RISC-V is rewriting the silicon industry just as surely as the Linux kernel rewrote the software industry.
If you worked in the tech industry in the mid 1990s, you’ll remember the incredible wave that crashed on the shore of industry when Linus Torvalds released version 1.0 of the Linux kernel—the core of the Linux operating system (OS). The idea that software could be open, not proprietary, seemed outrageous to many. Pundits and programmers alike predicted the quick demise of Linux because it did not come initially with a GUI interface like Microsoft Windows or the Mac OS.
Fast forward almost 30 years and Linux now powers the majority of web and cloud servers. The ability to easily modify the software for uses as diverse as Android phones, the Google Chrome web browser, and Kubernetes for porting legacy applications to cloud servers makes Linux a popular choice for developers and businesses alike. This is especially true because Linux distributions remain current and supported by communities of developers from across the industry. The result saves companies design time and effort by leveraging open source software from the global community—whether it’s for a tiny piece of their business or a central part of a technology offering.
A similar revolution is occurring in the silicon industry right now as companies embrace open standards to innovate faster and less expensively. No longer locked into proprietary processor architectures, businesses of all sizes are turning to alternatives to meet the challenging computing needs of artificial intelligence (AI), machine learning (ML), and other cutting-edge technologies. Enter the RISC-V instruction set architecture (ISA). The rise of RISC-V processors coincides with the decline in Moore’s Law coupled with the huge demand for more processing power. And so, RISC-V is rewriting the silicon industry just as surely as the Linux kernel rewrote the software industry.
An ISA for today’s workloads
Historically, the silicon industry was dominated by proprietary processor architectures, which limited companies’ design freedom. But RISC-V’s streamlined base ISA is built from the bottom up to handle the latest compute workloads. This ISA differs vastly from proprietary architectures that are decades old and burdened with legacy instructions. Collaboration drives the RISC-V ISA, which was created with flexibility, extensibility, and scalability in mind. By freeing organizations from a single vendor for their product roadmaps, RISC-V makes custom silicon far more accessible.
Similar to the Linux community, RISC-V also provides community-shared tools and development resources to reduce development time and effort. The shared nature of the RISC-V ecosystem gives companies endless opportunities to build more innovative solutions. Leveraging the shared ecosystem also enables a faster time-to-market (TTM) and reduces the need for an extensive in-house design team. Everyone benefits from the power of open standard collaboration. It inspires a surplus of innovation, which is especially important as computing requirements become more complex.
This culture of collaboration in the RISC-V community also means that companies you might not expect regularly work together to develop new RISC-V solutions and tools. As a result, many surprising collaborations take place.
Earlier this year, Intel and SiFive announced a partnership to build high-performance RISC-V platforms optimized for Intel’s process technology. While Intel has long advocated for its own x86 architecture, it is now embracing RISC-V as a key driver for the Intel Foundry Services (IFS) ecosystem. Intel explains that “RISC-V offers a level of scalability and customization that is unique in the industry” and “there is strong demand from foundry customers to support more RISC-V IP offerings.”
Another notable collaboration is NASA awarding a $50M contract to Microchip to develop aerospace solutions based on SiFive’s Performance X280 processor IP. These new RISC-V based solutions demonstrated 100X the performance of current space computers, opening up new possibilities for innovation in the technologies used in space exploration. One of the selection criteria was the vibrant RISC-V ecosystem and the belief that, with it, products will have support for 10 years or more, which is important to planning space expeditions.
From AI to automotive
In addition to aerospace, another industry where you might be surprised to hear that RISC-V is getting a lot of traction is the automotive market.
As this market transitions to zonal architectures, manufacturers are looking for the simplicity, security, and software flexibility RISC-V offers. Using a single ISA across all zonal architectures increases code portability and can greatly reduce cost and TTM. In addition, RISC-V vector extensions are bringing enhanced ML and DSP capabilities to automotive manufacturers. The robust global RISC-V ecosystem ensures that there is a wide choice of supported tools and solutions. Working without proprietary lock-in, automotive manufacturers can license from multiple vendors and have more flexibility to design their own IP where needed, all while maintaining ISA compatibility.
In August 2022, SiFive introduced its Automotive E6-A product series to address next-generation automotive needs for applications like infotainment, cockpit, connectivity, ADAS, and electrification. SiFive’s automotive solutions offer much more tailoring compared to other CPU vendors, providing IP options that are both area and performance optimized for different integrity levels like ASIL B, ASIL D, or mixed criticalities with split lock, in line with ISO26262. These tailored solutions will help automotive manufacturers customize solutions for specific markets as they design the next generation of digital vehicles.
Like NASA, automobile manufacturers want to know their products will be supported for five to 10 years or more, and the RISC-V ecosystem, with university students around the world now studying it in class, looks to stand the test of time.
The future market
RISC-V is already in billions of cores on the market, and promises to continue reshaping silicon design in almost every industry imaginable. Semico Research forecasts that RISC-V-based AI SoCs will see an impressive 73.6% compound annual growth rate (CAGR) by 2027. For the data center market, Semico also estimates that there will be a 98.1% growth rate for RISC-V AI SoCs by 2027. In another performance-driven market, SDDs, Semico predicts 88.5% growth for RISC-V AI SoCs during the same time frame.
While automotive and aerospace applications are new, RISC-V use is growing in the data center. Google recently highlighted how they are using SiFive RISC-V processors as companion cores in AI and ML applications, creating what they called a division of “technical and business labor”and allowing Google to focus on their expertise while taking advantage of the proven performance of RISC-V cores and the support of the wide ecosystem.
I look forward to seeing how RISC-V will continue to usher in exciting new innovations that transform the way we all compute and live.