Semiconductor News & Analysis Feed
20 articles
2026-08-21
developer.nvidia.com
2026-08-21
2026-08-12
developer.nvidia.com
2026-08-12
2026-08-11
developer.nvidia.com
2026-08-11
NVIDIA's NeMo Switchyard introduces a significant advancement in AI agent task routing across multiple models. As AI applications grow in complexity, single models struggle to meet diverse task requirements, each with distinct strengths, costs, and performance profiles. NeMo Switchyard addresses thi
2026-08-04
developer.nvidia.com
2026-08-04
2026-07-31
developer.nvidia.com
2026-07-31
2026-07-24
developer.nvidia.com
2026-07-24
2026-07-21
developer.nvidia.com
2026-07-21
2026-07-21
developer.nvidia.com
2026-07-21
2026-07-21
developer.nvidia.com
2026-07-21
2026-07-16
developer.nvidia.com
2026-07-16
2026-07-16
developer.nvidia.com
2026-07-16
2026-07-15
news.google.com
2026-07-15
2026-07-14
news.google.com
2026-07-14
2026-07-11
developer.nvidia.com
2026-07-11
This article explores kernel fusion techniques in NVIDIA CUDA and their impact on GPU performance optimization. Kernel fusion combines multiple GPU operations into a single device kernel, reducing memory traffic and kernel launch overhead. The post demonstrates how traditional two-kernel approaches
2026-07-11
developer.nvidia.com
2026-07-11
This article explores the co-design approach for large language models (LLMs) in the AI domain, emphasizing how model design can be aligned with hardware characteristics to enhance overall performance. It highlights that AI performance is determined by three key dimensions: accuracy, throughput, and
2026-07-10
developer.nvidia.com
2026-07-10
This article explores the use of NVIDIA NeMo tools to generate high-quality synthetic financial data, addressing the challenges of limited and imbalanced data in financial natural language processing (NLP). Real-world financial data tends to be skewed toward common events like earnings reports and s
2026-07-08
news.google.com
2026-07-08
2026-07-07
developer.nvidia.com
2026-07-07
NVIDIA's new Vera CPU enhances AI factory throughput by optimizing CPU performance for agentic workloads. Agentic systems rely on multi-step workflows combining inference, tool use, code execution, and retrieval, where the CPU plays a critical role in coordinating GPU resources and managing intermed
2026-07-07
developer.nvidia.com
2026-07-07
NVIDIA's technical blog highlights a novel approach to enhancing Goodput in large-scale LLM training through Nonuniform Tensor Parallelism (NTP). As AI model training increasingly relies on thousands of GPUs, interruptions and resource fluctuations pose significant challenges. NTP addresses these by
2026-07-01
developer.nvidia.com
2026-07-01
NVIDIA's technical blog explores the design and implementation of GPU-accelerated query engines (GQE), aimed at enhancing SQL query performance on large datasets using modern NVIDIA hardware. GQE leverages technologies such as high bandwidth memory (HBM), NVLink-C2C, and dedicated decompression engi