Google First to Upgrade Cloud Data Centers with Intel’s Latest Chips

Claims internal tests showed improved application performance by up to 30 percent.

Yevgeniy Sverdlik

February 24, 2017

1 Min Read
Google First to Upgrade Cloud Data Centers with Intel’s Latest Chips
Attendees at Google Cloud Platform Next 2016 conference in San Francisco in March viewing the 360-degree virtual tour of a Google data center (Photo: Yevgeniy Sverdlik)

Google has upgraded servers in cloud data centers across five availability regions with Intel’s latest Xeon processors, codenamed Skylake. The company claims it is the first cloud provider to do so.

Amazon said last year it expected to launch Skylake-powered C5 instances on its Amazon Web Services cloud sometime in early 2017. Microsoft has not revealed plans to upgrade to Skylake, but the blog AnandTech has deduced from a company blog post that Intel’s latest and greatest in data center tech is likely to appear in the next-generation Open Compute servers the giant said were in the works last November under the codename Project Olympus.

More on Project Olympus: Microsoft's New Cloud Server Design is Half-Baked, and That's the Point

The processors are geared for workloads that require high performance, such as scientific modeling, genomic research, 3D rendering, data analytics, and engineering simulations, Urs Hölzle, Google’s senior VP of cloud infrastructure, wrote in a blog post.

These applications will benefit from the new chips’ Advanced Vector Extensions (AVX-512) feature. In Google’s internal tests the feature improved application performance by up to 30 percent, Hölzle said.

Google optimized Skylake for all its Google Compute Engine VMs, including standard, highmem, highcpu, and Custom Machine Types. Cloud servers powered by Skylake are initially available in five Google cloud regions: Western US, Eastern US, Central US, Western Europe, and Eastern Asia Pacific.

This is a second major processor upgrade announcement from Google's cloud services division this week. On Tuesday, the company said it had added the option to spin up bare-metal GPUs along with cloud VMs for machine learning and other compute-heavy applications.

See alsoHow to Get a Data Center Job at Google

Read more about:

Google Alphabet
Subscribe to the Data Center Knowledge Newsletter
Get analysis and expert insight on the latest in data center business and technology delivered to your inbox daily.

You May Also Like