メインコンテンツに移動

Renesas' Next-Generation AI Solution – Single Board Computer with Cortex-A55 Core

画像
Abigail Sun
Senior Applications Engineer
掲載: 2021年10月26日

Nowadays, Artificial Intelligence (AI) is changing our life. This technology is enjoying a growing market acceptance resulting in an increasing investment into its technology and application. Instead of a centralized AI compute system, AI at the edge becomes more and more popular, reducing latency and increasing data security. Edge-AI brings a lot of benefits to the user:

  1. Local data processing ensures real-time AI and reduces or eliminates necessary cloud connectivity
  2. Ensures low latency AI processing and allows real-time data processing
  3. Edge-AI increases privacy protection and data security, as all data processing is performed locally and no cloud connection is required to process sensitive data
  4. Reduces the data sent to cloud and reduces the bandwidth requirement between local and cloud communication

Many smart devices have emerged over the last couple of years and the introduction of Edge-AI will further increase the systems features, capability and adoption.

Renesas Electronics provides a wide selection of AI-enabled devices (MPUs listed from high-end to low-end): RZ/V2M system, RZ/V2L system, RZ/G2L system, RZ/G2LC system, and RZ/G2UL system. The Single Board Computer winning combination is ideal for Edge-AI systems and covers RZ/V2L, RZ/G2L and RZ/G2LC.

画像
Single Board Computer

The RZ/V2L MPU embodies a multicore Cortex®-A55 core plus a DRP-AI engine. The DRP-AI engine is Renesas' special AI accelerator that boosts AI performance with less power needed. In addition, the DRP engine can be used for Vision AI applications and can process images 10x faster than general-purpose cores.

The RZ/G2L MPU is an entry-level pin-to-pin compatible solution to its higher performance RZ/V2L. RZ/G2L has a multicore Cortex-A55 core. The Arm® Cortex-A55 processor is a market-leading CPU that delivers the best combination of power efficiency and performance in its class. It is part of the first generation of application processors based on DynamIQ technology and features the latest Arm v8-A architecture extensions, with dedicated machine learning instructions (https://developer.arm.com/ip-products/processors/cortex-a/cortex-a55).

The RZ/G2LC MPU is even more optimized for small form factor applications than the RZ/G2L, utilizing a small package and limited peripheral interfaces.

As most MPU designs are complex, requiring special attention to power, clock and memory connectivity, Renesas' Single Board Computer winning combination provides a solution for Renesas’ next-generation MPUs, simplifying the design effort for the system designer and getting them to market faster.

The Single Board Computer solution uses one single power management IC (PMIC), RAA215300, to provide power for all MPU power rails, eliminating the tedious tasks of designing power sequencing, input voltage monitor, reset delay, real-time clock management, and watchdog monitoring. This significantly reduces design risk.

The Single Board Computer also incorporates an on-board single device multi-output clock, 5P35023. This innovative and programmable clock reduces board space and increases flexibility.

Besides the power and clock, the Single Board Computer solution provides multiple additional functions and features, including:

  1. Optional boot from QSPI or eMMC
  2. Gigabit Ethernet communication (2 ports directly supported by MPU)
  3. USB communication (2 ports directly supported by MPU and 2 additional via USB hub)
  4. General-purpose audio output
  5. DSI/HDMI graphic output support up to full HD
  6. Wireless communication
  7. CAN interface
  8. Multi-pin GPIO for external connection capability

With all these functions, this system can cover many use cases, and any system designer will be able to quickly customize their system based on this architecture.

For more information about our Edge-AI enabled RZ microprocessor product family, visit the Single Board Computer winning combination page.

この記事をシェアする