News
Improving GPU Energy Efficiency With Component-Level Power Management (AMD)
1+ hour, 50+ min ago (252+ words) Researchers from AMD released "Comp Pow: A Case for Component-level GPU Power Management. Abstract "The ever increasing demand for ML-driven intelligence in a wide spectrum of domains has led to ubiquity of GPUs. At the same time, GPUs are notorious…...
Large-scale, SRAM-based LLM Inference Deployment (Groq)
4+ day, 3+ hour ago (272+ words) A new technical paper, "SHIP: SRAM-Based Huge Inference Pipelines for Fast LLM Serving," was published by researchers at Nvidia, with work done while at Groq. Abstract "The proliferation of large language models (LLMs) demands inference systems with both low latency…...
Cost-Effective High-Performance Flip Chip Micro Lead Frame (fc MLF) Package Introduction
4+ day, 12+ hour ago (293+ words) Enabling electrical and thermal performance enhancements while maintaining the manufacturing efficiency and scalability of the MLF leadframe technology. The post Cost-Effective High-Performance Flip Chip Micro Lead Frame (fc MLF) Package Introduction appeared first on Semiconductor Engineering. Abstract "The demand for…...
Advancing Heterogeneous Integration Through Industry Roadmap Improvements
4+ day, 12+ hour ago (565+ words) More detailed roadmaps are needed to guide how the industry stacks, connects, powers, and cools tomorrow's chips. The post Advancing Heterogeneous Integration Through Industry Roadmap Improvements appeared first on Semiconductor Engineering....
Beyond Ideal Crystals: The Case For Scale In Atomistic Modeling
4+ day, 12+ hour ago (446+ words) Almost all computer simulations face the same trade-off: larger models can be more realistic and therefore more useful, but they also take longer to run. Engineers and scientists are therefore faced with an almost daily challenge of choosing a model…...
With Chiplets, What Role Does Economics Play?
4+ day, 12+ hour ago (1174+ words) Costs can rise with chiplets. Will that change? Will it matter? Chiplets are notoriously expensive and finicky today, and they raise questions about how costs can be managed and reduced. If chiplets are to reach their full potential, those issues…...
Enabling Production-Ready AI For Semiconductor Manufacturing
4+ day, 12+ hour ago (428+ words) Deep learning for inspection needs an operationalization layer that puts capability in the hands of engineers closest to the process. The inspection community caught on to deep learning early. CNNs outperformed classical classifiers on wafer defect binning. Object detection delivered…...
AI & Energy: Bending The Curve
4+ day, 12+ hour ago (287+ words) By Pushkar P. Apte and Melissa Grupen-Shemansky Artificial intelligence (AI) is scaling at a pace that is reshaping semiconductor roadmaps, data center design, and long-term infrastructure strategy. AI promises many economic and social benefits, but the growth comes with an escalating…...
Four-Tier Memory Hierarchy for LLM Reasoning (USC, UW)
5+ day, 1+ hour ago (371+ words) A new technical paper, "Not All Thoughts Need HBM: Semantics-Aware Memory Hierarchy for LLM Reasoning," was published by researchers at USC and University of Wisconsin-Madison. Abstract "Reasoning LLMs produce thousands of chain-of-thought tokens whose KV cache must reside in scarce…...
Options Grow For Standardizing Data Movement And Sharing Resources
5+ day, 12+ hour ago (335+ words) But figuring out which ones to use, and when to use them, isn't always clear. SE: In the real world, data movement is messier than the interconnect standards would suggest. Not all data moves at the same speed, there are…...