Blog Review: June 26

Cadence’s Neelabh Singh examines the Gen4 link recovery mechanism in USB4 Version 2.0, an autonomous process that is initiated by a router when it encounters uncorrectable error events, and identified verification challenges.

Synopsys’ Gary Ruggles and Priyank Shukla highlight improvements to PCIe 7.0 that will enable secure data transfers and boost bandwidth for the next generation of AI and HPC chips.

Siemens’ Fabrice Gallo and Chiel Verhoeven point to the importance of protecting EV battery packs, often integrated into the vehicle’s floor, from damage caused by collisions, rough roads, and speed bumps.

Arm’s Ravi Malhotra anticipates that the next big thing in AI will be a shift to small language models (SLMs) that are fine-tuned with context-specific data to focus on specific applications so they require less energy to train and can be run directly on mobile devices.

Ansys’ Raha Vafaei checks out how radar systems and wireless communications design can be improved though a combination of computer-aided simulations and on-demand synthetic data generation for training AI/ML prediction of coverage patterns and network design optimization.

Keysight’s Emily Yan finds that testing and validation are key to enabling software-defined vehicles and instilling confidence in new ADAS features, but the integration of many ECUs, wireless technologies, sensors, and automotive operating systems presents challenges.

In a blog for SEMI, Intel Capital’s Jen Ard highlights the nine finalists of the Startups for Sustainable Semiconductors program, with solutions ranging from PFAS destruction to epi-layer separation.

Plus, check out the blogs featured in the latest Manufacturing, Packaging & Materials newsletter:

Amkor’s BeomSeok Kim, SeongHwan Kim, Unki Kim, SeongBeom Cho, DongHo Seo, and SangHun Yun show how to measure the impedance, capacitance, and inductance of active and passive elements.

Lam Research’s Taeyon (TY) Oh looks at parasitic defects in DRAM and the effect of process variability on device performance at both the cell center and the cell edge.

Synopsys’ Dwight Hunter explains how fabs can ramp new processes more quickly.

Tignis’ Charlie Parker explores how AI/ML fits into the fab, from inventory management to institutional knowledge capture, but warns of potential pitfalls ahead.

eBeam Initiative’s Toru Fujimori examines continued pattern shrinkage and the impact of stochastics in lithography.

SEMI’s Clark Tseng suggests looking beyond chip production when commercializing advanced technologies.