Google added AMD’s second-generation EPYC processors to its virtual machine (VM)-hosting option list that is targeted at general purpose and high-memory bandwidth workloads. It also builds on hosting options added during the second half of last year.

The new N2D VMs can support machine types with up to 224 virtual CPUs (vCPUs), which is the largest general-purpose VM on Google’s Compute Engine. They also include local SSD, customer machine types, and live migration from the previous generation N2 VMs, though with lower costs and higher performance.

For high-performance computing (HPC) workloads, the N2D machine types can be configured with 128 and 224 vCPUs for 70% higher platform memory bandwidth than N2 instances, according to Google. And a higher core count can provide up to a doubling of performance compared with the previous iteration.

The N2D VMs build on Google’s N2 VMs launched last August. Those were based on second-generation Intel Xeon scalable processor machines that provided a performance boost compared with first-generation machines.

Google also offers its E2 general-purpose VMs that are targeted as a lower cost alternative to its N1 VMs. A Google spokesperson noted that E2 is for customers that don’t require large instance sizes, GPUs, or local SSD, and that they can run on either AMD or Intel processors.

AMD launched its second-generation 7nm EPYC processor architecture last August. They are designed to power servers, VMs, and data centers from many of the industry’s top vendors and hyperscale cloud providers. Google was one of more than 60 partners that were on board for the launch.

Patrick Moorhead, president and principal analyst at Moor Insights & Strategy, called the platform a “big step forward” for AMD in the data center and a “bigger leap forward” than he expected from the company overall.

“AMD gained low, single-digit share with first-generation EPYC but I expect the company to gain more share with second-generation EPYC with [communications service providers], enterprises, and [high-performance computing],” he wrote in an email to SDxCentral.

The launch was also part of a growing arms race among silicon providers looking to topple Intel’s x86 dominance in the data center space.

Google rival Amazon Web Services (AWS) late last year unveiled its purpose-built, Arm-based Graviton2 platform that will power instances within many of its core cloud services. AWS CEO Andrew Jassy claimed the new architecture provided a 40% better price performance than the latest generation of Intel-based x86 processors.

That price-performance gain is based on Arm’s 64-bit Neoverse N1 intellectual property (IP) platform that it initially unveiled in 2018. The platform was designed to be optimized for 7nm processor technology and focused on compute speed.

“While lots of companies including ourselves have worked with x86 processors for a long time … if we wanted to push the price performance envelope for you it meant we had to do some innovating ourselves,” Jassy said.

Moorhead noted that the traction being gained by AMD and Arm will put additional pressure on Intel, which is expected to launch a next-generation 10nm silicon architecture later this year. “Whatever the outcome, we now have even more competition in the data center processor market,” he explained in an email to SDxCentral tied to the AWS announcement.

That pressure will need to be intense if it's going to worry Intel. The chip giant recently reported record data center revenues for its most recent operating quarter.