Arm’s influence is growing across major public clouds around the world: AWS, Microsoft, Google, Alibaba, and Oracle. The Neoverse platform is gradually expanding into computing for high-performance computing (HPC) and next-generation infrastructure, as well as cloud infrastructure markets.

On the 15th, Arm announced new updates to its Arm Neoverse roadmap and next-generation infrastructure technologies at an online press conference. The presentation provided insights into Arm’s Neoverse updates and added details about its partner ecosystem.

■ Neoverse Platform, Accelerating the Expansion of Infrastructure Computing

To address the needs of hyperscale and HPC customers and to further enhance cloud workload performance without requiring more power or space, Arm has unveiled the Arm Neoverse V2 platform, featuring the latest V-series cores and the widely deployed Arm CMN-700 mesh interconnect.

“Recently, several partners have been working on Arm Neoverse-based designs,” said Hwang Sun-wook, country manager of Arm Korea, citing Nvidia, which is utilizing Neoverse V2 as the compute foundation for its Grace data center CPU. The Grace chip for AI HPC combines the power efficiency of Neoverse V2 with the power efficiency of LPDDR5X memory to deliver twice the performance per watt of servers running on existing architectures.

Additionally, as part of Arm’s continued investment in efficient performance and throughput, Arm introduced that the next-generation N-series products are in development and are scheduled to be available to partners in 2023. This next-generation N-series CPU is expected to provide generational improvements in both performance and efficiency, beyond the market-leading efficiency of the N2.

Arm's Neoverse E-series CPUs are versatile across data plane processing, 5G RAN, edge networking, and accelerators. To support this diversity, Arm offers the Neoverse E2 platform, which combines the Arm Cortex-A510 CPU with the scalable Neoverse CMN-700 and N2 system backplane.

The Neoverse platform is purpose-built to handle the full spectrum of accelerated infrastructure solutions. This specialized processing is the solution to reduce workload power consumption and meet the increasing compute demands of modern workloads.

“Arm is already a new leader in the infrastructure sector, and the future of infrastructure computing will be built on Arm,” said CEO Hwang. “Future infrastructure must be different from the past,” he said, emphasizing the need for ubiquity, acceleration, and power efficiency.

■ Cloud-native software running on Arm

Arm is working with major cloud providers to optimize cloud-native software infrastructure, frameworks, and workloads. This includes support for widely adopted projects such as Kubernetes, Istio, and several CI/CD tools to provide native builds for the Arm architecture.

It also includes improvements for cloud software workloads such as machine learning frameworks like TensorFlow, open source databases, big data and analytics, and media processing. By working with developers who contribute to the open source community to optimize full-stack software solutions running on Arm, we support all end users in building next-generation applications on Arm.

The strength of the Arm software ecosystem is the combination of the capabilities of the Neoverse platform with leading high-performance fabrics such as AMBA CHI, UCIe, and CXL, enabling Arm to enable partners to develop tailored solutions across the cloud-to-edge.

■ Customized solutions from cloud to edge

Neoverse-based platforms are redefining what’s possible in cloud computing by delivering industry-leading scalable efficiency and enabling the Arm ecosystem with the freedom to innovate across all infrastructure domains.

As a result, Neoverse has seen rapid adoption since its launch in 2018, with every major public cloud offering instances based on Neoverse today.

In 5G RAN and wireless infrastructure, Arm can deliver leading performance with high throughput at low power, providing significant total cost of ownership (TCO) benefits. This enables Arm's partners to customize their products for applications and use cases ranging from small cells and macro cells to private networks.

Through the Arm 5G Solutions Lab, Arm is working with leaders in the 5G ecosystem, including Dish Wireless, Google Cloud, Marvell, NXP, and Vodafone, to accelerate end-to-end 5G networks on Arm.

Moving forward, Arm is moving into more traditional enterprise territory with the launch of the HPE Proliant 11th generation platform based on Ampere Altra, as well as initiatives such as VMware’s Project Monterey, RedHat supporting OpenShift on Arm, and SAP HANA migrating its cloud infrastructure to AWS Graviton based on Arm.

“The industry is increasingly focused on performance, power efficiency, specialized processing and workload acceleration, enabling Arm Neoverse to redefine and transform the world’s compute infrastructure,” said Chris Bergey, senior vice president and general manager, Infrastructure Business Unit, Arm. “With the world’s largest compute footprint and unparalleled understanding of the entire compute spectrum, the next era of infrastructure technology relies on Arm’s performance and efficiency DNA.”