In conversation with Dipesh Patel from Arm
PragmatIC Semiconductor is delighted to welcome Dipesh Patel, Chief Technology Officer of Arm, to our Board of Directors. Dipesh has worked at Arm for 25 years, and now leads Arm’s Research and Digital IT teams. We spoke to Dipesh about the enduring relationship between PragmatIC and Arm, and how he believes it will continue to evolve over the coming years.
1) Please tell us a little about the partnership
Dipesh: Our relationship dates back to 2013, when Mike Muller (CTO of Arm at the time) first met Scott (CEO of PragmatIC) and discussed the exciting possibility of creating an ultra-low-cost, thin and flexible Arm processor. Early efforts were successful in producing simple prototype circuits, so in 2015 there was an attempt to build a complete PlasticArm, but unfortunately, we didn’t achieve a fully working part that time.
With hindsight, we were a little over ambitious in 2015 but we gained valuable lessons. For the next project we decided to create a chip with a machine learning (ML) processing engine paired with sensors to categorise odour. The result was reported in our joint Nature Electronics article back in 2020. This was a pivotal project for the two companies, where we undertook co-optimisation of hardware and software in order to produce the smallest custom ML engine for the task at hand.
Then, after continued work by PragmatIC on their unique technology and FlexLogIC® manufacturing system, we were delighted to produce the world’s first non-silicon 32-bit Arm processor, reported in Nature mid-2021. This represented the largest flexible integrated circuit ever made.
2) What do you think are the major advantages with PragmatIC's approach?
Dipesh: There is of course a very obvious advantage in addition to low cost – the flexible nature of the circuits is well suited to certain applications, for example non-rigid objects, or curved surfaces with small radii.
But there is another type of flexibility that makes the approach very attractive – development agility. The production cycle time of a chip through the FlexLogIC line is much shorter than typical cycle times for conventional silicon manufacturing fabrication plants. Additionally, the cost of taping out a design is also orders of magnitude lower. Both of these benefits combine to reduce the barrier to spinning new chips, making it feasible to hardware optimise devices for an application, reducing overheads and further lowering the cost of the final chip.
3) Will flexible chips take over from silicon?
Dipesh: I don’t expect to see flexible chips displace silicon. As a result of Moore’s law, silicon chips have advanced dramatically. When we compare the first Arm processor system, which came to life in 1985, to the leading edge now, there is a vast difference. The latest Arm based System on Chip (SoC), have multiple Arm cores spread across multiple sub-systems within the SoC with more than 30 billion transistors produced on 5nm technology. For many high-end applications, there is a continued drive to further increase complexity and performance. For the foreseeable future such systems will continue to rely on silicon.
Flexible integrated circuits on the other hand enable new possibilities. Due to the benefits outlined earlier, we can address new markets that cannot be reached by silicon chips. For example, the edge of the Internet of Things can be extended towards embedded electronics in everyday objects. They will follow their own version of Moore’s law, integrating greater functionality per device but without the associated increase in fabrication cost we have seen with silicon. I expect that one of the most common uses will be a compute engine paired with sensors in order to analyse real-world data directly within mass-market products, providing immediately actionable output to improve business processes or customer experience. There are many other possibilities – for example: monitoring temperature and humidity to improve supply chains for food and beverages through intelligent packaging, or enabling automated wound monitoring via a smart bandage.
4) What will Arm and PragmatIC be focusing on next?
Dipesh: Both companies are excited by the potential for technology to help solve some of the most pressing problems facing the world – climate change, health inequality, etc. We see a lot of opportunities to apply our combined solutions in these areas.
A key focus is exploring how we can improve healthcare outcomes, especially in low- and middle-income countries. For example, can we identify problems with wounds earlier, so that they can be more effectively treated? Approximately 85% of sepsis cases and related deaths are outside the developed world, so a low-cost wound monitoring solution could have a transformative impact.
Also in more developed countries, the growing proportion of adults living longer requires advances in technologies that enable them to maintain healthy and independent lives. Ubiquitously available low-cost wearable solutions could have a major impact, without disrupting healthcare budgets that are already under significant pressure.
Finally, our teams are excited about enabling new approaches to agile design of electronic systems. We want to encourage a new generation of engineers to create innovative solutions that improve everyday life, bringing them to market even faster than is possible with conventional electronics.