Jim Hogan and Paul McLellan, two EDA entrepreneurs wandering the forests of today’s hi-flux EDA industry, have just published the first of a two-part article on the EETimes Web site (The Evolution of Design Methodology (Part 1)) that strongly makes the case for the EDA360 idea of apps-driven design. This first article, attributed to Hogan, begins with a brief overview of the IC industry’s major transitions from semiconductor manufacturers (what we now call IDMs) to ASIC vendors to foundries. All of these transitions were driven, in one way or another, by Moore’s Law scaling of the raw silicon and EDA’s constant efforts to derive the most value from the silicon at each stage.
We’re in a new stage now, writes Hogan. We’re at the software stage. “Everything is now focused on the block level and the software load. Chips are now structured with a microprocessor-centric view of the world whereby IP blocks are considered peripherals attached to microprocessors through complex interconnect structures or Networks on Chips (NoCs). Design is done at the software level.”
That’s another way of saying that the apps drive the system—and the most value resides in the system design, as this article says again and again. “Value as always is at the system company. … But the IP available and the design flows mean that the system companies no longer need to share much of this value with the rest of the supply chain. … Apple built their own silicon for running their iOS and applications fast and at low power. In effect, they designed the A4 as an optimal silicon engine to run a particular software load and, as a result, kept more of the value for themselves.”