PS4 to use Cell, NOT PS3?

[-tkf-]

PS2 Laser

For those that have problems with reading games on the PS2 try this:

Tweak3D article
http://www.tweak3d.net/articles/ps2repair/2.shtml

Auto Laser Alignment
http://au.playstation.com/support/technical/ps2/ps2autodiagnosis.jhtml

 

[BenSkywalker]

With this talk of a 350mm^2 chip I have to ask, how many billions of dollars should Sony be willing to lose up front? If everything goes perfectly for them it could easily be in the $5Billion range for the first year(depending on how they mark it off, could start writing it down now as PS2 related costs but I digress). For year two is they were to come close to hitting the sales they did with the PS2 they will be out another couple of billion(yields will be horrible at that size no matter the build process not to mention the heat generated would be staggering, could top 100Watts fairly easily). Yields are not linear, chips in excess of 300mm^2 have always had massive yield issues not only at ramp up, but throughout production.

Sony doesn't have MS's money, nor are they turning the kind of profits to warrant that level of initial investment either. I think looking at a 250mm^2 chip is far more reasonable. It's one thing to run in to major manufacturing issues when you are the only game in town, if there are two competitive offerings with a like amount of titles on the shelf it could turn out to be a very bad situation for them to repeat their last launch. Compound that with the insane level of cash a 350mm^2 class chip would take to produce in quantity and it just doesn't sound like a viable solution.

 

[Gubbi]

Yields are not linear, chips in excess of 300mm^2 have always had massive yield issues not only at ramp up, but throughout production.

Yields are worse than exponentiel with size. A chip twice the size will have twice the amount failures, at the same time you only have half as many dies per wafer, so the number of good dies can drop dramatically.

However, I'm sure the eDRAM arrays will have built in redundancy to ensure very high yield (99%). Yield is then only dependent on the size of the logic part of the chip.

We're in violent agreement on power consumption.

Cheers
Gubbi

 

[MfA]

I dont see how you could expect power consumption not to go through the roof at ~3 GHz. With the same area but ~10 times the clock frequency and same average switching activity per clock of the EE even SOI wont save you (if I remember my rules of thumb correctly).

 

[Panajev2001a]

I was just wondering about a maximum ( and probably exagerated size )... still the chip might end up being bigger than 250 mm^2 as the 250 nm GS used in launch PlayStation 2s in Japan was 279 mm^2 ( its layout was not very optimized as seen in the micro-graphs of successive die-shrinks so I have to give that to you... ) ...

45 nm is right around the corner... listening to the Sony and Toshiba guys and this means that they are setting up the new fabs so that a smooth upgrade to 45 nm can happen as quickly as possible.

A die shrink from 65 nm to 45 nm should also bring the chip in much more reasonable range ( they know the chip is NOT going to be small... the reason using 300 mm wafers could also be this ) and I don't think Sony would stop the progress there and they would be researching with Toshiba even more advanced manufacturing processes... to shrink the die even further.


Gubbi makes a valid point about the e-DRAM and having yelds more dependent on the logic portion of the die.

Ben, I am glad you are coming into this thread... ( btw, with a firend, looking at H&P, she saw ~16% of yelds using a 300 mm wafer and 350 mm^2 chips... so yes that is not too high, but I think they would improove quicker than what many thinks ) 350 mm^2 is too big ? Well let's try to guess-estimate the actual size based on the patent specs for the Broadband Engine and the constraints on the process used ( gubbi piosted them from the Sony-Toshiba PR )...

I think Sony is indeed writing PlayStation 3's R&D costs in their actual budget and that might be pushing their shares down "a little", but at least they won't go DEEPLY in red the year when Playstation 3 ships in Japan and North America.

 

[jvd]

I think sony knows the cell chip will be very expensive. But they prob feel this is the only chip to give them an edge over xbox. Thus taking away the only real advantage ms had this gen. Of course that may fail if ms is able to launch a more powerfull console at about the same cost .

 

[V3]

With this talk of a 350mm^2 chip I have to ask, how many billions of dollars should Sony be willing to lose up front?

Its just talk, its like the worst case scenario.

But the chip design is modular, so they only need to get one Processor Element to be functional, and that chip wouldn't be wasted.

 

[pcostabel]

I think sony knows the cell chip will be very expensive. But they prob feel this is the only chip to give them an edge over xbox. Thus taking away the only real advantage ms had this gen. Of course that may fail if ms is able to launch a more powerfull console at about the same cost .

I doubt Sony strategy is dictated by MicroSoft. If anything, this generation (and the last one for that matter) proved that performances are irrelevant to the success of a console, with the most powerful (Xbox) in fourth place behind PS2 DC and GC.

The Cell strategy was designed well before MS entered the industry and goes well beyond providing better performances for PS3.

 

[jvd]

I think sony knows the cell chip will be very expensive. But they prob feel this is the only chip to give them an edge over xbox. Thus taking away the only real advantage ms had this gen. Of course that may fail if ms is able to launch a more powerfull console at about the same cost .

I doubt Sony strategy is dictated by MicroSoft. If anything, this generation (and the last one for that matter) proved that performances are irrelevant to the success of a console, with the most powerful (Xbox) in fourth place behind PS2 DC and GC.

The Cell strategy was designed well before MS entered the industry and goes well beyond providing better performances for PS3.

U sure about that ? I'm pretty sure cell came after the xbox .

 

[pcostabel]

I think sony knows the cell chip will be very expensive. But they prob feel this is the only chip to give them an edge over xbox. Thus taking away the only real advantage ms had this gen. Of course that may fail if ms is able to launch a more powerfull console at about the same cost .

I doubt Sony strategy is dictated by MicroSoft. If anything, this generation (and the last one for that matter) proved that performances are irrelevant to the success of a console, with the most powerful (Xbox) in fourth place behind PS2 DC and GC.

The Cell strategy was designed well before MS entered the industry and goes well beyond providing better performances for PS3.

U sure about that ? I'm pretty sure cell came after the xbox .

They might have announced it after, but I'm sure the big plan was in place well before.

 

[nonamer]

I think sony knows the cell chip will be very expensive. But they prob feel this is the only chip to give them an edge over xbox. Thus taking away the only real advantage ms had this gen. Of course that may fail if ms is able to launch a more powerfull console at about the same cost .

I doubt Sony strategy is dictated by MicroSoft. If anything, this generation (and the last one for that matter) proved that performances are irrelevant to the success of a console, with the most powerful (Xbox) in fourth place behind PS2 DC and GC.

How is DC ahead of the Xbox? It's a dead console. Also, the GC might ahead of the Xbox (depends on who you ask), but it certain won't hold that title very long if sales continue as they are.

 

[Saem]

I thought Sony made a statement saying the XBox made them pause. They've got some really stiff competition.

 

[Paul]

They did, a long time ago.

I guess their fears were proven wrong though as xbox doesn't come near as to threaten ps2 sales wise.

 

[Panajev2001a]

Their fears were proven right as they have seen Microsoft is able to dump Billion of Dollars into the Xbox group without worrying , for now at least, about the losses...

That is also why they have been so aggressive in their PlayStation 3 strategy so far... they want to make sure the conditions are right when they launch PlayStation 3 as the battle, they know it, will be tougher.

 

[Saem]

Well if MS was really smart, they'd develop the engine technologies they have now, give out engine lisences and have companies put out content for said engines. You could get quite a few games built on top of the Halo 1 engine which could be used to make tactical shooter and so on.

Just a matter of finding content creators who aren't out to make a some super title, just something that can be built on top of exisiting technology.

 

[pcostabel]

How is DC ahead of the Xbox? It's a dead console. Also, the GC might ahead of the Xbox (depends on who you ask), but it certain won't hold that title very long if sales continue as they are.

DC has sold ~10Mln units in the same timeframe in wich Xbox sold ~8Mln,
so I'd say DC is in 2nd place at this point, even if it's dead.
If you only want to count consoles still in production, then the Xbox is still in fourth place:

PS2
PSOne
GCN
XBOX

 

[Almasy]

Well if MS was really smart, they'd develop the engine technologies they have now, give out engine lisences and have companies put out content for said engines. You could get quite a few games built on top of the Halo 1 engine which could be used to make tactical shooter and so on.

I think that part of why that doesn´t happen is that companies want their games to differentiate themselves from others. It´s a matter of public perception, if they see the effects of Halo in another game, or a similar scale of battles, then the public will not think that company is "that" special, or there´s several similar companies that can produce what they can.

 

[nonamer]

How is DC ahead of the Xbox? It's a dead console. Also, the GC might ahead of the Xbox (depends on who you ask), but it certain won't hold that title very long if sales continue as they are.

DC has sold ~10Mln units in the same timeframe in wich Xbox sold ~8Mln,
so I'd say DC is in 2nd place at this point, even if it's dead.
If you only want to count consoles still in production, then the Xbox is still in fourth place:

PS2
PSOne
GCN
XBOX

So effectively it's second? A distant second but still second.

 

[V3]

U sure about that ? I'm pretty sure cell came after the xbox.

Sony worried with the launch of Xbox and GC being later than PS2, they are afraid, that people might dump PS2 like they did with Dreamcast, when PS2 came out.

But it doesn't seems to be so.

Sony pretty much started work for PS3, after they finished EE.

 

[Dural]

DC has sold ~10Mln units in the same timeframe in wich Xbox sold ~8Mln,
so I'd say DC is in 2nd place at this point, even if it's dead.

Um, no. DC sold 8.2 million units at the end of the fiscal year 2001. At that time it had been released for two and a half years in Japan and one year and nine months in the US. In the same timeframe as Xbox(one and a half years) DC would have sold less, not more. If you are talking total sales of DC since it's release in 1998, then yes it has sold more, but it has been out for four and a half years compared to Xbox being out for a year and a half.