Yes the cell can do some graphical effects, but it is not powerfull enough to handle an entire graphics pipeline.
Each SPE is say capable of about 25 Gflops, we chuck in 6 Cell CPU's with 8 SPE's each, that is still only 1.2 Teraflops.
Currently a Geforce GTX 570 is capable of 1405 GFLOPS. (Both running single precision calculations.)
So not only would having half a dozen Cell CPU's be more expensive (As it's not a single chip solution), but they would consume more power and will result in less performance.
Also when you have multiple chips connected but not part of the same chip, you need to factor in latency between the each chip, this will reduce the total throughput capability.

Now why did Sony design the cell? Simple.
For everything.
And it's not just Sony that designed it, IBM and Toshiba helped as well.
The idea was for use in servers, Televisions, Blu-ray players, consoles, everything! However that really has not come to fruition in full force as of yet.

And for the record, number of posts doesn't equate to me not being in the "know", that is just ignorant on your behalf.

It's just typical 'patent everything you think of whilst on the toilet incase it actually was brilliant' behaviour.

It wont work for exactly the reason the Wii HD would never have worked if it existed .

Game designers have the choice of;

Make no changes to code = no improvement

Make the game use more ram = severely reduces customer base

Make the game use more ram bundle the addon with your game = expensive

Program 2 games that have the same plot, gameplay etc but different graphics and put them on the same disc/on 2 seperate discs = expensive

Run the same game code but with minor tweeks, extra ram means you can increase draw distance if the person has the addon = a slight improvement but not really enough to justify the investment, either for Sony or for the customer.

We shall see the PS4 quite soon, and this 3.5 shit is a fantasy

First off, I think some of you are freaking out and thinking that just because they have filed a patent that means they are working on production units.  A patent just saves their IDEA.  This is probably no more than a concept at this point.

Secondly, I do not see how it could work as an add-on to current systems.  There just isn't an input that is capable of handling the necessary data transfer rates on current PS3's.  Not to mention they'd need to make it compatible with both the Fat and Slim models which means different adapters or different casing styles.

The only way I see this working is if they used the RJ45 port (LAN port) or multiple USB ports in conjunction and even then the latency of the data transfers back and forth to reach the multiple processors may hinder it more than help it.

An external add-on to increase power, I'm afraid, would do much more harm to their user base than good.

Do you know any impressive Pc game? Sometimes I'm drolling over Dx11 tech demos, but I'm yet to find someone something that feels truly next gen.

Alright folks, listen up (folks = Sony).

Step 1 - You invest in this add-on and have a set release date for it for either Christmas 2011 or March/September 2012.

Step 2 - Do not drop the PS3's price, keep it at $299 and when the time comes to release this new add-on you should sell the add-on cheaper than it is to produce it. There are two reasons for doing this. First reason is whatever you are losing  from this add-on you will be making much more back from the PS3 itself since you haven't dropped the price yet. The PS3 will seem, in comparison to other consoles, like a new system and if marketed right could make alot of people go "Wow look at those graphics, look at the size of this game". Second reason is it will help software sales since the better multiplatform game would presumably be on PS3.

Step 3 - Come Christmas 2012 or March 2013, drop the PS3 price down to $199 but keep the add-on the same price and by this time you should be making money on both and have the most killer stream of games gamers can ask for.

Seriously, which version of CoD would you guys play? 360/Wii version or the version that is 50% better graphics, smoother gameplay, higher framerate. Hey...WHAT THE FUCK!!!! Sony isn't a PC company!!! DAMMMITTTT!!

wow i fuckn called it

http://gamrconnect.vgchartz.com/thread.php?id=49440

just wait til this whole prediction comes true

what current gen HD consoles need is only one simple thing, memory, give more of that, and everything will change.

oh goodness sir. If I was implying that you weren't in the "know", then I apologize beforehand. But really though, you really don't have to make such a tremendous effort in your argument. Its not like we're betting money here. lol 

However, please allow some personality to come through your words. We welcome online personalities with humor all the time :) 

As for the GPU, yes GPU technologies have improved to the point of it being better than say gaming console processors released back in 2006/7. But a Geforce GTX 570 is a little expensive, is it not? Doesn't it cost more than a PS3? I wouldn't be surprised if it did better than the Cell for obvious reasons. 

Yes it is, but costs will only come down thank's to Moores law.
Secondly, it can't be any more expensive than say... Strapping half a dozen Cell processors together and calling it a day.
Thirdly Sony would be buying those chips in bulk, without the PCB, Cooling systems, regulators and such, so prices would be much lower with that factored into it.
Fourthly... 28nm fabrication processes will be upon us in 2011, so the cost of making that chip should be only a fraction of what it is now. (As it's still a fairly large chip, yields will affect costs as well, smaller fabrication processes means more working chips produced per wafer.)

his post pointed out that it was better than 6 Cells in raw compute performance he forgot to mention that as it's a GPU designed specifically to do rendering that in terms of rendering speed the GPU would be an order of magnitude better than CELLs at graphics than raw compute power suggests.