[ad_1]
Giant of e-commerce and cloud computing Amazon.com (NASDAQ: AMZN) just announced that the Amazon Alexa digital assistant runs on Amazon’s own hardware instead of chips designed by Nvidia (NASDAQ: NVDA). In a blog post for Amazon Web Services (AWS) developers on November 12, technical evangelist Seb Stormarcq said that “the vast majority” of Alexa’s machine learning workloads now run on Amazon’s AWS Inferentia chips.
What’s up?
To be clear, nothing has changed in Amazon Echo devices and other Alexa-powered gear that you might buy for the holidays. The silicon shift happened on the back-end of Alexa’s services, where data is sent to AWS cloud systems for final processing. Inferentia was explicitly designed to run neural network software, which is how Alexa learns to interpret voice commands.
According to the first tests from Amazon, the new Inferentia clusters provide the same results as Nvidia’s T4 chips, but with 25% lower latency and 30% lower cost. The lower latency will allow Alexa developers to perform more advanced analysis of incoming data without letting the user wait for a slow calculation.
Image source: Getty Images.
The backstory
Amazon launched the Inferentia line of processors two years ago, with the aim of maximizing the processing speeds of the company’s artificial intelligence workloads while reducing costs by removing the middleman in the design process. fleas. The original designs come from Annapurna Labs, a specialist chip designer that Amazon acquired in 2015.
Alexa is not the first Amazon product to rely on Inferentia-based AWS Inf1 instances. Amazon’s facial recognition tool, Rekognition, is also moving to Inf1 instances. AWS customers are also free to use Inf1 and Inferentia for their own projects. For example, the Snapchat parent Break (NYSE: SNAP), health insurance giant Anthem (NYSE: ANTM)and global publishing house Conde Nast are already using Amazon’s Inferentia-based neural network instances to boost their artificial intelligence projects.
[ad_2]
Source link