Global technology leader Arm is developing the tools and technologies required to enable internet of things (IoT) and artificial intelligence (AI) applications which will make smart cities and autonomous cars a reality in the coming years. They have a growing team based in Galway.
“AI and machine learning have been around for many years”, said Dave Murray, a distinguished engineer and solution architect with Arm.
“Over the last couple of years, a confluence of events has allowed for their real-world application. A combination of vastly increased chip density and computing power means we can now apply algorithms that existed before but couldn’t run efficiently in real-world scenarios. The computing power available to us now makes those very complex algorithms deployable.”
And Arm provides that computing power.
Arm’s presence in Galway dates back to 2014 when it acquired Duolog, an Irish company with a heritage dating back to the very beginning of the semiconductor industry in Ireland in the late 1980s.
Established in 1999 by founders with vast experience in the semiconductor design sector, Duolog went on to develop world leading expertise in the development of the design automation tools that assist with the production of complex chips.
The company initially had two areas of activity. “Back in the early 2000s there was huge pressure on companies to reduce costs and outsource services and activities”, Murray explained. This led to Duolog establishing an engineering centre in Budapest which offered solutions to customers around the world at highly competitive rates.
But this was low margin work. “We started to look at the design process here in Galway”, he added. “We were helping customers design chips for mobile platforms and other applications and we found a lot of inefficiencies in the process. We started to develop tools to help chip designers do it quicker and more efficiently. These value-add solutions offered huge efficiency gains and savings to our customers.”
That was Duolog’s entry to the electronic design automation (EDA) market. “No one else was doing that in Ireland, it was very much a start-up activity”, Murray continued. “Internationally, there were three big companies in that space and a lot of fledglings. It offered us a way to move up the value chain and our customers helped us invest in it by pre-licensing the tool sets.”
In essence, Duolog’s tools helped designers combine different functions, or IP blocks, together on a single chip, very efficiently.
“While everybody was talking about treating these IP Blocks like Lego and building something very quickly, we found that the standards and capabilities to do this just weren’t there. We focused on providing solutions for rapid IP integration into a system on a chip, produced tools and helped drive standards in this direction.”
That proved to be an inflection point for the company and it quickly earned an international reputation for itself. Duolog won three awards at the 2008 Design Automation Conference, the global gathering for EDA companies.
“That put us on the global stage”, said Murray. “We were travelling all over the world meeting companies in Europe, Asia and the US and were really punching above our weight. We continued to add more functions to the tools and soon had 60 to 80 semiconductor companies around the world as customers.”
Around this time the company looked to scale its technology and strategic relationships with its customers. This led to a partnership between Duolog and Arm. “The companies were natural partners”, Mike Lardner, design centre director with Arm Ireland pointed out.
“Arm was supplying IP blocks to the semiconductor industry but was looking to add value for its partners by accelerating the integration of this IP within the partners designs. Duolog was producing the tools that could enable them to do that very efficiently. Arm gave Duolog the scale it needed and Duolog provided Arm with this integration capability.”
That partnership led to the acquisition of Duolog by Arm in 2014. The focus for Arm Ireland has evolved to deeply integrating the core technologies and tools with Arm’s IP products. Arm’s partners use the tools to quickly configure and integrate Arm’s IP within their chip designs.
A dedicated AI and machine learning team has also been established in Galway with the support of IDA Ireland since the acquisition.
“In 2014 we had 30 people in Galway”, said Lardner. “We now have a team of around 60 here. We are continually hiring graduates from Irish universities. The team here is recognised as a highly skilled, dynamic and agile software group. We work with many groups and technologies across the wider Arm team”
Arm has been growing as well, according to Lardner.
“SoftBank acquired Arm in 2016 and it has grown to over 6,000 employees since. SoftBank takes a long-term strategic view of its investments and this allows Arm to invest in technologies and products which look far into the future,” he said. “One technology area becoming more important is the machine learning and AI space and we are investing heavily in that area.”
Looking to the future, Lardner says Arm will provide the leading-edge computing platforms that will enable machine learning, AI and other technologies to realise their full transformative potential.
He points to the concept of ‘machine learning at the edge’ in this context. Having computing power on servers and on the cloud with remote devices having to communicate back and forth can slow systems down to the point where the application cannot work effectively. For example, an autonomous car can’t wait for a distant chip to decide if it should slow down or avoid an obstacle – these decisions must be instantaneous.
Lardner concluded: “Sending signals back and forth from a data centre to an autonomous car doesn’t work because of latency and other issues and in a future with trillions of IoT devices deployed and connected, the scale of computing and communication power required means that these devices will need to run complex machine learning based applications locally.
“This is where Arm is currently focused, at the device level with the computing power at the edge. The AI and machine learning team here in Galway is working on multiple applications that enable the software and algorithms to run faster, more efficiently and use less processing power, across Arm’s compute products.”
You'll find us responsive to your needs, proactive, professional and willing to go the extra mile.
We have 28 Offices worldwide helping support companies expand their operations in Ireland.