Technology

Addressing Silent Data Corruption (SDC) with In-System Embedded Deterministic Testing

Source

SemiWiki

Published

Jan 1, 2026

TL;DR

AI Generated

Silent Data Corruption (SDC) is a critical issue in semiconductor design, especially in high-performance computing environments like AI data centers, where errors can occur without detection, leading to significant failures. In a presentation by Broadcom Inc. and Siemens EDA, it was revealed that SDC incidents can disrupt mission-critical operations, such as AI model training. To combat this, in-system testing capabilities are recommended for periodic checks without system downtime, with Siemens’ In-System Test (IST) solution being a key enabler. Challenges in design implementation, such as functional isolation and clock splitting, were addressed in Broadcom's methodology, showcasing the successful implementation and verification of the solution. The collaboration highlights the importance of embedded deterministic testing in improving reliability and reducing field failures in next-generation silicon.

Read Full Article

Dr. L.C. Lu on TSMC Advanced Technology Design Solutions

Dr. L.C. Lu, a key figure at TSMC, focuses on design-technology co-optimization, packaging innovations, and AI-driven methodologies for next-gen semiconductor systems. TSMC emphasizes DTCO and DDCL innovations for scaling from N5 to A14 nodes, with NanoFlex and NanoFlex Pro architectures offering efficiency gains. N2P and N2U nodes incorporate advanced DTCO and power delivery optimizations, with hybrid dual-rail architectures achieving significant energy savings. TSMC collaborates with EDA partners for AI integration, enhancing productivity and design quality. Advanced packaging technologies like CoWoS and SoIC play a crucial role in enabling AI scaling, with memory bandwidth and interconnect performance scaling aggressively. TSMC addresses power delivery and thermal management challenges in AI systems through advanced solutions. TSMC's advancements in design methodologies and AI-driven automation promise improved productivity and scalability in chip-package co-design.

SemiWiki•

5 hours ago

Intel details 18A-P process node, touts higher performance, lower power, and better thermals — 9% more performance, thermal conductivity improved by 50%

Intel is progressing with its 18A-P process node technology, boasting higher performance, lower power consumption, and improved thermals compared to its baseline 18A technology. The 18A-P process introduces new transistors, tighter process variability control, and enhanced thermals, potentially attracting interest from companies like Apple. This enhanced technology offers a 9% performance boost or an 18% power reduction compared to the base 18A process. Additionally, improvements in skew corner tightening, threshold voltage options, and metal stack resistance and capacitance contribute to better performance efficiency and yield. The 18A-P process also focuses on enhancing thermals, reliability, and voltage behavior critical for both client and data center applications.

Tom's Hardware•

7 hours ago

Solving the EDA tool fragmentation crisis

Design teams are facing challenges with EDA tool fragmentation as specialized tools for IC verification struggle to share design data efficiently. The Calibre Connectivity Interface (CCI) aims to bridge this gap by transforming Layout vs. Schematic (LVS) verification data into a universal source that downstream tools can access accurately. CCI operates on the Standard Verification Database (SVDB) to provide rich connectivity data for various analysis tools, including parasitic extraction, electromagnetic simulation, and power integrity analysis. Integration with third-party tools like Empyrean's PEX, Phlexing's GloryEX, Synopsys StarRC, and Cadence QRC showcases CCI's ability to streamline workflows and enhance design accuracy. The article emphasizes the importance of seamless multi-tool integration in modern IC design to improve efficiency and accelerate time-to-market for semiconductor innovation.

SemiWiki•

1 day ago

Chinese GPU maker Cambricon's Q1 revenue hits $423 million as country's homegrown AI chip market accelerates — Chinese chipmakers continue to leech market share from Nvidia

Chinese GPU maker Cambricon Technologies reported a Q1 revenue of $423 million, marking a significant increase over the previous year. The company's net profit also rose by 185%, reflecting its growth in the AI chip market. Cambricon's success comes as Chinese chipmakers, including MetaX Integrated Circuits, are gaining market share from Nvidia in China. Despite facing challenges, such as losing Huawei as a major customer, Cambricon has shown substantial growth and profitability. The company aims to increase its AI accelerator shipments in 2026 and is positioned as a key player in China's AI hardware supply chain.