Semiconductor News & Analysis Feed
12 articles
2026-09-12
news.google.com
2026-09-12
2026-09-11
2026-08-31
finance.biggo.com
2026-08-31
2026-08-29
www.indexbox.io
2026-08-29
2026-07-11
developer.nvidia.com
2026-07-11
This article explores how host offloading techniques can alleviate high-bandwidth memory (HBM) bottlenecks in large language model (LLM) training using the JAX framework. As model size, sequence length, and batch size increase, GPU memory becomes a critical constraint. The study highlights NVIDIA's
2026-06-17
tomshardware.com
2026-06-17
NVIDIA has released RTX Remix 1.5, introducing significant enhancements including improved RTX IO storage compression that reduces modded game file sizes by up to 37.5%. This update builds on the existing RTX IO technology introduced with the RTX 30 series in 2020, now more tightly integrated into t
2026-06-12
mezha.ua
2026-06-12
Analyst Mezha reports that TSMC is actively developing cost-reducing advanced packaging processes, highlighting the semiconductor industry's ongoing efforts to address rising manufacturing costs and competitive pressures. This development underscores TSMC's strategic focus on maintaining market lead
2026-06-08
news.google.com
2026-06-08
2026-06-04
www.cloudmagazin.com
2026-06-04
As AI models become more prevalent in production environments, inference costs are increasingly dominating cloud expenditure. Unlike one-time training costs, inference incurs daily charges with each request. This article explores how numerical format optimizations—such as FP8 and FP4—can significant
2026-06-01
finance.yahoo.com
2026-06-01
2026-05-27
www.donga.com
2026-05-27
SK Hynix's unveiling of innovative heat-reducing HBM technology represents a significant advancement in semiconductor cooling solutions. As data center and AI computing demands intensify, traditional memory thermal challenges have become critical bottlenecks affecting system performance and stabilit
2026-05-09
www.tomshardware.com
2026-05-09
NVIDIA's RTX Mega Geometry technology marks a significant advancement in path-traced rendering, offering improved performance and reduced VRAM consumption. By introducing a Cluster Acceleration Structure (CLAS), it dramatically speeds up BVH rebuilding and decreases CPU overhead. Testing in games li