인피니언 8월20일부터
This page was machine-translated and may differ from the original. View original

Broadcom Announces Server-Class ARMv8-A Multicore Processor Architecture

Google 우선 소스Published2013.11.04 14:12

- Optimized for next-generation networking and communication applications to achieve industry-leading performance
- 64-bit ARMv8-A core with quad-issue and quad-threading, providing ideal server-class performance with superscalar out-of-order execution capabilities
- Goal to establish and develop an ISA-independent open NFV software environment through a partnership with ARM

October 29, 2013 – Broadcom, a global wired and wireless semiconductor company, today announced the architecture for its next-generation multicore processor line featuring the industry's highest-performing 64-bit ARM cores. Using a license for the ARMv8-A architecture, Broadcom is developing new CPU cores that provide true server-class performance for networking, big data, communications, storage, and security applications, as well as for top-tier virtual accelerators and Network Function Virtualization (NFV) implementations.

Due to the exponential increase in global network traffic stemming from device-to-device connectivity and cloud services, service providers are adopting cloud-based networks that support increasingly fluid services and tasks. Broadcom plans to introduce a new line of processors to meet these next-generation virtualization trends. This new line of processors provides superior computational performance for control and data plane processing and is optimized to deliver the maximum throughput required by networking functions such as security and packet processing.

Ron Jankov, Senior Vice President of Processors and Wireless Infrastructure at Broadcom, stated, “Broadcom will maintain its position as a leader in multicore processing, which it has held for generations, through the design of industry-leading ARM-based multicore processors,” adding, “Broadcom’s innovations in CPU cores and comprehensive virtualization, along with the introduction of 16-nm FINFET technology, will further differentiate Broadcom from its competitors.”

Furthermore, Broadcom is collaborating with ARM to establish and develop an open standards-based NFV software environment for the ARM ecosystem. As members of the Linaro Networking Group and the European Telecommunications Standards Institute (ETSI), as well as leading service providers, Broadcom and ARM are working together to standardize programming models, toolchains, application programming interfaces, and detailed networking libraries for the NFV environment. Equipment manufacturers designing NFV solutions to meet next-generation carrier-grade requirements are expected to reap significant benefits, as enabling the software environment to be easily moved across various platforms and instruction sets improves code portability and reduces development time.

Tom Cronk, Senior Vice President of ARM’s Processor Business Unit, stated, “ARM and Broadcom have consistently maintained a close relationship,” adding, “ARM is very pleased that Broadcom has actively leveraged ARMv8-A designs to introduce high-performance processing solutions to the telecommunications market. Broadcom is actively cooperating to help ARM build an open standards-based NFV software environment.” “In today’s world, where everything is connected and requires vast amounts of data, this environment enables the development of intelligent and highly efficient technologies needed by businesses and consumers,” he stated.

Broadcom customers can easily migrate existing products based on the market-leading XLP II multi-core processors to the new ARMv8-A processor architecture without new software investment through an interoperable open software environment. The XLP980, the latest model in the XLP II family, delivers top-tier performance, including 160 Gbps of information throughput, comprehensive end-to-end virtualization of the CPU, accelerators, and I/O, and secure acceleration capabilities of up to 100 Gbps.

Linley Gwennap, Senior Analyst at Linley Group, stated, “As network bandwidth continues to increase and various services are launched, network managers are seeking to improve the efficiency and scalability of their networks through virtualization capabilities.” He added, “Broadcom’s design, featuring out-of-order instruction processing, quad-threading, and quad-issue, enables a new level of performance through the ARM architecture, giving it a competitive edge over other multi-core manufacturers. We expect this design, derived from 16-nm process and design innovations, to deliver top-tier performance in both single and multi-threaded workloads.”

본 기사에 대한 정정·반론·추후보도 청구는 보도 청구 안내를, 그간 게재된 보도문은 정정·반론보도 모아보기를 참고해 주세요.
명세환 기자
명세환 기자