User Manual

6 AI accelerator card
1.1 Introduction
ASUS AI accelerator card is a full height, half length, PCIe Gen3 AI
acceleration card based on Google Coral Edge TPU processor. It allows
users to build intelligent devices that are ecient, private and ideal for
edge AI inference. With twin Axial-tech turbofans which propel cool air
over a Copper heatsink in order to dispatch the most heat from the cores,
the optimized thermal solution guarantees the best performance at high
workload.
The heart of ASUS Edge AI accelerators is the Edge TPU coprocessor.
It’s a small-yet-mighty, low-power ASIC (Application-specic integrated
circuit) that provides high performance neural net inferencing. Edge TPU is
optimized for TensorFlow Lite models, which means common ML models
can be easily compiled and run on the ASUS AI accelerator card.
1.2 Features
CRL-G18U-P3DF/CRL-G116U-P3DF features are listed below
• Eight Google Coral Edge TPU
• Interface: PCIe Gen. 3 x16
• Dual Active Turbofan
• High performance: INT8, 32TOPS/64TOPS
• Supports TensorFlow Lite framework for on-device deep learning
inference
• Easy IoT integration between edge AI solution and cloud TPU