The Coral System-on-Module (SoM) is a fully-integrated system that helps you build embedded systems that demand fast machine learning (ML) inferencing.
Provides a complete system: The Coral SoM is a fully-integrated Linux system that includes NXP's iMX8M system-on-chip (SoC), eMMC memory, LPDDR4 RAM, Wi-Fi, and Bluetooth, and the Edge TPU coprocessor for ML acceleration. It runs a derivative of Debian Linux we call Mendel.
Performs high-speed ML inferencing: The on-board Edge TPU coprocessor is capable of performing 4 trillion operations (tera-operations) per second (TOPS), using 0.5 watts for each TOPS (2 TOPS per watt). For example, it can execute state-of-the-art mobile vision models such as MobileNet v2 at 400 FPS, in a power efficient manner.
Integrates with your custom hardware: The SoM connects to your own baseboard hardware with three 100-pin connectors.
Supports TensorFlow Lite: No need to build models from the ground up. TensorFlow Lite models can be compiled to run on the Edge TPU.
Supports AutoML Vision Edge: Easily build and deploy fast, high-accuracy custom image classification models to your device with AutoML Vision Edge. Also available with a baseboard as part of the Coral Dev Board.
The Coral System-on-Module (SoM) is a fully-integrated system that helps you build embedded systems that demand fast machine learning (ML) inferencing. It contains NXP's iMX8M system-on-chip (SoC), eMMC memory, LPDDR4 RAM, Wi-Fi, and Bluetooth, but its unique power comes from Google's Edge TPU coprocessor.
The Edge TPU is a small ASIC designed by Google that provides high-performance ML inferencing with a low power cost. For example, it can execute state-of-the-art mobile vision models such as MobileNet v2 at almost 400 FPS, in a power-efficient manner. This on-device processing reduces latency, increases data privacy, and removes the need for a high-bandwidth connection used to perform ML inferencing in the cloud.
Key benefits of the SoM:
High-speed and low-power ML inferencing (4 TOPS @2 W)
A complete Linux system (running Mendel, a Debian derivative)
Small footprint (40 x 48 mm)
The SoM is also included in the Coral Dev Board, which is a single-board computer that enables fast prototyping and evaluation of the standalone SoM.
NXP i.MX 8M SoC
Quad-core ARM Cortex-A53, plus Cortex-M4F
2D/3D Vivante GC7000 Lite GPU and VPU
Google Edge TPU ML accelerator
Wi-Fi 2x2 MIMO (802.11b/g/n/ac 2.4/5 GHz)
HDMI and MIPI-DSI
Up to 95x GPIO (including SPI, I2C, PWM, UART, SAI, and SDIO)
Main system-on-chip (i.MX8M)
Arm Cortex-A53 MPCore platform
Quad symmetric Cortex-A53 processors:
32 KB L1 Instruction Cache
32 KB L1 Data Cache
Support L1 cache RAMs protection with parity/ECC
Support of 64-bit Armv8-A architecture:
1 MB unified L2 cache
Support L2 cache RAMs protection with ECC
Frequency of 1.5 GHz
Arm Cortex-M4 core platform
16 KB L1 Instruction Cache
16 KB L1 Data Cache
256 KB tightly coupled memory (TCM)
Graphic Processing Unit (GPU)
267 million triangles/sec
32 GFLOPs 32-bit or 64 GFLOPs 16-bit
Supports OpenGL ES 1.1, 2.0, 3.0, 3.1, Open CL 1.2, and Vulkan
Video Processing Unit (VPU)
4Kp60 HEVC/H.265 main, and main 10 decoder
4Kp60 VP9 and 4Kp30 AVC/H.264 decoder (requires full system resources)