Overview
Synopsys UCIe PHY IP enables high-bandwidth, low-power and
low-latency die-to-die connectivity in a package for hyperscale data centers, AI, and networking applications. The PHY’s flexible architecture supports standard and advanced package technologies and delivers up to 4Tbps bandwidth
in a multi-module configuration. Supporting widely used protocols such as PCI Express and CXL and enabling latency-optimized NoC-to-NoC links with streaming protocols, the IP offers maximum performance, minimum latency and implementation flexibility. Synopsys UCIe PHY IP delivers high energy- efficiency with an optimized architecture that uses clock forwarding and
low-voltage signaling. The IP implements a comprehensive set of testability features to ensure known good dies and offers test and repair capabilities to improve package assembly yield. Robust die-to-die link operation is ensured with embedded training and calibration algorithms.
Synopsys UCIe PHY IP interoperates with Synopsys UCIe Controller IP to deliver a complete, low-latency solution for die-to-die links in any package.
Learn more about UCIe IP core
Discover Synopsys 64G UCIe IP for energy-efficient, high-bandwidth die-to-die connectivity in advanced multi-die AI and HPC designs.
In keeping with a rapidly broadening portfolio of die-to-die connectivity solutions, Cadence has taped out its IP subsystem for 32GT/s UCIe solution on Samsung's 4nm (SF4X) process technology.
Understand the role of the UCIe D2D Adapter in enabling reliable, scalable, multi-protocol die-to-die communication for chiplet architectures.
Cadence announces that it has taped out its IP subsystem for the 32G UCIe advanced package on TSMC’s 3nm (N3P) process technology.
Cadence has successfully demonstrated first-pass silicon success of its UCIe™ standard package IP on Samsung Foundry's 5nm automotive process. This milestone underscores Cadence's commitment to delivering high-performance, automotive-grade IP solutions that meet the stringent requirements of next-generation automotive and high-performance computing applications.
For AI workloads to be processed reliably at a fast rate, the die-to-die interface in multi-die designs must be robust, low latency, and most importantly high bandwidth. This article outlines the need for 40G UCIe IP in AI data center chips leveraging multi-die designs.