Power IS Top Priority, Isn’t It?

The question of power taking top priority boils down to a matter of definition.

popularity

While I don’t mean to start a battle – or maybe I do! – I heard something last week during DAC that gave me pause. The person I was speaking with – who told me they ‘got in a little bit of trouble’ for saying this – reminded me of an interesting subject we had talked about previously, namely, that they did not believe power is the number one concern of engineering teams today. Yes, power is important, like timing, area and performance, but making sure the chip works is the most important thing, right?

Well, the more I’ve thought about this, the more I am convinced it boils down to a matter of definition. The power must be right in the context of a correctly functioning chip. Of course the chip has to work AND the power has to be within spec.

We know that the amount of wiggle room in the system power spec is just as painful now for any application, not just the battery powered devices we would have suspected in the past. Be it in a datacenter, in an industrial plant, in an automobile – the power being out of whack is more than just a bigger energy bill. In the datacenter, it means higher cooling costs; in an industrial plant, it could be the same; in an automobile, the consequences of the power spec not being met could equate to fried electronics or other parts of the electronic system that leads to catastrophic consequences nobody wants to think about.

Then, there are thermal issues that must be considered, which the big players in the EDA industry are finally starting to do something about, joining the smaller players that have been sounding the alarm for years now.

So what do you think, engineers? What are your experiences with power in the context of making decisions and how does the system architect and the design manager look at it?