template<class Base >

void CppAD::local::reverse_erf_op ( size_t  d, size_t  i_z, const addr_t *  arg, const Base *  parameter, size_t  cap_order, const Base *  taylor, size_t  nc_partial, Base *  partial  )

inline

Compute reverse mode partial derivatives for result of op = ErfOp.

The C++ source code corresponding to this operation is

     z = erf(x)

Template Parameters

Base	base type for the operator; i.e., this operation was recorded using AD< Base > and computations by this routine are done using type Base.

Parameters

d	highest order Taylor of the Taylor coefficients that we are computing the partial derivatives with respect to.
i_z	variable index corresponding to the last (primary) result for this operation; i.e. the row index in taylor corresponding to z. The auxillary results are called y_j have index i_z - j.
arg	arg[0]: is the variable index corresponding to x. arg[1]: is the parameter index corresponding to the value zero. [2]: is the parameter index correspodning to the value 2 / sqrt(pi).
parameter	parameter[ arg[1] ] is the value zero, and parameter[ arg[2] ] is the value 2 / sqrt(pi).
cap_order	maximum number of orders that will fit in the taylor array.
taylor	Input: taylor [ arg[0] * cap_order + k ] for k = 0 , ... , d, is the k-th order Taylor coefficient corresponding to x. taylor [ (i_z - j) * cap_order + k ] for k = 0 , ... , d, and for j = 0 , ... , 4, is the k-th order Taylor coefficient corresponding to the j-th result for this operation.
nc_partial	number of columns in the matrix containing all the partial derivatives
partial	Input: partial [ arg[0] * nc_partial + k ] for k = 0 , ... , d, is the partial derivative of G( z , x , w , u , ... ) with respect to the k-th order Taylor coefficient for x. Input: partial [ (i_z - j) * nc_partial + k ] for k = 0 , ... , d, and for j = 0 , ... , 4, is the partial derivative of G( z , x , w , u , ... ) with respect to the k-th order Taylor coefficient for the j-th result of this operation. Output: partial [ arg[0] * nc_partial + k ] for k = 0 , ... , d, is the partial derivative of H( x , w , u , ... ) with respect to the k-th order Taylor coefficient for x. Output: partial [ (i_z-j) * nc_partial + k ] for k = 0 , ... , d, and for j = 0 , ... , 4, may be used as work space; i.e., may change in an unspecified manner.

Checked Assertions

• NumArg(op) == 3
• NumRes(op) == 5
• q < cap_order
• p <= q

Definition at line 479 of file erf_op.hpp.

Referenced by reverse_sweep().