good first issuehelp wantedperformance
Metriche repository
- Star
- (56 stelle)
- Metriche merge PR
- (Metriche PR in attesa)
Descrizione
I haven't had the time to extensively test performance of each individual function of the new DynamicalSystems.jl v3.0 API. I've checked the over-arching performance, e.g., lyapunovspectrum has approximately same speed as before.
But one should be methodical here. All functions that participate in the DynamicalSystem api should be non allocating. E.g. set_deviations(tands, Q) should be non allocating regardless of what type of tangent dynamical system we have.
Someone should check these functions with the @benchmark from BenchmarkTools.jl.