Siemens and NVIDIA Collaborate on AI Workflows for Chip Design

Chip and PCB Design: The Role of AI in Modern Engineering

Engineering, particularly in the realms of chip and PCB design, is undergoing rapid transformation. At the heart of this change is artificial intelligence (AI), which is revolutionizing traditional workflows and enhancing productivity. This shift was highlighted during two live conversations with Siemens and NVIDIA at the Design Automation Conference (DAC), where experts dug deep into how AI is being integrated into engineering processes.

Context / Why This Matters

AI is no longer a futuristic concept; it's a present-day reality that is fundamentally altering how engineers approach design and development. In the highly competitive field of chip and PCB design, the ability to leverage AI for self-verifying and agentic workflows can provide significant advantages. These technologies not only streamline processes but also enhance the accuracy and efficiency of design tasks.

Exploring AI in Chip and PCB Design

Self-Verifying Agentic AI Workflows

The first conversation featured Amit Gupta, SVP, GM & Chief AI Strategy Officer at Siemens EDA, and Timothy Costa, VP & GM of Industrial Engineering and Quantum at NVIDIA. They discussed how Siemens and NVIDIA are teaming up on self-verifying agentic AI workflows for chip and PCB design. This collaboration is crucial as it focuses on creating systems that can validate their own operations, reducing the need for manual verification and speeding up the design process.

Self-verifying workflows ensure that every step in the design process is checked and confirmed, minimizing errors and improving reliability. This is particularly important in the complex world of chip design, where even small mistakes can have significant impacts.

AI-native EDA Workflows

One of the key areas highlighted was AI-native EDA (Electronic Design Automation) workflows. This approach integrates AI from the ground up, ensuring that every aspect of the design process benefits from advanced algorithms and machine learning. Siemens' AI-native EDA workflows are designed to handle the intricacies of semiconductor, 3D IC, and PCB design, offering a robust framework for modern engineering challenges.

The Technical Side: NeMo Gym, OpenShell, and Nemotron Models

For those who delve deeper into the technical side, the second conversation provided an in-depth look. Niranjan Sitapure, PhD, Central AI Product Manager at Siemens EDA, and Joshua Mabry, Senior Technical Lead for Enterprise AI Partnerships at NVIDIA, discussed how various tools and models come together to enhance AI capabilities.

• NeMo Gym is a framework for training and evaluating AI models, crucial for developing accurate and efficient design solutions.
• OpenShell provides a versatile platform for AI development, allowing engineers to create and test new models.
• Nemotron models are specifically designed for AI-native workflows, offering specialized tools for chip and PCB design.

Fuse EDA AI Agent and Accelerated Computing

The integration of these tools within the Fuse EDA AI Agent is pivotal. The Fuse EDA AI Agent is a comprehensive platform that brings together Siemens' expertise in design automation with NVIDIA's powerful computing capabilities. This includes:

• Accelerated computing libraries from NVIDIA, which provide the necessary horsepower for AI-driven design tasks.
• CUDA-X libraries, which optimize performance and ensure that AI models run efficiently.

Practical Tips for Implementing AI in Engineering

Implementing AI in engineering workflows requires a strategic approach. Here are some practical tips to get started:

1. Understand Your Needs: Identify specific areas in your design process that could benefit from AI. Whether it's verification, optimization, or automation, pinpointing the right use case is crucial.

2. Invest in the Right Tools: Leverage frameworks like NeMo Gym, OpenShell, and Nemotron models. These tools are designed to streamline AI integration and enhance productivity.

3. Leverage Partnerships: Collaborations with industry leaders like Siemens and NVIDIA can provide access to cutting-edge technology and expertise, accelerating your AI adoption.

4. Train Your Team: Ensure that your team is well-versed in AI technologies. This includes not only technical training but also a shift in mindset towards AI-driven solutions.

5. Iterate and Improve: AI is an evolving field. Continuously update your workflows and models to take advantage of the latest advancements.

Important Takeaways

AI is transforming engineering, particularly in chip and PCB design, by offering self-verifying and agentic workflows. These advancements not only enhance productivity but also ensure that designs are accurate and reliable. By integrating AI-native EDA workflows and leveraging tools like NeMo Gym, OpenShell, and Nemotron models, engineers can stay ahead in a rapidly evolving field.

Conclusion

The future of engineering is undeniably tied to AI. As we continue to see advancements in self-verifying agentic AI workflows and AI-native EDA workflows, the design process will become more efficient and precise. Whether you're a seasoned engineer or just starting out, understanding and integrating AI into your workflows is essential for staying competitive in the modern engineering landscape.