Alibaba Group has unveiled its first AI inference chip, which it says will enhance cloud computing power and boost search, recommendation and customer service for e-commerce.
Developed by T-Head under the Alibaba DAMO Academy, the AI inference chip – a neural processing unit (NPU) named Hanguang 800, that specializes in the acceleration of machine learning tasks – is currently being used internally within Alibaba’s business operations, especially in product search and automatic translation on e-commerce sites, personalized recommendations, advertising, and intelligent customer services.
These areas require extensive computing power for the AI tasks to optimize the shopping experience, said Jeff Zhang, Alibaba Group CTO and president of Alibaba Cloud Intelligence.
“The launch of Hanguang 800 is an important step in our pursuit of next-generation technologies, boosting computing capabilities that will drive both our current and emerging businesses while improving energy-efficiency, ” said Zhang.
Propelled by a self-developed hardware framework, as well as highly-optimized algorithm designs that are tailored for business applications such as retail and logistics in the Alibaba ecosystem, the Hanguang 800 has recorded computing performance reaching 78,563 IPS at peak, while the computation efficiency was 500 IPS/W during the Resnet-50 Inference test.
In practical terms, that means the AI chip can categorize the large volume of product images uploaded to Taobao (around a billion per day), tailor search and personalize recommendations for hundreds of millions of consumers in five minutes. The same task now takes one hour.