The NPU (Neural Processing Unit), as a dedicated AI acceleration chip, requires its hardware interface design to meet the demands of high bandwidth, low latency, and flexible scalability. Below are the commonly used interface signals and corresponding protocols for NPUs at the hardware level, explained with actual chip examples: 1. Memory Interface Function: Connects to … Read more
CPC (Co-Packaged Copper Interconnect) is a high-speed interconnect technology widely used in AI computing power and data center scenarios, supporting high-speed data transmission from 224Gbps to 448Gbps. With the explosive growth in demand for AI servers, CPC technology has become key to enhancing data transmission efficiency, with the market size expected to reach $2 billion … Read more
As the market revels in the computing power of AI chips and marvels at each appearance of humanoid robots, attention often focuses on the flashy brands. However, the true revolution lies behind the scenes with countless invisible giants providing the “infrastructure.” Luxshare Precision (002475.SZ), a traditional leader in consumer electronics, is quietly positioning itself in … Read more
Source: LightSense Technology Original Author: LIG This article introduces the Waveguide Embedded System (WES) based on optical PCBs, designed for AI/HPC data centers to overcome the challenges of CPO integration. WES achieves high-density, low-loss, fiber-free optical interconnects between devices by integrating an optical engine with precise coupling structures. The Waveguide Embedded System (WES) is designed … Read more