Although I suspect they will get a bit more than just a 2x cpu performance jump...but we'll see.

Imagine how good of a Linux box this would be

 

[Carl B]

Why would they want to up CELL frequency? I don't see any reasons to do it.

My view on it is that if you're in an environment where now the Cell and GPU would be able to utilize a common RAM pool (I guess depending on what actually is the deal there with memory controller shifts), you might also begin to see the chip play more directly into the graphics sub-system. With Larrabee coming from one side and GPUs from the other to encroach on each others space, Cell (in line with some original aspirations) could become more of the GPU as well, with the actual GPU hopefully serving a complementary role in terms of hardware acceleration rather than reflecting the general trend towards increased functionality/programmability that GPUs are arching towards. In that case increased frequency should be of some use for the raw compute advantages... and hey it's a free upgrade anyway.

 

[kyetech]

I would lmao if Nintendo released another repackaged Gamecube and sold another 100 million units @ $250. It would surely be the greatest con of any company of all time.

 

[RudeCurve]

are you a ps3 owner?

I will be when KZ2 comes out.