Photonics die or integrated photonics modules co-packaged with compute engines have the potential to deliver significant improvements in power, bandwidth and reach needed to meet the computing and communication demands of data centers and other high-performance computing (HPC) systems. The challenges and solutions in co-packaging photonics modules are described through two case studies; one of a network-switch die co-packaged with socketable photonics modules and another of a Field Programmable Gate Array (FPGA) co-packaged with optical dies (tiles). The technical requirements to deliver the promise of co-packaged photonics in high volume are outlined.
22 April 2025 in IEEE Antennas and Propagation Magazine
This article describes current initiatives to form benchmarks for the characterization of electronic packages. Four available benchmarks representing typical cases for signal integrity (SI) and power integrity (PI) analysis and varying in complexity from simple microstrip line to a full package model are intended to serve as standardized cases for testing the performance and accuracy of the modeling tools. While the benchmarks, consisting of CAD model files as well as simulated and measured network parameters, are publicly available and described in the accompanying manual, the emphasis in this article is on the description of challenges encountered in creating measured data, the development of modeling capabilities in the computational tools essential for accurate and expedient electromagnetic (EM) analysis of the benchmarks, and common practices going into matching of the simulated to the measured data. Ongoing efforts toward the standardization of densely packed interconnects for die-to-die interfacing on advanced packages, also known as heterogeneous integration (HI), are discussed in the context of the Open Compute Project (OCP). Emerging needs for benchmarking and standardization of test cases and training datasets for machine learning (ML)-assisted characterization of high-speed channels are also outlined.
30 July 2021 in IEEE Transactions on Components, Packaging and Manufacturing Technology
The ever-increasing demand for higher bandwidth and lower loss makes the predictability of high-speed interconnect performance increasingly challenging. Validating models against measurements of manufactured test structures requires not only accurate methodology but also understanding of the uncertainty impact. S-parameter measurements, characterization of materials, manufacturing processes, and models constructed based on measured inputs all cause uncertainty in the performance metrics of interest. Therefore, it is critical to anticipate the results with the associated uncertainty to assess the correlation quality more objectively. This paper investigates the reproducibility of measurements required for high-speed package interconnect validation through rigorous analysis and presents a methodology to quantify the impact of measurement uncertainty on commonly used performance metrics.