NVIDIA expands Agent Toolkit for engineering workflows

Key Takeaways

• NVIDIA expands Agent Toolkit with NVIDIA PhysicsNeMo and CUDA-X libraries for engineering workflows
• AI physics models and accelerated computing tools enhance chip design, verification, packaging, and systems engineering
• NVIDIA Agent Toolkit enables developers to build engineering AI assistants with domain-specific tools and data
• Key capabilities include AI physics skills, iterative sparse solvers, direct sparse solvers, and quantum chemistry tools

Introduction to NVIDIA Agent Toolkit Expansion

NVIDIA has announced an expansion of the NVIDIA Agent Toolkit, adding NVIDIA PhysicsNeMo and CUDA-X libraries to support engineering workflows. This expansion enables AI engineering agents to use specialized tools, run simulations, and generate data for chip design, verification, packaging, and systems engineering.

NVIDIA PhysicsNeMo and CUDA-X Libraries

NVIDIA PhysicsNeMo provides AI physics tools for training and deploying models, while CUDA-X libraries add accelerated solvers and quantum chemistry capabilities for engineering workflows. The re-architected PhysicsNeMo libraries and updated CUDA-X libraries enable agents to use AI physics models, accelerated solvers, and quantum chemistry tools for chip, system, and industrial engineering applications.

Key Capabilities of NVIDIA Agent Toolkit

The NVIDIA Agent Toolkit offers the following key capabilities:

• AI physics skills: NVIDIA PhysicsNeMo libraries help agents train and deploy customizable AI physics models for design and simulation tasks
• Iterative sparse solvers: NVIDIA cuISS library accelerates large sparse linear systems used in physics-based and engineering simulations
• Direct sparse solvers: NVIDIA cuDSS library accelerates large sparse linear systems used in electronic design automation and scientific simulation
• Quantum chemistry: NVIDIA cuEST library supports quantum chemistry simulations for device-scale applications

Comparison of NVIDIA Libraries

Bottom Line

The expansion of the NVIDIA Agent Toolkit with NVIDIA PhysicsNeMo and CUDA-X libraries enhances the capabilities of AI engineering agents in chip design, verification, packaging, and systems engineering. With the ability to use AI physics models, accelerated solvers, and quantum chemistry tools, developers can build more efficient and effective engineering AI assistants. The NVIDIA Agent Toolkit is a powerful tool for engineering workflows, and its expansion is expected to have a significant impact on the industry.