Evaluate a survival model's S, C++ fast path for the built-in families

Internal dispatcher used by the Gibbs hot path (StEM() through simulateD() and the per-sweep survivalpSameH refresh): built-in models evaluate through the fused, OpenMP-parallel survivalS_cpp() kernel; type = "custom" models keep calling the user's R function, unchanged.

Usage

survival_S(model, X, alpha, times, threads = 4)

Arguments

model

Object built by survival_model().

X

Covariate matrix, intercept column first (same contract as model$S).

alpha

Parameter vector (layout per family, see survival_model()).

times

Numeric vector of time gaps.

threads

Integer, number of OpenMP threads for the C++ kernel (default 4; ignored without OpenMP; identical results for any value).

Value

Numeric vector of survival probabilities, one per row of X.