CEO Rupert Baines comments on recent viewpoints from NXP and Imagination Technologies
At embedded and EDA/chip design conferences it’s not unusual for me to pick up on things I hear and relate closely to what we’re doing here at UltraSoC. I find that plenty of major design and EDA companies are like-minded in their belief in the value of data science embedded deep within chip and system design; and in the value of having monitoring providing insights into the inner workings of SoCs, or entire technology devices even when they’re deployed in-field. It’s nice to see the wider industry not only ‘getting it’ but espousing the same values. It’s even better when you see it come from two of the most senior spokespeople from two very prominent technology companies within the same week!
CTO, Lars Reger – actually, he’s not a new CTO… but his role has expanded from being CTO of the automotive division at NXP, to now overseeing the entire NXP business. The decision to select Reger for the overall CTO role reflects, in my mind, a strategic shift in the industry that reflects our own experience. Automotive silicon and system design focuses the mind on adherence to the strictest safety standards and tolerances; it also focuses the mind on performance and reliability at the highest levels; and, increasingly due to ADAS, it focuses the mind on the critical need for functional safety to be designed in and throughout the system – and for that system to be holistically designed. The process, the tools and the technological approach fine-tuned in automotive design can and should be applied to improve system development in a broad range of disciplines: AI, ML and high performance computing, for example. So, when Reger started talking about his biggest technology agenda being filtering data, my ears pricked up.
“…we need to be smart about how best to filter data. We must extract the right data at the right stage. Filtering will determine the type of network we rely on. Filtering also helps our customers to look at the most important data.
Equally important is to apply the right level of deep learning to data. We need smarter algorithms for anomaly detection. I see [that] machine learning applies to every thing from radars and camera systems to predictions/planning by autonomous vehicles.”
Let’s pick up that one key phrase from Reger: “We need smarter algorithms for anomaly detection.”
Meanwhile, Imagination Technologies has also made a new appointment. Its new CEO, Ron Black, was also interviewed last week by EETimes to find out how he plans to ensure the company stays relevant in the face of evolving applications and needs of the high growth technologies. What jumped out at me from this interview was his vision of a future with: “rich data generated in systems…..” and “turning data into informed insights”.
So ‘smarter anomaly detection’, and ‘informed insights from rich data’. If only there was a technology provider able to provide such embedded analytics and monitoring capabilities….I don’t like to sit and here and say ‘I told you so’ – but it’s fantastic and very significant in my opinion that we do see more and tech companies, including those at the higher system level, picking up the sticks to bang the same drum we have been banging for some time…
I’m interested to hear what others think, and I’ll be at Embedded World in Nürnberg from 26-28 February 2019. We will be exhibiting at the event as part of the RISC-V Foundation booth (3A-536). If you’re attending the event and would like to book an appointment to find out more about UltraSoC and how we can help your product development, please email firstname.lastname@example.org.