Shall I compare thee to a…Rolls Royce jet engine?
‘There is a new era dawning whereby deeply embedded sensing within all technology will bring about great benefit for the reliability and performance of semiconductor-based products.’ These were my words during a presentation to an industry audience in China back in September 2015. During that same presentation, somewhat to the consternation of the technology veterans in the room, I also drew comparisons between semiconductor design and an aspect of aviation technology being offered by Rolls Royce. Why on earth would I do that?
I could envision real-time, high accuracy, embedded monitoring becoming ubiquitous in all technology. Plus, understanding that valuable insights can be gained from gathering large amounts of data across entire product ranges could enable a revolution within the semiconductor industry.
To explain my comparison for a moment – A core principle of Rolls Royce’s R2 Labs Intelligent Engine is ‘data to insight.’ The technology offered by the aviation giant involves gathering mechanical, electronic and system level data for each jet engine in operation, wherever that may be in the world. Through centralised, large data analysis, Rolls Royce have enabled the ability to predict reliability issues, schedule engine maintenance and also allow for trends across fleets of aircraft to be assessed. My point back in Sept 2015 was, in the near future we shall be applying the same approaches and analytics principles to semiconductor devices. This ‘near future’ has now become our reality.
Gathering information from the physical world and acting upon it has been fundamental to human evolution.
In the modern day, how does this correlation to jet engines relate to semiconductors? The answer is within some of the challenges we can identify today – there is undeniable value in: predicting the failure of a critical automotive chip; or finding the operational sweet spot for a processor in terms of clock speed or power, steering an entire product range of data center chips to consume less power while achieving operational performance, such that carbon footprints are reduced by a power station or two.
I’m not the first to make comments of benefits of having an enlightened position through deeper observation. In 1665, Robert Hooke’s book ‘Micrographia’, (the first scientific best-selling book!) provides us with a good example the discovery of ‘Minute Bodies,’ or cells, through the use of magnifying glasses.
‘If you can’t measure it, you can’t improve it’
The famous management consultant Peter Drucker points out that, “If you can’t measure it, you can’t improve it.” A personal favourite is the inspired observation from Beyoncé, “You try and fix something, but you can’t fix what you can’t see.” I am sure that big data and a desire to seek patterns within dynamic semiconductor device behaviours was at the forefront of her mind as she wrote those lyrics!
Embedded monitoring within semiconductor devices will evolve, bringing with it a greater opportunity to consume less power, increase speed performance, enhance reliability and reduce design re-spin costs. As technology evolves, as with my earlier jet engine analogy, expectations upon the semiconductor industry will increase from our vertical-market masters.
Today, solutions are available that will monitor the rapidly changing conditions within a chip, alongside assessments of how it has been fabricated, looking at variation from one chip to the next. This is all for the benefit of developing stable, reliable and optimized products. So my message to chip designers is that your chip is always saying something …. the question is are you listening?