Assuming that 100 MHz would be achievable would be a mistake, but assuming that 10 MHz, when a classic 12 MHz-clocked part operates at 1 MHz, is achievable might not be so much of a stretch.

Operating a typical 805x 12-clocker at 24 MHz, which is not uncommon, on a 2-layer board, should be quite straightforward for a properly prepared hobbyist/designer who's done his homework.

The same task using any number of layers but done by a fellow with a PhD but who's too lazy to read the fine print and completely unversed in PCB layout strategies and supply bypass techniques, is doomed. It may "sort-of" work, like Mike's circuit.

I agree that it's unwise to encourage people to try the extremely difficult before they're thoroughly experienced in the more common. It's long been know to be possible to succeed, though, with proper preparation and diligence.

RE