Simulate Light Faster with Ansys Lumerical’s HPC Workflows

The next episode of Ansys’ photonics webinar series presents the latest workflows for high-performance computing (HPC) with Ansys Lumerical tools. HPC allows you to meet aggressive deadlines on time and on budget, and provides you with the ability to verify designs too large to practically simulate other ways.

Support for Photonic Inverse Design

Compute-intensive optimization techniques such as photonic inverse design (PID) are used by engineers to stretch their ideas to the performance limit. Design improvements that would otherwise be conceptually impossible can be achieved with PID, often requiring access to large-scale computing clusters to meet ambitious deadlines. Ansys Lumerical supports both PID and cloud deployment, supporting your next-generation photonics designs.

Whether on-premise or in the cloud, using HPC helps you meet aggressive deadlines. HPC workflows distribute the simulation effort across multiple nodes, allowing a job that would take 10 hours on a single computer to run in one hour on 10 computers for the same cost.

Cloud platforms enable on-demand access to massive computing resources during peak times of the development cycle. They provide secure, flexible, elastic computing with leading-edge hardware, minimal upfront cost and no maintenance.
 

Simulate Light Faster with Ansys Lumerical

Ansys supports workflows for its Lumerical photonic simulation tools that allow you to quickly and seamlessly migrate your simulation workflows to in-house and cloud platforms.

Key HPC Features available for Lumerical products include:

• Support for accelerating very large simulations by distributing them across multiple nodes
• Support for accelerating parameter sweeps by concurrently running them on multiple cores
• Job check pointing to reduce computing costs, enabling users to recover from hardware failures or to access market pricing from cloud providers
• Optimized Python API data transfer and meshing efficiency to maximize HPC performance
• Support for third-party cluster management and job scheduler tools 

Ansys Lumerical has continued to improve HPC simulation performance for its meshing and Python API (left) and now supports third-party cluster and job scheduler tools (right).

To date, leading photonics designers have used Ansys's HPC solution for Lumerical products to achieve results not possible otherwise. In some cases, HPC has been coupled with photonics inverse design to generate non-intuitive designs that perform better and have a smaller area footprint than designs generated by hand. In other cases, HPC has been used for verification on very large designs that would normally be difficult to simulate because of their memory requirements and unreasonably long run times.