Arm Flexible Access Reflects the State Today’s Semiconductor Market

September 27, 2019 Brandon Lewis

If you’ve been monitoring the semiconductor market over the past few years, you’ll know that it’s ripe for change.

The end of Moore’s law has prompted chip designers to look elsewhere for improvements in performance, power consumption, and cost. The IoT is promoting the deployment of compute intelligence everywhere. The desire for custom, application-specific processors only continues to increase. The momentum around open-source hardware is now palpable thanks to Raspberry Pi, Arduino, and BeagleBoard in the Maker community, and the RISC-V ISA in the academic, research, and commercial realms.

These trends are also leading to a world where engineers want – and need – access to lower cost, more accessible processor IP, if only just for evaluation and prototyping.

It also serves IP vendors to lower the barriers to entry for their technology. Doing so may cost them in the short term, but lead to more high-volume, higher-margin business over the long haul.

All of these dynamics and more led Arm, the largest semiconductor IP provider in the world, to adjust its traditional IP licensing model to include Arm Flexible Access.

More “Flexible Access”

Arm Flexible Access is a new way of allowing SoC design teams to access a broad set of Arm IP for initial evaluation and development. But more importantly, they only have to pay licensing fees on what is used in production.

It’s not free, of course. Annual access fees start at $75,000, which includes one tape-out per year. They scale up to a $200,000-per-year option that allows for unlimited tape-outs, as well as additional tools, training, and design services.

Licensing fees are applied on a per-product basis, and standard royalty fees also apply. But for established SoC engineering teams that want to tinker around with Arm technology and have the resources to spare, it sure beats the alternative (Figure 1).

Figure 1. Arm Flexible Access departs from the traditional Arm licensing model by postponing up-front licensing fees until tape-out.

Arm Flexible Access can be leveraged for a wide range of processor cores, interconnects, system controllers, security IP, and tools. These include, but are not limited to:

  • The majority of Arm Cortex-M, -R, and -A-class CPUs (which have accounted for 75 percent of all Cortex CPU licenses in the last two years)
  • Mali GPUs
  • Corstone foundation IP
  • Artisan Physical IP & Libraries
  • Arm DS Gold One
  • Arm Socrates
  • CoreSight Debug & Trace
  • Virtual System Models

A full list of the technologies covered under Arm Flexible Access can be found at www.arm.com/why-arm/how-licensing-works.

Early participants in the Flexible Access program include Nordic Semiconductor, Invecas, and AlphaICs.

Even more flexibility for research and academia

Shortly after the Flexible Access launch in July, Arm continued to build on its new model with the unveiling of Arm Flexible Access for Research.

This sector-specific offering is, you guessed it, only available to members of the academic and research community. It offers the majority of Arm Cortex-M, -R, and -A CPU cores; complete RTL access; and other subsystems and tools for completely free. No fees or costs of any kind are associated with this flavor of the Flexible Access offering.

This is obviously designed to insert Arm IP into the hearts, minds, and early designs of advanced project developers. But it may also purchase the company some good will on the university circuit, where RISC-V has grown increasingly popular in recent years.

To gain access users just have to complete a click-through end user license agreement (EULA) at www.arm.com/resources/research/enablement/contact-us. Arm Research SoC Labs is also developing an online community that will help foster technology sharing, reuse, and peer review for Arm Flexible Access for Research participants.

Still get your “DesignStart”

It’s important to note that Arm Flexible Access will have no impact on the popular Arm DesignStart program, which offers a $0 upfront access fee for Arm Cortex-M0, -M1, and -M3 CPUs and a $75,000 access fee for the Arm Cortex-A5. In fact, Dipti Vachani, Senior Vice President and General Manager of the Automotive and IoT business units at Arm says that “Arm Flexible Access complements DesignStart.”

“DesignStart will continue to be ideal for partners who want the lowest possible upfront costs,” she asserts. “Arm Flexible Access provides access to a wider set of technology and tools – all the essential IP to design an SoC.

“They both also offer a portal to the world’s largest compute ecosystem, which helps enable lower SoC build costs, smaller risk profiles, and faster time-to-market.”

Semiconductor innovation or just a different way to pay?

Given all of the market forces currently at play, it was inevitable that the winds of change would begin blowing in the semiconductor industry. It’s a sector that’s been ready for some flux (or maybe some “flex”) for several years, at least.

But what do programs like Arm Flexible Access really accomplish? Are they truly going to foster semiconductor innovation?

Or are they just another way for a company that has dominated a market segment for nearly two decades to extract revenue earlier from small-to-medium-sized businesses that are on the fence about technology investments? And does it say anything that this announcement comes at a time when a somewhat viable alternative appears to be taking shape in the form of RISC-V?

“It's a convergence of multiple factors, with the biggest factor being a new wave of data-driven computing that is pushing more data processing to the edge,” says Vachani. “This, of course, has led to new computing models, specifically around machine learning and artificial intelligence. These new computing models have resulted in increased demand for custom silicon.

“The demand for custom silicon has prompted new entrants into the silicon market, such as startups, system providers, and OEMs seeking more agile development processes,” she continues. “And these new entrants want to work within a pricing model they are familiar with. That pricing model is similar to the SaaS and cloud computing pricing models that have become commonplace in the technology industry.”

I think the answer is a soft “yes” to all of the questions posed. But whatever the answer is, if it means things get better for semiconductor engineers, I’ll take it.

About the Author

Brandon Lewis

Brandon Lewis, Editor-in-Chief of Embedded Computing Design, is responsible for guiding the property's content strategy, editorial direction, and engineering community engagement, which includes IoT Design, Automotive Embedded Systems, the Power Page, Industrial AI & Machine Learning, and other publications. As an experienced technical journalist, editor, and reporter with an aptitude for identifying key technologies, products, and market trends in the embedded technology sector, he enjoys covering topics that range from development kits and tools to cyber security and technology business models. Brandon received a BA in English Literature from Arizona State University, where he graduated cum laude. He can be reached by email at brandon.lewis@opensysmedia.com.

Follow on Twitter Follow on Linkedin Visit Website More Content by Brandon Lewis
Previous Article
Life on the Edge: How AI and the NVIDIA Jetson Ecosystem are Transforming Industry, Transportation, and Logistics

The following case study describes how one AI engineering company leveraged the NVIDIA technology ecosystem...

Next Article
Talking About SIPs with Gene Frantz and Greg Sheridan of Octavo Systems

We recently had the opportunity to talk to Octavo Systems’ CTO Gene Frantz and Marketing Manager Greg Sheri...