English
In advanced processes, the process name (e.g., 7nm, 5nm, 3nm) does not correspond to the actual physical dimensions. Instead, chip manufacturers use them to differentiate generations for marketing purposes, highlighting performance improvements to consumers. In reality, recent process node names no longer refer to any specific circuit dimension. By
English
Part 1: The Beginning of Implementation (Front-end Bridge) Foreword: The First Gateway Before RTL Code Becomes Silicon In modern semiconductor design, particularly in the process of designing SoCs (Systems on Chip) integrating billions of transistors, 'RTL coding' is merely the starting point of a monumental journey. A common
English
The New Order of the Silicon Empire and the Value of Engineers As of 2025, semiconductors stand at the center of global economic and technological hegemony. Once called the 'rice of industry,' semiconductors have now been elevated to the 'the 'brain of industry' and a
![[VLSI CAD] Why Optimization is Hard: NP-hard, heuristic, AI EDA](https://images.unsplash.com/photo-1581089778245-3ce67677f718?crop=entropy&cs=tinysrgb&fit=max&fm=jpg&ixid=M3wxMTc3M3wwfDF8c2VhcmNofDE3fHxtYXRofGVufDB8fHx8MTc2NzI0MzQ1MXww&ixlib=rb-4.1.0&q=80&w=600)
English
TL;DR * Many of the core optimization problems in semiconductor design are NP-hard / NP-Complete / PSPACE-complete, so algorithms that guarantee an "optimal solution" are virtually unprofitable at scale (e.g., even if there are only 40 cases of 2, there are 1 trillion cases) * This is why EDA tools
English
Executive Summary * Small Market and High R&D Burden: The EDA industry is a niche market with an extremely limited customer base, making research and development (R&D) for software development extremely costly. * The limited number of chip design companies in the world makes it difficult to achieve
English
It used to be that "how many nanometers of process or how many cores" seemed to describe most of the performance of semiconductors. Nowadays, words like chiplet, 2.5D, 3D IC, and UCIe are the first things you'll hear. It's not just because they&
English
One of the most frustrating bugs after semiconductor tape-out is this one:Function simulation is perfect, Silicon Bring-up works most of the time, but "sometimes" it behaves erratically. When I try to reproduce it in front of a professor, it works fine.... The debug logs are also faint
English
5. Why March seems intuitive from a memory structure perspective 5.1 How to Access SRAM Consider a bank: * Select a specific row (wordline) with the address bus * Read or write the value of a cell through the bitline associated with that row * Can only control "one address"
English
On December 1, 2025, NVIDIA announced an approximately $2 billion equity investment in Synopsys, one of the world’s leading EDA (Electronic Design Automation) companies. TL;DR: The collaboration between Synopsys and NVIDIA aims to tightly integrate NVIDIA’s CUDA GPU-accelerated computing with Synopsys’ EDA and engineering-simulation technologies, enabling ultra-high-speed
English
electronic design automation (EDA) and semiconductor IP sectors have been recognized for their high growth and strategic importance in the AI era. Big techs are doing well, but certain semiconductor companies are still facing challenges. These are the companies that used to be the best in the world and are
English
Executive summary * HBM4 changes the scaling axis: instead of pushing only pin rate, JEDEC doubles the bus width to 2048-bit and channels to 32 per stack, delivering ~2 TB/s per stack at 8 Gb/s—a structural shift from HBM3(E). (Tom's Hardware) * Vendors are already exceeding
English
In digital semiconductors, the most critical signal is the clock. The delay from the clock source to the clock pin of a sequential cell is called Clock Latency. The difference in latency between sequential cells is referred to as Clock Skew. From Zero Skew to Useful Skew In the past,