There's a simple rule of thumb that can be used. SMP is great for brute force computation, though it more or less runs out of steam at eight-way unless multithreading and task parallelism gets better. Not many embedded tasks are brute-force symmetric. But some MPU vendors will sell symmetric multicore the same way they'll sell Ferraris - if someone is silly enough to buy a 16-way i7 because it seems to be macho, well, that's their error. Better to look for a Freescale, Cavium, etc.
Seems like this is just one more example pointing up the growing requirement for engineers to shore up their interdisciplnary skills. As Loring well points out, designers evaluating multicore processors for embedded applications are going to need to understand a whole lot more about software.
To me, the line is getting very blurred between so-called "embedded" devices and microprocessors. At the end of the day, I guess it's the application which defines whether the part supplying the computing cycles is embedded or not, but it sure seems to be that vendors are very happy to upsell (maybe that should be upCELL, or upcycle) engineers, when lower cost, lower power consuming processors would work just as well. The apps you cite, such as routers, clearly benefit from these beefier embedded multicore parts. And robotics apps do too. But somewhere along the apps continuum, the needs of the vendors (to sell higher ASP processors) and those of engineers (staying within cost and power budgets) start to diverge.
Transfers the control of a large number of motion axes from one numerical control kernel to another within a CNC system, using multiple NCKs, and enables implement control schemes for virtually any type of machine tool.
Focus on Fundamentals consists of 45-minute on-line classes that cover a host of technologies. You learn without leaving the comfort of your desk. All classes are taught by subject-matter experts and all are archived. So if you can't attend live, attend at your convenience.