As the semiconductor industry continues to demand efficient and accurate designs, AI technology is becoming an important force in driving this change. Recently, Ansys announced that it is leveraging NVIDIA's AI hardware and software to drive innovation in electronic design automation (EDA). This collaboration will not only accelerate the process of semiconductor design, but also bring higher precision and performance to chip design, further promoting technological breakthroughs in the industry.
This integration will dramatically accelerate design optimization, enabling engineers to quickly iterate on designs and expand the broader design space using customized generative AI replacement models. In particular, the collaboration will help improve design efficiency in areas such as GPUs, high-performance computing (HPC) chips, AI chips, smartphone processors, and advanced analog integrated circuits.
NVIDIA Modulus Framework: Accelerate AI-driven design optimization
NVIDIA Modulus is an AI framework that combines physics domain knowledge with simulation data to train and deploy AI models that reflect physics. The framework allows users to create customized AI engines for complex semiconductor design tasks. In this collaboration, Ansys seamlessly integrates the Modulus framework with its SeaScape platform to empower engineers with powerful AI. Through this integration, the high-fidelity data generated by Ansys' solvers will be the basis for training AI models, helping designers explore and optimize designs in less time.
AI enables rapid design iteration and optimization
By integrating NVIDIA Modulus, Ansys will enable engineers to accelerate the design process by training AI models. For example, designers can use existing design libraries in Ansys RedHawk-SC to train AI models in the integrated Modulus framework. Once the AI model is trained, engineers can use it to perform rapid design optimizations, automatically identifying the best design to meet requirements such as performance, power, and size. This approach not only greatly improves design efficiency, but also reduces the iteration of the traditional design process.
Pictured: Ansys and NVIDIA join forces to advance AI-accelerated semiconductor design (Source: PRNewswire)
Ansys also plans to integrate the AI accelerators created by Modulus into several of its semiconductor simulation tools, including RedHawk-SC, Totem-SC, PathFinder-SC and RedHawk-SC Electrothermal. These integrations will further speed up thermal simulations and simplify the power calculation process. With AI-accelerated simulation pipelines, Ansys and NVIDIA have demonstrated that thermal simulations can be faster by more than 100 times with this new technology.
Collaboration Perspective: Driving the Future of EDA Tools
Ansys and NVIDIA's deep collaboration goes beyond technology integration and will play an important role in driving innovation in electronic design automation (EDA) tools. NVIDIA has long been a valued partner of Ansys and one of its key customers for its products. With NVIDIA's powerful silicon technology and computing power, Ansys is able to continuously optimize its semiconductor design solutions to provide customers with more efficient and accurate design tools.
"NVIDIA Modulus' ability to generate physics-based AI models through a simple training and deployment process is critical to speeding up simulations and discovering the best design options," said Tim Costa, senior director at NVIDIA. The combination with Ansys' multiphysics simulation products will provide customers with a faster and more efficient design process. ”
Industry Impact: AI Accelerates the Future of Semiconductor Design
As AI technology continues to evolve, the revolution in semiconductor design is accelerating. From thermal simulation to power calculation to design optimization, the application of AI is gradually penetrating all aspects of semiconductor design. Ansys' collaboration with NVIDIA further demonstrates the potential of AI in this space. By combining physics simulation with AI, engineers can not only improve design accuracy, but also dramatically improve design efficiency, driving rapid innovation for the next generation of high-performance chips.
AI-accelerated design optimization technology will become a key competitiveness in future semiconductor R&D, helping companies launch advanced products that meet market needs more quickly.
