Skip to main content

Arm reveals its latest server chips for data centers, edge, HPC and more

servers
(Image credit: Shutterstock.com / Gorodenkoff)

Arm has announced a completely new chip architecture as it aims to continue its progress into the data center and high-performance computing (HPC) space.

Dubbed the Neoverse V1, Arm claims the new chip will outpace its current server space contender, the Neoverse N1, offering twice the floating-point performance as well as a four-fold increase in machine learning performance.

Arm first introduced the Neoverse N1 back in early 2019 in the form of Amazon’s Graviton2 and Ampere’s Altra server processors. As per AnandTech, the Graviton2 became the dominant hardware deployment in Amazon’s AWS additions in 2020, picking up the majority of market share being lost by Intel.

Talking to TechCrunch, Chris Bergey, SVP and GM of Arm’s Infrastructure Line of Business, shared that while the new Neoverse V1 wasn’t specifically built for the HPC market, it was definitely a target market. 

“The V1 is about how much performance can we bring — and that was the goal,” shared Bergey, noting that the V1 is Arm’s widest architecture to date.

Server performance

The current Neoverse V1 platform isn’t based on the new Armv9 architecture, but Bergey shares that the next generation of the chip will be.

Alongside the V1, Arm also announced the Neoverse N2, which it pitches as the next-generation general compute platform that is meant for a wide variety of use cases, ranging from hyperscale clouds to SmartNICs and edge devices. 

The neoverse N2 is also notable for being the first chip that’s based on the new Armv9 architecture.

Bergey pointed out that with the N2 the objective was to get the most performance per watt. He illustrates this by sharing that in-house testing revealed that the Nginx web server showed a 1.3 times performance increase as compared to the Neoverse N1.

Via TechCrunch