0.1.4 More Numba optimizations and 4x simulation speedup (2026-04-21)

We made another round of performance-focused improvements to the simulation core, including tighter hot-loop kernels, reduced temporary allocations, and cleaner data flow between attitude dynamics and actuator updates.

Result: depending on the scenario, simulations now run up to about 4x faster compared to the previous baseline.

Highlights

  • Further optimized Numba-compiled math and dynamics paths.

  • Reduced Python overhead in per-step simulation loops.

  • Improved memory behavior by avoiding unnecessary intermediate arrays.

Speed Comparison

Before

After
Terminal output before latest Numba optimizations Terminal output after latest Numba optimizations