I doubt they would've got that sort of performance from a processor.
The reason that CPU manufacturers have moved from a traditional single core to multiple core solutions is because with their current development techniques they'd hit a plateu in terms of what they could put in to the chip.
Good example is the research paper that AMD released about 2years ago, where they took their then "new" Athlon64 processor and tried to over-clock it to the maximum possible.
While it could reach 8GHz (bear in mind this is a 2.5GHz variant) that was the maximum speed they could reach without instability.
As the stability of the processor doesn't just depend on the heat produced from the over-clocking but the powerwaste in general.
Single Cores, hit their limits because the more power they put in; the more power got wasted (which is what produces the heat) so there is a limit to how much power you can phsyically put into them before you reach that limit.
Easiest way to imagine what I'm talking about (provided you all have a basic phsyics understanding) is when you try to reach the speed of light. The faster you go, the more force is being put against you until you reach the speed of light when the force you put in will always be equaled by the reaction force. Thus making it currently impossible with our level of technology to break that speed barrier. It's the same with processors. The manufacturers' know full well that it's the way they're being designed that causes this limitation, but until they can figure out how to compensate or break this limitation then they've decided to go with adding more cores and get extra performance from more devices in a smaller space.
The Cell and Intel C80 honestly aren't really the same as the multi-core processors hitting the market right now. They're actually ridiculously specialised making them useless for general computer uses; but if programmed for correctly they can be quite useful for specialised applications.
Before Cell was even announced ATI (AMD) had already created a Unified Shader Core, which is almost identical to Cell but with between 6-12 Shader Cores. Physical data wise, working with Floating-Points which is what they seem to prefer to boast about now when talking about performance the performance from GPUs has always been far beyond that of CPUs.
Good example is PS3 is claimed to have 2TFlop performance.
1.8 of that comes from the RSX GPU.. which ment that the CPU was actually slightly more powerful than the Xbox CPU. Despite the claims comming from IBM and Sony.
Just hope Intel have thought ahead unlike IBM. Because the Cell is a nightmare to get any sort of reasonable performance from; It's not like multi-core processors where you can either set-up threads to do different tasks but hope that the control chip splits up data correctly to be processed.
There's also no fall-back, meaning everything is converted to floating-point rather than having a seperate integer and operator access through the CPU. It's all very poorly done and messy.
