R package TMB. With this tool, a user should be able to quickly implement complex random effect models through simple C++ templates. The package combines CppAD (C++ automatic differentiation), Eigen (templated matrix-vector library) and CHOLMOD (sparse matrix routines available from R) to obtain an efficient implementation of the applied Laplace approximation with exact derivatives. Key features are: Automatic sparseness detection, parallelism through BLAS and parallel user templates.