The cell was a bust it cost sony millons and it was a pain to code for.
I hope sony dont use the cell in the ps4 and also the cell could not do graphics
Good thatswhy they used ninvida awesome gpu

Speaking of Larrabee, here's what came out of that idea at the end - Intel Xeon Phi:

http://www.tomshardware.com/reviews/xeon-phi-larrabee-stampede-hpc,3342.html

Not bad when compared to new offerings from nVidia and AMD:

http://translate.googleusercontent.com/translate_c?depth=1&hl=en&rurl=translate.google.com&sl=de&tl=en&u=http://www.pcgameshardware.de/Xeon-Phi-Hardware-256199/News/Intel-Larrabee-Xeon-Phi-5110P-Xeon-Phi-3100-Serie-1035298/&usg=ALkJrhj2GcbFRuYHUEI-bNMqSkB_7rSxsw

The main compute is effectively the same and the design follows the same paradigms so it can be considered the successor I guess. The main parts of larrabee like the memory architecture and cache are still intact as well as the many in-order X86 cores.

All I said was that they stripped out the GPU components, other than that it is the Larrabee architecture too a fault. But Larrabee as the Intel GPU project was still scraped as it was not competitive with AMD/Nvidia GPUs for real-time rendering, which was the point. 

Despite Larrabee having Intel's R&D muscle behind it, it still offered only a fraction of the game performance of a high end GPU from AMD/Nvidia. And such it was never used as a GPU outside of R&D labs. For much the same reasons that the crazy plan of using multiple CELL processors instead of using a dedicated GPU in the PS3 was scrapped. A dedicated GPU is still just much more effeciant at doing a GPU's job than a heterogeneous hybrid CPU is, even with some dedicated graphics components like Larrabee had.

Not to say that a future where the flexibility of a heterogeneous processor that uses hundreds of simple compute units for all computing tasks will never be viable, the industry does seem to be going that way. But so far it looks like that will be more GPU architecture bassed than CPU. We are a long way away from that point IMO tho.

I was really sad when that happened - I remember that Project Offset game they bought (along with its developer Offset Software) - not only it looked superb, but gameplay itself was looking really promising. The game was to show power of Larrabee, and when it got canceled, it was lost probably for all times in dark dungeons of Intel's empire.

IBM released the PowerXCell 8i in 2008. The one and only upgrade (aside from die shrinks) to the Cell Broadband Engine.

The Cell is simply too difficult to work with, isn't well suited for general purpose use and has already been greatly surpassed in the fields it was designed for.

Don't forget that the Xbox 360 CPU, Xenon, is actually based on the Cell itself.

To really nail the coffin down, the facility that Sony obtained from Toshiba that developed the Cell has largely been converted to a camera CMOS manufacturing facility.

Project Offset was an interesting project in many ways, the work they were doing with the custom Larrabee APIs was very much a fore runner for modern DX11 compute shaders, and were developing some great rendering tech. Some of the devs that worked on the project went on to make HAWKEN which explains why that game looks as good as it does lol. The Visuals still hold up today

1. You need dev kits to make next-gen games. Would be pretty stupid to launch a console with no games.
2. No, they can't. Well...technically they can but its not something they would do on purpose. It would hurt the Xbox brand and cost them loads of money.

Yeah, really shame game didn't live to see release, I really liked gameplay from those few videos they posted. Still, there is hope that one day Intel might decide to give it another shot at high-end GPU market (remember all those rumours about Larrabee and next-gen consoles back in the day?), so perhaps something will come out of that game after all.

...... are you guys serious? If anything Sony should reuse Cell Architecture.

1. The technology has already been broken down.. little to no learning curve for existing developers.
2. Companies that built Engines for PS3's Cell can continue to use that Engine with little to no modifications..
3. Cheap Manufacturing cost since they've been pumping those out for 8 years.
4. Cell was originally designed to be scalable, adding an SPU with more SPEs should be a simple redesign task to boost its processing power to match next generation.

Anything outside of GPU design would result a less efficient processing unit.