We hear a lot of talk these days about three-dimensional integrated circuits (3D ICs), but there's still a lot of confusion in this area, so who better than yours truly to rip the veils asunder? To set the scene, let's consider the way of the world before 3D ICs.
Lots of small, individually packaged dice
Not so long ago, different functions like high-performance logic, lower-performance logic, memory, and analog/RF were presented as discrete dice in their own chip packages, as illustrated below. (Dice is the accepted plural form of die in the semiconductor industry.)
A bird's-eye view of a circuit board with individually packaged chips.
One advantage of this scenario was that different chip companies could concentrate on creating devices that fell into their realm of expertise. Another benefit was that each die could be implemented at the most appropriate technology node. The high-performance digital logic chip could be created at the latest and greatest (and more expensive) technology node. The lower-performance digital logic device could be created using an earlier (and more affordable) technology node. Yet another advantage was that the yield of smaller silicon chips was significantly higher than that for larger dice.
On the downside, the resulting circuit board was relatively large and heavy and consumed a lot of power. Also, every soldered joint on a circuit board was a potential point of failure. Furthermore, it took a relatively long time for signals to propagate across a circuit board from one chip package to another. This hurt the performance of the system as a whole.