Multivariate normal log density using a precomputed upper-Cholesky factor (R such that Sigma = t(R) %*% R, i.e. chol(Sigma)).

Always returns the log density: every call site wants log_prob()'s contract, none of dnorm()'s log = FALSE, so there is no log argument to forget to set.

Usage

dmvnorm_chol(x, mean, R)