LONDON – Fabless network processor company Netronome Systems Inc. (Santa Clara, Calif.) has announced details of the NFP-6xxx family of flow processors, which operate at up to 200-Gbits per second. The chips are made using Intel's 22-nm FinFET manufacturing process technology. Netronome claims they provide more than six times the packet processing performance at less than half the power of alternatives made using 28-nm CMOS.
The NFP-6XXX product line includes an ARM11 multiprocessing core together with 256-kbytes of L2 cache as its housekeeping processor. As a result the most advanced ARM processor yet made, at least in terms of process technology, has been made by Intel, often seen as a rival to ARM.
The NFP-6xxx product line is designed for networking applications that require programmable, line-rate processing on millions of simultaneous, complex flows. It is aimed at equipment with data rates of 10- to 400-Gbits per second. The NFP-6xxx provides a software-compatible migration for customers from the NFP-32XX range which also used the ARM11MP core.
Click on image to enlarge.
Block diagram of the NFP-6XXX. Source Netronome
However, the NFP-6XXX does most of its network processing work using 216 programmable cores – 120 flow processing cores with 960 threads, and 96 packet processing cores – delivering over 307 billion instructions per second to 384 million packets per second, the company states.
It is suitable for line cards, service blades and PCIe cards in applications such as carrier networking, cyber security, mobile networking, virtualized datacenters and cloud computing, the company said.
The processor also includes 100 programmable hardware accelerators for deep packet inspection, regular expression matching, optimized I/O packet transfers, traffic management, security processing and bulk cryptography and over 30-Mbytes of on-chip storage.
The NFP-6xxx is supported by software development tools, including a graphical integrated development environment, C-compiler, packet generator, and cycle- and timing- accurate simulators.
"We experienced tremendous growth last year as a direct result of our early commitment to flow processing and its supporting building blocks. The NFP-6xxx strengthens this commitment and extends the benefits of high throughput flow aware processing to our leading edge customers," said Howard Bubb, CEO of Netronome, in a statement. "With a 10x increase in flow processing capabilities, the NFP-6xxx will accelerate the movement from packet to flow-aware processing for a wide range of advanced network services." Related links and articles:
News articles: Netronome to use Intel 22-nm process in foundry deal
Netronome is revealed as ARM multiprocessor licensor
Netronome reduces SoC power use with timing tricks
Report: ARM aims to take 20 percent of notebook PC market
Intel vying for Apple's foundry business