NVIDIA cuDSS
AI overview
Sign in and the AI will write an overview from our coverage.
Headlines · 2
- NVIDIA expands physical AI collaborations in South Korea
NVIDIA said South Korean manufacturers are applying accelerated computing, AI and simulation across semiconductor production, robotics and autonomous vehicles. The company highlighted expanded work with Samsung Electronics on accelerated semiconductor engineering, Kakao Mobility on a planned Level 4 autonomous-driving system, and LG Electronics on robot-learning and factory-simulation workflows. Samsung and NVIDIA reported selected performance gains, including up to fivefold faster quantum-chemical calculations and a roughly 1,600x AI-homogenization speedup, but these vendor claims were not independently verified.
NVIDIA Blog · 🔥 0 - NVIDIA outlines an AI-defined semiconductor fab workflow
NVIDIA described an “AI-defined fab” workflow combining accelerated simulation, physics-based AI, AI agents and digital twins across semiconductor manufacturing. The report cited vendor- and partner-reported speedups of up to 10x for COMSOL simulations, 10x for Applied Materials’ Ginestra platform, 17x for key Siemens Calibre workloads using cuLitho and 35x for Applied Materials’ ACE+ topography simulations. Lam Research intends to integrate NanoVDB into SEMulator3D, with the addition expected to accelerate selected wafer-etch modeling by up to 30x. Applied Materials also uses NVIDIA Omniverse libraries for fab-level digital twins, while Advantest uses NeMo Switchyard agents to automate model selection. NVIDIA said engineers remain responsible for validation and final decisions; the performance claims were not independently verified in the supplied reports.
NVIDIA Blog · 🔥 0
Experience and discussion from the community
Share my NVIDIA cuDSS experienceAsk about NVIDIA cuDSS
Nobody has shared their experience with NVIDIA cuDSS yet.