Read without download here.

Photon Design has published a white paper comparing the processing speed and accuracy of the company’s EigenMode Expansion (EME) simulation tools, with the more commonly used, Finite Difference Time Domain (FDTD) alternative. The white paper uses actual design examples to objectively benchmark both computational methods. It concludes that Photon Design’s EME results were equally rigorous as those of leading FDTD vendors, but simulations were completed in around 15 seconds, on a workstation PC, whereas FDTD simulations took many hours. Even when expensive, state-of-the-art, cloud-based GPU processing was used to accelerate the FDTD simulations, runtime for a 12um radius ring resonator still took 20 minutes compared to just 11s for Photon Design’s example at over twice the size.

EME results converge to similar values as FDTD found from publication in a fraction of the time for directional coupler example.

Dominic Gallagher, CEO of Photon Design, said: “FDTD is the most prevalent computational method used by photonic simulation CAD vendors, but it’s slow to process. The misconception within the industry that FDTD’s longer runtime leads to higher rigour, is clearly dispelled in our white paper. EME and FDTD deliver similar simulation results across a range of devices, however, EME achieves these results significantly faster.

“Processing speed is as important as simulation accuracy, when designing photonic devices. To achieve optimal results, engineers often run more than a hundred separate simulations during the photonic design process, whilst fine tuning a design. EME runtimes make this workflow practical, whereas one hundred FDTD simulations would take many days. To meet time restraints and project deadlines, engineers are often forced to produce a design with fewer iterations with FDTD, compromising the quality of the results.”

Processing speed and accuracy are both key when selecting photonics simulation software, which Photon Design’s EME-based tools clearly provide.