Dealing With AI/ML Uncertainty


Despite their widespread popularity, large language models (LLMs) have several well-known design issues, the most notorious being hallucinations, in which an LLM tries to pass off its statistics-based concoctions as real-world facts. Hallucinations are examples of a fundamental, underlying issue with LLMs. The inner workings of LLMs, as well as other deep neural nets (DNNs), are only partly kno... » read more

Blog Review: April 24


Cadence's Vatsal Patel notes the factors that make high-bandwidth memory ideal for AI, such as improved bandwidth and area from vertical stacking and power reduction features like data bus inversion. Synopsys' Rob van Blommestein points to early power network analysis as a way to ensure that enough power is delivered to each transistor to mitigate potential power-related issues within the ch... » read more

Chip Industry Week In Review


SK hynix and TSMC plan to collaborate on HBM4 development and next-generation packaging technology, with plans to mass produce HBM4 chips in 2026. The agreement is an early indicator for just how competitive, and potentially lucrative, the HBM market is becoming. SK hynix said the collaboration will enable breakthroughs in memory performance with increased density of the memory controller at t... » read more

Blog Review: April 17


Siemens' Sumit Vishwakarma highlights the importance of crystal oscillators to the proper functioning of many semiconductor devices and applications, from clock signals to transmission and reception of radio waves. Cadence's Jay Domadia introduces some of the new features in GDDR7, such as a semi-independent row and column command address bus and two modes of data signaling, enabling PAM3 fo... » read more

Architecting Chips For High-Performance Computing


The world’s leading hyperscaler cloud data center companies — Amazon, Google, Meta, Microsoft, Oracle, and Akamai — are launching heterogeneous, multi-core architectures specifically for the cloud, and the impact is being felt in high-performance CPU development across the chip industry. It's unlikely that any these chips will ever be sold commercially. They are optimized for specific ... » read more

Chip Industry Week In Review


Applied Materials may scale back or cancel its $4 billion new Silicon Valley R&D facility in light of the U.S. government's recent announcement to reduce funding for construction, modernization, or expansion of semiconductor research and development (R&D) facilities in the United States, according to the San Francisco Chronicle. TSMC could receive up to $6.6 billion in direct funding... » read more

BOLT Optimization Technology Could Bring Obvious Performance Uplift On Arm Server


BOLT is a post-link optimization technology which builds on LLVM framework, which leverages perf tool to collection sampling data and convert the executable into an optimized version. After evaluating BOLT on several workloads such as MySQL, Redis, memcached and nginx on Arm server, we could see obvious performance uplift. This blog post illustrates the methods used to enable BOLT and per... » read more

Performance Boost In Powerful Real-Time Cortex-R Processor Using Data Prefetch Control


High-performance processors employ hardware data prefetching to reduce the negative performance impact of large main memory latencies. An effective prefetching mechanism can improve cache hit rate significantly. Data prefetching boosts the execution performance by fetching data before it is needed. While prefetching improves performance substantially on many programs, it can significantly red... » read more

Blog Review: April 10


Cadence's Shyam Sharma looks at the evolution of the LPDDR standard and finds that LPDDR5X is opening new specialized markets for low-power DRAMs beyond the traditional areas of mobile, IoT, and automotive. Siemens' Hossam Sarhan and Dusan Petranovic find that new physical verification approaches are needed to ensure the performance and reliability of superconducting ICs and introduce a hybr... » read more

Chip Industry Week In Review


By Jesse Allen, Gregory Haley, and Liz Allan. The Japanese government approved $3.9 billion in funding for chipmaker Rapidus to expand its foundry business, of which 10% will be invested in advanced packaging. This is in addition to the previously announced $2.18 billion in funding. In a meeting next week, the U.S. and Japan are expected to cooperate on increasing semiconductor development a... » read more

← Older posts