Advancing Semiconductor Innovation Across Materials Engineering and Manufacturing | NVIDIA Technical Blog

What this post added

This post details the collaboration between Applied Materials and NVIDIA to create an end-to-end digital development model for semiconductor innovation. It highlights the use of NVIDIA CUDA-X libraries (cuDSS, cuEST) to accelerate atomic-scale materials modeling (Ginestra) and density functional theory (DFT) workflows, achieving up to 55x speedups in quantum chemistry. It also describes the integration of NVIDIA PhysicsNeMo with Applied Materials' ACE+ platform for multiphysics process modeling, resulting in up to 35x faster chamber simulations. The post further mentions the use of NVIDIA Omniverse for creating fab digital twins to predict factory performance, unifying atomic-scale discovery, process engineering, and factory optimization.