Accelerating Photonic Design for Next Generation Computing with Multiphysics Simulation

Enabling the era of AI and machine learning is contingent on our ability to process an astronomical amount of data using technology that is fast and sustainable. Optical interconnects packaged within the same 3DIC stack as the computational core holds the key to achieving this. The next generation of system engineers require tools that provide multiphysics solutions for evaluating the optical coupling as well as the signal, power, and structural integrity of these densely packaged, multi-die, electro-optical systems to ensure high-bandwidth, low power, and low latency. In this talk, we will show how recent developments at Ansys are enabling the design and optimization of co-packaged optics, programmable PICs for AI, and photonic circuits for quantum computing. We will demonstrate how GPU-accelerated FDTD simulations are unlocking new scales for photonic device design in advanced technologies, highlighting innovations that are crucial for meeting the demands of modern AI and future computing technologies.