As we look forward to DAC, and UltraSoC’s first public outing as a member of the Constellations initiative, I thought I’d write a quick post to explain more about our company and (without indulging in too much back-slapping) share some thoughts on why we think Constellations is so important.
Part of the Constellations “elevator pitch” is that “a healthy semiconductor industry is best served by an IP ecosystem composed of many companies, large and small”. It’s hard to disagree with that, because I believe that in many ways it’s a description of the very situation we find ourselves in today. SoC designers routinely have to combine standard cores from multiple vendors, purchased IP from domain specialists, foundation IP sourced from EDA companies, and more. According to Semico the average SoC now has more than 100 IP blocks. There’s already an ecosystem that generates and supplies that IP, and also offers robust, capable tools to help with the integration, verification and debugging process: our goal in participating in Constellations is to make that ecosystem even stronger, and give it a voice.
In many ways, UltraSoC sits on two sides of the fence in this. Yes, we are a semiconductor IP supplier. But what we supply is analytics and debug IP – one of the benefits of our UltraDebug product is to massively speed up the process of SoC development and accelerate time-to-market. So what we do fits very neatly with a concept like Constellations: we are independent of any particular core or IP vendor, so we allow a holistic approach to debugging in heterogenous multicore SoCs. Our IP can benefit any SoC designer, and helps IP from different parts of the ecosystem to work together.
Of course, on-chip analytics accelerates SoC time to market, but UltraSoC’s business is also a powerful illustration of how semiconductor IP can enable really powerful innovations in the systems in which our products are so deeply embedded. This is one aspect that really excites me about my new job as UltraSoC’s CEO.
By adding hardware-based universal debug capabilities to SoCs, we enable an integrated, coherent view of the SoC hardware and software that can be viewed by a human engineer, or captured for further analysis. With these capabilities “baked in” to the silicon, the system designer who integrates the chip into the system can use them to build-in features like in-field analytics, forensics and dynamics. System development can effectively continue even after the end product has shipped and products can be performance-optimized in-service. In a world of 50 billion connected devices, such “in life” optimization could be hugely valuable. Functionality can be improved based on hard data about real-life usage, not lab-based predictions.
When I talk about performance optimization, of course, that could mean a range of attributes: it could mean performance in the sense of throughput, for example; we could reduce power consumption or increase battery life, because the SoC is capable of delivering a system-level understanding of how to execute a task most efficiently. Alternatively, where safety is important or there are legal implications or SLAs, it might mean remote diagnostics or forensics to report on the root cause of a problem.
So here at UltraSoC I think we’re a great example of “why people should care” about the IP ecosystem. Getting the message out – that our ecosystem is an engine room for real innovation – is a primary motivation behind our joining Constellations.
I’ve got personal experience in past roles of what can be achieved by mobilizing the shared resources of many organizations: I know that next month’s DAC is just the start for UltraSoC within Constellations, an opportunity as a group to “punch above our weight” at one of the industry’s premier events. It’s great to feel part of this group.