Honey, I helped save the planet today…

I’m sitting in Munich airport. Snowflakes swirl softly in the air outside. I am waiting for my connecting flight. In front of me, an extra large flat screen TV hangs off a pillar, in the waiting area by the boarding gate. On the news, images from Copenhagen and the Climate Change Conference: protesters and diplomats in HD pixels; close-ups of nuclear plants and melting icebergs mirrored throughout the airport on countless other instances of that flat screen TV model.

It’s late; I am tired and jetlagged but can’t help wondering: how many such TVs are there in this terminal? And what do these marvels of semiconductor technology have to do with the Copenhagen talks? After all, aren’t we talking about a hundreds fifty watts or so per LCD?

Preparing Recommendations

Here is an excerpt of what efficientproducts.org has to say about this: 

• “TVs consume about 90 billion kilowatt hours (kWh) of electricity per year in the U.S. or about 23% of household Miscelleneous Electric Loads, which in turn represent 24% of total houshold electricity consumption. TVs use over 90% of this energy when the set is turned on. The remaining amount is used in standby mode when the set has been turned off by the consumer. TV use in the U.S. results in a greater than $10 billion annual electricity bill and releases over 55 million tons of CO2 from power plants. Unless TV efficiency improves, the rapid growth in TV sales, increased hours of TV viewing, and multiple TVs per household will expand TV energy consumption to about 170 billion kilowatt hours (kWh) by 2014.”

Power consumption compared

Now imagine the impact on carbon dioxide emissions and electricity bills if you could reduce power consumption by 20% on such products… Or course, such an improvement doesn’t happen by simple magic: it must be engineered. It is no wonder if power consumption rose to be such a hot topic in our industry. Not so long ago, square microns and picoseconds were the two key metrics of a design success. Today, it’s all about watts and how much (or how few, really) there are.

But one of the challenges facing design teams is that low-power optimization requires specific skills which are not yet wide-spread. Chasing watts is a new kind of art, and one quite different from the area and timing optimization techniques which designers lived by for so many years. Even more challenging, low-power design isn’t a single task. Rather, it must happen at all stages of the development cycle. It is a transversal discipline, impacting all tasks from system-level (ESL) to physical design, and the earlier in the process, the greater impact you can have on your power consumption.

Take high-level synthesis (HLS) for instance. Through the process of generating RTL from abstract specification in C++, HLS tools gain full insight over the design and its architecture. From there advanced tools can optimize the results, not only for timing and area, but also for power by automating prevailing low-power design techniques, such as multiple clock domains, clock-gating, multi Vth, etc… A tool like Catapult C as shown to reduce power by up to 40% on some designs - click here for a demo. And if you step up from the implementation to the architecture level, virtual platforms based on transaction-level models (TLM) can provide invaluable information about how architectural decisions impact power consumption. Mentor’s Vista solution provides a effective and automated way to do just that.

Admittedly, change is always difficult. But today the stakes are high and the good news is that all of us have an incentive to make lower-power designs a reality - whether for business, personal or moral reasons, whether to improve competitiveness of a product, develop new skills or simply to be able to say tonigh: “Honey, I helped save the planet today…”

More Blog Posts

Preparing Recommendations