Marvell has joined the Universal Chiplet Interconnect Express (UCIe) Consortium as part of its ongoing development of open chiplet interconnect standards.
Marvell said its contributions to the UCIe standard will leverage the company’s advanced chiplet interconnect and packaging experience to help further the consortium’s objective of developing open standards that create a more robust ecosystem for inter-operable chiplets.
Marvell’s UCIe participation is complementary to the company’s existing work in the Open Compute Project (OCP) and Optical Internetworking Forum (OIF) which is aimed at the development of open standards-based solutions for connecting single or multi-node 5nm and 3nm chiplets. Through these experiences, Marvell brings advanced domain knowledge and a unique perspective that will help align and optimize global chiplet interconnect standards for a range of leading-edge applications with differing requirements such as CXL, Ethernet and custom low-latency connectivity while fostering chiplet interoperability.
“Marvell has been an industry pioneer in chiplet connectivity and continues to push the envelope of performance optimization for a wide range of multi-chiplet applications in advanced packaging architectures,” said Noam Mizrahi, Chief Technology Officer and Senior Fellow at Marvell. “We see the great value in aligning interconnect standards across the industry and look forward to contributing towards that goal as a member of the UCIe consortium.”
“UCIe represents the culmination of years of learning and implementation experience with on-package interconnects at a time that is right for industry standardization,” said Dr. Debendra Das Sharma, Intel Senior Fellow. “We are excited that Marvell has joined the consortium and is working to develop cloud-optimized multi-die solutions that are compatible with UCIe.”
https://www.marvell.com/products/custom-asic.html
NVIDIA opens its NVLink die-to-die and chip-to-chip
NVIDIA will expand the use of its NVLink chip-to-chip and die-to-die interconnect technology in its GPUs, CPUs, DPUs, NICs and SOCs. The company also plans to open the technolgy to other for for custom chip and chiplet integrations. NVIDIA NVLink-C2C is built on top of NVIDIA’s world-class SERDES and LINK design technology, and it is extensible from PCB-level integrations and multichip modules to silicon interposer and wafer-level connections,…
