Try our mobile app

Microchip Extends Leadership in Data Center Connectivity with Industry’s Lowest Latency PCI Express® 5.0 and CXL™ 2.0 Retimers

Published: 2020-11-10 14:00:00 ET
<<<  go to MCHP company page

XpressConnect™ retimers cost-effectively triple high-speed signal reach

CHANDLER, Ariz., Nov. 10, 2020 (GLOBE NEWSWIRE) -- As the high-performance computing demands of data center workloads increase, new ultra-low-latency signal transmission technology is required to advance the performance in Artificial Intelligence (AI), Machine Learning (ML), Advanced Driver Assisted Systems (ADAS) and other computational workload applications. To address this need, Microchip Technology Inc. (Nasdaq: MCHP) announced today the XpressConnect family of low-latency PCI Express (PCIe®) 5.0 and Compute Express Link™ (CXLTM) 1.1/2.0 retimers. 

XpressConnect retimers triple the reach of PCIe Gen 5 electrical signals, which enables data center equipment providers to harness the next generational advancement in compute IO performance while providing the required flexibility and connectivity for advanced hardware architectures. XpressConnect retimers deliver ultra-low-latency signal transmission that enables the most demanding computational workloads in AI, ML, communication systems and high-performance computing applications.

“With increasing storage, compute and memory bandwidth demands, hyperscale data center, server and storage providers have continued to look to Microchip for leading compute, storage and memory connectivity solutions,” said Andrew Dieckmann, associate vice president of marketing and applications engineering for Microchip’s data center solutions business unit. “Microchip is uniquely positioned to deliver seamless interoperability with XpressConnect retimers given our extensive portfolio of PCIe SwitchtecTM and Flashtec® products and our close relationships with industry partners, providing lower engineering costs and faster time to market for our customers.”

The XpressConnect family delivers extended reach at >80% lower latency than the PCIe specification, with a pin-to-pin latency of