Skip to main content

Banana Pi BPI-M6 SBC Showcases SenaryTech SN3680 Quad-Core Cortex-A73 AI Processor

The Banana Pi BPI-M6 is a single board computer (SBC) that integrates the SenaryTech SN3680 SoC, which features a quad-core Arm Cortex-A73 processor, an Arm Cortex-M3 real-time core, an Imagination GE9920 GPU, and a Neural Processing Unit (NPU) capable of delivering up to 6.75 TOPS for AI tasks. This makes it an interesting choice for developers and enthusiasts interested in AI, for example.

The BPI-M6 comes equipped with 4GB LPDDR4 RAM and 16GB of eMMC flash storage, expandable up to 64GB. It offers a form factor and port layout reminiscent of the Raspberry Pi 4, including four USB ports, Gigabit Ethernet, a 40-pin GPIO header, a USB Type-C power port, and two micro HDMI ports. A standout feature is the provision of an HDMI input alongside the HDMI output, adding to its versatility.

Here are the detailed specifications of the Banana Pi BPI-M6:

  • SoC: SenaryTech SN3680 with:
    • CPU: Quad-core Arm Cortex-A73 processor up to 2.1GHz
    • MCU: Arm Cortex-M3 real-time security core @ 250MHz
    • GPU: Imagination PowerVR Series9XE GE9920 GPU
    • VPU: Supports 4Kp60 H265, H264, VP9, ​​VP8, AV1, MPEG-2 video decoding and dual 1080p60 H.264/VP8 video encoding
    • NPU: Up to 6.75 TOPS
    • Manufactured using a 12nm process
  • Memory: 4 GB LPDDR4
  • Storage: 16GB eMMC flash, expandable via microSD and SPI flash
  • Video & Audio Interfaces: Micro HDMI 2.1 output (4Kp60), MIPI DSI interface, Micro HDMI input
  • Networking: Gigabit Ethernet RJ45 port, optional WiFi via USB dongle
  • USB: 4x USB 3.0 ports
  • Expansion: M.2 Key E socket (PCIe + MIPI CSI), 40-pin header
  • Misc: SPI BOOT, UBOOT, Reset buttons, power and activity LEDs
  • Power Supply: 5V/3A via USB Type-C port
  • Dimensions: 92 x 60mm
  • Weight: 48 grams

Banana Pi provides support for Android and Ubuntu 20.04, along with a Linux SDK that includes Kernel 5.4 and Buildroot 2019.10. Despite showcasing impressive AI capabilities, the board’s NPU utilization is somewhat limited due to a lack of publicly available development tools from Synaptics for the Vivante VIP9000 NPU integrated within the SoC.

Source: CNX Software – Embedded Systems News.