Qualcomm has launched its high-performance AI inference accelerator Cloud AI 100, the company said.
It will also be providing the edge development kit.
Cloud AI 100 will provide advanced signal processing and power efficiency to support for AI solutions from data centers, cloud edge, edge appliance and 5G infrastructure, the US chip giant said.
The chip will offer AI inference efficiency capability to process up to 400 trillion operations per second (TOPS) per Watt.
Cloud AI 100 supports multiple form factors with card level power profiles ranging from 15W to 75W.
It will support Tensorflow, PyTorch and Caffe tools as well as Glow and Onnx runtimes.
The software suite includes the Cloud AI 100 Applications SDK with compiler and simulator __ along with the Platform SDK with runtimes, APIs, kernel drivers and tools.
This software suite is designed to enable the development, testing and deployment of Cloud AI 100 platforms into a multitude of production environments, Qualcomm said.
The development kit is engineered to accelerate adoption of edge applications by offering a complete system solution for AI processing up to 24 simultaneous 1080p video streams along with 5G connectivity. The three modules of the kit are: Cloud AI 100 for low-power and high-performance AI inference processing, Snapdragon 865 Modular Platform for application and video processing as well as Snapdragon X55 ModemRF for 5G
Qualcomm has started shipment of Cloud AI 100 for some customers. The commercial version will launch in the first half of 2020.