Edge Computing for IoT AI with RISC-V

Edge Computing for IoT AI: Computer-CPU-designer job is one of the highest paid in the world. Designing a CPU requires design expertise in several specialties: electronic digital logic, cross-compilers, Power-on, and operating systems. It is impossible to find such a team outside of a professional engineering organization. Therefore, all commercial vendors of computer designs in the market today, such as ARM and MIPS charge processor manufacturers royalties for the use of their plans, patents, and copyrights.

RISC-V was started in the 1980s to solve these problems by academicians at USC – Berkley. The goal was to make an ISA that was available as open-source, usable in any hardware or software design without royalties. Many companies started using these RISC-V in their processor designs about a decade back.

Another significant RISC-V benefit is that it enables companies to develop a product that is modified specially to the contextual workload, so they start with the RISC-V core and add need-based special functions, saving both money and time. These savings can result in a lower cost to consumers, or give a longterm benefit of lower energy consumption. However, the most significant advantage of RISC-V is the internal security built-in instructions and the resultant peace of mind to both consumers and businesses alike.

RISC-V has now been going for around four years since the public announcement. Berkeley released it to the world, the original developers of RISC-V then produced a VC backed startup SiFive, which has been doing much work in the RISC-V ecosystem, but there are also several other vendors who are creating cores and producing tooling around it.

On the operating system side, there’s support for RISC-V in FreeRTOS, Zephyr, seL4 some initial work on Tock. Tock is an exciting RTOS implemented using Rust.

RISC-V is taking mainstream now with the entire semiconductor industry chasing the technology. Nvidia plans to use a RISC-V processor onboard their GPU design. Western Digital is going to ship a billion RISC-V units in the coming years. If you are someone who loves small board computers such as Raspberry Pi’s or Arduino’s, SiFive sells a RISC-V developer board right now.

SiFive’s HiFive1

Shakti RISC-V is the First Open Source Chip from India. Shakti is an open-source processor initiative by the prestigious IIT Madras aimed at developing industrial-grade processors based on RISC-V. Students in India produce this open source RISC-V processor with assistance from Intel and its 22nm FinFET Process Technology.

The latest is the Grove AI HAT for Edge Computing that features a Sipeed MAix M1 AI MODULE with Kendryte K210 processor. A Hat is an add-on to a raspberry pi. However, it’s a very powerful low-cost raspberry pi AI hat which assists raspberry pi to run the AI workloads at the edge. The Grove AI HAT is capable enough to work independently in edge computing.

Grove AI Hat

The Sipeed’s MAix M1 is a powerful RISC-V 600MHz AI processor module that features the following
A Dual-core 64-bit CPU,
A 230GMULps 16-bit KPU(Neural Network Processor),
An FPU(Float Point Unit) supports DP&SP, and
APU(Audio Processor) supports eight microphones.

With a powerful Kendryte K210 CPU at the core of the module, the Grove AI HAT finds its application as Edge Computing board that provides a wealth of peripherals interfaces for connecting the various device: I2C/UART/SPI/I2S/PWM/GPIO interfaces. The hat also offers an LCD and a camera interface, which supports the Sipeed 2.4inch QVGA LCD and DVP camera, it will be helpful and convenience with your AI vision project.

The MAIX M1 AI Module is equivalent to Google’s Edge TPU accelerator. However, MAIX M1 AI acts as a master controller, not a CPU accelerator, thus making it a lower-cost and lower-power solution. MAIX finds it’s usage for a growing number of industrial use-cases such as predictive maintenance, anomaly detection, machine vision, robotics, voice recognition, and many more. Its applications are in manufacturing, on-premise, healthcare, retail, smart spaces, transportation.

Edge Computing for IoT AI is leveraging RISC-V architecture to its best. Edge Computing for IoT AI basically means machine-learning models are required to be trained using Edge computing instead of powerful machines or cloud-based infrastructure. However, Grove AI Hat for Raspberry Pi will accelerate the speed at which models can be trained to run and be used to infer information from data like to identify a specific make of car in a video or to perform speech recognition in a live or recorded audio. While AI-related tasks like image recognition used to run in the cloud, RISC-V computing is pushing for machine-learning models to run locally on low-power devices such as the Pi.


  • Mobodexter, Inc based in Redmond, WA builds IOT Edge solutions for enterprises applications on Kubernetes & Dockers.
  • Register now at developers.paasmer.co and enjoy free Edge trial license that can run on Raspberry Pi.
  • Follow our all our weekly IoT Blogs:
  • Join our IoT Hub of 3000+ subscribers to get these blogs in Email –  Subscribe.
  • Become our affiliate partner today.
%d bloggers like this: