There Can Be Only One

Standards can be good, but so can competition. The challenge is striking the right balance.

popularity

By Cary Chin
The tagline of the 1986 fantasy film “Highlander” implies that, at least in some instances, we eventually will arrive at a single, best solution for our problems. In the case of low-power design, the most obvious application of the phrase is in the standardization of low power intent formats, where the Unified Power Format (UPF) and the Common Power Format (CPF) have been locked in one of those eternal struggles for all of about six years. About a week ago, on March 6, there was a “Quickening” in Highlander terms, or a “disturbance in the Force” if you prefer alternate science-fiction film metaphors. The IEEE approved IEEE 1801-2013, also known as UPF 2.1, with the full participation of all of the major EDA companies, as well as many computer/chip/IP design firms.

So are the format wars now officially over? Time will tell, but the history in our industry doesn’t necessarily point us toward a singular conclusion. After all, Verilog and VHDL have existed side-by-side in our industry for more than 20 years, and we still have both Accellera and Si2 looking to “standardize” everything from verification methodologies to system-level design. In fact, our industry (like many others) actually might feed more off the energy of competition than the strength of standardization.

Competition (in its “healthy” form) drives innovation and encourages creativity. These are certainly traits that we associate with our industry at its best, so why standardize at all? The main reasons are that “unhealthy” competition can lead to inefficiency and destabilization. With the pace of change and improvement in our industry set by the exponential growth of Moore’s Law, we simply can’t afford to make changes slowly. We must do everything right to make the most of competition as a driving force, while still considering the potential efficiencies of standardization. As usual, the decision is hard, but this is what engineers are good at! We need to look at standards as a mechanism we can use to raise the level of efficiency in our processes, where there’s little to be gained by continually exploring alternative solutions. That might mean picking the right time for standardization—after we’ve passed the 80/20 point of fast innovation and change. UPF 2.1 sounds good, but 11.2 would be scary.

And I still say that Betamax was in every way superior to VHS! After all, there can be only one. Or can there?

—Cary Chin is director of marketing for low-power solutions at Synopsys.