Google Unveils ARM-based AI CPU, Google Axion
Google announced first ARM-based CPU Google Axion, as well as upgraded TPU AI chips, but it won’t be selling them to customers
Alphabet’s Google on Tuesday has unveiled an update of its data centre AI chip, as well as an ARM-based central processor.
Google said Axion is its first custom ARM-based CPU designed for the data centre, and it has also released new TPU AI chips (TPU v5p) that are used as alternatives to Nvidia’s GPUs for AI acceleration tasks.
The demand for AI chips is increasing and other players besides Nvidia are seeking to bolster their portfolio. This week Intel took the wraps off its latest artificial intelligence chip, called the Gaudi 3 AI accelerator.
Image credit Intel
Google CPU
Last month Nvidia introduced the next-generation AI accelerator chips. The new platform, “Blackwell”, powers the B200 GPU that is the successor to the current generation of “Hopper”-based H100.
Amazon and Microsoft have developed ARM CPUs as well.
Now on Tuesday Amin Vahdat, VP/GM, Machine Learning, Systems, and Cloud AI at Google blogged the unveiling of Google’s first ARM-based CPU, called Google Axion.
Image credit Google
It should be noted that Google has previously built custom chips for YouTube, AI and its smartphones, but until now had not developed a CPU. It is understood that the Google Axion chips are already powering YouTube ads, the Google Earth Engine, and other Google services.
“Today, we are thrilled to announce … Google Axion Processors, our first custom ARM-based CPUs designed for the data centre,” wrote Vahdat. “Axion delivers industry-leading performance and energy efficiency and will be available to Google Cloud customers later this year.”
Google TPU
Vahdat noted that Axion is but the latest in a long line of custom Google silicon, as Google since 2015 has released five generations of Tensor Processing Units (TPU).
Reuters reported that Google’s tensor processing units (TPUs) are one of the few viable alternatives to the advanced AI chips made by Nvidia, though developers can only access them through Google’s Cloud Platform and not buy them directly.
A single TPU v5p pod contains 8,960 chips, which is more than double the amount of chips found on the TPU v4 pod.
“We’re making it easy for customers to bring their existing workloads to ARM,” Mark Lohmeyer, Google Cloud’s vice president and general manager of compute and machine learning infrastructure told Reuters. “Axion is built on open foundations but customers using ARM anywhere can easily adopt Axion without re-architecting or re-writing their apps.”
Google claims Axion processors combine Google’s silicon expertise with ARM’s highest performing CPU cores to deliver instances with up to 30 percent better performance than the fastest general-purpose ARM-based instances available in the cloud today.
It also claimed that Axion also offers up to 50 percent better performance and up to 60 percent better energy-efficiency than comparable current-generation x86-based instances
Google plans to offer Axion via the Google Cloud.
