No it wasn't Sony and Microsoft both tried to sell their CPU's as the greatest thing under the sun. But in reality they made very simple CPU's with very simple proccessing capabilities and then clocked them up very high. A a P4-D or an A64X2 would have ran circles around them. An Atom at 1GHz didn't have the computing power to topple a 1GHz Tbird (also known as a 2001 CPU), but is faster than Xenon or Cell compute unit. Add on Brazos is faster considerably per core even to this day (but maybe not for long) than any Atom core and that Jaguar is supposed to be considerably faster, It's easy to A.) see that it will be much much much faster than the previous gen, B.) That while it might be very far from the front line it will still be a potent chip with the 8 cores and all.
Actually these APU units are probably going to closer to the future and performance of general computing then the last gen CPU's anyways. Markets spiraling quicker and quicker to "just enough computing". OS's have less overhead. People are actually slimming down applications. CPU performance hasn't been needed in a long time and its the things that are needed for CPU performance that are holding up a lot of markets.
Without benchmarks, it's hard to tell.
However, the Xenon and Cell were large and power hungry cpus. The Xenon was about as large as an Athlon X2, and the Cell was about as large as a Core 2 Quad. Per core performance wasn't great, but it's hard to see it being worse than a low power optimized and tiny Atom core.
Early reports by developers stated that a Xenon core was about on par with a 2.4ghz Pentium 4. Maybe that was already including SMP, so two threads on a core.
Still, Jaguar is a low power, low die size core. It's per mhz performance is certainly higher, but is it >2x as high? I doubt it has the same floating point performance of a xenon core either, although it does have more cores plus gpu compute to help.