Back to home
Technology

Chip Industry Technical Paper Roundup: Sept 8

Source

SemiEngineering

Published

TL;DR

AI Generated

The article highlights new technical papers added to Semiconductor Engineering's library, covering various topics in the chip industry. Papers include discussions on memory limits and opportunities by Stanford University, leveraging 3D technologies for hardware security by UCSB and Columbia University, large language models for electronic design automation by multiple universities, and more. Other topics covered include simulating hardware for high-level synthesis designs, balancing power and performance in multi-core systems, and exploring new tools for processing-in-memory architectures. Additionally, there are papers on modulators for high-speed communications, MRAM bitcell scaling, and cybersecurity threats in AI system development.

Read Full Article

Similar Articles

SemiEngineering

What Does Semiconductor Disruption Look Like?

The article discusses the potential disruption AI could bring to chip design and EDA tools, with insights from Anand Thiruvengadam of Synopsys. It explores how AI could transform the EDA flow, emphasizing the need for true disruption rather than just productivity aids. The piece delves into the impact of AI on bias and errors in systems, as well as the complexities of chip design and the historical evolution of technology like parallel processing. It concludes by speculating on how EDA disruption might manifest, particularly in high-level synthesis, and the importance of AI advancements in the semiconductor industry.

SemiEngineering
Intel could sell up to 49% of its foundry business to external investors, but a full IPO or spin-off is unlikely

Intel could sell up to 49% of its foundry business to external investors, but a full IPO or spin-off is unlikely

Intel is considering selling up to 49% of its foundry business to external investors, but a full IPO or spin-off is unlikely. The company's CFO mentioned that selling a stake below 51% could trigger issues with the U.S. government. Intel's ownership agreements with the government require it to maintain at least 51% control of its foundry unit for the next five years. The Semiconductor Co-Investment Program allows Intel to raise capital without violating ownership clauses. While an IPO is still possible, the partial ownership of key fabs by third parties may complicate the process and reduce potential valuation.

Tom's Hardware
Simulating HW with C Speed and RTL Accuracy for HLS Designs (Georgia Tech)

Simulating HW with C Speed and RTL Accuracy for HLS Designs (Georgia Tech)

Researchers at Georgia Tech have introduced a new framework called OmniSim that enhances the simulation capabilities of high-level synthesis (HLS) tools for hardware design. OmniSim enables fast and accurate simulation of complex dataflow designs that were previously unsupported by commercial tools. By utilizing software multi-threading and FIFO tables to record hardware timing, OmniSim achieves near-C simulation speed with near-RTL accuracy. The framework successfully simulates designs that were not supported by any HLS tool before, achieving significant speedups over traditional simulation methods. This advancement addresses the challenges in simulating HLS designs accurately and efficiently.

SemiEngineering
SemiEngineering

Challenges In Stacking HBM

AI data centers are aiming for denser high-bandwidth memory, with stacking layers expected to increase from 8 to 24 by 2030. The main challenge lies in the interconnects and aligning microbumps, particularly as bump pitch decreases to less than 10 microns at 16 layers. Damon Tsai from Onto Innovation discusses strategies to reduce stress-induced warpage, necessary changes in HBM architectures, and the implications of incorporating hybrid bonding and co-packaged optics in these devices. The article highlights the complexities and advancements needed in stacking HBM to meet future demands in semiconductor engineering.

SemiEngineering

We use cookies

We use cookies to ensure you get the best experience on our website. For more information on how we use cookies, please see our cookie policy.