New Google Cloud VMs With AMD EPYC Promise Top Price-Performance
The cloud service provider says the new class of virtual machines provides better performance and price-performance than Arm- and Intel-powered VMs from its top competitors. This is not the first time Google Cloud has chosen AMD over Intel for a new class of products. ‘With Tau VMs, Google Cloud has set a new bar for price-performance,’ says one Google Cloud consulting partner.
Google Cloud said a new class of virtual machines powered by AMD’s latest EPYC processors provide better performance and price-performance than VMs powered by Intel and Arm from other top cloud service providers by roughly 50 percent.
The Mountain View, Calif.-based cloud provider announced the new Tau VM product family as an extension of the Google Compute Engine’s VM offerings, saying they will arrive in the third quarter and take advantage of AMD’s new third-generation EPYC CPUs, code-named Milan.
[Related: AMD: Enterprise Server Sales Will Catch Up With Overall EPYC Growth]
This is not the first time Google Cloud has chosen AMD over Intel for a new class of products. Last summer, the cloud provider launched Confidential VMs as part of its new Confidential Computing portfolio, and it turned to EPYC CPUs to power the product’s memory encryption capabilities.
“We work closely with Google Cloud and are proud they selected AMD to exclusively power the new Tau VM T2D instance which provides customers with powerful new options to run their most demanding scale-out workloads,” AMD CEO Lisa Su said in a statement.
In a blog post, Google Cloud executives Urs Hölzle and June Yang said AMD’s EPYC Milan CPUs will power T2D, the first instance in the new Tau VM family, and it “leapfrogs the VMs for scale-out workloads of any leading public cloud provider available today, both in terms of performance and workload total cost of ownership.”
To make their point, they said the new EPYC-powered instance provides 56 percent higher “absolute performance” and 42 percent higher price-performance when compared to general-purpose VMs from two unnamed leading cloud vendors, one with an Arm-based CPUs and the other with Intel’s second-generation Xeon Scalable CPUs, which is one generation behind Intel’s new Ice Lake CPUs. The comparisons were based on estimates from the SPECRATe2017_int-base benchmark.
The Google Cloud executives also shared a price-performance comparison using the CoreMark benchmark, showing that the T2D instance has 38 percent better price-performance than the VM from the first cloud vendor using Arm-based processors.
While Google Cloud didn’t identify the names of the “leading cloud vendors” in the comparison, Amazon Web Services is the only player out of the top three cloud providers that has publicly available instances based on Arm-based CPUs. AWS has been designing its own Arm-based CPUs, and they power several instances now.
Hölzle and Yang said the new EPYC-powered T2D instance has another benefit over Arm-powered instances: no need to port applications to a new processor architecture. The point was echoed by Thomas Kurian, CEO of Google Cloud, in a press release from AMD.
“At Google Cloud, our customers’ compute needs are evolving,” Kurian said in a statement. “By collaborating with AMD, Google Cloud customers can now leverage amazing performance for scale-out applications, with great price-performance, all without compromising x86 compatibility.”
To show support for the new Tau VM family, Google Cloud gathered quotes from two customers, Twitter, and a cloud consulting partner, DoIT International.
“In our preliminary testing of Google’s new Tau VMs with the CoreMark benchmark, we were thrilled to see the incredible performance at 50 percent better than a comparable Arm-based offering from another leading public cloud,” said Yoav Toussia-Cohen, CEO of DoIT International, in a statement. “With Tau VMs, Google Cloud has set a new bar for price-performance, making the cloud even more accessible to digital-native companies. We are excited to bring Google’s Tau VMs to our joint customers.”