| Model (DOF) | v.2.20 Time | v.2.21 Time | Speedup | |-------------|-------------|-------------|---------| | Cantilever beam (3.2k) | 1.2s | 0.4s | 3.0x | | Perforated plate (18k) | 48s (est.) | 2.1s | ~22x | | Geomechanics slope (45k) | OOM | 11.4s | – |
(OOM = Out of Memory)
No tool is without faults, and FF2D v.2.21 had its limitations. Being a 2D solver, it assumed infinite invariance in the third dimension, an idealization that never truly exists in reality. Furthermore, relying on MATLAB created a financial barrier; the software was inaccessible to those without a university license. ff2d v.2.21
However, the legacy of FF2D v.2.21 is profound. It served as an educational platform. Because the code was often open-source (or readable MATLAB scripts), students could look "under the hood" to see exactly how the Finite Difference method was implemented. It demystified the black box of commercial solvers. Many current electromagnetic simulation packages owe their intuitive workflows to the standards set by early academic codes like FF2D. | Model (DOF) | v
Perhaps the most critical fixes in v.2.21 related to the physics solver: However, the legacy of FF2D v